1 accept_memory= [MM] 2 Format: { eager | lazy } 3 default: lazy 4 By default, unaccepted memory is accepted lazily to 5 avoid prolonged boot times. The lazy option will add 6 some runtime overhead until all memory is eventually 7 accepted. In most cases the overhead is negligible. 8 For some workloads or for debugging purposes 9 accept_memory=eager can be used to accept all memory 10 at once during boot. 11 12 acpi= [HW,ACPI,X86,ARM64,RISCV64,EARLY] 13 Advanced Configuration and Power Interface 14 Format: { force | on | off | strict | noirq | rsdt | 15 copy_dsdt | nospcr } 16 force -- enable ACPI if default was off 17 on -- enable ACPI but allow fallback to DT [arm64,riscv64] 18 off -- disable ACPI if default was on 19 noirq -- do not use ACPI for IRQ routing 20 strict -- Be less tolerant of platforms that are not 21 strictly ACPI specification compliant. 22 rsdt -- prefer RSDT over (default) XSDT 23 copy_dsdt -- copy DSDT to memory 24 nospcr -- disable console in ACPI SPCR table as 25 default _serial_ console on ARM64 26 For ARM64, ONLY "acpi=off", "acpi=on", "acpi=force" or 27 "acpi=nospcr" are available 28 For RISCV64, ONLY "acpi=off", "acpi=on" or "acpi=force" 29 are available 30 31 See also Documentation/power/runtime_pm.rst, pci=noacpi 32 33 acpi_apic_instance= [ACPI,IOAPIC,EARLY] 34 Format: <int> 35 2: use 2nd APIC table, if available 36 1,0: use 1st APIC table 37 default: 0 38 39 acpi_backlight= [HW,ACPI] 40 { vendor | video | native | none } 41 If set to vendor, prefer vendor-specific driver 42 (e.g. thinkpad_acpi, sony_acpi, etc.) instead 43 of the ACPI video.ko driver. 44 If set to video, use the ACPI video.ko driver. 45 If set to native, use the device's native backlight mode. 46 If set to none, disable the ACPI backlight interface. 47 48 acpi_force_32bit_fadt_addr [ACPI,EARLY] 49 force FADT to use 32 bit addresses rather than the 50 64 bit X_* addresses. Some firmware have broken 64 51 bit addresses for force ACPI ignore these and use 52 the older legacy 32 bit addresses. 53 54 acpica_no_return_repair [HW, ACPI] 55 Disable AML predefined validation mechanism 56 This mechanism can repair the evaluation result to make 57 the return objects more ACPI specification compliant. 58 This option is useful for developers to identify the 59 root cause of an AML interpreter issue when the issue 60 has something to do with the repair mechanism. 61 62 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG] 63 acpi.debug_level= [HW,ACPI,ACPI_DEBUG] 64 Format: <int> 65 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI 66 debug output. Bits in debug_layer correspond to a 67 _COMPONENT in an ACPI source file, e.g., 68 #define _COMPONENT ACPI_EVENTS 69 Bits in debug_level correspond to a level in 70 ACPI_DEBUG_PRINT statements, e.g., 71 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ... 72 The debug_level mask defaults to "info". See 73 Documentation/firmware-guide/acpi/debug.rst for more information about 74 debug layers and levels. 75 76 Enable processor driver info messages: 77 acpi.debug_layer=0x20000000 78 Enable AML "Debug" output, i.e., stores to the Debug 79 object while interpreting AML: 80 acpi.debug_layer=0xffffffff acpi.debug_level=0x2 81 Enable all messages related to ACPI hardware: 82 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff 83 84 Some values produce so much output that the system is 85 unusable. The "log_buf_len" parameter may be useful 86 if you need to capture more output. 87 88 acpi_enforce_resources= [ACPI] 89 { strict | lax | no } 90 Check for resource conflicts between native drivers 91 and ACPI OperationRegions (SystemIO and SystemMemory 92 only). IO ports and memory declared in ACPI might be 93 used by the ACPI subsystem in arbitrary AML code and 94 can interfere with legacy drivers. 95 strict (default): access to resources claimed by ACPI 96 is denied; legacy drivers trying to access reserved 97 resources will fail to bind to device using them. 98 lax: access to resources claimed by ACPI is allowed; 99 legacy drivers trying to access reserved resources 100 will bind successfully but a warning message is logged. 101 no: ACPI OperationRegions are not marked as reserved, 102 no further checks are performed. 103 104 acpi_force_table_verification [HW,ACPI,EARLY] 105 Enable table checksum verification during early stage. 106 By default, this is disabled due to x86 early mapping 107 size limitation. 108 109 acpi_irq_balance [HW,ACPI] 110 ACPI will balance active IRQs 111 default in APIC mode 112 113 acpi_irq_nobalance [HW,ACPI] 114 ACPI will not move active IRQs (default) 115 default in PIC mode 116 117 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 118 Format: <irq>,<irq>... 119 120 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 121 use by PCI 122 Format: <irq>,<irq>... 123 124 acpi_mask_gpe= [HW,ACPI] 125 Due to the existence of _Lxx/_Exx, some GPEs triggered 126 by unsupported hardware/firmware features can result in 127 GPE floodings that cannot be automatically disabled by 128 the GPE dispatcher. 129 This facility can be used to prevent such uncontrolled 130 GPE floodings. 131 Format: <byte> or <bitmap-list> 132 133 acpi_no_auto_serialize [HW,ACPI] 134 Disable auto-serialization of AML methods 135 AML control methods that contain the opcodes to create 136 named objects will be marked as "Serialized" by the 137 auto-serialization feature. 138 This feature is enabled by default. 139 This option allows to turn off the feature. 140 141 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 142 kernels. 143 144 acpi_no_static_ssdt [HW,ACPI,EARLY] 145 Disable installation of static SSDTs at early boot time 146 By default, SSDTs contained in the RSDT/XSDT will be 147 installed automatically and they will appear under 148 /sys/firmware/acpi/tables. 149 This option turns off this feature. 150 Note that specifying this option does not affect 151 dynamic table installation which will install SSDT 152 tables to /sys/firmware/acpi/tables/dynamic. 153 154 acpi_no_watchdog [HW,ACPI,WDT] 155 Ignore the ACPI-based watchdog interface (WDAT) and let 156 a native driver control the watchdog device instead. 157 158 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY] 159 Pass the RSDP address to the kernel, mostly used 160 on machines running EFI runtime service to boot the 161 second kernel for kdump. 162 163 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 164 Format: To spoof as Windows 98: ="Microsoft Windows" 165 166 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 167 of 2 which is mandated by ACPI 6) as the supported ACPI 168 specification revision (when using this switch, it may 169 be necessary to carry out a cold reboot _twice_ in a 170 row to make it take effect on the platform firmware). 171 172 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 173 acpi_osi="string1" # add string1 174 acpi_osi="!string2" # remove string2 175 acpi_osi=!* # remove all strings 176 acpi_osi=! # disable all built-in OS vendor 177 strings 178 acpi_osi=!! # enable all built-in OS vendor 179 strings 180 acpi_osi= # disable all strings 181 182 'acpi_osi=!' can be used in combination with single or 183 multiple 'acpi_osi="string1"' to support specific OS 184 vendor string(s). Note that such command can only 185 affect the default state of the OS vendor strings, thus 186 it cannot affect the default state of the feature group 187 strings and the current state of the OS vendor strings, 188 specifying it multiple times through kernel command line 189 is meaningless. This command is useful when one do not 190 care about the state of the feature group strings which 191 should be controlled by the OSPM. 192 Examples: 193 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 194 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 195 can make '_OSI("Windows 2000")' TRUE. 196 197 'acpi_osi=' cannot be used in combination with other 198 'acpi_osi=' command lines, the _OSI method will not 199 exist in the ACPI namespace. NOTE that such command can 200 only affect the _OSI support state, thus specifying it 201 multiple times through kernel command line is also 202 meaningless. 203 Examples: 204 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 205 FALSE. 206 207 'acpi_osi=!*' can be used in combination with single or 208 multiple 'acpi_osi="string1"' to support specific 209 string(s). Note that such command can affect the 210 current state of both the OS vendor strings and the 211 feature group strings, thus specifying it multiple times 212 through kernel command line is meaningful. But it may 213 still not able to affect the final state of a string if 214 there are quirks related to this string. This command 215 is useful when one want to control the state of the 216 feature group strings to debug BIOS issues related to 217 the OSPM features. 218 Examples: 219 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 220 '_OSI("Module Device")' FALSE. 221 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 222 '_OSI("Module Device")' TRUE. 223 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 224 equivalent to 225 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 226 and 227 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 228 they all will make '_OSI("Windows 2000")' TRUE. 229 230 acpi_pm_good [X86] 231 Override the pmtimer bug detection: force the kernel 232 to assume that this machine's pmtimer latches its value 233 and always returns good values. 234 235 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode 236 Format: { level | edge | high | low } 237 238 acpi_skip_timer_override [HW,ACPI,EARLY] 239 Recognize and ignore IRQ0/pin2 Interrupt Override. 240 For broken nForce2 BIOS resulting in XT-PIC timer. 241 242 acpi_sleep= [HW,ACPI] Sleep options 243 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig, 244 s4_nohwsig, old_ordering, nonvs, 245 sci_force_enable, nobl } 246 See Documentation/power/video.rst for information on 247 s3_bios and s3_mode. 248 s3_beep is for debugging; it makes the PC's speaker beep 249 as soon as the kernel's real-mode entry point is called. 250 s4_hwsig causes the kernel to check the ACPI hardware 251 signature during resume from hibernation, and gracefully 252 refuse to resume if it has changed. This complies with 253 the ACPI specification but not with reality, since 254 Windows does not do this and many laptops do change it 255 on docking. So the default behaviour is to allow resume 256 and simply warn when the signature changes, unless the 257 s4_hwsig option is enabled. 258 s4_nohwsig prevents ACPI hardware signature from being 259 used (or even warned about) during resume. 260 old_ordering causes the ACPI 1.0 ordering of the _PTS 261 control method, with respect to putting devices into 262 low power states, to be enforced (the ACPI 2.0 ordering 263 of _PTS is used by default). 264 nonvs prevents the kernel from saving/restoring the 265 ACPI NVS memory during suspend/hibernation and resume. 266 sci_force_enable causes the kernel to set SCI_EN directly 267 on resume from S1/S3 (which is against the ACPI spec, 268 but some broken systems don't work without it). 269 nobl causes the internal blacklist of systems known to 270 behave incorrectly in some ways with respect to system 271 suspend and resume to be ignored (use wisely). 272 273 acpi_use_timer_override [HW,ACPI,EARLY] 274 Use timer override. For some broken Nvidia NF5 boards 275 that require a timer override, but don't have HPET 276 277 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in 278 kernel's map of available physical RAM. 279 280 agp= [AGP] 281 { off | try_unsupported } 282 off: disable AGP support 283 try_unsupported: try to drive unsupported chipsets 284 (may crash computer or cause data corruption) 285 286 ALSA [HW,ALSA] 287 See Documentation/sound/alsa-configuration.rst 288 289 alignment= [KNL,ARM] 290 Allow the default userspace alignment fault handler 291 behaviour to be specified. Bit 0 enables warnings, 292 bit 1 enables fixups, and bit 2 sends a segfault. 293 294 align_va_addr= [X86-64] 295 Align virtual addresses by clearing slice [14:12] when 296 allocating a VMA at process creation time. This option 297 gives you up to 3% performance improvement on AMD F15h 298 machines (where it is enabled by default) for a 299 CPU-intensive style benchmark, and it can vary highly in 300 a microbenchmark depending on workload and compiler. 301 302 32: only for 32-bit processes 303 64: only for 64-bit processes 304 on: enable for both 32- and 64-bit processes 305 off: disable for both 32- and 64-bit processes 306 307 alloc_snapshot [FTRACE] 308 Allocate the ftrace snapshot buffer on boot up when the 309 main buffer is allocated. This is handy if debugging 310 and you need to use tracing_snapshot() on boot up, and 311 do not want to use tracing_snapshot_alloc() as it needs 312 to be done where GFP_KERNEL allocations are allowed. 313 314 allow_mismatched_32bit_el0 [ARM64,EARLY] 315 Allow execve() of 32-bit applications and setting of the 316 PER_LINUX32 personality on systems where only a strict 317 subset of the CPUs support 32-bit EL0. When this 318 parameter is present, the set of CPUs supporting 32-bit 319 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0 320 and hot-unplug operations may be restricted. 321 322 See Documentation/arch/arm64/asymmetric-32bit.rst for more 323 information. 324 325 amd_iommu= [HW,X86-64] 326 Pass parameters to the AMD IOMMU driver in the system. 327 Possible values are: 328 fullflush - Deprecated, equivalent to iommu.strict=1 329 off - do not initialize any AMD IOMMU found in 330 the system 331 force_isolation - Force device isolation for all 332 devices. The IOMMU driver is not 333 allowed anymore to lift isolation 334 requirements as needed. This option 335 does not override iommu=pt 336 force_enable - Force enable the IOMMU on platforms known 337 to be buggy with IOMMU enabled. Use this 338 option with care. 339 pgtbl_v1 - Use v1 page table for DMA-API (Default). 340 pgtbl_v2 - Use v2 page table for DMA-API. 341 irtcachedis - Disable Interrupt Remapping Table (IRT) caching. 342 nohugepages - Limit page-sizes used for v1 page-tables 343 to 4 KiB. 344 v2_pgsizes_only - Limit page-sizes used for v1 page-tables 345 to 4KiB/2Mib/1GiB. 346 347 348 amd_iommu_dump= [HW,X86-64] 349 Enable AMD IOMMU driver option to dump the ACPI table 350 for AMD IOMMU. With this option enabled, AMD IOMMU 351 driver will print ACPI tables for AMD IOMMU during 352 IOMMU initialization. 353 354 amd_iommu_intr= [HW,X86-64] 355 Specifies one of the following AMD IOMMU interrupt 356 remapping modes: 357 legacy - Use legacy interrupt remapping mode. 358 vapic - Use virtual APIC mode, which allows IOMMU 359 to inject interrupts directly into guest. 360 This mode requires kvm-amd.avic=1. 361 (Default when IOMMU HW support is present.) 362 363 amd_pstate= [X86,EARLY] 364 disable 365 Do not enable amd_pstate as the default 366 scaling driver for the supported processors 367 passive 368 Use amd_pstate with passive mode as a scaling driver. 369 In this mode autonomous selection is disabled. 370 Driver requests a desired performance level and platform 371 tries to match the same performance level if it is 372 satisfied by guaranteed performance level. 373 active 374 Use amd_pstate_epp driver instance as the scaling driver, 375 driver provides a hint to the hardware if software wants 376 to bias toward performance (0x0) or energy efficiency (0xff) 377 to the CPPC firmware. then CPPC power algorithm will 378 calculate the runtime workload and adjust the realtime cores 379 frequency. 380 guided 381 Activate guided autonomous mode. Driver requests minimum and 382 maximum performance level and the platform autonomously 383 selects a performance level in this range and appropriate 384 to the current workload. 385 386 amd_prefcore= 387 [X86] 388 disable 389 Disable amd-pstate preferred core. 390 391 amijoy.map= [HW,JOY] Amiga joystick support 392 Map of devices attached to JOY0DAT and JOY1DAT 393 Format: <a>,<b> 394 See also Documentation/input/joydev/joystick.rst 395 396 analog.map= [HW,JOY] Analog joystick and gamepad support 397 Specifies type or capabilities of an analog joystick 398 connected to one of 16 gameports 399 Format: <type1>,<type2>,..<type16> 400 401 apc= [HW,SPARC] 402 Power management functions (SPARCstation-4/5 + deriv.) 403 Format: noidle 404 Disable APC CPU standby support. SPARCstation-Fox does 405 not play well with APC CPU idle - disable it if you have 406 APC and your system crashes randomly. 407 408 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller 409 Change the output verbosity while booting 410 Format: { quiet (default) | verbose | debug } 411 Change the amount of debugging information output 412 when initialising the APIC and IO-APIC components. 413 For X86-32, this can also be used to specify an APIC 414 driver name. 415 Format: apic=driver_name 416 Examples: apic=bigsmp 417 418 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting 419 Format: { bsp (default) | all | none } 420 bsp: External NMI is delivered only to CPU 0 421 all: External NMIs are broadcast to all CPUs as a 422 backup of CPU 0 423 none: External NMI is masked for all CPUs. This is 424 useful so that a dump capture kernel won't be 425 shot down by NMI 426 427 autoconf= [IPV6] 428 See Documentation/networking/ipv6.rst. 429 430 apm= [APM] Advanced Power Management 431 See header of arch/x86/kernel/apm_32.c. 432 433 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 434 Format: { "0" | "1" } 435 See security/apparmor/Kconfig help text 436 0 -- disable. 437 1 -- enable. 438 Default value is set via kernel config option. 439 440 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards 441 Format: <io>,<irq>,<nodeID> 442 443 arm64.no32bit_el0 [ARM64] Unconditionally disable the execution of 444 32 bit applications. 445 446 arm64.nobti [ARM64] Unconditionally disable Branch Target 447 Identification support 448 449 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory 450 Set instructions support 451 452 arm64.nompam [ARM64] Unconditionally disable Memory Partitioning And 453 Monitoring support 454 455 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension 456 support 457 458 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication 459 support 460 461 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix 462 Extension support 463 464 arm64.nosve [ARM64] Unconditionally disable Scalable Vector 465 Extension support 466 467 ataflop= [HW,M68k] 468 469 atarimouse= [HW,MOUSE] Atari Mouse 470 471 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 472 EzKey and similar keyboards 473 474 atkbd.reset= [HW] Reset keyboard during initialization 475 476 atkbd.set= [HW] Select keyboard code set 477 Format: <int> (2 = AT (default), 3 = PS/2) 478 479 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 480 keyboards 481 482 atkbd.softraw= [HW] Choose between synthetic and real raw mode 483 Format: <bool> (0 = real, 1 = synthetic (default)) 484 485 atkbd.softrepeat= [HW] 486 Use software keyboard repeat 487 488 audit= [KNL] Enable the audit sub-system 489 Format: { "0" | "1" | "off" | "on" } 490 0 | off - kernel audit is disabled and can not be 491 enabled until the next reboot 492 unset - kernel audit is initialized but disabled and 493 will be fully enabled by the userspace auditd. 494 1 | on - kernel audit is initialized and partially 495 enabled, storing at most audit_backlog_limit 496 messages in RAM until it is fully enabled by the 497 userspace auditd. 498 Default: unset 499 500 audit_backlog_limit= [KNL] Set the audit queue size limit. 501 Format: <int> (must be >=0) 502 Default: 64 503 504 bau= [X86_UV] Enable the BAU on SGI UV. The default 505 behavior is to disable the BAU (i.e. bau=0). 506 Format: { "0" | "1" } 507 0 - Disable the BAU. 508 1 - Enable the BAU. 509 unset - Disable the BAU. 510 511 baycom_epp= [HW,AX25] 512 Format: <io>,<mode> 513 514 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem 515 Format: <io>,<mode> 516 See header of drivers/net/hamradio/baycom_par.c. 517 518 baycom_ser_fdx= [HW,AX25] 519 BayCom Serial Port AX.25 Modem (Full Duplex Mode) 520 Format: <io>,<irq>,<mode>[,<baud>] 521 See header of drivers/net/hamradio/baycom_ser_fdx.c. 522 523 baycom_ser_hdx= [HW,AX25] 524 BayCom Serial Port AX.25 Modem (Half Duplex Mode) 525 Format: <io>,<irq>,<mode> 526 See header of drivers/net/hamradio/baycom_ser_hdx.c. 527 528 bdev_allow_write_mounted= 529 Format: <bool> 530 Control the ability to open a mounted block device 531 for writing, i.e., allow / disallow writes that bypass 532 the FS. This was implemented as a means to prevent 533 fuzzers from crashing the kernel by overwriting the 534 metadata underneath a mounted FS without its awareness. 535 This also prevents destructive formatting of mounted 536 filesystems by naive storage tooling that don't use 537 O_EXCL. Default is Y and can be changed through the 538 Kconfig option CONFIG_BLK_DEV_WRITE_MOUNTED. 539 540 bert_disable [ACPI] 541 Disable BERT OS support on buggy BIOSes. 542 543 bgrt_disable [ACPI,X86,EARLY] 544 Disable BGRT to avoid flickering OEM logo. 545 546 blkdevparts= Manual partition parsing of block device(s) for 547 embedded devices based on command line input. 548 See Documentation/block/cmdline-partition.rst 549 550 boot_delay= [KNL,EARLY] 551 Milliseconds to delay each printk during boot. 552 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled, 553 and you may also have to specify "lpj=". Boot_delay 554 values larger than 10 seconds (10000) are assumed 555 erroneous and ignored. 556 Format: integer 557 558 bootconfig [KNL,EARLY] 559 Extended command line options can be added to an initrd 560 and this will cause the kernel to look for it. 561 562 See Documentation/admin-guide/bootconfig.rst 563 564 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 565 bttv.radio= Most important insmod options are available as 566 kernel args too. 567 bttv.pll= See Documentation/admin-guide/media/bttv.rst 568 bttv.tuner= 569 570 bulk_remove=off [PPC] This parameter disables the use of the pSeries 571 firmware feature for flushing multiple hpte entries 572 at a time. 573 574 c101= [NET] Moxa C101 synchronous serial card 575 576 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 577 Sometimes CPU hardware bugs make them report the cache 578 size incorrectly. The kernel will attempt work arounds 579 to fix known problems, but for some CPUs it is not 580 possible to determine what the correct size should be. 581 This option provides an override for these situations. 582 583 carrier_timeout= 584 [NET] Specifies amount of time (in seconds) that 585 the kernel should wait for a network carrier. By default 586 it waits 120 seconds. 587 588 ca_keys= [KEYS] This parameter identifies a specific key(s) on 589 the system trusted keyring to be used for certificate 590 trust validation. 591 format: { id:<keyid> | builtin } 592 593 cca= [MIPS,EARLY] Override the kernel pages' cache coherency 594 algorithm. Accepted values range from 0 to 7 595 inclusive. See arch/mips/include/asm/pgtable-bits.h 596 for platform specific values (SB1, Loongson3 and 597 others). 598 599 ccw_timeout_log [S390] 600 See Documentation/arch/s390/common_io.rst for details. 601 602 cgroup_disable= [KNL] Disable a particular controller or optional feature 603 Format: {name of the controller(s) or feature(s) to disable} 604 The effects of cgroup_disable=foo are: 605 - foo isn't auto-mounted if you mount all cgroups in 606 a single hierarchy 607 - foo isn't visible as an individually mountable 608 subsystem 609 - if foo is an optional feature then the feature is 610 disabled and corresponding cgroup files are not 611 created 612 {Currently only "memory" controller deal with this and 613 cut the overhead, others just disable the usage. So 614 only cgroup_disable=memory is actually worthy} 615 Specifying "pressure" disables per-cgroup pressure 616 stall information accounting feature 617 618 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 619 Format: { { controller | "all" | "named" } 620 [,{ controller | "all" | "named" }...] } 621 Like cgroup_disable, but only applies to cgroup v1; 622 the blacklisted controllers remain available in cgroup2. 623 "all" blacklists all controllers and "named" disables 624 named mounts. Specifying both "all" and "named" disables 625 all v1 hierarchies. 626 627 cgroup_favordynmods= [KNL] Enable or Disable favordynmods. 628 Format: { "true" | "false" } 629 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS. 630 631 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 632 Format: <string> 633 nosocket -- Disable socket memory accounting. 634 nokmem -- Disable kernel memory accounting. 635 nobpf -- Disable BPF memory accounting. 636 637 checkreqprot= [SELINUX] Set initial checkreqprot flag value. 638 Format: { "0" | "1" } 639 See security/selinux/Kconfig help text. 640 0 -- check protection applied by kernel (includes 641 any implied execute protection). 642 1 -- check protection requested by application. 643 Default value is set via a kernel config option. 644 Value can be changed at runtime via 645 /sys/fs/selinux/checkreqprot. 646 Setting checkreqprot to 1 is deprecated. 647 648 cio_ignore= [S390] 649 See Documentation/arch/s390/common_io.rst for details. 650 651 clearcpuid=X[,X...] [X86] 652 Disable CPUID feature X for the kernel. See 653 arch/x86/include/asm/cpufeatures.h for the valid bit 654 numbers X. Note the Linux-specific bits are not necessarily 655 stable over kernel options, but the vendor-specific 656 ones should be. 657 X can also be a string as appearing in the flags: line 658 in /proc/cpuinfo which does not have the above 659 instability issue. However, not all features have names 660 in /proc/cpuinfo. 661 Note that using this option will taint your kernel. 662 Also note that user programs calling CPUID directly 663 or using the feature without checking anything 664 will still see it. This just prevents it from 665 being used by the kernel or shown in /proc/cpuinfo. 666 Also note the kernel might malfunction if you disable 667 some critical bits. 668 669 clk_ignore_unused 670 [CLK] 671 Prevents the clock framework from automatically gating 672 clocks that have not been explicitly enabled by a Linux 673 device driver but are enabled in hardware at reset or 674 by the bootloader/firmware. Note that this does not 675 force such clocks to be always-on nor does it reserve 676 those clocks in any way. This parameter is useful for 677 debug and development, but should not be needed on a 678 platform with proper driver support. For more 679 information, see Documentation/driver-api/clk.rst. 680 681 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 682 [Deprecated] 683 Forces specified clocksource (if available) to be used 684 when calculating gettimeofday(). If specified 685 clocksource is not available, it defaults to PIT. 686 Format: { pit | tsc | cyclone | pmtmr } 687 688 clocksource= Override the default clocksource 689 Format: <string> 690 Override the default clocksource and use the clocksource 691 with the name specified. 692 Some clocksource names to choose from, depending on 693 the platform: 694 [all] jiffies (this is the base, fallback clocksource) 695 [ACPI] acpi_pm 696 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 697 pxa_timer,timer3,32k_counter,timer0_1 698 [X86-32] pit,hpet,tsc; 699 scx200_hrt on Geode; cyclone on IBM x440 700 [MIPS] MIPS 701 [PARISC] cr16 702 [S390] tod 703 [SH] SuperH 704 [SPARC64] tick 705 [X86-64] hpet,tsc 706 707 clocksource.arm_arch_timer.evtstrm= 708 [ARM,ARM64,EARLY] 709 Format: <bool> 710 Enable/disable the eventstream feature of the ARM 711 architected timer so that code using WFE-based polling 712 loops can be debugged more effectively on production 713 systems. 714 715 clocksource.verify_n_cpus= [KNL] 716 Limit the number of CPUs checked for clocksources 717 marked with CLOCK_SOURCE_VERIFY_PERCPU that 718 are marked unstable due to excessive skew. 719 A negative value says to check all CPUs, while 720 zero says not to check any. Values larger than 721 nr_cpu_ids are silently truncated to nr_cpu_ids. 722 The actual CPUs are chosen randomly, with 723 no replacement if the same CPU is chosen twice. 724 725 clocksource-wdtest.holdoff= [KNL] 726 Set the time in seconds that the clocksource 727 watchdog test waits before commencing its tests. 728 Defaults to zero when built as a module and to 729 10 seconds when built into the kernel. 730 731 cma=nn[MG]@[start[MG][-end[MG]]] 732 [KNL,CMA,EARLY] 733 Sets the size of kernel global memory area for 734 contiguous memory allocations and optionally the 735 placement constraint by the physical address range of 736 memory allocations. A value of 0 disables CMA 737 altogether. For more information, see 738 kernel/dma/contiguous.c 739 740 cma_pernuma=nn[MG] 741 [KNL,CMA,EARLY] 742 Sets the size of kernel per-numa memory area for 743 contiguous memory allocations. A value of 0 disables 744 per-numa CMA altogether. And If this option is not 745 specified, the default value is 0. 746 With per-numa CMA enabled, DMA users on node nid will 747 first try to allocate buffer from the pernuma area 748 which is located in node nid, if the allocation fails, 749 they will fallback to the global default memory area. 750 751 numa_cma=<node>:nn[MG][,<node>:nn[MG]] 752 [KNL,CMA,EARLY] 753 Sets the size of kernel numa memory area for 754 contiguous memory allocations. It will reserve CMA 755 area for the specified node. 756 757 With numa CMA enabled, DMA users on node nid will 758 first try to allocate buffer from the numa area 759 which is located in node nid, if the allocation fails, 760 they will fallback to the global default memory area. 761 762 cmo_free_hint= [PPC] Format: { yes | no } 763 Specify whether pages are marked as being inactive 764 when they are freed. This is used in CMO environments 765 to determine OS memory pressure for page stealing by 766 a hypervisor. 767 Default: yes 768 769 coherent_pool=nn[KMG] [ARM,KNL,EARLY] 770 Sets the size of memory pool for coherent, atomic dma 771 allocations, by default set to 256K. 772 773 com20020= [HW,NET] ARCnet - COM20020 chipset 774 Format: 775 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]] 776 777 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers) 778 Format: <io>[,<irq>] 779 780 com90xx= [HW,NET] 781 ARCnet - COM90xx chipset (memory-mapped buffers) 782 Format: <io>[,<irq>[,<memstart>]] 783 784 condev= [HW,S390] console device 785 conmode= 786 787 con3215_drop= [S390,EARLY] 3215 console drop mode. 788 Format: y|n|Y|N|1|0 789 When set to true, drop data on the 3215 console when 790 the console buffer is full. In this case the 791 operator using a 3270 terminal emulator (for example 792 x3270) does not have to enter the clear key for the 793 console output to advance and the kernel to continue. 794 This leads to a much faster boot time when a 3270 795 terminal emulator is active. If no 3270 terminal 796 emulator is used, this parameter has no effect. 797 798 console= [KNL] Output console device and options. 799 800 tty<n> Use the virtual console device <n>. 801 802 ttyS<n>[,options] 803 ttyUSB0[,options] 804 Use the specified serial port. The options are of 805 the form "bbbbpnf", where "bbbb" is the baud rate, 806 "p" is parity ("n", "o", or "e"), "n" is number of 807 bits, and "f" is flow control ("r" for RTS or 808 omit it). Default is "9600n8". 809 810 See Documentation/admin-guide/serial-console.rst for more 811 information. See 812 Documentation/networking/netconsole.rst for an 813 alternative. 814 815 <DEVNAME>:<n>.<n>[,options] 816 Use the specified serial port on the serial core bus. 817 The addressing uses DEVNAME of the physical serial port 818 device, followed by the serial core controller instance, 819 and the serial port instance. The options are the same 820 as documented for the ttyS addressing above. 821 822 The mapping of the serial ports to the tty instances 823 can be viewed with: 824 825 $ ls -d /sys/bus/serial-base/devices/*:*.*/tty/* 826 /sys/bus/serial-base/devices/00:04:0.0/tty/ttyS0 827 828 In the above example, the console can be addressed with 829 console=00:04:0.0. Note that a console addressed this 830 way will only get added when the related device driver 831 is ready. The use of an earlycon parameter in addition to 832 the console may be desired for console output early on. 833 834 uart[8250],io,<addr>[,options] 835 uart[8250],mmio,<addr>[,options] 836 uart[8250],mmio16,<addr>[,options] 837 uart[8250],mmio32,<addr>[,options] 838 uart[8250],0x<addr>[,options] 839 Start an early, polled-mode console on the 8250/16550 840 UART at the specified I/O port or MMIO address, 841 switching to the matching ttyS device later. 842 MMIO inter-register address stride is either 8-bit 843 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 844 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 845 to be equivalent to 'mmio'. 'options' are specified in 846 the same format described for ttyS above; if unspecified, 847 the h/w is not re-initialized. 848 849 hvc<n> Use the hypervisor console device <n>. This is for 850 both Xen and PowerPC hypervisors. 851 852 { null | "" } 853 Use to disable console output, i.e., to have kernel 854 console messages discarded. 855 This must be the only console= parameter used on the 856 kernel command line. 857 858 If the device connected to the port is not a TTY but a braille 859 device, prepend "brl," before the device type, for instance 860 console=brl,ttyS0 861 For now, only VisioBraille is supported. 862 863 console_msg_format= 864 [KNL] Change console messages format 865 default 866 By default we print messages on consoles in 867 "[time stamp] text\n" format (time stamp may not be 868 printed, depending on CONFIG_PRINTK_TIME or 869 `printk_time' param). 870 syslog 871 Switch to syslog format: "<%u>[time stamp] text\n" 872 IOW, each message will have a facility and loglevel 873 prefix. The format is similar to one used by syslog() 874 syscall, or to executing "dmesg -S --raw" or to reading 875 from /proc/kmsg. 876 877 consoleblank= [KNL] The console blank (screen saver) timeout in 878 seconds. A value of 0 disables the blank timer. 879 Defaults to 0. 880 881 coredump_filter= 882 [KNL] Change the default value for 883 /proc/<pid>/coredump_filter. 884 See also Documentation/filesystems/proc.rst. 885 886 coresight_cpu_debug.enable 887 [ARM,ARM64] 888 Format: <bool> 889 Enable/disable the CPU sampling based debugging. 890 0: default value, disable debugging 891 1: enable debugging at boot time 892 893 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 894 Format: 895 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 896 897 cpuidle.off=1 [CPU_IDLE] 898 disable the cpuidle sub-system 899 900 cpuidle.governor= 901 [CPU_IDLE] Name of the cpuidle governor to use. 902 903 cpufreq.off=1 [CPU_FREQ] 904 disable the cpufreq sub-system 905 906 cpufreq.default_governor= 907 [CPU_FREQ] Name of the default cpufreq governor or 908 policy to use. This governor must be registered in the 909 kernel before the cpufreq driver probes. 910 911 cpu_init_udelay=N 912 [X86,EARLY] Delay for N microsec between assert and de-assert 913 of APIC INIT to start processors. This delay occurs 914 on every CPU online, such as boot, and resume from suspend. 915 Default: 10000 916 917 cpuhp.parallel= 918 [SMP] Enable/disable parallel bringup of secondary CPUs 919 Format: <bool> 920 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise 921 the parameter has no effect. 922 923 crash_kexec_post_notifiers 924 Run kdump after running panic-notifiers and dumping 925 kmsg. This only for the users who doubt kdump always 926 succeeds in any situation. 927 Note that this also increases risks of kdump failure, 928 because some panic notifiers can make the crashed 929 kernel more unstable. 930 931 crashkernel=size[KMG][@offset[KMG]] 932 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel' 933 upon panic. This parameter reserves the physical 934 memory region [offset, offset + size] for that kernel 935 image. If '@offset' is omitted, then a suitable offset 936 is selected automatically. 937 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region 938 under 4G first, and fall back to reserve region above 939 4G when '@offset' hasn't been specified. 940 See Documentation/admin-guide/kdump/kdump.rst for further details. 941 942 crashkernel=range1:size1[,range2:size2,...][@offset] 943 [KNL] Same as above, but depends on the memory 944 in the running system. The syntax of range is 945 start-[end] where start and end are both 946 a memory unit (amount[KMG]). See also 947 Documentation/admin-guide/kdump/kdump.rst for an example. 948 949 crashkernel=size[KMG],high 950 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be 951 above 4G. 952 Allow kernel to allocate physical memory region from top, 953 so could be above 4G if system have more than 4G ram 954 installed. Otherwise memory region will be allocated 955 below 4G, if available. 956 It will be ignored if crashkernel=X is specified. 957 crashkernel=size[KMG],low 958 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G. 959 When crashkernel=X,high is passed, kernel could allocate 960 physical memory region above 4G, that cause second kernel 961 crash on system that require some amount of low memory, 962 e.g. swiotlb requires at least 64M+32K low memory, also 963 enough extra low memory is needed to make sure DMA buffers 964 for 32-bit devices won't run out. Kernel would try to allocate 965 default size of memory below 4G automatically. The default 966 size is platform dependent. 967 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB) 968 --> arm64: 128MiB 969 --> riscv: 128MiB 970 --> loongarch: 128MiB 971 This one lets the user specify own low range under 4G 972 for second kernel instead. 973 0: to disable low allocation. 974 It will be ignored when crashkernel=X,high is not used 975 or memory reserved is below 4G. 976 977 cryptomgr.notests 978 [KNL] Disable crypto self-tests 979 980 cs89x0_dma= [HW,NET] 981 Format: <dma> 982 983 cs89x0_media= [HW,NET] 984 Format: { rj45 | aui | bnc } 985 986 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU 987 function call handling. When switched on, 988 additional debug data is printed to the console 989 in case a hanging CPU is detected, and that 990 CPU is pinged again in order to try to resolve 991 the hang situation. The default value of this 992 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT 993 Kconfig option. 994 995 dasd= [HW,NET] 996 See header of drivers/s390/block/dasd_devmap.c. 997 998 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 999 (one device per port) 1000 Format: <port#>,<type> 1001 See also Documentation/input/devices/joystick-parport.rst 1002 1003 debug [KNL,EARLY] Enable kernel debugging (events log level). 1004 1005 debug_boot_weak_hash 1006 [KNL,EARLY] Enable printing [hashed] pointers early in the 1007 boot sequence. If enabled, we use a weak hash instead 1008 of siphash to hash pointers. Use this option if you are 1009 seeing instances of '(___ptrval___)') and need to see a 1010 value (hashed pointer) instead. Cryptographically 1011 insecure, please do not use on production kernels. 1012 1013 debug_locks_verbose= 1014 [KNL] verbose locking self-tests 1015 Format: <int> 1016 Print debugging info while doing the locking API 1017 self-tests. 1018 Bitmask for the various LOCKTYPE_ tests. Defaults to 0 1019 (no extra messages), setting it to -1 (all bits set) 1020 will print _a_lot_ more information - normally only 1021 useful to lockdep developers. 1022 1023 debug_objects [KNL,EARLY] Enable object debugging 1024 1025 debug_guardpage_minorder= 1026 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this 1027 parameter allows control of the order of pages that will 1028 be intentionally kept free (and hence protected) by the 1029 buddy allocator. Bigger value increase the probability 1030 of catching random memory corruption, but reduce the 1031 amount of memory for normal system use. The maximum 1032 possible value is MAX_PAGE_ORDER/2. Setting this 1033 parameter to 1 or 2 should be enough to identify most 1034 random memory corruption problems caused by bugs in 1035 kernel or driver code when a CPU writes to (or reads 1036 from) a random memory location. Note that there exists 1037 a class of memory corruptions problems caused by buggy 1038 H/W or F/W or by drivers badly programming DMA 1039 (basically when memory is written at bus level and the 1040 CPU MMU is bypassed) which are not detectable by 1041 CONFIG_DEBUG_PAGEALLOC, hence this option will not 1042 help tracking down these problems. 1043 1044 debug_pagealloc= 1045 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 1046 enables the feature at boot time. By default, it is 1047 disabled and the system will work mostly the same as a 1048 kernel built without CONFIG_DEBUG_PAGEALLOC. 1049 Note: to get most of debug_pagealloc error reports, it's 1050 useful to also enable the page_owner functionality. 1051 on: enable the feature 1052 1053 debugfs= [KNL,EARLY] This parameter enables what is exposed to 1054 userspace and debugfs internal clients. 1055 Format: { on, no-mount, off } 1056 on: All functions are enabled. 1057 no-mount: 1058 Filesystem is not registered but kernel clients can 1059 access APIs and a crashkernel can be used to read 1060 its content. There is nothing to mount. 1061 off: Filesystem is not registered and clients 1062 get a -EPERM as result when trying to register files 1063 or directories within debugfs. 1064 This is equivalent of the runtime functionality if 1065 debugfs was not enabled in the kernel at all. 1066 Default value is set in build-time with a kernel configuration. 1067 1068 debugpat [X86] Enable PAT debugging 1069 1070 default_hugepagesz= 1071 [HW] The size of the default HugeTLB page. This is 1072 the size represented by the legacy /proc/ hugepages 1073 APIs. In addition, this is the default hugetlb size 1074 used for shmget(), mmap() and mounting hugetlbfs 1075 filesystems. If not specified, defaults to the 1076 architecture's default huge page size. Huge page 1077 sizes are architecture dependent. See also 1078 Documentation/admin-guide/mm/hugetlbpage.rst. 1079 Format: size[KMG] 1080 1081 deferred_probe_timeout= 1082 [KNL] Debugging option to set a timeout in seconds for 1083 deferred probe to give up waiting on dependencies to 1084 probe. Only specific dependencies (subsystems or 1085 drivers) that have opted in will be ignored. A timeout 1086 of 0 will timeout at the end of initcalls. If the time 1087 out hasn't expired, it'll be restarted by each 1088 successful driver registration. This option will also 1089 dump out devices still on the deferred probe list after 1090 retrying. 1091 1092 delayacct [KNL] Enable per-task delay accounting 1093 1094 dell_smm_hwmon.ignore_dmi= 1095 [HW] Continue probing hardware even if DMI data 1096 indicates that the driver is running on unsupported 1097 hardware. 1098 1099 dell_smm_hwmon.force= 1100 [HW] Activate driver even if SMM BIOS signature does 1101 not match list of supported models and enable otherwise 1102 blacklisted features. 1103 1104 dell_smm_hwmon.power_status= 1105 [HW] Report power status in /proc/i8k 1106 (disabled by default). 1107 1108 dell_smm_hwmon.restricted= 1109 [HW] Allow controlling fans only if SYS_ADMIN 1110 capability is set. 1111 1112 dell_smm_hwmon.fan_mult= 1113 [HW] Factor to multiply fan speed with. 1114 1115 dell_smm_hwmon.fan_max= 1116 [HW] Maximum configurable fan speed. 1117 1118 dfltcc= [HW,S390] 1119 Format: { on | off | def_only | inf_only | always } 1120 on: s390 zlib hardware support for compression on 1121 level 1 and decompression (default) 1122 off: No s390 zlib hardware support 1123 def_only: s390 zlib hardware support for deflate 1124 only (compression on level 1) 1125 inf_only: s390 zlib hardware support for inflate 1126 only (decompression) 1127 always: Same as 'on' but ignores the selected compression 1128 level always using hardware support (used for debugging) 1129 1130 dhash_entries= [KNL] 1131 Set number of hash buckets for dentry cache. 1132 1133 disable_1tb_segments [PPC,EARLY] 1134 Disables the use of 1TB hash page table segments. This 1135 causes the kernel to fall back to 256MB segments which 1136 can be useful when debugging issues that require an SLB 1137 miss to occur. 1138 1139 disable_dma32= [KNL] 1140 Dynamically disable ZONE_DMA32 on kernels compiled with 1141 CONFIG_ZONE_DMA32=y. 1142 1143 disable= [IPV6] 1144 See Documentation/networking/ipv6.rst. 1145 1146 disable_radix [PPC,EARLY] 1147 Disable RADIX MMU mode on POWER9 1148 1149 disable_tlbie [PPC] 1150 Disable TLBIE instruction. Currently does not work 1151 with KVM, with HASH MMU, or with coherent accelerators. 1152 1153 disable_ddw [PPC/PSERIES,EARLY] 1154 Disable Dynamic DMA Window support. Use this 1155 to workaround buggy firmware. 1156 1157 disable_ipv6= [IPV6] 1158 See Documentation/networking/ipv6.rst. 1159 1160 disable_mtrr_cleanup [X86,EARLY] 1161 The kernel tries to adjust MTRR layout from continuous 1162 to discrete, to make X server driver able to add WB 1163 entry later. This parameter disables that. 1164 1165 disable_mtrr_trim [X86, Intel and AMD only,EARLY] 1166 By default the kernel will trim any uncacheable 1167 memory out of your available memory pool based on 1168 MTRR settings. This parameter disables that behavior, 1169 possibly causing your machine to run very slowly. 1170 1171 disable_timer_pin_1 [X86,EARLY] 1172 Disable PIN 1 of APIC timer 1173 Can be useful to work around chipset bugs. 1174 1175 dis_ucode_ldr [X86] Disable the microcode loader. 1176 1177 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 1178 this option disables the debugging code at boot. 1179 1180 dma_debug_entries=<number> 1181 This option allows to tune the number of preallocated 1182 entries for DMA-API debugging code. One entry is 1183 required per DMA-API allocation. Use this if the 1184 DMA-API debugging code disables itself because the 1185 architectural default is too low. 1186 1187 dma_debug_driver=<driver_name> 1188 With this option the DMA-API debugging driver 1189 filter feature can be enabled at boot time. Just 1190 pass the driver to filter for as the parameter. 1191 The filter can be disabled or changed to another 1192 driver later using sysfs. 1193 1194 reg_file_data_sampling= 1195 [X86] Controls mitigation for Register File Data 1196 Sampling (RFDS) vulnerability. RFDS is a CPU 1197 vulnerability which may allow userspace to infer 1198 kernel data values previously stored in floating point 1199 registers, vector registers, or integer registers. 1200 RFDS only affects Intel Atom processors. 1201 1202 on: Turns ON the mitigation. 1203 off: Turns OFF the mitigation. 1204 1205 This parameter overrides the compile time default set 1206 by CONFIG_MITIGATION_RFDS. Mitigation cannot be 1207 disabled when other VERW based mitigations (like MDS) 1208 are enabled. In order to disable RFDS mitigation all 1209 VERW based mitigations need to be disabled. 1210 1211 For details see: 1212 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst 1213 1214 driver_async_probe= [KNL] 1215 List of driver names to be probed asynchronously. * 1216 matches with all driver names. If * is specified, the 1217 rest of the listed driver names are those that will NOT 1218 match the *. 1219 Format: <driver_name1>,<driver_name2>... 1220 1221 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1222 Broken monitors, graphic adapters, KVMs and EDIDless 1223 panels may send no or incorrect EDID data sets. 1224 This parameter allows to specify an EDID data sets 1225 in the /lib/firmware directory that are used instead. 1226 An EDID data set will only be used for a particular 1227 connector, if its name and a colon are prepended to 1228 the EDID name. Each connector may use a unique EDID 1229 data set by separating the files with a comma. An EDID 1230 data set with no connector name will be used for 1231 any connectors not explicitly specified. 1232 1233 dscc4.setup= [NET] 1234 1235 dt_cpu_ftrs= [PPC,EARLY] 1236 Format: {"off" | "known"} 1237 Control how the dt_cpu_ftrs device-tree binding is 1238 used for CPU feature discovery and setup (if it 1239 exists). 1240 off: Do not use it, fall back to legacy cpu table. 1241 known: Do not pass through unknown features to guests 1242 or userspace, only those that the kernel is aware of. 1243 1244 dump_apple_properties [X86] 1245 Dump name and content of EFI device properties on 1246 x86 Macs. Useful for driver authors to determine 1247 what data is available or for reverse-engineering. 1248 1249 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1250 <module>.dyndbg[="val"] 1251 Enable debug messages at boot time. See 1252 Documentation/admin-guide/dynamic-debug-howto.rst 1253 for details. 1254 1255 early_ioremap_debug [KNL,EARLY] 1256 Enable debug messages in early_ioremap support. This 1257 is useful for tracking down temporary early mappings 1258 which are not unmapped. 1259 1260 earlycon= [KNL,EARLY] Output early console device and options. 1261 1262 When used with no options, the early console is 1263 determined by stdout-path property in device tree's 1264 chosen node or the ACPI SPCR table if supported by 1265 the platform. 1266 1267 cdns,<addr>[,options] 1268 Start an early, polled-mode console on a Cadence 1269 (xuartps) serial port at the specified address. Only 1270 supported option is baud rate. If baud rate is not 1271 specified, the serial port must already be setup and 1272 configured. 1273 1274 uart[8250],io,<addr>[,options[,uartclk]] 1275 uart[8250],mmio,<addr>[,options[,uartclk]] 1276 uart[8250],mmio32,<addr>[,options[,uartclk]] 1277 uart[8250],mmio32be,<addr>[,options[,uartclk]] 1278 uart[8250],0x<addr>[,options] 1279 Start an early, polled-mode console on the 8250/16550 1280 UART at the specified I/O port or MMIO address. 1281 MMIO inter-register address stride is either 8-bit 1282 (mmio) or 32-bit (mmio32 or mmio32be). 1283 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1284 to be equivalent to 'mmio'. 'options' are specified 1285 in the same format described for "console=ttyS<n>"; if 1286 unspecified, the h/w is not initialized. 'uartclk' is 1287 the uart clock frequency; if unspecified, it is set 1288 to 'BASE_BAUD' * 16. 1289 1290 pl011,<addr> 1291 pl011,mmio32,<addr> 1292 Start an early, polled-mode console on a pl011 serial 1293 port at the specified address. The pl011 serial port 1294 must already be setup and configured. Options are not 1295 yet supported. If 'mmio32' is specified, then only 1296 the driver will use only 32-bit accessors to read/write 1297 the device registers. 1298 1299 liteuart,<addr> 1300 Start an early console on a litex serial port at the 1301 specified address. The serial port must already be 1302 setup and configured. Options are not yet supported. 1303 1304 meson,<addr> 1305 Start an early, polled-mode console on a meson serial 1306 port at the specified address. The serial port must 1307 already be setup and configured. Options are not yet 1308 supported. 1309 1310 msm_serial,<addr> 1311 Start an early, polled-mode console on an msm serial 1312 port at the specified address. The serial port 1313 must already be setup and configured. Options are not 1314 yet supported. 1315 1316 msm_serial_dm,<addr> 1317 Start an early, polled-mode console on an msm serial 1318 dm port at the specified address. The serial port 1319 must already be setup and configured. Options are not 1320 yet supported. 1321 1322 owl,<addr> 1323 Start an early, polled-mode console on a serial port 1324 of an Actions Semi SoC, such as S500 or S900, at the 1325 specified address. The serial port must already be 1326 setup and configured. Options are not yet supported. 1327 1328 rda,<addr> 1329 Start an early, polled-mode console on a serial port 1330 of an RDA Micro SoC, such as RDA8810PL, at the 1331 specified address. The serial port must already be 1332 setup and configured. Options are not yet supported. 1333 1334 sbi 1335 Use RISC-V SBI (Supervisor Binary Interface) for early 1336 console. 1337 1338 smh Use ARM semihosting calls for early console. 1339 1340 s3c2410,<addr> 1341 s3c2412,<addr> 1342 s3c2440,<addr> 1343 s3c6400,<addr> 1344 s5pv210,<addr> 1345 exynos4210,<addr> 1346 Use early console provided by serial driver available 1347 on Samsung SoCs, requires selecting proper type and 1348 a correct base address of the selected UART port. The 1349 serial port must already be setup and configured. 1350 Options are not yet supported. 1351 1352 lantiq,<addr> 1353 Start an early, polled-mode console on a lantiq serial 1354 (lqasc) port at the specified address. The serial port 1355 must already be setup and configured. Options are not 1356 yet supported. 1357 1358 lpuart,<addr> 1359 lpuart32,<addr> 1360 Use early console provided by Freescale LP UART driver 1361 found on Freescale Vybrid and QorIQ LS1021A processors. 1362 A valid base address must be provided, and the serial 1363 port must already be setup and configured. 1364 1365 ec_imx21,<addr> 1366 ec_imx6q,<addr> 1367 Start an early, polled-mode, output-only console on the 1368 Freescale i.MX UART at the specified address. The UART 1369 must already be setup and configured. 1370 1371 ar3700_uart,<addr> 1372 Start an early, polled-mode console on the 1373 Armada 3700 serial port at the specified 1374 address. The serial port must already be setup 1375 and configured. Options are not yet supported. 1376 1377 qcom_geni,<addr> 1378 Start an early, polled-mode console on a Qualcomm 1379 Generic Interface (GENI) based serial port at the 1380 specified address. The serial port must already be 1381 setup and configured. Options are not yet supported. 1382 1383 efifb,[options] 1384 Start an early, unaccelerated console on the EFI 1385 memory mapped framebuffer (if available). On cache 1386 coherent non-x86 systems that use system memory for 1387 the framebuffer, pass the 'ram' option so that it is 1388 mapped with the correct attributes. 1389 1390 linflex,<addr> 1391 Use early console provided by Freescale LINFlexD UART 1392 serial driver for NXP S32V234 SoCs. A valid base 1393 address must be provided, and the serial port must 1394 already be setup and configured. 1395 1396 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY] 1397 earlyprintk=vga 1398 earlyprintk=sclp 1399 earlyprintk=xen 1400 earlyprintk=serial[,ttySn[,baudrate]] 1401 earlyprintk=serial[,0x...[,baudrate]] 1402 earlyprintk=ttySn[,baudrate] 1403 earlyprintk=dbgp[debugController#] 1404 earlyprintk=pciserial[,force],bus:device.function[,baudrate] 1405 earlyprintk=xdbc[xhciController#] 1406 earlyprintk=bios 1407 1408 earlyprintk is useful when the kernel crashes before 1409 the normal console is initialized. It is not enabled by 1410 default because it has some cosmetic problems. 1411 1412 Append ",keep" to not disable it when the real console 1413 takes over. 1414 1415 Only one of vga, serial, or usb debug port can 1416 be used at a time. 1417 1418 Currently only ttyS0 and ttyS1 may be specified by 1419 name. Other I/O ports may be explicitly specified 1420 on some architectures (x86 and arm at least) by 1421 replacing ttySn with an I/O port address, like this: 1422 earlyprintk=serial,0x1008,115200 1423 You can find the port for a given device in 1424 /proc/tty/driver/serial: 1425 2: uart:ST16650V2 port:00001008 irq:18 ... 1426 1427 Interaction with the standard serial driver is not 1428 very good. 1429 1430 The VGA output is eventually overwritten by 1431 the real console. 1432 1433 The xen option can only be used in Xen domains. 1434 1435 The sclp output can only be used on s390. 1436 1437 The bios output can only be used on SuperH. 1438 1439 The optional "force" to "pciserial" enables use of a 1440 PCI device even when its classcode is not of the 1441 UART class. 1442 1443 edac_report= [HW,EDAC] Control how to report EDAC event 1444 Format: {"on" | "off" | "force"} 1445 on: enable EDAC to report H/W event. May be overridden 1446 by other higher priority error reporting module. 1447 off: disable H/W event reporting through EDAC. 1448 force: enforce the use of EDAC to report H/W event. 1449 default: on. 1450 1451 edd= [EDD] 1452 Format: {"off" | "on" | "skip[mbr]"} 1453 1454 efi= [EFI,EARLY] 1455 Format: { "debug", "disable_early_pci_dma", 1456 "nochunk", "noruntime", "nosoftreserve", 1457 "novamap", "no_disable_early_pci_dma" } 1458 debug: enable misc debug output. 1459 disable_early_pci_dma: disable the busmaster bit on all 1460 PCI bridges while in the EFI boot stub. 1461 nochunk: disable reading files in "chunks" in the EFI 1462 boot stub, as chunking can cause problems with some 1463 firmware implementations. 1464 noruntime : disable EFI runtime services support 1465 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1466 attribute may cause the kernel to reserve the 1467 memory range for a memory mapping driver to 1468 claim. Specify efi=nosoftreserve to disable this 1469 reservation and treat the memory by its base type 1470 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1471 novamap: do not call SetVirtualAddressMap(). 1472 no_disable_early_pci_dma: Leave the busmaster bit set 1473 on all PCI bridges while in the EFI boot stub 1474 1475 efi_no_storage_paranoia [EFI,X86,EARLY] 1476 Using this parameter you can use more than 50% of 1477 your efi variable storage. Use this parameter only if 1478 you are really sure that your UEFI does sane gc and 1479 fulfills the spec otherwise your board may brick. 1480 1481 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1482 that is to be dynamically loaded by Linux. If there are 1483 multiple variables with the same name but with different 1484 vendor GUIDs, all of them will be loaded. See 1485 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1486 1487 1488 eisa_irq_edge= [PARISC,HW] 1489 See header of drivers/parisc/eisa.c. 1490 1491 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging 1492 Format: ekgdboc=kbd 1493 1494 This is designed to be used in conjunction with 1495 the boot argument: earlyprintk=vga 1496 1497 This parameter works in place of the kgdboc parameter 1498 but can only be used if the backing tty is available 1499 very early in the boot process. For early debugging 1500 via a serial port see kgdboc_earlycon instead. 1501 1502 elanfreq= [X86-32] 1503 See comment before function elanfreq_setup() in 1504 arch/x86/kernel/cpu/cpufreq/elanfreq.c. 1505 1506 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY] 1507 Specifies physical address of start of kernel core 1508 image elf header and optionally the size. Generally 1509 kexec loader will pass this option to capture kernel. 1510 See Documentation/admin-guide/kdump/kdump.rst for details. 1511 1512 enable_mtrr_cleanup [X86,EARLY] 1513 The kernel tries to adjust MTRR layout from continuous 1514 to discrete, to make X server driver able to add WB 1515 entry later. This parameter enables that. 1516 1517 enable_timer_pin_1 [X86] 1518 Enable PIN 1 of APIC timer 1519 Can be useful to work around chipset bugs 1520 (in particular on some ATI chipsets). 1521 The kernel tries to set a reasonable default. 1522 1523 enforcing= [SELINUX] Set initial enforcing status. 1524 Format: {"0" | "1"} 1525 See security/selinux/Kconfig help text. 1526 0 -- permissive (log only, no denials). 1527 1 -- enforcing (deny and log). 1528 Default value is 0. 1529 Value can be changed at runtime via 1530 /sys/fs/selinux/enforce. 1531 1532 erst_disable [ACPI] 1533 Disable Error Record Serialization Table (ERST) 1534 support. 1535 1536 ether= [HW,NET] Ethernet cards parameters 1537 This option is obsoleted by the "netdev=" option, which 1538 has equivalent usage. See its documentation for details. 1539 1540 evm= [EVM] 1541 Format: { "fix" } 1542 Permit 'security.evm' to be updated regardless of 1543 current integrity status. 1544 export_pmu_events 1545 [KNL,ARM64] Sets the PMU export bit (PMCR_EL0.X), which enables 1546 the exporting of events over an IMPLEMENTATION DEFINED PMU event 1547 export bus to another device. 1548 1549 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier 1550 stages so cover more early boot allocations. 1551 Please note that as side effect some optimizations 1552 might be disabled to achieve that (e.g. parallelized 1553 memory initialization is disabled) so the boot process 1554 might take longer, especially on systems with a lot of 1555 memory. Available with CONFIG_PAGE_EXTENSION=y. 1556 1557 failslab= 1558 fail_usercopy= 1559 fail_page_alloc= 1560 fail_make_request=[KNL] 1561 General fault injection mechanism. 1562 Format: <interval>,<probability>,<space>,<times> 1563 See also Documentation/fault-injection/. 1564 1565 fb_tunnels= [NET] 1566 Format: { initns | none } 1567 See Documentation/admin-guide/sysctl/net.rst for 1568 fb_tunnels_only_for_init_ns 1569 1570 floppy= [HW] 1571 See Documentation/admin-guide/blockdev/floppy.rst. 1572 1573 forcepae [X86-32] 1574 Forcefully enable Physical Address Extension (PAE). 1575 Many Pentium M systems disable PAE but may have a 1576 functionally usable PAE implementation. 1577 Warning: use of this parameter will taint the kernel 1578 and may cause unknown problems. 1579 1580 fred= [X86-64] 1581 Enable/disable Flexible Return and Event Delivery. 1582 Format: { on | off } 1583 on: enable FRED when it's present. 1584 off: disable FRED, the default setting. 1585 1586 ftrace=[tracer] 1587 [FTRACE] will set and start the specified tracer 1588 as early as possible in order to facilitate early 1589 boot debugging. 1590 1591 ftrace_boot_snapshot 1592 [FTRACE] On boot up, a snapshot will be taken of the 1593 ftrace ring buffer that can be read at: 1594 /sys/kernel/tracing/snapshot. 1595 This is useful if you need tracing information from kernel 1596 boot up that is likely to be overridden by user space 1597 start up functionality. 1598 1599 Optionally, the snapshot can also be defined for a tracing 1600 instance that was created by the trace_instance= command 1601 line parameter. 1602 1603 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo 1604 1605 The above will cause the "foo" tracing instance to trigger 1606 a snapshot at the end of boot up. 1607 1608 ftrace_dump_on_oops[=2(orig_cpu) | =<instance>][,<instance> | 1609 ,<instance>=2(orig_cpu)] 1610 [FTRACE] will dump the trace buffers on oops. 1611 If no parameter is passed, ftrace will dump global 1612 buffers of all CPUs, if you pass 2 or orig_cpu, it 1613 will dump only the buffer of the CPU that triggered 1614 the oops, or the specific instance will be dumped if 1615 its name is passed. Multiple instance dump is also 1616 supported, and instances are separated by commas. Each 1617 instance supports only dump on CPU that triggered the 1618 oops by passing 2 or orig_cpu to it. 1619 1620 ftrace_dump_on_oops=foo=orig_cpu 1621 1622 The above will dump only the buffer of "foo" instance 1623 on CPU that triggered the oops. 1624 1625 ftrace_dump_on_oops,foo,bar=orig_cpu 1626 1627 The above will dump global buffer on all CPUs, the 1628 buffer of "foo" instance on all CPUs and the buffer 1629 of "bar" instance on CPU that triggered the oops. 1630 1631 ftrace_filter=[function-list] 1632 [FTRACE] Limit the functions traced by the function 1633 tracer at boot up. function-list is a comma-separated 1634 list of functions. This list can be changed at run 1635 time by the set_ftrace_filter file in the debugfs 1636 tracing directory. 1637 1638 ftrace_notrace=[function-list] 1639 [FTRACE] Do not trace the functions specified in 1640 function-list. This list can be changed at run time 1641 by the set_ftrace_notrace file in the debugfs 1642 tracing directory. 1643 1644 ftrace_graph_filter=[function-list] 1645 [FTRACE] Limit the top level callers functions traced 1646 by the function graph tracer at boot up. 1647 function-list is a comma-separated list of functions 1648 that can be changed at run time by the 1649 set_graph_function file in the debugfs tracing directory. 1650 1651 ftrace_graph_notrace=[function-list] 1652 [FTRACE] Do not trace from the functions specified in 1653 function-list. This list is a comma-separated list of 1654 functions that can be changed at run time by the 1655 set_graph_notrace file in the debugfs tracing directory. 1656 1657 ftrace_graph_max_depth=<uint> 1658 [FTRACE] Used with the function graph tracer. This is 1659 the max depth it will trace into a function. This value 1660 can be changed at run time by the max_graph_depth file 1661 in the tracefs tracing directory. default: 0 (no limit) 1662 1663 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier 1664 devices by scanning the firmware to infer the 1665 consumer/supplier relationships. This feature is 1666 especially useful when drivers are loaded as modules as 1667 it ensures proper ordering of tasks like device probing 1668 (suppliers first, then consumers), supplier boot state 1669 clean up (only after all consumers have probed), 1670 suspend/resume & runtime PM (consumers first, then 1671 suppliers). 1672 Format: { off | permissive | on | rpm } 1673 off -- Don't create device links from firmware info. 1674 permissive -- Create device links from firmware info 1675 but use it only for ordering boot state clean 1676 up (sync_state() calls). 1677 on -- Create device links from firmware info and use it 1678 to enforce probe and suspend/resume ordering. 1679 rpm -- Like "on", but also use to order runtime PM. 1680 1681 fw_devlink.strict=<bool> 1682 [KNL,EARLY] Treat all inferred dependencies as mandatory 1683 dependencies. This only applies for fw_devlink=on|rpm. 1684 Format: <bool> 1685 1686 fw_devlink.sync_state = 1687 [KNL,EARLY] When all devices that could probe have finished 1688 probing, this parameter controls what to do with 1689 devices that haven't yet received their sync_state() 1690 calls. 1691 Format: { strict | timeout } 1692 strict -- Default. Continue waiting on consumers to 1693 probe successfully. 1694 timeout -- Give up waiting on consumers and call 1695 sync_state() on any devices that haven't yet 1696 received their sync_state() calls after 1697 deferred_probe_timeout has expired or by 1698 late_initcall() if !CONFIG_MODULES. 1699 1700 gamecon.map[2|3]= 1701 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1702 support via parallel port (up to 5 devices per port) 1703 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1704 See also Documentation/input/devices/joystick-parport.rst 1705 1706 gamma= [HW,DRM] 1707 1708 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART 1709 Format: off | on 1710 default: on 1711 1712 gather_data_sampling= 1713 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS) 1714 mitigation. 1715 1716 Gather Data Sampling is a hardware vulnerability which 1717 allows unprivileged speculative access to data which was 1718 previously stored in vector registers. 1719 1720 This issue is mitigated by default in updated microcode. 1721 The mitigation may have a performance impact but can be 1722 disabled. On systems without the microcode mitigation 1723 disabling AVX serves as a mitigation. 1724 1725 force: Disable AVX to mitigate systems without 1726 microcode mitigation. No effect if the microcode 1727 mitigation is present. Known to cause crashes in 1728 userspace with buggy AVX enumeration. 1729 1730 off: Disable GDS mitigation. 1731 1732 gcov_persist= [GCOV] When non-zero (default), profiling data for 1733 kernel modules is saved and remains accessible via 1734 debugfs, even when the module is unloaded/reloaded. 1735 When zero, profiling data is discarded and associated 1736 debugfs files are removed at module unload time. 1737 1738 goldfish [X86] Enable the goldfish android emulator platform. 1739 Don't use this when you are not running on the 1740 android emulator 1741 1742 gpio-mockup.gpio_mockup_ranges 1743 [HW] Sets the ranges of gpiochip of for this device. 1744 Format: <start1>,<end1>,<start2>,<end2>... 1745 gpio-mockup.gpio_mockup_named_lines 1746 [HW] Let the driver know GPIO lines should be named. 1747 1748 gpt [EFI] Forces disk with valid GPT signature but 1749 invalid Protective MBR to be treated as GPT. If the 1750 primary GPT is corrupted, it enables the backup/alternate 1751 GPT to be used instead. 1752 1753 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1754 the "Enable 0" bit of the configuration register. 1755 Format: 0 | 1 1756 Default: 0 1757 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1758 the "Enable 0" bit of the configuration register. 1759 Format: 0 | 1 1760 Default: 0 1761 grcan.select= [HW] Select which physical interface to use. 1762 Format: 0 | 1 1763 Default: 0 1764 grcan.txsize= [HW] Sets the size of the tx buffer. 1765 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1766 Default: 1024 1767 grcan.rxsize= [HW] Sets the size of the rx buffer. 1768 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1769 Default: 1024 1770 1771 hardened_usercopy= 1772 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 1773 hardening is enabled for this boot. Hardened 1774 usercopy checking is used to protect the kernel 1775 from reading or writing beyond known memory 1776 allocation boundaries as a proactive defense 1777 against bounds-checking flaws in the kernel's 1778 copy_to_user()/copy_from_user() interface. 1779 on Perform hardened usercopy checks (default). 1780 off Disable hardened usercopy checks. 1781 1782 hardlockup_all_cpu_backtrace= 1783 [KNL] Should the hard-lockup detector generate 1784 backtraces on all cpus. 1785 Format: 0 | 1 1786 1787 hashdist= [KNL,NUMA] Large hashes allocated during boot 1788 are distributed across NUMA nodes. Defaults on 1789 for 64-bit NUMA, off otherwise. 1790 Format: 0 | 1 (for off | on) 1791 1792 hd= [EIDE] (E)IDE hard drive subsystem geometry 1793 Format: <cyl>,<head>,<sect> 1794 1795 hest_disable [ACPI] 1796 Disable Hardware Error Source Table (HEST) support; 1797 corresponding firmware-first mode error processing 1798 logic will be disabled. 1799 1800 hibernate= [HIBERNATION] 1801 noresume Don't check if there's a hibernation image 1802 present during boot. 1803 nocompress Don't compress/decompress hibernation images. 1804 no Disable hibernation and resume. 1805 protect_image Turn on image protection during restoration 1806 (that will set all pages holding image data 1807 during restoration read-only). 1808 1809 hibernate.compressor= [HIBERNATION] Compression algorithm to be 1810 used with hibernation. 1811 Format: { lzo | lz4 } 1812 Default: lzo 1813 1814 lzo: Select LZO compression algorithm to 1815 compress/decompress hibernation image. 1816 1817 lz4: Select LZ4 compression algorithm to 1818 compress/decompress hibernation image. 1819 1820 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact 1821 size of <nn>. This works even on boxes that have no 1822 highmem otherwise. This also works to reduce highmem 1823 size on bigger boxes. 1824 1825 highres= [KNL] Enable/disable high resolution timer mode. 1826 Valid parameters: "on", "off" 1827 Default: "on" 1828 1829 hlt [BUGS=ARM,SH] 1830 1831 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename). 1832 Format: <string> 1833 This allows setting the system's hostname during early 1834 startup. This sets the name returned by gethostname. 1835 Using this parameter to set the hostname makes it 1836 possible to ensure the hostname is correctly set before 1837 any userspace processes run, avoiding the possibility 1838 that a process may call gethostname before the hostname 1839 has been explicitly set, resulting in the calling 1840 process getting an incorrect result. The string must 1841 not exceed the maximum allowed hostname length (usually 1842 64 characters) and will be truncated otherwise. 1843 1844 hpet= [X86-32,HPET] option to control HPET usage 1845 Format: { enable (default) | disable | force | 1846 verbose } 1847 disable: disable HPET and use PIT instead 1848 force: allow force enabled of undocumented chips (ICH4, 1849 VIA, nVidia) 1850 verbose: show contents of HPET registers during setup 1851 1852 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 1853 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 1854 1855 hugepages= [HW] Number of HugeTLB pages to allocate at boot. 1856 If this follows hugepagesz (below), it specifies 1857 the number of pages of hugepagesz to be allocated. 1858 If this is the first HugeTLB parameter on the command 1859 line, it specifies the number of pages to allocate for 1860 the default huge page size. If using node format, the 1861 number of pages to allocate per-node can be specified. 1862 See also Documentation/admin-guide/mm/hugetlbpage.rst. 1863 Format: <integer> or (node format) 1864 <node>:<integer>[,<node>:<integer>] 1865 1866 hugepagesz= 1867 [HW] The size of the HugeTLB pages. This is used in 1868 conjunction with hugepages (above) to allocate huge 1869 pages of a specific size at boot. The pair 1870 hugepagesz=X hugepages=Y can be specified once for 1871 each supported huge page size. Huge page sizes are 1872 architecture dependent. See also 1873 Documentation/admin-guide/mm/hugetlbpage.rst. 1874 Format: size[KMG] 1875 1876 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation 1877 of gigantic hugepages. Or using node format, the size 1878 of a CMA area per node can be specified. 1879 Format: nn[KMGTPE] or (node format) 1880 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]] 1881 1882 Reserve a CMA area of given size and allocate gigantic 1883 hugepages using the CMA allocator. If enabled, the 1884 boot-time allocation of gigantic hugepages is skipped. 1885 1886 hugetlb_free_vmemmap= 1887 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP 1888 enabled. 1889 Control if HugeTLB Vmemmap Optimization (HVO) is enabled. 1890 Allows heavy hugetlb users to free up some more 1891 memory (7 * PAGE_SIZE for each 2MB hugetlb page). 1892 Format: { on | off (default) } 1893 1894 on: enable HVO 1895 off: disable HVO 1896 1897 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, 1898 the default is on. 1899 1900 Note that the vmemmap pages may be allocated from the added 1901 memory block itself when memory_hotplug.memmap_on_memory is 1902 enabled, those vmemmap pages cannot be optimized even if this 1903 feature is enabled. Other vmemmap pages not allocated from 1904 the added memory block itself do not be affected. 1905 1906 hung_task_panic= 1907 [KNL] Should the hung task detector generate panics. 1908 Format: 0 | 1 1909 1910 A value of 1 instructs the kernel to panic when a 1911 hung task is detected. The default value is controlled 1912 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time 1913 option. The value selected by this boot parameter can 1914 be changed later by the kernel.hung_task_panic sysctl. 1915 1916 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 1917 terminal devices. Valid values: 0..8 1918 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 1919 If specified, z/VM IUCV HVC accepts connections 1920 from listed z/VM user IDs only. 1921 1922 hvc_dcc.enable= [ARM,ARM64] Enable DCC driver at runtime. For GKI, 1923 disabled at runtime by default to prevent 1924 crashes in devices which do not support DCC. 1925 1926 hv_nopvspin [X86,HYPER_V,EARLY] 1927 Disables the paravirt spinlock optimizations 1928 which allow the hypervisor to 'idle' the guest 1929 on lock contention. 1930 1931 i2c_bus= [HW] Override the default board specific I2C bus speed 1932 or register an additional I2C bus that is not 1933 registered from board initialization code. 1934 Format: 1935 <bus_id>,<clkrate> 1936 1937 i2c_touchscreen_props= [HW,ACPI,X86] 1938 Set device-properties for ACPI-enumerated I2C-attached 1939 touchscreen, to e.g. fix coordinates of upside-down 1940 mounted touchscreens. If you need this option please 1941 submit a drivers/platform/x86/touchscreen_dmi.c patch 1942 adding a DMI quirk for this. 1943 1944 Format: 1945 <ACPI_HW_ID>:<prop_name>=<val>[:prop_name=val][:...] 1946 Where <val> is one of: 1947 Omit "=<val>" entirely Set a boolean device-property 1948 Unsigned number Set a u32 device-property 1949 Anything else Set a string device-property 1950 1951 Examples (split over multiple lines): 1952 i2c_touchscreen_props=GDIX1001:touchscreen-inverted-x: 1953 touchscreen-inverted-y 1954 1955 i2c_touchscreen_props=MSSL1680:touchscreen-size-x=1920: 1956 touchscreen-size-y=1080:touchscreen-inverted-y: 1957 firmware-name=gsl1680-vendor-model.fw:silead,home-button 1958 1959 i8042.debug [HW] Toggle i8042 debug mode 1960 i8042.unmask_kbd_data 1961 [HW] Enable printing of interrupt data from the KBD port 1962 (disabled by default, and as a pre-condition 1963 requires that i8042.debug=1 be enabled) 1964 i8042.direct [HW] Put keyboard port into non-translated mode 1965 i8042.dumbkbd [HW] Pretend that controller can only read data from 1966 keyboard and cannot control its state 1967 (Don't attempt to blink the leds) 1968 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 1969 i8042.nokbd [HW] Don't check/create keyboard port 1970 i8042.noloop [HW] Disable the AUX Loopback command while probing 1971 for the AUX port 1972 i8042.nomux [HW] Don't check presence of an active multiplexing 1973 controller 1974 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 1975 controllers 1976 i8042.notimeout [HW] Ignore timeout condition signalled by controller 1977 i8042.reset [HW] Reset the controller during init, cleanup and 1978 suspend-to-ram transitions, only during s2r 1979 transitions, or never reset 1980 Format: { 1 | Y | y | 0 | N | n } 1981 1, Y, y: always reset controller 1982 0, N, n: don't ever reset controller 1983 Default: only on s2r transitions on x86; most other 1984 architectures force reset to be always executed 1985 i8042.unlock [HW] Unlock (ignore) the keylock 1986 i8042.kbdreset [HW] Reset device connected to KBD port 1987 i8042.probe_defer 1988 [HW] Allow deferred probing upon i8042 probe errors 1989 1990 i810= [HW,DRM] 1991 1992 i915.invert_brightness= 1993 [DRM] Invert the sense of the variable that is used to 1994 set the brightness of the panel backlight. Normally a 1995 brightness value of 0 indicates backlight switched off, 1996 and the maximum of the brightness value sets the backlight 1997 to maximum brightness. If this parameter is set to 0 1998 (default) and the machine requires it, or this parameter 1999 is set to 1, a brightness value of 0 sets the backlight 2000 to maximum brightness, and the maximum of the brightness 2001 value switches the backlight off. 2002 -1 -- never invert brightness 2003 0 -- machine default 2004 1 -- force brightness inversion 2005 2006 ia32_emulation= [X86-64] 2007 Format: <bool> 2008 When true, allows loading 32-bit programs and executing 32-bit 2009 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at 2010 boot time. When false, unconditionally disables IA32 emulation. 2011 2012 icn= [HW,ISDN] 2013 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]] 2014 2015 2016 idle= [X86,EARLY] 2017 Format: idle=poll, idle=halt, idle=nomwait 2018 Poll forces a polling idle loop that can slightly 2019 improve the performance of waking up a idle CPU, but 2020 will use a lot of power and make the system run hot. 2021 Not recommended. 2022 idle=halt: Halt is forced to be used for CPU idle. 2023 In such case C2/C3 won't be used again. 2024 idle=nomwait: Disable mwait for CPU C-states 2025 2026 idxd.sva= [HW] 2027 Format: <bool> 2028 Allow force disabling of Shared Virtual Memory (SVA) 2029 support for the idxd driver. By default it is set to 2030 true (1). 2031 2032 idxd.tc_override= [HW] 2033 Format: <bool> 2034 Allow override of default traffic class configuration 2035 for the device. By default it is set to false (0). 2036 2037 ieee754= [MIPS] Select IEEE Std 754 conformance mode 2038 Format: { strict | legacy | 2008 | relaxed | emulated } 2039 Default: strict 2040 2041 Choose which programs will be accepted for execution 2042 based on the IEEE 754 NaN encoding(s) supported by 2043 the FPU and the NaN encoding requested with the value 2044 of an ELF file header flag individually set by each 2045 binary. Hardware implementations are permitted to 2046 support either or both of the legacy and the 2008 NaN 2047 encoding mode. 2048 2049 Available settings are as follows: 2050 strict accept binaries that request a NaN encoding 2051 supported by the FPU 2052 legacy only accept legacy-NaN binaries, if supported 2053 by the FPU 2054 2008 only accept 2008-NaN binaries, if supported 2055 by the FPU 2056 relaxed accept any binaries regardless of whether 2057 supported by the FPU 2058 emulated accept any binaries but enable FPU emulator 2059 if binary mode is unsupported by the FPU. 2060 2061 The FPU emulator is always able to support both NaN 2062 encodings, so if no FPU hardware is present or it has 2063 been disabled with 'nofpu', then the settings of 2064 'legacy' and '2008' strap the emulator accordingly, 2065 'relaxed' straps the emulator for both legacy-NaN and 2066 2008-NaN, whereas 'strict' enables legacy-NaN only on 2067 legacy processors and both NaN encodings on MIPS32 or 2068 MIPS64 CPUs. 2069 2070 The setting for ABS.fmt/NEG.fmt instruction execution 2071 mode generally follows that for the NaN encoding, 2072 except where unsupported by hardware. 2073 2074 ignore_loglevel [KNL,EARLY] 2075 Ignore loglevel setting - this will print /all/ 2076 kernel messages to the console. Useful for debugging. 2077 We also add it as printk module parameter, so users 2078 could change it dynamically, usually by 2079 /sys/module/printk/parameters/ignore_loglevel. 2080 2081 ignore_rlimit_data 2082 Ignore RLIMIT_DATA setting for data mappings, 2083 print warning at first misuse. Can be changed via 2084 /sys/module/kernel/parameters/ignore_rlimit_data. 2085 2086 ihash_entries= [KNL] 2087 Set number of hash buckets for inode cache. 2088 2089 ima_appraise= [IMA] appraise integrity measurements 2090 Format: { "off" | "enforce" | "fix" | "log" } 2091 default: "enforce" 2092 2093 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 2094 The builtin appraise policy appraises all files 2095 owned by uid=0. 2096 2097 ima_canonical_fmt [IMA] 2098 Use the canonical format for the binary runtime 2099 measurements, instead of host native format. 2100 2101 ima_hash= [IMA] 2102 Format: { md5 | sha1 | rmd160 | sha256 | sha384 2103 | sha512 | ... } 2104 default: "sha1" 2105 2106 The list of supported hash algorithms is defined 2107 in crypto/hash_info.h. 2108 2109 ima_policy= [IMA] 2110 The builtin policies to load during IMA setup. 2111 Format: "tcb | appraise_tcb | secure_boot | 2112 fail_securely | critical_data" 2113 2114 The "tcb" policy measures all programs exec'd, files 2115 mmap'd for exec, and all files opened with the read 2116 mode bit set by either the effective uid (euid=0) or 2117 uid=0. 2118 2119 The "appraise_tcb" policy appraises the integrity of 2120 all files owned by root. 2121 2122 The "secure_boot" policy appraises the integrity 2123 of files (eg. kexec kernel image, kernel modules, 2124 firmware, policy, etc) based on file signatures. 2125 2126 The "fail_securely" policy forces file signature 2127 verification failure also on privileged mounted 2128 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 2129 flag. 2130 2131 The "critical_data" policy measures kernel integrity 2132 critical data. 2133 2134 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 2135 Load a policy which meets the needs of the Trusted 2136 Computing Base. This means IMA will measure all 2137 programs exec'd, files mmap'd for exec, and all files 2138 opened for read by uid=0. 2139 2140 ima_template= [IMA] 2141 Select one of defined IMA measurements template formats. 2142 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" | 2143 "ima-sigv2" } 2144 Default: "ima-ng" 2145 2146 ima_template_fmt= 2147 [IMA] Define a custom template format. 2148 Format: { "field1|...|fieldN" } 2149 2150 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage 2151 Format: <min_file_size> 2152 Set the minimal file size for using asynchronous hash. 2153 If left unspecified, ahash usage is disabled. 2154 2155 ahash performance varies for different data sizes on 2156 different crypto accelerators. This option can be used 2157 to achieve the best performance for a particular HW. 2158 2159 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size 2160 Format: <bufsize> 2161 Set hashing buffer size. Default: 4k. 2162 2163 ahash performance varies for different chunk sizes on 2164 different crypto accelerators. This option can be used 2165 to achieve best performance for particular HW. 2166 2167 indirect_target_selection= [X86,Intel] Mitigation control for Indirect 2168 Target Selection(ITS) bug in Intel CPUs. Updated 2169 microcode is also required for a fix in IBPB. 2170 2171 on: Enable mitigation (default). 2172 off: Disable mitigation. 2173 force: Force the ITS bug and deploy default 2174 mitigation. 2175 vmexit: Only deploy mitigation if CPU is affected by 2176 guest/host isolation part of ITS. 2177 stuff: Deploy RSB-fill mitigation when retpoline is 2178 also deployed. Otherwise, deploy the default 2179 mitigation. 2180 2181 For details see: 2182 Documentation/admin-guide/hw-vuln/indirect-target-selection.rst 2183 2184 init= [KNL] 2185 Format: <full_path> 2186 Run specified binary instead of /sbin/init as init 2187 process. 2188 2189 initcall_debug [KNL] Trace initcalls as they are executed. Useful 2190 for working out where the kernel is dying during 2191 startup. 2192 2193 initcall_blacklist= [KNL] Do not execute a comma-separated list of 2194 initcall functions. Useful for debugging built-in 2195 modules and initcalls. 2196 2197 initramfs_async= [KNL] 2198 Format: <bool> 2199 Default: 1 2200 This parameter controls whether the initramfs 2201 image is unpacked asynchronously, concurrently 2202 with devices being probed and 2203 initialized. This should normally just work, 2204 but as a debugging aid, one can get the 2205 historical behaviour of the initramfs 2206 unpacking being completed before device_ and 2207 late_ initcalls. 2208 2209 initrd= [BOOT,EARLY] Specify the location of the initial ramdisk 2210 2211 initrdmem= [KNL,EARLY] Specify a physical address and size from which to 2212 load the initrd. If an initrd is compiled in or 2213 specified in the bootparams, it takes priority over this 2214 setting. 2215 Format: ss[KMG],nn[KMG] 2216 Default is 0, 0 2217 2218 init_on_alloc= [MM,EARLY] Fill newly allocated pages and heap objects with 2219 zeroes. 2220 Format: 0 | 1 2221 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 2222 2223 init_on_free= [MM,EARLY] Fill freed pages and heap objects with zeroes. 2224 Format: 0 | 1 2225 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 2226 2227 init_pkru= [X86] Specify the default memory protection keys rights 2228 register contents for all processes. 0x55555554 by 2229 default (disallow access to all but pkey 0). Can 2230 override in debugfs after boot. 2231 2232 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 2233 Format: <irq> 2234 2235 int_pln_enable [X86] Enable power limit notification interrupt 2236 2237 integrity_audit=[IMA] 2238 Format: { "0" | "1" } 2239 0 -- basic integrity auditing messages. (Default) 2240 1 -- additional integrity auditing messages. 2241 2242 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 2243 on 2244 Enable intel iommu driver. 2245 off 2246 Disable intel iommu driver. 2247 igfx_off [Default Off] 2248 By default, gfx is mapped as normal device. If a gfx 2249 device has a dedicated DMAR unit, the DMAR unit is 2250 bypassed by not enabling DMAR with this option. In 2251 this case, gfx device will use physical address for 2252 DMA. 2253 strict [Default Off] 2254 Deprecated, equivalent to iommu.strict=1. 2255 sp_off [Default Off] 2256 By default, super page will be supported if Intel IOMMU 2257 has the capability. With this option, super page will 2258 not be supported. 2259 sm_on 2260 Enable the Intel IOMMU scalable mode if the hardware 2261 advertises that it has support for the scalable mode 2262 translation. 2263 sm_off 2264 Disallow use of the Intel IOMMU scalable mode. 2265 tboot_noforce [Default Off] 2266 Do not force the Intel IOMMU enabled under tboot. 2267 By default, tboot will force Intel IOMMU on, which 2268 could harm performance of some high-throughput 2269 devices like 40GBit network cards, even if identity 2270 mapping is enabled. 2271 Note that using this option lowers the security 2272 provided by tboot because it makes the system 2273 vulnerable to DMA attacks. 2274 2275 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 2276 0 disables intel_idle and fall back on acpi_idle. 2277 1 to 9 specify maximum depth of C-state. 2278 2279 intel_pstate= [X86,EARLY] 2280 disable 2281 Do not enable intel_pstate as the default 2282 scaling driver for the supported processors 2283 active 2284 Use intel_pstate driver to bypass the scaling 2285 governors layer of cpufreq and provides it own 2286 algorithms for p-state selection. There are two 2287 P-state selection algorithms provided by 2288 intel_pstate in the active mode: powersave and 2289 performance. The way they both operate depends 2290 on whether or not the hardware managed P-states 2291 (HWP) feature has been enabled in the processor 2292 and possibly on the processor model. 2293 passive 2294 Use intel_pstate as a scaling driver, but configure it 2295 to work with generic cpufreq governors (instead of 2296 enabling its internal governor). This mode cannot be 2297 used along with the hardware-managed P-states (HWP) 2298 feature. 2299 force 2300 Enable intel_pstate on systems that prohibit it by default 2301 in favor of acpi-cpufreq. Forcing the intel_pstate driver 2302 instead of acpi-cpufreq may disable platform features, such 2303 as thermal controls and power capping, that rely on ACPI 2304 P-States information being indicated to OSPM and therefore 2305 should be used with caution. This option does not work with 2306 processors that aren't supported by the intel_pstate driver 2307 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 2308 no_hwp 2309 Do not enable hardware P state control (HWP) 2310 if available. 2311 hwp_only 2312 Only load intel_pstate on systems which support 2313 hardware P state control (HWP) if available. 2314 support_acpi_ppc 2315 Enforce ACPI _PPC performance limits. If the Fixed ACPI 2316 Description Table, specifies preferred power management 2317 profile as "Enterprise Server" or "Performance Server", 2318 then this feature is turned on by default. 2319 per_cpu_perf_limits 2320 Allow per-logical-CPU P-State performance control limits using 2321 cpufreq sysfs interface 2322 2323 intremap= [X86-64,Intel-IOMMU,EARLY] 2324 on enable Interrupt Remapping (default) 2325 off disable Interrupt Remapping 2326 nosid disable Source ID checking 2327 no_x2apic_optout 2328 BIOS x2APIC opt-out request will be ignored 2329 nopost disable Interrupt Posting 2330 posted_msi 2331 enable MSIs delivered as posted interrupts 2332 2333 iomem= Disable strict checking of access to MMIO memory 2334 strict regions from userspace. 2335 relaxed 2336 2337 iommu= [X86,EARLY] 2338 off 2339 force 2340 noforce 2341 biomerge 2342 panic 2343 nopanic 2344 merge 2345 nomerge 2346 soft 2347 pt [X86] 2348 nopt [X86] 2349 nobypass [PPC/POWERNV] 2350 Disable IOMMU bypass, using IOMMU for PCI devices. 2351 2352 iommu.forcedac= [ARM64,X86,EARLY] Control IOVA allocation for PCI devices. 2353 Format: { "0" | "1" } 2354 0 - Try to allocate a 32-bit DMA address first, before 2355 falling back to the full range if needed. 2356 1 - Allocate directly from the full usable range, 2357 forcing Dual Address Cycle for PCI cards supporting 2358 greater than 32-bit addressing. 2359 2360 iommu.strict= [ARM64,X86,S390,EARLY] Configure TLB invalidation behaviour 2361 Format: { "0" | "1" } 2362 0 - Lazy mode. 2363 Request that DMA unmap operations use deferred 2364 invalidation of hardware TLBs, for increased 2365 throughput at the cost of reduced device isolation. 2366 Will fall back to strict mode if not supported by 2367 the relevant IOMMU driver. 2368 1 - Strict mode. 2369 DMA unmap operations invalidate IOMMU hardware TLBs 2370 synchronously. 2371 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}. 2372 Note: on x86, strict mode specified via one of the 2373 legacy driver-specific options takes precedence. 2374 2375 iommu.passthrough= 2376 [ARM64,X86,EARLY] Configure DMA to bypass the IOMMU by default. 2377 Format: { "0" | "1" } 2378 0 - Use IOMMU translation for DMA. 2379 1 - Bypass the IOMMU for DMA. 2380 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 2381 2382 io7= [HW] IO7 for Marvel-based Alpha systems 2383 See comment before marvel_specify_io7 in 2384 arch/alpha/kernel/core_marvel.c. 2385 2386 io_delay= [X86,EARLY] I/O delay method 2387 0x80 2388 Standard port 0x80 based delay 2389 0xed 2390 Alternate port 0xed based delay (needed on some systems) 2391 udelay 2392 Simple two microseconds delay 2393 none 2394 No delay 2395 2396 ip= [IP_PNP] 2397 See Documentation/admin-guide/nfs/nfsroot.rst. 2398 2399 ipcmni_extend [KNL,EARLY] Extend the maximum number of unique System V 2400 IPC identifiers from 32,768 to 16,777,216. 2401 2402 ipe.enforce= [IPE] 2403 Format: <bool> 2404 Determine whether IPE starts in permissive (0) or 2405 enforce (1) mode. The default is enforce. 2406 2407 ipe.success_audit= 2408 [IPE] 2409 Format: <bool> 2410 Start IPE with success auditing enabled, emitting 2411 an audit event when a binary is allowed. The default 2412 is 0. 2413 2414 irqaffinity= [SMP] Set the default irq affinity mask 2415 The argument is a cpu list, as described above. 2416 2417 irqchip.gicv2_force_probe= 2418 [ARM,ARM64,EARLY] 2419 Format: <bool> 2420 Force the kernel to look for the second 4kB page 2421 of a GICv2 controller even if the memory range 2422 exposed by the device tree is too small. 2423 2424 irqchip.gicv3_nolpi= 2425 [ARM,ARM64,EARLY] 2426 Force the kernel to ignore the availability of 2427 LPIs (and by consequence ITSs). Intended for system 2428 that use the kernel as a bootloader, and thus want 2429 to let secondary kernels in charge of setting up 2430 LPIs. 2431 2432 irqchip.gicv3_pseudo_nmi= [ARM64,EARLY] 2433 Enables support for pseudo-NMIs in the kernel. This 2434 requires the kernel to be built with 2435 CONFIG_ARM64_PSEUDO_NMI. 2436 2437 irqfixup [HW] 2438 When an interrupt is not handled search all handlers 2439 for it. Intended to get systems with badly broken 2440 firmware running. 2441 2442 irqpoll [HW] 2443 When an interrupt is not handled search all handlers 2444 for it. Also check all handlers each timer 2445 interrupt. Intended to get systems with badly broken 2446 firmware running. 2447 2448 isapnp= [ISAPNP] 2449 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2450 2451 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2452 [Deprecated - use cpusets instead] 2453 Format: [flag-list,]<cpu-list> 2454 2455 Specify one or more CPUs to isolate from disturbances 2456 specified in the flag list (default: domain): 2457 2458 nohz 2459 Disable the tick when a single task runs. 2460 2461 A residual 1Hz tick is offloaded to workqueues, which you 2462 need to affine to housekeeping through the global 2463 workqueue's affinity configured via the 2464 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2465 by using the 'domain' flag described below. 2466 2467 NOTE: by default the global workqueue runs on all CPUs, 2468 so to protect individual CPUs the 'cpumask' file has to 2469 be configured manually after bootup. 2470 2471 domain 2472 Isolate from the general SMP balancing and scheduling 2473 algorithms. Note that performing domain isolation this way 2474 is irreversible: it's not possible to bring back a CPU to 2475 the domains once isolated through isolcpus. It's strongly 2476 advised to use cpusets instead to disable scheduler load 2477 balancing through the "cpuset.sched_load_balance" file. 2478 It offers a much more flexible interface where CPUs can 2479 move in and out of an isolated set anytime. 2480 2481 You can move a process onto or off an "isolated" CPU via 2482 the CPU affinity syscalls or cpuset. 2483 <cpu number> begins at 0 and the maximum value is 2484 "number of CPUs in system - 1". 2485 2486 managed_irq 2487 2488 Isolate from being targeted by managed interrupts 2489 which have an interrupt mask containing isolated 2490 CPUs. The affinity of managed interrupts is 2491 handled by the kernel and cannot be changed via 2492 the /proc/irq/* interfaces. 2493 2494 This isolation is best effort and only effective 2495 if the automatically assigned interrupt mask of a 2496 device queue contains isolated and housekeeping 2497 CPUs. If housekeeping CPUs are online then such 2498 interrupts are directed to the housekeeping CPU 2499 so that IO submitted on the housekeeping CPU 2500 cannot disturb the isolated CPU. 2501 2502 If a queue's affinity mask contains only isolated 2503 CPUs then this parameter has no effect on the 2504 interrupt routing decision, though interrupts are 2505 only delivered when tasks running on those 2506 isolated CPUs submit IO. IO submitted on 2507 housekeeping CPUs has no influence on those 2508 queues. 2509 2510 The format of <cpu-list> is described above. 2511 2512 iucv= [HW,NET] 2513 2514 ivrs_ioapic [HW,X86-64] 2515 Provide an override to the IOAPIC-ID<->DEVICE-ID 2516 mapping provided in the IVRS ACPI table. 2517 By default, PCI segment is 0, and can be omitted. 2518 2519 For example, to map IOAPIC-ID decimal 10 to 2520 PCI segment 0x1 and PCI device 00:14.0, 2521 write the parameter as: 2522 ivrs_ioapic=10@0001:00:14.0 2523 2524 Deprecated formats: 2525 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0 2526 write the parameter as: 2527 ivrs_ioapic[10]=00:14.0 2528 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and 2529 PCI device 00:14.0 write the parameter as: 2530 ivrs_ioapic[10]=0001:00:14.0 2531 2532 ivrs_hpet [HW,X86-64] 2533 Provide an override to the HPET-ID<->DEVICE-ID 2534 mapping provided in the IVRS ACPI table. 2535 By default, PCI segment is 0, and can be omitted. 2536 2537 For example, to map HPET-ID decimal 10 to 2538 PCI segment 0x1 and PCI device 00:14.0, 2539 write the parameter as: 2540 ivrs_hpet=10@0001:00:14.0 2541 2542 Deprecated formats: 2543 * To map HPET-ID decimal 0 to PCI device 00:14.0 2544 write the parameter as: 2545 ivrs_hpet[0]=00:14.0 2546 * To map HPET-ID decimal 10 to PCI segment 0x1 and 2547 PCI device 00:14.0 write the parameter as: 2548 ivrs_ioapic[10]=0001:00:14.0 2549 2550 ivrs_acpihid [HW,X86-64] 2551 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2552 mapping provided in the IVRS ACPI table. 2553 By default, PCI segment is 0, and can be omitted. 2554 2555 For example, to map UART-HID:UID AMD0020:0 to 2556 PCI segment 0x1 and PCI device ID 00:14.5, 2557 write the parameter as: 2558 ivrs_acpihid=AMD0020:0@0001:00:14.5 2559 2560 Deprecated formats: 2561 * To map UART-HID:UID AMD0020:0 to PCI segment is 0, 2562 PCI device ID 00:14.5, write the parameter as: 2563 ivrs_acpihid[00:14.5]=AMD0020:0 2564 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and 2565 PCI device ID 00:14.5, write the parameter as: 2566 ivrs_acpihid[0001:00:14.5]=AMD0020:0 2567 2568 js= [HW,JOY] Analog joystick 2569 See Documentation/input/joydev/joystick.rst. 2570 2571 kasan_multi_shot 2572 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2573 report on every invalid memory access. Without this 2574 parameter KASAN will print report only for the first 2575 invalid access. 2576 2577 keep_bootcon [KNL,EARLY] 2578 Do not unregister boot console at start. This is only 2579 useful for debugging when something happens in the window 2580 between unregistering the boot console and initializing 2581 the real console. 2582 2583 keepinitrd [HW,ARM] See retain_initrd. 2584 2585 kernelcore= [KNL,X86,PPC,EARLY] 2586 Format: nn[KMGTPE] | nn% | "mirror" 2587 This parameter specifies the amount of memory usable by 2588 the kernel for non-movable allocations. The requested 2589 amount is spread evenly throughout all nodes in the 2590 system as ZONE_NORMAL. The remaining memory is used for 2591 movable memory in its own zone, ZONE_MOVABLE. In the 2592 event, a node is too small to have both ZONE_NORMAL and 2593 ZONE_MOVABLE, kernelcore memory will take priority and 2594 other nodes will have a larger ZONE_MOVABLE. 2595 2596 ZONE_MOVABLE is used for the allocation of pages that 2597 may be reclaimed or moved by the page migration 2598 subsystem. Note that allocations like PTEs-from-HighMem 2599 still use the HighMem zone if it exists, and the Normal 2600 zone if it does not. 2601 2602 It is possible to specify the exact amount of memory in 2603 the form of "nn[KMGTPE]", a percentage of total system 2604 memory in the form of "nn%", or "mirror". If "mirror" 2605 option is specified, mirrored (reliable) memory is used 2606 for non-movable allocations and remaining memory is used 2607 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2608 are exclusive, so you cannot specify multiple forms. 2609 2610 kgdbdbgp= [KGDB,HW,EARLY] kgdb over EHCI usb debug port. 2611 Format: <Controller#>[,poll interval] 2612 The controller # is the number of the ehci usb debug 2613 port as it is probed via PCI. The poll interval is 2614 optional and is the number seconds in between 2615 each poll cycle to the debug port in case you need 2616 the functionality for interrupting the kernel with 2617 gdb or control-c on the dbgp connection. When 2618 not using this parameter you use sysrq-g to break into 2619 the kernel debugger. 2620 2621 kgdboc= [KGDB,HW] kgdb over consoles. 2622 Requires a tty driver that supports console polling, 2623 or a supported polling keyboard driver (non-usb). 2624 Serial only format: <serial_device>[,baud] 2625 keyboard only format: kbd 2626 keyboard and serial format: kbd,<serial_device>[,baud] 2627 Optional Kernel mode setting: 2628 kms, kbd format: kms,kbd 2629 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 2630 2631 kgdboc_earlycon= [KGDB,HW,EARLY] 2632 If the boot console provides the ability to read 2633 characters and can work in polling mode, you can use 2634 this parameter to tell kgdb to use it as a backend 2635 until the normal console is registered. Intended to 2636 be used together with the kgdboc parameter which 2637 specifies the normal console to transition to. 2638 2639 The name of the early console should be specified 2640 as the value of this parameter. Note that the name of 2641 the early console might be different than the tty 2642 name passed to kgdboc. It's OK to leave the value 2643 blank and the first boot console that implements 2644 read() will be picked. 2645 2646 kgdbwait [KGDB,EARLY] Stop kernel execution and enter the 2647 kernel debugger at the earliest opportunity. 2648 2649 kmac= [MIPS] Korina ethernet MAC address. 2650 Configure the RouterBoard 532 series on-chip 2651 Ethernet adapter MAC address. 2652 2653 kmemleak= [KNL,EARLY] Boot-time kmemleak enable/disable 2654 Valid arguments: on, off 2655 Default: on 2656 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 2657 the default is off. 2658 2659 kprobe_event=[probe-list] 2660 [FTRACE] Add kprobe events and enable at boot time. 2661 The probe-list is a semicolon delimited list of probe 2662 definitions. Each definition is same as kprobe_events 2663 interface, but the parameters are comma delimited. 2664 For example, to add a kprobe event on vfs_read with 2665 arg1 and arg2, add to the command line; 2666 2667 kprobe_event=p,vfs_read,$arg1,$arg2 2668 2669 See also Documentation/trace/kprobetrace.rst "Kernel 2670 Boot Parameter" section. 2671 2672 kpti= [ARM64,EARLY] Control page table isolation of 2673 user and kernel address spaces. 2674 Default: enabled on cores which need mitigation. 2675 0: force disabled 2676 1: force enabled 2677 2678 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires 2679 CONFIG_KUNIT to be set to be fully enabled. The 2680 default value can be overridden via 2681 KUNIT_DEFAULT_ENABLED. 2682 Default is 0 (disabled) 2683 2684 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 2685 Default is 0 (don't ignore, but inject #GP) 2686 2687 kvm.eager_page_split= 2688 [KVM,X86] Controls whether or not KVM will try to 2689 proactively split all huge pages during dirty logging. 2690 Eager page splitting reduces interruptions to vCPU 2691 execution by eliminating the write-protection faults 2692 and MMU lock contention that would otherwise be 2693 required to split huge pages lazily. 2694 2695 VM workloads that rarely perform writes or that write 2696 only to a small region of VM memory may benefit from 2697 disabling eager page splitting to allow huge pages to 2698 still be used for reads. 2699 2700 The behavior of eager page splitting depends on whether 2701 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If 2702 disabled, all huge pages in a memslot will be eagerly 2703 split when dirty logging is enabled on that memslot. If 2704 enabled, eager page splitting will be performed during 2705 the KVM_CLEAR_DIRTY ioctl, and only for the pages being 2706 cleared. 2707 2708 Eager page splitting is only supported when kvm.tdp_mmu=Y. 2709 2710 Default is Y (on). 2711 2712 kvm.enable_virt_at_load=[KVM,ARM64,LOONGARCH,MIPS,RISCV,X86] 2713 If enabled, KVM will enable virtualization in hardware 2714 when KVM is loaded, and disable virtualization when KVM 2715 is unloaded (if KVM is built as a module). 2716 2717 If disabled, KVM will dynamically enable and disable 2718 virtualization on-demand when creating and destroying 2719 VMs, i.e. on the 0=>1 and 1=>0 transitions of the 2720 number of VMs. 2721 2722 Enabling virtualization at module lode avoids potential 2723 latency for creation of the 0=>1 VM, as KVM serializes 2724 virtualization enabling across all online CPUs. The 2725 "cost" of enabling virtualization when KVM is loaded, 2726 is that doing so may interfere with using out-of-tree 2727 hypervisors that want to "own" virtualization hardware. 2728 2729 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 2730 Default is false (don't support). 2731 2732 kvm.nx_huge_pages= 2733 [KVM] Controls the software workaround for the 2734 X86_BUG_ITLB_MULTIHIT bug. 2735 force : Always deploy workaround. 2736 off : Never deploy workaround. 2737 auto : Deploy workaround based on the presence of 2738 X86_BUG_ITLB_MULTIHIT. 2739 2740 Default is 'auto'. 2741 2742 If the software workaround is enabled for the host, 2743 guests do need not to enable it for nested guests. 2744 2745 kvm.nx_huge_pages_recovery_ratio= 2746 [KVM] Controls how many 4KiB pages are periodically zapped 2747 back to huge pages. 0 disables the recovery, otherwise if 2748 the value is N KVM will zap 1/Nth of the 4KiB pages every 2749 period (see below). The default is 60. 2750 2751 kvm.nx_huge_pages_recovery_period_ms= 2752 [KVM] Controls the time period at which KVM zaps 4KiB pages 2753 back to huge pages. If the value is a non-zero N, KVM will 2754 zap a portion (see ratio above) of the pages every N msecs. 2755 If the value is 0 (the default), KVM will pick a period based 2756 on the ratio, such that a page is zapped after 1 hour on average. 2757 2758 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in 2759 KVM/SVM. Default is 1 (enabled). 2760 2761 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables, 2762 a.k.a. Two-Dimensional Page Tables. Default is 1 2763 (enabled). Disable by KVM if hardware lacks support 2764 for NPT. 2765 2766 kvm-arm.hyp_lm_size_mb= 2767 [KVM,ARM,EARLY] Maximum amount of contiguous memory mappable in 2768 the pKVM hypervisor linear map, in MB. Any attempt to map more 2769 memory than this into pKVM stage-1 at run-time may be fatal. 2770 2771 kvm-arm.mode= 2772 [KVM,ARM,EARLY] Select one of KVM/arm64's modes of 2773 operation. 2774 2775 none: Forcefully disable KVM. 2776 2777 nvhe: Standard nVHE-based mode, without support for 2778 protected guests. 2779 2780 protected: nVHE-based mode with support for guests whose 2781 state is kept private from the host. See 2782 Documentation/virt/kvm/arm/pkvm.rst for more 2783 information about this mode of operation. 2784 2785 nested: VHE-based mode with support for nested 2786 virtualization. Requires at least ARMv8.3 2787 hardware. 2788 2789 Defaults to VHE/nVHE based on hardware support. Setting 2790 mode to "protected" will disable kexec and hibernation 2791 for the host. "nested" is experimental and should be 2792 used with extreme caution. 2793 2794 kvm-arm.protected_modules= 2795 [KVM,ARM] List of pKVM modules to load before the host 2796 is deprevileged. 2797 2798 This option only applies when booting with 2799 kvm-arm.mode=protected. 2800 2801 kvm-arm.protected_prefault= 2802 [KVM,ARM] Minimum order for each protected VM fault. 2803 This range from 0 (default) to 9. 2804 2805 kvm-arm.vgic_v3_group0_trap= 2806 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0 2807 system registers 2808 2809 kvm-arm.vgic_v3_group1_trap= 2810 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1 2811 system registers 2812 2813 kvm-arm.vgic_v3_common_trap= 2814 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common 2815 system registers 2816 2817 kvm-arm.vgic_v4_enable= 2818 [KVM,ARM,EARLY] Allow use of GICv4 for direct 2819 injection of LPIs. 2820 2821 kvm-arm.wfe_trap_policy= 2822 [KVM,ARM] Control when to set WFE instruction trap for 2823 KVM VMs. Traps are allowed but not guaranteed by the 2824 CPU architecture. 2825 2826 trap: set WFE instruction trap 2827 2828 notrap: clear WFE instruction trap 2829 2830 kvm-arm.wfi_trap_policy= 2831 [KVM,ARM] Control when to set WFI instruction trap for 2832 KVM VMs. Traps are allowed but not guaranteed by the 2833 CPU architecture. 2834 2835 trap: set WFI instruction trap 2836 2837 notrap: clear WFI instruction trap 2838 2839 kvm_cma_resv_ratio=n [PPC,EARLY] 2840 Reserves given percentage from system memory area for 2841 contiguous memory allocation for KVM hash pagetable 2842 allocation. 2843 By default it reserves 5% of total system memory. 2844 Format: <integer> 2845 Default: 5 2846 2847 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables, 2848 a.k.a. Two-Dimensional Page Tables. Default is 1 2849 (enabled). Disable by KVM if hardware lacks support 2850 for EPT. 2851 2852 kvm-intel.emulate_invalid_guest_state= 2853 [KVM,Intel] Control whether to emulate invalid guest 2854 state. Ignored if kvm-intel.enable_unrestricted_guest=1, 2855 as guest state is never invalid for unrestricted 2856 guests. This param doesn't apply to nested guests (L2), 2857 as KVM never emulates invalid L2 guest state. 2858 Default is 1 (enabled). 2859 2860 kvm-intel.flexpriority= 2861 [KVM,Intel] Control KVM's use of FlexPriority feature 2862 (TPR shadow). Default is 1 (enabled). Disable by KVM if 2863 hardware lacks support for it. 2864 2865 kvm-intel.nested= 2866 [KVM,Intel] Control nested virtualization feature in 2867 KVM/VMX. Default is 1 (enabled). 2868 2869 kvm-intel.unrestricted_guest= 2870 [KVM,Intel] Control KVM's use of unrestricted guest 2871 feature (virtualized real and unpaged mode). Default 2872 is 1 (enabled). Disable by KVM if EPT is disabled or 2873 hardware lacks support for it. 2874 2875 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 2876 CVE-2018-3620. 2877 2878 Valid arguments: never, cond, always 2879 2880 always: L1D cache flush on every VMENTER. 2881 cond: Flush L1D on VMENTER only when the code between 2882 VMEXIT and VMENTER can leak host memory. 2883 never: Disables the mitigation 2884 2885 Default is cond (do L1 cache flush in specific instances) 2886 2887 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor 2888 Identification feature (tagged TLBs). Default is 1 2889 (enabled). Disable by KVM if hardware lacks support 2890 for it. 2891 2892 l1d_flush= [X86,INTEL,EARLY] 2893 Control mitigation for L1D based snooping vulnerability. 2894 2895 Certain CPUs are vulnerable to an exploit against CPU 2896 internal buffers which can forward information to a 2897 disclosure gadget under certain conditions. 2898 2899 In vulnerable processors, the speculatively 2900 forwarded data can be used in a cache side channel 2901 attack, to access data to which the attacker does 2902 not have direct access. 2903 2904 This parameter controls the mitigation. The 2905 options are: 2906 2907 on - enable the interface for the mitigation 2908 2909 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on 2910 affected CPUs 2911 2912 The kernel PTE inversion protection is unconditionally 2913 enabled and cannot be disabled. 2914 2915 full 2916 Provides all available mitigations for the 2917 L1TF vulnerability. Disables SMT and 2918 enables all mitigations in the 2919 hypervisors, i.e. unconditional L1D flush. 2920 2921 SMT control and L1D flush control via the 2922 sysfs interface is still possible after 2923 boot. Hypervisors will issue a warning 2924 when the first VM is started in a 2925 potentially insecure configuration, 2926 i.e. SMT enabled or L1D flush disabled. 2927 2928 full,force 2929 Same as 'full', but disables SMT and L1D 2930 flush runtime control. Implies the 2931 'nosmt=force' command line option. 2932 (i.e. sysfs control of SMT is disabled.) 2933 2934 flush 2935 Leaves SMT enabled and enables the default 2936 hypervisor mitigation, i.e. conditional 2937 L1D flush. 2938 2939 SMT control and L1D flush control via the 2940 sysfs interface is still possible after 2941 boot. Hypervisors will issue a warning 2942 when the first VM is started in a 2943 potentially insecure configuration, 2944 i.e. SMT enabled or L1D flush disabled. 2945 2946 flush,nosmt 2947 2948 Disables SMT and enables the default 2949 hypervisor mitigation. 2950 2951 SMT control and L1D flush control via the 2952 sysfs interface is still possible after 2953 boot. Hypervisors will issue a warning 2954 when the first VM is started in a 2955 potentially insecure configuration, 2956 i.e. SMT enabled or L1D flush disabled. 2957 2958 flush,nowarn 2959 Same as 'flush', but hypervisors will not 2960 warn when a VM is started in a potentially 2961 insecure configuration. 2962 2963 off 2964 Disables hypervisor mitigations and doesn't 2965 emit any warnings. 2966 It also drops the swap size and available 2967 RAM limit restriction on both hypervisor and 2968 bare metal. 2969 2970 Default is 'flush'. 2971 2972 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 2973 2974 l2cr= [PPC] 2975 2976 l3cr= [PPC] 2977 2978 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS 2979 disabled it. 2980 2981 lapic= [X86,APIC] Do not use TSC deadline 2982 value for LAPIC timer one-shot implementation. Default 2983 back to the programmable timer unit in the LAPIC. 2984 Format: notscdeadline 2985 2986 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer 2987 in C2 power state. 2988 2989 libata.dma= [LIBATA] DMA control 2990 libata.dma=0 Disable all PATA and SATA DMA 2991 libata.dma=1 PATA and SATA Disk DMA only 2992 libata.dma=2 ATAPI (CDROM) DMA only 2993 libata.dma=4 Compact Flash DMA only 2994 Combinations also work, so libata.dma=3 enables DMA 2995 for disks and CDROMs, but not CFs. 2996 2997 libata.ignore_hpa= [LIBATA] Ignore HPA limit 2998 libata.ignore_hpa=0 keep BIOS limits (default) 2999 libata.ignore_hpa=1 ignore limits, using full disk 3000 3001 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 3002 when set. 3003 Format: <int> 3004 3005 libata.force= [LIBATA] Force configurations. The format is a comma- 3006 separated list of "[ID:]VAL" where ID is PORT[.DEVICE]. 3007 PORT and DEVICE are decimal numbers matching port, link 3008 or device. Basically, it matches the ATA ID string 3009 printed on console by libata. If the whole ID part is 3010 omitted, the last PORT and DEVICE values are used. If 3011 ID hasn't been specified yet, the configuration applies 3012 to all ports, links and devices. 3013 3014 If only DEVICE is omitted, the parameter applies to 3015 the port and all links and devices behind it. DEVICE 3016 number of 0 either selects the first device or the 3017 first fan-out link behind PMP device. It does not 3018 select the host link. DEVICE number of 15 selects the 3019 host link and device attached to it. 3020 3021 The VAL specifies the configuration to force. As long 3022 as there is no ambiguity, shortcut notation is allowed. 3023 For example, both 1.5 and 1.5G would work for 1.5Gbps. 3024 The following configurations can be forced. 3025 3026 * Cable type: 40c, 80c, short40c, unk, ign or sata. 3027 Any ID with matching PORT is used. 3028 3029 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 3030 3031 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 3032 udma[/][16,25,33,44,66,100,133] notation is also 3033 allowed. 3034 3035 * nohrst, nosrst, norst: suppress hard, soft and both 3036 resets. 3037 3038 * rstonce: only attempt one reset during hot-unplug 3039 link recovery. 3040 3041 * [no]dbdelay: Enable or disable the extra 200ms delay 3042 before debouncing a link PHY and device presence 3043 detection. 3044 3045 * [no]ncq: Turn on or off NCQ. 3046 3047 * [no]ncqtrim: Enable or disable queued DSM TRIM. 3048 3049 * [no]ncqati: Enable or disable NCQ trim on ATI chipset. 3050 3051 * [no]trim: Enable or disable (unqueued) TRIM. 3052 3053 * trim_zero: Indicate that TRIM command zeroes data. 3054 3055 * max_trim_128m: Set 128M maximum trim size limit. 3056 3057 * [no]dma: Turn on or off DMA transfers. 3058 3059 * atapi_dmadir: Enable ATAPI DMADIR bridge support. 3060 3061 * atapi_mod16_dma: Enable the use of ATAPI DMA for 3062 commands that are not a multiple of 16 bytes. 3063 3064 * [no]dmalog: Enable or disable the use of the 3065 READ LOG DMA EXT command to access logs. 3066 3067 * [no]iddevlog: Enable or disable access to the 3068 identify device data log. 3069 3070 * [no]logdir: Enable or disable access to the general 3071 purpose log directory. 3072 3073 * max_sec_128: Set transfer size limit to 128 sectors. 3074 3075 * max_sec_1024: Set or clear transfer size limit to 3076 1024 sectors. 3077 3078 * max_sec_lba48: Set or clear transfer size limit to 3079 65535 sectors. 3080 3081 * external: Mark port as external (hotplug-capable). 3082 3083 * [no]lpm: Enable or disable link power management. 3084 3085 * [no]setxfer: Indicate if transfer speed mode setting 3086 should be skipped. 3087 3088 * [no]fua: Disable or enable FUA (Force Unit Access) 3089 support for devices supporting this feature. 3090 3091 * dump_id: Dump IDENTIFY data. 3092 3093 * disable: Disable this device. 3094 3095 If there are multiple matching configurations changing 3096 the same attribute, the last one is used. 3097 3098 load_ramdisk= [RAM] [Deprecated] 3099 3100 lockd.nlm_grace_period=P [NFS] Assign grace period. 3101 Format: <integer> 3102 3103 lockd.nlm_tcpport=N [NFS] Assign TCP port. 3104 Format: <integer> 3105 3106 lockd.nlm_timeout=T [NFS] Assign timeout value. 3107 Format: <integer> 3108 3109 lockd.nlm_udpport=M [NFS] Assign UDP port. 3110 Format: <integer> 3111 3112 lockdown= [SECURITY,EARLY] 3113 { integrity | confidentiality } 3114 Enable the kernel lockdown feature. If set to 3115 integrity, kernel features that allow userland to 3116 modify the running kernel are disabled. If set to 3117 confidentiality, kernel features that allow userland 3118 to extract confidential information from the kernel 3119 are also disabled. 3120 3121 locktorture.acq_writer_lim= [KNL] 3122 Set the time limit in jiffies for a lock 3123 acquisition. Acquisitions exceeding this limit 3124 will result in a splat once they do complete. 3125 3126 locktorture.bind_readers= [KNL] 3127 Specify the list of CPUs to which the readers are 3128 to be bound. 3129 3130 locktorture.bind_writers= [KNL] 3131 Specify the list of CPUs to which the writers are 3132 to be bound. 3133 3134 locktorture.call_rcu_chains= [KNL] 3135 Specify the number of self-propagating call_rcu() 3136 chains to set up. These are used to ensure that 3137 there is a high probability of an RCU grace period 3138 in progress at any given time. Defaults to 0, 3139 which disables these call_rcu() chains. 3140 3141 locktorture.long_hold= [KNL] 3142 Specify the duration in milliseconds for the 3143 occasional long-duration lock hold time. Defaults 3144 to 100 milliseconds. Select 0 to disable. 3145 3146 locktorture.nested_locks= [KNL] 3147 Specify the maximum lock nesting depth that 3148 locktorture is to exercise, up to a limit of 8 3149 (MAX_NESTED_LOCKS). Specify zero to disable. 3150 Note that this parameter is ineffective on types 3151 of locks that do not support nested acquisition. 3152 3153 locktorture.nreaders_stress= [KNL] 3154 Set the number of locking read-acquisition kthreads. 3155 Defaults to being automatically set based on the 3156 number of online CPUs. 3157 3158 locktorture.nwriters_stress= [KNL] 3159 Set the number of locking write-acquisition kthreads. 3160 3161 locktorture.onoff_holdoff= [KNL] 3162 Set time (s) after boot for CPU-hotplug testing. 3163 3164 locktorture.onoff_interval= [KNL] 3165 Set time (s) between CPU-hotplug operations, or 3166 zero to disable CPU-hotplug testing. 3167 3168 locktorture.rt_boost= [KNL] 3169 Do periodic testing of real-time lock priority 3170 boosting. Select 0 to disable, 1 to boost 3171 only rt_mutex, and 2 to boost unconditionally. 3172 Defaults to 2, which might seem to be an 3173 odd choice, but which should be harmless for 3174 non-real-time spinlocks, due to their disabling 3175 of preemption. Note that non-realtime mutexes 3176 disable boosting. 3177 3178 locktorture.rt_boost_factor= [KNL] 3179 Number that determines how often and for how 3180 long priority boosting is exercised. This is 3181 scaled down by the number of writers, so that the 3182 number of boosts per unit time remains roughly 3183 constant as the number of writers increases. 3184 On the other hand, the duration of each boost 3185 increases with the number of writers. 3186 3187 locktorture.shuffle_interval= [KNL] 3188 Set task-shuffle interval (jiffies). Shuffling 3189 tasks allows some CPUs to go into dyntick-idle 3190 mode during the locktorture test. 3191 3192 locktorture.shutdown_secs= [KNL] 3193 Set time (s) after boot system shutdown. This 3194 is useful for hands-off automated testing. 3195 3196 locktorture.stat_interval= [KNL] 3197 Time (s) between statistics printk()s. 3198 3199 locktorture.stutter= [KNL] 3200 Time (s) to stutter testing, for example, 3201 specifying five seconds causes the test to run for 3202 five seconds, wait for five seconds, and so on. 3203 This tests the locking primitive's ability to 3204 transition abruptly to and from idle. 3205 3206 locktorture.torture_type= [KNL] 3207 Specify the locking implementation to test. 3208 3209 locktorture.verbose= [KNL] 3210 Enable additional printk() statements. 3211 3212 locktorture.writer_fifo= [KNL] 3213 Run the write-side locktorture kthreads at 3214 sched_set_fifo() real-time priority. 3215 3216 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 3217 Format: <irq> 3218 3219 loglevel= [KNL,EARLY] 3220 All Kernel Messages with a loglevel smaller than the 3221 console loglevel will be printed to the console. It can 3222 also be changed with klogd or other programs. The 3223 loglevels are defined as follows: 3224 3225 0 (KERN_EMERG) system is unusable 3226 1 (KERN_ALERT) action must be taken immediately 3227 2 (KERN_CRIT) critical conditions 3228 3 (KERN_ERR) error conditions 3229 4 (KERN_WARNING) warning conditions 3230 5 (KERN_NOTICE) normal but significant condition 3231 6 (KERN_INFO) informational 3232 7 (KERN_DEBUG) debug-level messages 3233 3234 log_buf_len=n[KMG] [KNL,EARLY] 3235 Sets the size of the printk ring buffer, in bytes. 3236 n must be a power of two and greater than the 3237 minimal size. The minimal size is defined by 3238 LOG_BUF_SHIFT kernel config parameter. There 3239 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config 3240 parameter that allows to increase the default size 3241 depending on the number of CPUs. See init/Kconfig 3242 for more details. 3243 3244 logo.nologo [FB] Disables display of the built-in Linux logo. 3245 This may be used to provide more screen space for 3246 kernel log messages and is useful when debugging 3247 kernel boot problems. 3248 3249 lp=0 [LP] Specify parallel ports to use, e.g, 3250 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 3251 lp=reset first parallel port). 'lp=0' disables the 3252 lp=auto printer driver. 'lp=reset' (which can be 3253 specified in addition to the ports) causes 3254 attached printers to be reset. Using 3255 lp=port1,port2,... specifies the parallel ports 3256 to associate lp devices with, starting with 3257 lp0. A port specification may be 'none' to skip 3258 that lp device, or a parport name such as 3259 'parport0'. Specifying 'lp=auto' instead of a 3260 port specification list means that device IDs 3261 from each port should be examined, to see if 3262 an IEEE 1284-compliant printer is attached; if 3263 so, the driver will manage that printer. 3264 See also header of drivers/char/lp.c. 3265 3266 lpj=n [KNL] 3267 Sets loops_per_jiffy to given constant, thus avoiding 3268 time-consuming boot-time autodetection (up to 250 ms per 3269 CPU). 0 enables autodetection (default). To determine 3270 the correct value for your kernel, boot with normal 3271 autodetection and see what value is printed. Note that 3272 on SMP systems the preset will be applied to all CPUs, 3273 which is likely to cause problems if your CPUs need 3274 significantly divergent settings. An incorrect value 3275 will cause delays in the kernel to be wrong, leading to 3276 unpredictable I/O errors and other breakage. Although 3277 unlikely, in the extreme case this might damage your 3278 hardware. 3279 3280 lsm.debug [SECURITY] Enable LSM initialization debugging output. 3281 3282 lsm=lsm1,...,lsmN 3283 [SECURITY] Choose order of LSM initialization. This 3284 overrides CONFIG_LSM, and the "security=" parameter. 3285 3286 machtype= [Loongson] Share the same kernel image file between 3287 different yeeloong laptops. 3288 Example: machtype=lemote-yeeloong-2f-7inch 3289 3290 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 3291 will bring up during bootup. maxcpus=n : n >= 0 limits 3292 the kernel to bring up 'n' processors. Surely after 3293 bootup you can bring up the other plugged cpu by executing 3294 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 3295 only takes effect during system bootup. 3296 While n=0 is a special case, it is equivalent to "nosmp", 3297 which also disables the IO APIC. 3298 3299 max_loop= [LOOP] The number of loop block devices that get 3300 (loop.max_loop) unconditionally pre-created at init time. The default 3301 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 3302 of statically allocating a predefined number, loop 3303 devices can be requested on-demand with the 3304 /dev/loop-control interface. 3305 3306 mce [X86-32] Machine Check Exception 3307 3308 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst 3309 3310 md= [HW] RAID subsystems devices and level 3311 See Documentation/admin-guide/md.rst. 3312 3313 mdacon= [MDA] 3314 Format: <first>,<last> 3315 Specifies range of consoles to be captured by the MDA. 3316 3317 mds= [X86,INTEL,EARLY] 3318 Control mitigation for the Micro-architectural Data 3319 Sampling (MDS) vulnerability. 3320 3321 Certain CPUs are vulnerable to an exploit against CPU 3322 internal buffers which can forward information to a 3323 disclosure gadget under certain conditions. 3324 3325 In vulnerable processors, the speculatively 3326 forwarded data can be used in a cache side channel 3327 attack, to access data to which the attacker does 3328 not have direct access. 3329 3330 This parameter controls the MDS mitigation. The 3331 options are: 3332 3333 full - Enable MDS mitigation on vulnerable CPUs 3334 full,nosmt - Enable MDS mitigation and disable 3335 SMT on vulnerable CPUs 3336 off - Unconditionally disable MDS mitigation 3337 3338 On TAA-affected machines, mds=off can be prevented by 3339 an active TAA mitigation as both vulnerabilities are 3340 mitigated with the same mechanism so in order to disable 3341 this mitigation, you need to specify tsx_async_abort=off 3342 too. 3343 3344 Not specifying this option is equivalent to 3345 mds=full. 3346 3347 For details see: Documentation/admin-guide/hw-vuln/mds.rst 3348 3349 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size. 3350 Must be specified, otherwise memory size will be 0. 3351 3352 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount 3353 of memory Amount of memory to be used in cases 3354 as follows: 3355 3356 1 for test; 3357 2 when the kernel is not able to see the whole system memory; 3358 3 memory that lies after 'mem=' boundary is excluded from 3359 the hypervisor, then assigned to KVM guests. 3360 4 to limit the memory available for kdump kernel. 3361 3362 [ARC,MICROBLAZE] - the limit applies only to low memory, 3363 high memory is not affected. 3364 3365 [ARM64] - only limits memory covered by the linear 3366 mapping. The NOMAP regions are not affected. 3367 3368 [X86] Work as limiting max address. Use together 3369 with memmap= to avoid physical address space collisions. 3370 Without memmap= PCI devices could be placed at addresses 3371 belonging to unused RAM. 3372 3373 Note that this only takes effects during boot time since 3374 in above case 3, memory may need be hot added after boot 3375 if system memory of hypervisor is not sufficient. 3376 3377 mem=nn[KMG]@ss[KMG] 3378 [ARM,MIPS,EARLY] - override the memory layout 3379 reported by firmware. 3380 Define a memory region of size nn[KMG] starting at 3381 ss[KMG]. 3382 Multiple different regions can be specified with 3383 multiple mem= parameters on the command line. 3384 3385 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 3386 memory. 3387 3388 memblock=debug [KNL,EARLY] Enable memblock debug messages. 3389 3390 memchunk=nn[KMG] 3391 [KNL,SH] Allow user to override the default size for 3392 per-device physically contiguous DMA buffers. 3393 3394 memhp_default_state=online/offline/online_kernel/online_movable 3395 [KNL] Set the initial state for the memory hotplug 3396 onlining policy. If not specified, the default value is 3397 set according to the 3398 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config 3399 option. 3400 See Documentation/admin-guide/mm/memory-hotplug.rst. 3401 3402 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact 3403 E820 memory map, as specified by the user. 3404 Such memmap=exactmap lines can be constructed based on 3405 BIOS output or other requirements. See the memmap=nn@ss 3406 option description. 3407 3408 memmap=nn[KMG]@ss[KMG] 3409 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory. 3410 Region of memory to be used is from ss to ss+nn. 3411 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 3412 which limits max address to nn[KMG]. 3413 Multiple different regions can be specified, 3414 comma delimited. 3415 Example: 3416 memmap=100M@2G,100M#3G,1G!1024G 3417 3418 memmap=nn[KMG]#ss[KMG] 3419 [KNL,ACPI,EARLY] Mark specific memory as ACPI data. 3420 Region of memory to be marked is from ss to ss+nn. 3421 3422 memmap=nn[KMG]$ss[KMG] 3423 [KNL,ACPI,EARLY] Mark specific memory as reserved. 3424 Region of memory to be reserved is from ss to ss+nn. 3425 Example: Exclude memory from 0x18690000-0x1869ffff 3426 memmap=64K$0x18690000 3427 or 3428 memmap=0x10000$0x18690000 3429 Some bootloaders may need an escape character before '$', 3430 like Grub2, otherwise '$' and the following number 3431 will be eaten. 3432 3433 memmap=nn[KMG]!ss[KMG,EARLY] 3434 [KNL,X86] Mark specific memory as protected. 3435 Region of memory to be used, from ss to ss+nn. 3436 The memory region may be marked as e820 type 12 (0xc) 3437 and is NVDIMM or ADR memory. 3438 3439 memmap=<size>%<offset>-<oldtype>+<newtype> 3440 [KNL,ACPI,EARLY] Convert memory within the specified region 3441 from <oldtype> to <newtype>. If "-<oldtype>" is left 3442 out, the whole region will be marked as <newtype>, 3443 even if previously unavailable. If "+<newtype>" is left 3444 out, matching memory will be removed. Types are 3445 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 3446 3 = ACPI, 12 = PRAM. 3447 3448 memory_corruption_check=0/1 [X86,EARLY] 3449 Some BIOSes seem to corrupt the first 64k of 3450 memory when doing things like suspend/resume. 3451 Setting this option will scan the memory 3452 looking for corruption. Enabling this will 3453 both detect corruption and prevent the kernel 3454 from using the memory being corrupted. 3455 However, its intended as a diagnostic tool; if 3456 repeatable BIOS-originated corruption always 3457 affects the same memory, you can use memmap= 3458 to prevent the kernel from using that memory. 3459 3460 memory_corruption_check_size=size [X86,EARLY] 3461 By default it checks for corruption in the low 3462 64k, making this memory unavailable for normal 3463 use. Use this parameter to scan for 3464 corruption in more or less memory. 3465 3466 memory_corruption_check_period=seconds [X86,EARLY] 3467 By default it checks for corruption every 60 3468 seconds. Use this parameter to check at some 3469 other rate. 0 disables periodic checking. 3470 3471 memory_hotplug.memmap_on_memory 3472 [KNL,X86,ARM] Boolean flag to enable this feature. 3473 Format: {on | off (default)} 3474 When enabled, runtime hotplugged memory will 3475 allocate its internal metadata (struct pages, 3476 those vmemmap pages cannot be optimized even 3477 if hugetlb_free_vmemmap is enabled) from the 3478 hotadded memory which will allow to hotadd a 3479 lot of memory without requiring additional 3480 memory to do so. 3481 This feature is disabled by default because it 3482 has some implication on large (e.g. GB) 3483 allocations in some configurations (e.g. small 3484 memory blocks). 3485 The state of the flag can be read in 3486 /sys/module/memory_hotplug/parameters/memmap_on_memory. 3487 Note that even when enabled, there are a few cases where 3488 the feature is not effective. 3489 3490 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest 3491 Format: <integer> 3492 default : 0 <disable> 3493 Specifies the number of memtest passes to be 3494 performed. Each pass selects another test 3495 pattern from a given set of patterns. Memtest 3496 fills the memory with this pattern, validates 3497 memory contents and reserves bad memory 3498 regions that are detected. 3499 3500 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 3501 Valid arguments: on, off 3502 Default: off 3503 mem_encrypt=on: Activate SME 3504 mem_encrypt=off: Do not activate SME 3505 3506 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst 3507 for details on when memory encryption can be activated. 3508 3509 mem_sleep_default= [SUSPEND] Default system suspend mode: 3510 s2idle - Suspend-To-Idle 3511 shallow - Power-On Suspend or equivalent (if supported) 3512 deep - Suspend-To-RAM or equivalent (if supported) 3513 See Documentation/admin-guide/pm/sleep-states.rst. 3514 3515 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 3516 the BIOS has incorrectly applied a workaround. TinyBIOS 3517 version 0.98 is known to be affected, 0.99 fixes the 3518 problem by letting the user disable the workaround. 3519 3520 mga= [HW,DRM] 3521 3522 microcode.force_minrev= [X86] 3523 Format: <bool> 3524 Enable or disable the microcode minimal revision 3525 enforcement for the runtime microcode loader. 3526 3527 mini2440= [ARM,HW,KNL] 3528 Format:[0..2][b][c][t] 3529 Default: "0tb" 3530 MINI2440 configuration specification: 3531 0 - The attached screen is the 3.5" TFT 3532 1 - The attached screen is the 7" TFT 3533 2 - The VGA Shield is attached (1024x768) 3534 Leaving out the screen size parameter will not load 3535 the TFT driver, and the framebuffer will be left 3536 unconfigured. 3537 b - Enable backlight. The TFT backlight pin will be 3538 linked to the kernel VESA blanking code and a GPIO 3539 LED. This parameter is not necessary when using the 3540 VGA shield. 3541 c - Enable the s3c camera interface. 3542 t - Reserved for enabling touchscreen support. The 3543 touchscreen support is not enabled in the mainstream 3544 kernel as of 2.6.30, a preliminary port can be found 3545 in the "bleeding edge" mini2440 support kernel at 3546 https://repo.or.cz/w/linux-2.6/mini2440.git 3547 3548 mitigations= 3549 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for 3550 CPU vulnerabilities. This is a set of curated, 3551 arch-independent options, each of which is an 3552 aggregation of existing arch-specific options. 3553 3554 Note, "mitigations" is supported if and only if the 3555 kernel was built with CPU_MITIGATIONS=y. 3556 3557 off 3558 Disable all optional CPU mitigations. This 3559 improves system performance, but it may also 3560 expose users to several CPU vulnerabilities. 3561 Equivalent to: if nokaslr then kpti=0 [ARM64] 3562 gather_data_sampling=off [X86] 3563 indirect_target_selection=off [X86] 3564 kvm.nx_huge_pages=off [X86] 3565 l1tf=off [X86] 3566 mds=off [X86] 3567 mmio_stale_data=off [X86] 3568 no_entry_flush [PPC] 3569 no_uaccess_flush [PPC] 3570 nobp=0 [S390] 3571 nopti [X86,PPC] 3572 nospectre_bhb [ARM64] 3573 nospectre_v1 [X86,PPC] 3574 nospectre_v2 [X86,PPC,S390,ARM64] 3575 reg_file_data_sampling=off [X86] 3576 retbleed=off [X86] 3577 spec_rstack_overflow=off [X86] 3578 spec_store_bypass_disable=off [X86,PPC] 3579 spectre_bhi=off [X86] 3580 spectre_v2_user=off [X86] 3581 srbds=off [X86,INTEL] 3582 ssbd=force-off [ARM64] 3583 tsx_async_abort=off [X86] 3584 vmscape=off [X86] 3585 3586 Exceptions: 3587 This does not have any effect on 3588 kvm.nx_huge_pages when 3589 kvm.nx_huge_pages=force. 3590 3591 auto (default) 3592 Mitigate all CPU vulnerabilities, but leave SMT 3593 enabled, even if it's vulnerable. This is for 3594 users who don't want to be surprised by SMT 3595 getting disabled across kernel upgrades, or who 3596 have other ways of avoiding SMT-based attacks. 3597 Equivalent to: (default behavior) 3598 3599 auto,nosmt 3600 Mitigate all CPU vulnerabilities, disabling SMT 3601 if needed. This is for users who always want to 3602 be fully mitigated, even if it means losing SMT. 3603 Equivalent to: l1tf=flush,nosmt [X86] 3604 mds=full,nosmt [X86] 3605 tsx_async_abort=full,nosmt [X86] 3606 mmio_stale_data=full,nosmt [X86] 3607 retbleed=auto,nosmt [X86] 3608 3609 mminit_loglevel= 3610 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this 3611 parameter allows control of the logging verbosity for 3612 the additional memory initialisation checks. A value 3613 of 0 disables mminit logging and a level of 4 will 3614 log everything. Information is printed at KERN_DEBUG 3615 so loglevel=8 may also need to be specified. 3616 3617 mmio_stale_data= 3618 [X86,INTEL,EARLY] Control mitigation for the Processor 3619 MMIO Stale Data vulnerabilities. 3620 3621 Processor MMIO Stale Data is a class of 3622 vulnerabilities that may expose data after an MMIO 3623 operation. Exposed data could originate or end in 3624 the same CPU buffers as affected by MDS and TAA. 3625 Therefore, similar to MDS and TAA, the mitigation 3626 is to clear the affected CPU buffers. 3627 3628 This parameter controls the mitigation. The 3629 options are: 3630 3631 full - Enable mitigation on vulnerable CPUs 3632 3633 full,nosmt - Enable mitigation and disable SMT on 3634 vulnerable CPUs. 3635 3636 off - Unconditionally disable mitigation 3637 3638 On MDS or TAA affected machines, 3639 mmio_stale_data=off can be prevented by an active 3640 MDS or TAA mitigation as these vulnerabilities are 3641 mitigated with the same mechanism so in order to 3642 disable this mitigation, you need to specify 3643 mds=off and tsx_async_abort=off too. 3644 3645 Not specifying this option is equivalent to 3646 mmio_stale_data=full. 3647 3648 For details see: 3649 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 3650 3651 <module>.async_probe[=<bool>] [KNL] 3652 If no <bool> value is specified or if the value 3653 specified is not a valid <bool>, enable asynchronous 3654 probe on this module. Otherwise, enable/disable 3655 asynchronous probe on this module as indicated by the 3656 <bool> value. See also: module.async_probe 3657 3658 module.async_probe=<bool> 3659 [KNL] When set to true, modules will use async probing 3660 by default. To enable/disable async probing for a 3661 specific module, use the module specific control that 3662 is documented under <module>.async_probe. When both 3663 module.async_probe and <module>.async_probe are 3664 specified, <module>.async_probe takes precedence for 3665 the specific module. 3666 3667 module.enable_dups_trace 3668 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set, 3669 this means that duplicate request_module() calls will 3670 trigger a WARN_ON() instead of a pr_warn(). Note that 3671 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s 3672 will always be issued and this option does nothing. 3673 module.sig_enforce 3674 [KNL] When CONFIG_MODULE_SIG is set, this means that 3675 modules without (valid) signatures will fail to load. 3676 Note that if CONFIG_MODULE_SIG_FORCE is set, that 3677 is always true, so this option does nothing. 3678 3679 module_blacklist= [KNL] Do not load a comma-separated list of 3680 modules. Useful for debugging problem modules. 3681 3682 mousedev.tap_time= 3683 [MOUSE] Maximum time between finger touching and 3684 leaving touchpad surface for touch to be considered 3685 a tap and be reported as a left button click (for 3686 touchpads working in absolute mode only). 3687 Format: <msecs> 3688 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 3689 reporting absolute coordinates, such as tablets 3690 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 3691 reporting absolute coordinates, such as tablets 3692 3693 movablecore= [KNL,X86,PPC,EARLY] 3694 Format: nn[KMGTPE] | nn% 3695 This parameter is the complement to kernelcore=, it 3696 specifies the amount of memory used for migratable 3697 allocations. If both kernelcore and movablecore is 3698 specified, then kernelcore will be at *least* the 3699 specified value but may be more. If movablecore on its 3700 own is specified, the administrator must be careful 3701 that the amount of memory usable for all allocations 3702 is not too small. 3703 3704 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory 3705 NUMA nodes to be movable. This means that the memory 3706 of such nodes will be usable only for movable 3707 allocations which rules out almost all kernel 3708 allocations. Use with caution! 3709 3710 MTD_Partition= [MTD] 3711 Format: <name>,<region-number>,<size>,<offset> 3712 3713 MTD_Region= [MTD] Format: 3714 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 3715 3716 mtdparts= [MTD] 3717 See drivers/mtd/parsers/cmdlinepart.c 3718 3719 mtouchusb.raw_coordinates= 3720 [HW] Make the MicroTouch USB driver use raw coordinates 3721 ('y', default) or cooked coordinates ('n') 3722 3723 mtrr=debug [X86,EARLY] 3724 Enable printing debug information related to MTRR 3725 registers at boot time. 3726 3727 mtrr_chunk_size=nn[KMG,X86,EARLY] 3728 used for mtrr cleanup. It is largest continuous chunk 3729 that could hold holes aka. UC entries. 3730 3731 mtrr_gran_size=nn[KMG,X86,EARLY] 3732 Used for mtrr cleanup. It is granularity of mtrr block. 3733 Default is 1. 3734 Large value could prevent small alignment from 3735 using up MTRRs. 3736 3737 mtrr_spare_reg_nr=n [X86,EARLY] 3738 Format: <integer> 3739 Range: 0,7 : spare reg number 3740 Default : 1 3741 Used for mtrr cleanup. It is spare mtrr entries number. 3742 Set to 2 or more if your graphical card needs more. 3743 3744 multitce=off [PPC] This parameter disables the use of the pSeries 3745 firmware feature for updating multiple TCE entries 3746 at a time. 3747 3748 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 3749 3750 netdev= [NET] Network devices parameters 3751 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 3752 Note that mem_start is often overloaded to mean 3753 something different and driver-specific. 3754 This usage is only documented in each driver source 3755 file if at all. 3756 3757 netpoll.carrier_timeout= 3758 [NET] Specifies amount of time (in seconds) that 3759 netpoll should wait for a carrier. By default netpoll 3760 waits 4 seconds. 3761 3762 nf_conntrack.acct= 3763 [NETFILTER] Enable connection tracking flow accounting 3764 0 to disable accounting 3765 1 to enable accounting 3766 Default value is 0. 3767 3768 nfs.cache_getent= 3769 [NFS] sets the pathname to the program which is used 3770 to update the NFS client cache entries. 3771 3772 nfs.cache_getent_timeout= 3773 [NFS] sets the timeout after which an attempt to 3774 update a cache entry is deemed to have failed. 3775 3776 nfs.callback_nr_threads= 3777 [NFSv4] set the total number of threads that the 3778 NFS client will assign to service NFSv4 callback 3779 requests. 3780 3781 nfs.callback_tcpport= 3782 [NFS] set the TCP port on which the NFSv4 callback 3783 channel should listen. 3784 3785 nfs.delay_retrans= 3786 [NFS] specifies the number of times the NFSv4 client 3787 retries the request before returning an EAGAIN error, 3788 after a reply of NFS4ERR_DELAY from the server. 3789 Only applies if the softerr mount option is enabled, 3790 and the specified value is >= 0. 3791 3792 nfs.enable_ino64= 3793 [NFS] enable 64-bit inode numbers. 3794 If zero, the NFS client will fake up a 32-bit inode 3795 number for the readdir() and stat() syscalls instead 3796 of returning the full 64-bit number. 3797 The default is to return 64-bit inode numbers. 3798 3799 nfs.idmap_cache_timeout= 3800 [NFS] set the maximum lifetime for idmapper cache 3801 entries. 3802 3803 nfs.max_session_cb_slots= 3804 [NFSv4.1] Sets the maximum number of session 3805 slots the client will assign to the callback 3806 channel. This determines the maximum number of 3807 callbacks the client will process in parallel for 3808 a particular server. 3809 3810 nfs.max_session_slots= 3811 [NFSv4.1] Sets the maximum number of session slots 3812 the client will attempt to negotiate with the server. 3813 This limits the number of simultaneous RPC requests 3814 that the client can send to the NFSv4.1 server. 3815 Note that there is little point in setting this 3816 value higher than the max_tcp_slot_table_limit. 3817 3818 nfs.nfs4_disable_idmapping= 3819 [NFSv4] When set to the default of '1', this option 3820 ensures that both the RPC level authentication 3821 scheme and the NFS level operations agree to use 3822 numeric uids/gids if the mount is using the 3823 'sec=sys' security flavour. In effect it is 3824 disabling idmapping, which can make migration from 3825 legacy NFSv2/v3 systems to NFSv4 easier. 3826 Servers that do not support this mode of operation 3827 will be autodetected by the client, and it will fall 3828 back to using the idmapper. 3829 To turn off this behaviour, set the value to '0'. 3830 3831 nfs.nfs4_unique_id= 3832 [NFS4] Specify an additional fixed unique ident- 3833 ification string that NFSv4 clients can insert into 3834 their nfs_client_id4 string. This is typically a 3835 UUID that is generated at system install time. 3836 3837 nfs.recover_lost_locks= 3838 [NFSv4] Attempt to recover locks that were lost due 3839 to a lease timeout on the server. Please note that 3840 doing this risks data corruption, since there are 3841 no guarantees that the file will remain unchanged 3842 after the locks are lost. 3843 If you want to enable the kernel legacy behaviour of 3844 attempting to recover these locks, then set this 3845 parameter to '1'. 3846 The default parameter value of '0' causes the kernel 3847 not to attempt recovery of lost locks. 3848 3849 nfs.send_implementation_id= 3850 [NFSv4.1] Send client implementation identification 3851 information in exchange_id requests. 3852 If zero, no implementation identification information 3853 will be sent. 3854 The default is to send the implementation identification 3855 information. 3856 3857 nfs4.layoutstats_timer= 3858 [NFSv4.2] Change the rate at which the kernel sends 3859 layoutstats to the pNFS metadata server. 3860 3861 Setting this to value to 0 causes the kernel to use 3862 whatever value is the default set by the layout 3863 driver. A non-zero value sets the minimum interval 3864 in seconds between layoutstats transmissions. 3865 3866 nfsd.inter_copy_offload_enable= 3867 [NFSv4.2] When set to 1, the server will support 3868 server-to-server copies for which this server is 3869 the destination of the copy. 3870 3871 nfsd.nfs4_disable_idmapping= 3872 [NFSv4] When set to the default of '1', the NFSv4 3873 server will return only numeric uids and gids to 3874 clients using auth_sys, and will accept numeric uids 3875 and gids from such clients. This is intended to ease 3876 migration from NFSv2/v3. 3877 3878 nfsd.nfsd4_ssc_umount_timeout= 3879 [NFSv4.2] When used as the destination of a 3880 server-to-server copy, knfsd temporarily mounts 3881 the source server. It caches the mount in case 3882 it will be needed again, and discards it if not 3883 used for the number of milliseconds specified by 3884 this parameter. 3885 3886 nfsaddrs= [NFS] Deprecated. Use ip= instead. 3887 See Documentation/admin-guide/nfs/nfsroot.rst. 3888 3889 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 3890 See Documentation/admin-guide/nfs/nfsroot.rst. 3891 3892 nfsrootdebug [NFS] enable nfsroot debugging messages. 3893 See Documentation/admin-guide/nfs/nfsroot.rst. 3894 3895 nmi_backtrace.backtrace_idle [KNL] 3896 Dump stacks even of idle CPUs in response to an 3897 NMI stack-backtrace request. 3898 3899 nmi_debug= [KNL,SH] Specify one or more actions to take 3900 when a NMI is triggered. 3901 Format: [state][,regs][,debounce][,die] 3902 3903 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 3904 Format: [panic,][nopanic,][rNNN,][num] 3905 Valid num: 0 or 1 3906 0 - turn hardlockup detector in nmi_watchdog off 3907 1 - turn hardlockup detector in nmi_watchdog on 3908 rNNN - configure the watchdog with raw perf event 0xNNN 3909 3910 When panic is specified, panic when an NMI watchdog 3911 timeout occurs (or 'nopanic' to not panic on an NMI 3912 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 3913 To disable both hard and soft lockup detectors, 3914 please see 'nowatchdog'. 3915 This is useful when you use a panic=... timeout and 3916 need the box quickly up again. 3917 3918 These settings can be accessed at runtime via 3919 the nmi_watchdog and hardlockup_panic sysctls. 3920 3921 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths 3922 emulation library even if a 387 maths coprocessor 3923 is present. 3924 3925 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes. 3926 Forces kernel to use 3-level paging instead. 3927 3928 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces 3929 kernel to use 4-level paging instead. 3930 3931 noalign [KNL,ARM] 3932 3933 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any 3934 IOAPICs that may be present in the system. 3935 3936 noautogroup Disable scheduler automatic task group creation. 3937 3938 nocache [ARM,EARLY] 3939 3940 no_console_suspend 3941 [HW] Never suspend the console 3942 Disable suspending of consoles during suspend and 3943 hibernate operations. Once disabled, debugging 3944 messages can reach various consoles while the rest 3945 of the system is being put to sleep (ie, while 3946 debugging driver suspend/resume hooks). This may 3947 not work reliably with all consoles, but is known 3948 to work with serial and VGA consoles. 3949 To facilitate more flexible debugging, we also add 3950 console_suspend, a printk module parameter to control 3951 it. Users could use console_suspend (usually 3952 /sys/module/printk/parameters/console_suspend) to 3953 turn on/off it dynamically. 3954 3955 no_debug_objects 3956 [KNL,EARLY] Disable object debugging 3957 3958 nodsp [SH] Disable hardware DSP at boot time. 3959 3960 noefi [EFI,EARLY] Disable EFI runtime services support. 3961 3962 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel. 3963 3964 noexec32 [X86-64] 3965 This affects only 32-bit executables. 3966 noexec32=on: enable non-executable mappings (default) 3967 read doesn't imply executable mappings 3968 noexec32=off: disable non-executable mappings 3969 read implies executable mappings 3970 3971 no_file_caps Tells the kernel not to honor file capabilities. The 3972 only way then for a file to be executed with privilege 3973 is to be setuid root or executed by root. 3974 3975 nofpu [MIPS,SH] Disable hardware FPU at boot time. 3976 3977 nofsgsbase [X86] Disables FSGSBASE instructions. 3978 3979 nofxsr [BUGS=X86-32] Disables x86 floating point extended 3980 register save and restore. The kernel will only save 3981 legacy floating-point registers on task switch. 3982 3983 no_hash_pointers 3984 [KNL,EARLY] 3985 Force pointers printed to the console or buffers to be 3986 unhashed. By default, when a pointer is printed via %p 3987 format string, that pointer is "hashed", i.e. obscured 3988 by hashing the pointer value. This is a security feature 3989 that hides actual kernel addresses from unprivileged 3990 users, but it also makes debugging the kernel more 3991 difficult since unequal pointers can no longer be 3992 compared. However, if this command-line option is 3993 specified, then all normal pointers will have their true 3994 value printed. This option should only be specified when 3995 debugging the kernel. Please do not use on production 3996 kernels. 3997 3998 nohibernate [HIBERNATION] Disable hibernation and resume. 3999 4000 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,RISCV,SH] Forces the kernel to 4001 busy wait in do_idle() and not use the arch_cpu_idle() 4002 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP 4003 to be effective. This is useful on platforms where the 4004 sleep(SH) or wfi(ARM,ARM64) instructions do not work 4005 correctly or when doing power measurements to evaluate 4006 the impact of the sleep instructions. This is also 4007 useful when using JTAG debugger. 4008 4009 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings. 4010 4011 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings. 4012 4013 nohz= [KNL] Boottime enable/disable dynamic ticks 4014 Valid arguments: on, off 4015 Default: on 4016 4017 nohz_full= [KNL,BOOT,SMP,ISOL] 4018 The argument is a cpu list, as described above. 4019 In kernels built with CONFIG_NO_HZ_FULL=y, set 4020 the specified list of CPUs whose tick will be stopped 4021 whenever possible. The boot CPU will be forced outside 4022 the range to maintain the timekeeping. Any CPUs 4023 in this list will have their RCU callbacks offloaded, 4024 just as if they had also been called out in the 4025 rcu_nocbs= boot parameter. 4026 4027 Note that this argument takes precedence over 4028 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 4029 4030 noinitrd [RAM] Tells the kernel not to load any configured 4031 initial RAM disk. 4032 4033 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt 4034 remapping. 4035 [Deprecated - use intremap=off] 4036 4037 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature. 4038 4039 noiotrap [SH] Disables trapped I/O port accesses. 4040 4041 noirqdebug [X86-32] Disables the code which attempts to detect and 4042 disable unhandled interrupt sources. 4043 4044 noisapnp [ISAPNP] Disables ISA PnP code. 4045 4046 nokaslr [KNL,EARLY] 4047 When CONFIG_RANDOMIZE_BASE is set, this disables 4048 kernel and module base offset ASLR (Address Space 4049 Layout Randomization). 4050 4051 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page 4052 fault handling. 4053 4054 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver 4055 4056 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC. 4057 4058 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer. 4059 4060 nomce [X86-32] Disable Machine Check Exception 4061 4062 nomfgpt [X86-32] Disable Multi-Function General Purpose 4063 Timer usage (for AMD Geode machines). 4064 4065 nomodeset Disable kernel modesetting. Most systems' firmware 4066 sets up a display mode and provides framebuffer memory 4067 for output. With nomodeset, DRM and fbdev drivers will 4068 not load if they could possibly displace the pre- 4069 initialized output. Only the system framebuffer will 4070 be available for use. The respective drivers will not 4071 perform display-mode changes or accelerated rendering. 4072 4073 Useful as error fallback, or for testing and debugging. 4074 4075 nomodule Disable module load 4076 4077 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 4078 shutdown the other cpus. Instead use the REBOOT_VECTOR 4079 irq. 4080 4081 nopat [X86,EARLY] Disable PAT (page attribute table extension of 4082 pagetables) support. 4083 4084 nopcid [X86-64,EARLY] Disable the PCID cpu feature. 4085 4086 nopku [X86] Disable Memory Protection Keys CPU feature found 4087 in some Intel CPUs. 4088 4089 nopti [X86-64,EARLY] 4090 Equivalent to pti=off 4091 4092 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY] 4093 Disables the PV optimizations forcing the guest to run 4094 as generic guest with no PV drivers. Currently support 4095 XEN HVM, KVM, HYPER_V and VMWARE guest. 4096 4097 nopvspin [X86,XEN,KVM,EARLY] 4098 Disables the qspinlock slow path using PV optimizations 4099 which allow the hypervisor to 'idle' the guest on lock 4100 contention. 4101 4102 norandmaps Don't use address space randomization. Equivalent to 4103 echo 0 > /proc/sys/kernel/randomize_va_space 4104 4105 noreplace-smp [X86-32,SMP] Don't replace SMP instructions 4106 with UP alternatives 4107 4108 noresume [SWSUSP] Disables resume and restores original swap 4109 space. 4110 4111 no-scroll [VGA] Disables scrollback. 4112 This is required for the Braillex ib80-piezo Braille 4113 reader made by F.H. Papenmeier (Germany). 4114 4115 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support. 4116 4117 nosmap [PPC,EARLY] 4118 Disable SMAP (Supervisor Mode Access Prevention) 4119 even if it is supported by processor. 4120 4121 nosmep [PPC64s,EARLY] 4122 Disable SMEP (Supervisor Mode Execution Prevention) 4123 even if it is supported by processor. 4124 4125 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel, 4126 and disable the IO APIC. legacy for "maxcpus=0". 4127 4128 nosmt [KNL,MIPS,PPC,S390,EARLY] Disable symmetric multithreading (SMT). 4129 Equivalent to smt=1. 4130 4131 [KNL,X86,PPC] Disable symmetric multithreading (SMT). 4132 nosmt=force: Force disable SMT, cannot be undone 4133 via the sysfs control file. 4134 4135 nosoftlockup [KNL] Disable the soft-lockup detector. 4136 4137 nospec_store_bypass_disable 4138 [HW,EARLY] Disable all mitigations for the Speculative 4139 Store Bypass vulnerability 4140 4141 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch 4142 history injection) vulnerability. System may allow data leaks 4143 with this option. 4144 4145 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1 4146 (bounds check bypass). With this option data leaks are 4147 possible in the system. 4148 4149 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations 4150 for the Spectre variant 2 (indirect branch 4151 prediction) vulnerability. System may allow data 4152 leaks with this option. 4153 4154 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,LOONGARCH,EARLY] 4155 Disable paravirtualized steal time accounting. steal time 4156 is computed, but won't influence scheduler behaviour 4157 4158 nosync [HW,M68K] Disables sync negotiation for all devices. 4159 4160 no_timer_check [X86,APIC] Disables the code which tests for 4161 broken timer IRQ sources. 4162 4163 no_uaccess_flush 4164 [PPC,EARLY] Don't flush the L1-D cache after accessing user data. 4165 4166 novmcoredd [KNL,KDUMP] 4167 Disable device dump. Device dump allows drivers to 4168 append dump data to vmcore so you can collect driver 4169 specified debug info. Drivers can append the data 4170 without any limit and this data is stored in memory, 4171 so this may cause significant memory stress. Disabling 4172 device dump can help save memory but the driver debug 4173 data will be no longer available. This parameter 4174 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 4175 is set. 4176 4177 no-vmw-sched-clock 4178 [X86,PV_OPS,EARLY] Disable paravirtualized VMware 4179 scheduler clock and use the default one. 4180 4181 nowatchdog [KNL] Disable both lockup detectors, i.e. 4182 soft-lockup and NMI watchdog (hard-lockup). 4183 4184 nowb [ARM,EARLY] 4185 4186 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode. 4187 4188 NOTE: this parameter will be ignored on systems with the 4189 LEGACY_XAPIC_DISABLED bit set in the 4190 IA32_XAPIC_DISABLE_STATUS MSR. 4191 4192 noxsave [BUGS=X86] Disables x86 extended register state save 4193 and restore using xsave. The kernel will fallback to 4194 enabling legacy floating-point and sse state. 4195 4196 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 4197 register states. The kernel will fall back to use 4198 xsave to save the states. By using this parameter, 4199 performance of saving the states is degraded because 4200 xsave doesn't support modified optimization while 4201 xsaveopt supports it on xsaveopt enabled systems. 4202 4203 noxsaves [X86] Disables xsaves and xrstors used in saving and 4204 restoring x86 extended register state in compacted 4205 form of xsave area. The kernel will fall back to use 4206 xsaveopt and xrstor to save and restore the states 4207 in standard form of xsave area. By using this 4208 parameter, xsave area per process might occupy more 4209 memory on xsaves enabled systems. 4210 4211 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel 4212 could support. nr_cpus=n : n >= 1 limits the kernel to 4213 support 'n' processors. It could be larger than the 4214 number of already plugged CPU during bootup, later in 4215 runtime you can physically add extra cpu until it reaches 4216 n. So during boot up some boot time memory for per-cpu 4217 variables need be pre-allocated for later physical cpu 4218 hot plugging. 4219 4220 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 4221 4222 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY] 4223 Disable NUMA, Only set up a single NUMA node 4224 spanning all memory. 4225 4226 numa=fake=<size>[MG] 4227 [KNL, ARM64, RISCV, X86, EARLY] 4228 If given as a memory unit, fills all system RAM with 4229 nodes of size interleaved over physical nodes. 4230 4231 numa=fake=<N> 4232 [KNL, ARM64, RISCV, X86, EARLY] 4233 If given as an integer, fills all system RAM with N 4234 fake nodes interleaved over physical nodes. 4235 4236 numa=fake=<N>U 4237 [KNL, ARM64, RISCV, X86, EARLY] 4238 If given as an integer followed by 'U', it will 4239 divide each physical node into N emulated nodes. 4240 4241 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic 4242 NUMA balancing. 4243 Allowed values are enable and disable 4244 4245 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 4246 'node', 'default' can be specified 4247 This can be set from sysctl after boot. 4248 See Documentation/admin-guide/sysctl/vm.rst for details. 4249 4250 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver. 4251 See Documentation/core-api/debugging-via-ohci1394.rst for more 4252 info. 4253 4254 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 4255 Rather than timing out after 20 ms if an EC 4256 command is not properly ACKed, override the length 4257 of the timeout. We have interrupts disabled while 4258 waiting for the ACK, so if this is set too high 4259 interrupts *may* be lost! 4260 4261 omap_mux= [OMAP] Override bootloader pin multiplexing. 4262 Format: <mux_mode0.mode_name=value>... 4263 For example, to override I2C bus2: 4264 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 4265 4266 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 4267 4268 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 4269 4270 boundary - index of last SLC block on Flex-OneNAND. 4271 The remaining blocks are configured as MLC blocks. 4272 lock - Configure if Flex-OneNAND boundary should be locked. 4273 Once locked, the boundary cannot be changed. 4274 1 indicates lock status, 0 indicates unlock status. 4275 4276 oops=panic [KNL,EARLY] 4277 Always panic on oopses. Default is to just kill the 4278 process, but there is a small probability of 4279 deadlocking the machine. 4280 This will also cause panics on machine check exceptions. 4281 Useful together with panic=30 to trigger a reboot. 4282 4283 page_alloc.shuffle= 4284 [KNL] Boolean flag to control whether the page allocator 4285 should randomize its free lists. This parameter can be 4286 used to enable/disable page randomization. The state of 4287 the flag can be read from sysfs at: 4288 /sys/module/page_alloc/parameters/shuffle. 4289 This parameter is only available if CONFIG_SHUFFLE_PAGE_ALLOCATOR=y. 4290 4291 page_owner= [KNL,EARLY] Boot-time page_owner enabling option. 4292 Storage of the information about who allocated 4293 each page is disabled in default. With this switch, 4294 we can turn it on. 4295 on: enable the feature 4296 4297 page_poison= [KNL,EARLY] Boot-time parameter changing the state of 4298 poisoning on the buddy allocator, available with 4299 CONFIG_PAGE_POISONING=y. 4300 off: turn off poisoning (default) 4301 on: turn on poisoning 4302 4303 page_reporting.page_reporting_order= 4304 [KNL] Minimal page reporting order 4305 Format: <integer> 4306 Adjust the minimal page reporting order. The page 4307 reporting is disabled when it exceeds MAX_PAGE_ORDER. 4308 4309 panic= [KNL] Kernel behaviour on panic: delay <timeout> 4310 timeout > 0: seconds before rebooting 4311 timeout = 0: wait forever 4312 timeout < 0: reboot immediately 4313 Format: <timeout> 4314 4315 panic_on_taint= [KNL,EARLY] 4316 Bitmask for conditionally calling panic() in add_taint() 4317 Format: <hex>[,nousertaint] 4318 Hexadecimal bitmask representing the set of TAINT flags 4319 that will cause the kernel to panic when add_taint() is 4320 called with any of the flags in this set. 4321 The optional switch "nousertaint" can be utilized to 4322 prevent userspace forced crashes by writing to sysctl 4323 /proc/sys/kernel/tainted any flagset matching with the 4324 bitmask set on panic_on_taint. 4325 See Documentation/admin-guide/tainted-kernels.rst for 4326 extra details on the taint flags that users can pick 4327 to compose the bitmask to assign to panic_on_taint. 4328 4329 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump 4330 on a WARN(). 4331 4332 panic_print= Bitmask for printing system info when panic happens. 4333 User can chose combination of the following bits: 4334 bit 0: print all tasks info 4335 bit 1: print system memory info 4336 bit 2: print timer info 4337 bit 3: print locks info if CONFIG_LOCKDEP is on 4338 bit 4: print ftrace buffer 4339 bit 5: print all printk messages in buffer 4340 bit 6: print all CPUs backtrace (if available in the arch) 4341 bit 7: print only tasks in uninterruptible (blocked) state 4342 *Be aware* that this option may print a _lot_ of lines, 4343 so there are risks of losing older messages in the log. 4344 Use this option carefully, maybe worth to setup a 4345 bigger log buffer with "log_buf_len" along with this. 4346 4347 parkbd.port= [HW] Parallel port number the keyboard adapter is 4348 connected to, default is 0. 4349 Format: <parport#> 4350 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 4351 0 for XT, 1 for AT (default is AT). 4352 Format: <mode> 4353 4354 parport= [HW,PPT] Specify parallel ports. 0 disables. 4355 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 4356 Use 'auto' to force the driver to use any 4357 IRQ/DMA settings detected (the default is to 4358 ignore detected IRQ/DMA settings because of 4359 possible conflicts). You can specify the base 4360 address, IRQ, and DMA settings; IRQ and DMA 4361 should be numbers, or 'auto' (for using detected 4362 settings on that particular port), or 'nofifo' 4363 (to avoid using a FIFO even if it is detected). 4364 Parallel ports are assigned in the order they 4365 are specified on the command line, starting 4366 with parport0. 4367 4368 parport_init_mode= [HW,PPT] 4369 Configure VIA parallel port to operate in 4370 a specific mode. This is necessary on Pegasos 4371 computer where firmware has no options for setting 4372 up parallel port mode and sets it to spp. 4373 Currently this function knows 686a and 8231 chips. 4374 Format: [spp|ps2|epp|ecp|ecpepp] 4375 4376 pata_legacy.all= [HW,LIBATA] 4377 Format: <int> 4378 Set to non-zero to probe primary and secondary ISA 4379 port ranges on PCI systems where no PCI PATA device 4380 has been found at either range. Disabled by default. 4381 4382 pata_legacy.autospeed= [HW,LIBATA] 4383 Format: <int> 4384 Set to non-zero if a chip is present that snoops speed 4385 changes. Disabled by default. 4386 4387 pata_legacy.ht6560a= [HW,LIBATA] 4388 Format: <int> 4389 Set to 1, 2, or 3 for HT 6560A on the primary channel, 4390 the secondary channel, or both channels respectively. 4391 Disabled by default. 4392 4393 pata_legacy.ht6560b= [HW,LIBATA] 4394 Format: <int> 4395 Set to 1, 2, or 3 for HT 6560B on the primary channel, 4396 the secondary channel, or both channels respectively. 4397 Disabled by default. 4398 4399 pata_legacy.iordy_mask= [HW,LIBATA] 4400 Format: <int> 4401 IORDY enable mask. Set individual bits to allow IORDY 4402 for the respective channel. Bit 0 is for the first 4403 legacy channel handled by this driver, bit 1 is for 4404 the second channel, and so on. The sequence will often 4405 correspond to the primary legacy channel, the secondary 4406 legacy channel, and so on, but the handling of a PCI 4407 bus and the use of other driver options may interfere 4408 with the sequence. By default IORDY is allowed across 4409 all channels. 4410 4411 pata_legacy.opti82c46x= [HW,LIBATA] 4412 Format: <int> 4413 Set to 1, 2, or 3 for Opti 82c611A on the primary 4414 channel, the secondary channel, or both channels 4415 respectively. Disabled by default. 4416 4417 pata_legacy.opti82c611a= [HW,LIBATA] 4418 Format: <int> 4419 Set to 1, 2, or 3 for Opti 82c465MV on the primary 4420 channel, the secondary channel, or both channels 4421 respectively. Disabled by default. 4422 4423 pata_legacy.pio_mask= [HW,LIBATA] 4424 Format: <int> 4425 PIO mode mask for autospeed devices. Set individual 4426 bits to allow the use of the respective PIO modes. 4427 Bit 0 is for mode 0, bit 1 is for mode 1, and so on. 4428 All modes allowed by default. 4429 4430 pata_legacy.probe_all= [HW,LIBATA] 4431 Format: <int> 4432 Set to non-zero to probe tertiary and further ISA 4433 port ranges on PCI systems. Disabled by default. 4434 4435 pata_legacy.probe_mask= [HW,LIBATA] 4436 Format: <int> 4437 Probe mask for legacy ISA PATA ports. Depending on 4438 platform configuration and the use of other driver 4439 options up to 6 legacy ports are supported: 0x1f0, 4440 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing 4441 of individual ports can be disabled by setting the 4442 corresponding bits in the mask to 1. Bit 0 is for 4443 the first port in the list above (0x1f0), and so on. 4444 By default all supported ports are probed. 4445 4446 pata_legacy.qdi= [HW,LIBATA] 4447 Format: <int> 4448 Set to non-zero to probe QDI controllers. By default 4449 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise. 4450 4451 pata_legacy.winbond= [HW,LIBATA] 4452 Format: <int> 4453 Set to non-zero to probe Winbond controllers. Use 4454 the standard I/O port (0x130) if 1, otherwise the 4455 value given is the I/O port to use (typically 0x1b0). 4456 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE, 4457 0 otherwise. 4458 4459 pata_platform.pio_mask= [HW,LIBATA] 4460 Format: <int> 4461 Supported PIO mode mask. Set individual bits to allow 4462 the use of the respective PIO modes. Bit 0 is for 4463 mode 0, bit 1 is for mode 1, and so on. Mode 0 only 4464 allowed by default. 4465 4466 pause_on_oops=<int> 4467 Halt all CPUs after the first oops has been printed for 4468 the specified number of seconds. This is to be used if 4469 your oopses keep scrolling off the screen. 4470 4471 pcbit= [HW,ISDN] 4472 4473 pci=option[,option...] [PCI,EARLY] various PCI subsystem options. 4474 4475 Some options herein operate on a specific device 4476 or a set of devices (<pci_dev>). These are 4477 specified in one of the following formats: 4478 4479 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 4480 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 4481 4482 Note: the first format specifies a PCI 4483 bus/device/function address which may change 4484 if new hardware is inserted, if motherboard 4485 firmware changes, or due to changes caused 4486 by other kernel parameters. If the 4487 domain is left unspecified, it is 4488 taken to be zero. Optionally, a path 4489 to a device through multiple device/function 4490 addresses can be specified after the base 4491 address (this is more robust against 4492 renumbering issues). The second format 4493 selects devices using IDs from the 4494 configuration space which may match multiple 4495 devices in the system. 4496 4497 earlydump dump PCI config space before the kernel 4498 changes anything 4499 off [X86] don't probe for the PCI bus 4500 bios [X86-32] force use of PCI BIOS, don't access 4501 the hardware directly. Use this if your machine 4502 has a non-standard PCI host bridge. 4503 nobios [X86-32] disallow use of PCI BIOS, only direct 4504 hardware access methods are allowed. Use this 4505 if you experience crashes upon bootup and you 4506 suspect they are caused by the BIOS. 4507 conf1 [X86] Force use of PCI Configuration Access 4508 Mechanism 1 (config address in IO port 0xCF8, 4509 data in IO port 0xCFC, both 32-bit). 4510 conf2 [X86] Force use of PCI Configuration Access 4511 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 4512 the function, IO port 0xCFA, also 8-bit, sets 4513 bus number. The config space is then accessed 4514 through ports 0xC000-0xCFFF). 4515 See http://wiki.osdev.org/PCI for more info 4516 on the configuration access mechanisms. 4517 noaer [PCIE] If the PCIEAER kernel config parameter is 4518 enabled, this kernel boot option can be used to 4519 disable the use of PCIE advanced error reporting. 4520 nodomains [PCI] Disable support for multiple PCI 4521 root domains (aka PCI segments, in ACPI-speak). 4522 nommconf [X86] Disable use of MMCONFIG for PCI 4523 Configuration 4524 check_enable_amd_mmconf [X86] check for and enable 4525 properly configured MMIO access to PCI 4526 config space on AMD family 10h CPU 4527 nomsi [MSI] If the PCI_MSI kernel config parameter is 4528 enabled, this kernel boot option can be used to 4529 disable the use of MSI interrupts system-wide. 4530 noioapicquirk [APIC] Disable all boot interrupt quirks. 4531 Safety option to keep boot IRQs enabled. This 4532 should never be necessary. 4533 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 4534 primary IO-APIC for bridges that cannot disable 4535 boot IRQs. This fixes a source of spurious IRQs 4536 when the system masks IRQs. 4537 noioapicreroute [APIC] Disable workaround that uses the 4538 boot IRQ equivalent of an IRQ that connects to 4539 a chipset where boot IRQs cannot be disabled. 4540 The opposite of ioapicreroute. 4541 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 4542 routing table. These calls are known to be buggy 4543 on several machines and they hang the machine 4544 when used, but on other computers it's the only 4545 way to get the interrupt routing table. Try 4546 this option if the kernel is unable to allocate 4547 IRQs or discover secondary PCI buses on your 4548 motherboard. 4549 rom [X86] Assign address space to expansion ROMs. 4550 Use with caution as certain devices share 4551 address decoders between ROMs and other 4552 resources. 4553 norom [X86] Do not assign address space to 4554 expansion ROMs that do not already have 4555 BIOS assigned address ranges. 4556 nobar [X86] Do not assign address space to the 4557 BARs that weren't assigned by the BIOS. 4558 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 4559 assigned automatically to PCI devices. You can 4560 make the kernel exclude IRQs of your ISA cards 4561 this way. 4562 pirqaddr=0xAAAAA [X86] Specify the physical address 4563 of the PIRQ table (normally generated 4564 by the BIOS) if it is outside the 4565 F0000h-100000h range. 4566 lastbus=N [X86] Scan all buses thru bus #N. Can be 4567 useful if the kernel is unable to find your 4568 secondary buses and you want to tell it 4569 explicitly which ones they are. 4570 assign-busses [X86] Always assign all PCI bus 4571 numbers ourselves, overriding 4572 whatever the firmware may have done. 4573 usepirqmask [X86] Honor the possible IRQ mask stored 4574 in the BIOS $PIR table. This is needed on 4575 some systems with broken BIOSes, notably 4576 some HP Pavilion N5400 and Omnibook XE3 4577 notebooks. This will have no effect if ACPI 4578 IRQ routing is enabled. 4579 noacpi [X86] Do not use ACPI for IRQ routing 4580 or for PCI scanning. 4581 use_crs [X86] Use PCI host bridge window information 4582 from ACPI. On BIOSes from 2008 or later, this 4583 is enabled by default. If you need to use this, 4584 please report a bug. 4585 nocrs [X86] Ignore PCI host bridge windows from ACPI. 4586 If you need to use this, please report a bug. 4587 use_e820 [X86] Use E820 reservations to exclude parts of 4588 PCI host bridge windows. This is a workaround 4589 for BIOS defects in host bridge _CRS methods. 4590 If you need to use this, please report a bug to 4591 <linux-pci@vger.kernel.org>. 4592 no_e820 [X86] Ignore E820 reservations for PCI host 4593 bridge windows. This is the default on modern 4594 hardware. If you need to use this, please report 4595 a bug to <linux-pci@vger.kernel.org>. 4596 routeirq Do IRQ routing for all PCI devices. 4597 This is normally done in pci_enable_device(), 4598 so this option is a temporary workaround 4599 for broken drivers that don't call it. 4600 skip_isa_align [X86] do not align io start addr, so can 4601 handle more pci cards 4602 noearly [X86] Don't do any early type 1 scanning. 4603 This might help on some broken boards which 4604 machine check when some devices' config space 4605 is read. But various workarounds are disabled 4606 and some IOMMU drivers will not work. 4607 bfsort Sort PCI devices into breadth-first order. 4608 This sorting is done to get a device 4609 order compatible with older (<= 2.4) kernels. 4610 nobfsort Don't sort PCI devices into breadth-first order. 4611 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 4612 tuning and use the BIOS-configured MPS defaults. 4613 pcie_bus_safe Set every device's MPS to the largest value 4614 supported by all devices below the root complex. 4615 pcie_bus_perf Set device MPS to the largest allowable MPS 4616 based on its parent bus. Also set MRRS (Max 4617 Read Request Size) to the largest supported 4618 value (no larger than the MPS that the device 4619 or bus can support) for best performance. 4620 pcie_bus_peer2peer Set every device's MPS to 128B, which 4621 every device is guaranteed to support. This 4622 configuration allows peer-to-peer DMA between 4623 any pair of devices, possibly at the cost of 4624 reduced performance. This also guarantees 4625 that hot-added devices will work. 4626 cbiosize=nn[KMG] The fixed amount of bus space which is 4627 reserved for the CardBus bridge's IO window. 4628 The default value is 256 bytes. 4629 cbmemsize=nn[KMG] The fixed amount of bus space which is 4630 reserved for the CardBus bridge's memory 4631 window. The default value is 64 megabytes. 4632 resource_alignment= 4633 Format: 4634 [<order of align>@]<pci_dev>[; ...] 4635 Specifies alignment and device to reassign 4636 aligned memory resources. How to 4637 specify the device is described above. 4638 If <order of align> is not specified, 4639 PAGE_SIZE is used as alignment. 4640 A PCI-PCI bridge can be specified if resource 4641 windows need to be expanded. 4642 To specify the alignment for several 4643 instances of a device, the PCI vendor, 4644 device, subvendor, and subdevice may be 4645 specified, e.g., 12@pci:8086:9c22:103c:198f 4646 for 4096-byte alignment. 4647 ecrc= Enable/disable PCIe ECRC (transaction layer 4648 end-to-end CRC checking). Only effective if 4649 OS has native AER control (either granted by 4650 ACPI _OSC or forced via "pcie_ports=native") 4651 bios: Use BIOS/firmware settings. This is the 4652 the default. 4653 off: Turn ECRC off 4654 on: Turn ECRC on. 4655 hpiosize=nn[KMG] The fixed amount of bus space which is 4656 reserved for hotplug bridge's IO window. 4657 Default size is 256 bytes. 4658 hpmmiosize=nn[KMG] The fixed amount of bus space which is 4659 reserved for hotplug bridge's MMIO window. 4660 Default size is 2 megabytes. 4661 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 4662 reserved for hotplug bridge's MMIO_PREF window. 4663 Default size is 2 megabytes. 4664 hpmemsize=nn[KMG] The fixed amount of bus space which is 4665 reserved for hotplug bridge's MMIO and 4666 MMIO_PREF window. 4667 Default size is 2 megabytes. 4668 hpbussize=nn The minimum amount of additional bus numbers 4669 reserved for buses below a hotplug bridge. 4670 Default is 1. 4671 realloc= Enable/disable reallocating PCI bridge resources 4672 if allocations done by BIOS are too small to 4673 accommodate resources required by all child 4674 devices. 4675 off: Turn realloc off 4676 on: Turn realloc on 4677 realloc same as realloc=on 4678 noari do not use PCIe ARI. 4679 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 4680 do not use PCIe ATS (and IOMMU device IOTLB). 4681 pcie_scan_all Scan all possible PCIe devices. Otherwise we 4682 only look for one device below a PCIe downstream 4683 port. 4684 big_root_window Try to add a big 64bit memory window to the PCIe 4685 root complex on AMD CPUs. Some GFX hardware 4686 can resize a BAR to allow access to all VRAM. 4687 Adding the window is slightly risky (it may 4688 conflict with unreported devices), so this 4689 taints the kernel. 4690 disable_acs_redir=<pci_dev>[; ...] 4691 Specify one or more PCI devices (in the format 4692 specified above) separated by semicolons. 4693 Each device specified will have the PCI ACS 4694 redirect capabilities forced off which will 4695 allow P2P traffic between devices through 4696 bridges without forcing it upstream. Note: 4697 this removes isolation between devices and 4698 may put more devices in an IOMMU group. 4699 config_acs= 4700 Format: 4701 <ACS flags>@<pci_dev>[; ...] 4702 Specify one or more PCI devices (in the format 4703 specified above) optionally prepended with flags 4704 and separated by semicolons. The respective 4705 capabilities will be enabled, disabled or 4706 unchanged based on what is specified in 4707 flags. 4708 4709 ACS Flags is defined as follows: 4710 bit-0 : ACS Source Validation 4711 bit-1 : ACS Translation Blocking 4712 bit-2 : ACS P2P Request Redirect 4713 bit-3 : ACS P2P Completion Redirect 4714 bit-4 : ACS Upstream Forwarding 4715 bit-5 : ACS P2P Egress Control 4716 bit-6 : ACS Direct Translated P2P 4717 Each bit can be marked as: 4718 '0' – force disabled 4719 '1' – force enabled 4720 'x' – unchanged 4721 For example, 4722 pci=config_acs=10x 4723 would configure all devices that support 4724 ACS to enable P2P Request Redirect, disable 4725 Translation Blocking, and leave Source 4726 Validation unchanged from whatever power-up 4727 or firmware set it to. 4728 4729 Note: this may remove isolation between devices 4730 and may put more devices in an IOMMU group. 4731 force_floating [S390] Force usage of floating interrupts. 4732 nomio [S390] Do not use MIO instructions. 4733 norid [S390] ignore the RID field and force use of 4734 one PCI domain per PCI function 4735 4736 pcie_aspm= [PCIE] Forcibly enable or ignore PCIe Active State Power 4737 Management. 4738 off Don't touch ASPM configuration at all. Leave any 4739 configuration done by firmware unchanged. 4740 force Enable ASPM even on devices that claim not to support it. 4741 WARNING: Forcing ASPM on may cause system lockups. 4742 4743 pcie_ports= [PCIE] PCIe port services handling: 4744 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 4745 even if the platform doesn't give the OS permission to 4746 use them. This may cause conflicts if the platform 4747 also tries to use these services. 4748 dpc-native Use native PCIe service for DPC only. May 4749 cause conflicts if firmware uses AER or DPC. 4750 compat Disable native PCIe services (PME, AER, DPC, PCIe 4751 hotplug). 4752 4753 pcie_port_pm= [PCIE] PCIe port power management handling: 4754 off Disable power management of all PCIe ports 4755 force Forcibly enable power management of all PCIe ports 4756 4757 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 4758 nomsi Do not use MSI for native PCIe PME signaling (this makes 4759 all PCIe root ports use INTx for all services). 4760 4761 pcmv= [HW,PCMCIA] BadgePAD 4 4762 4763 pcp_thp_order= [MM] 4764 Specify the order of the pcp used by THP. 4765 The specified value must be no greater than HPAGE_PMD_ORDER 4766 (default) and greater than 3 (PAGE_ALLOC_COSTLY_ORDER). 4767 4768 pd_ignore_unused 4769 [PM] 4770 Keep all power-domains already enabled by bootloader on, 4771 even if no driver has claimed them. This is useful 4772 for debug and development, but should not be 4773 needed on a platform with proper driver support. 4774 4775 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 4776 boot time. 4777 Format: { 0 | 1 } 4778 See arch/parisc/kernel/pdc_chassis.c 4779 4780 percpu_alloc= [MM,EARLY] 4781 Select which percpu first chunk allocator to use. 4782 Currently supported values are "embed" and "page". 4783 Archs may support subset or none of the selections. 4784 See comments in mm/percpu.c for details on each 4785 allocator. This parameter is primarily for debugging 4786 and performance comparison. 4787 4788 pirq= [SMP,APIC] Manual mp-table setup 4789 See Documentation/arch/x86/i386/IO-APIC.rst. 4790 4791 plip= [PPT,NET] Parallel port network link 4792 Format: { parport<nr> | timid | 0 } 4793 See also Documentation/admin-guide/parport.rst. 4794 4795 pmtmr= [X86] Manual setup of pmtmr I/O Port. 4796 Override pmtimer IOPort with a hex value. 4797 e.g. pmtmr=0x508 4798 4799 pmu_override= [PPC] Override the PMU. 4800 This option takes over the PMU facility, so it is no 4801 longer usable by perf. Setting this option starts the 4802 PMU counters by setting MMCR0 to 0 (the FC bit is 4803 cleared). If a number is given, then MMCR1 is set to 4804 that number, otherwise (e.g., 'pmu_override=on'), MMCR1 4805 remains 0. 4806 4807 pm_debug_messages [SUSPEND,KNL] 4808 Enable suspend/resume debug messages during boot up. 4809 4810 pnp.debug=1 [PNP] 4811 Enable PNP debug messages (depends on the 4812 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 4813 via /sys/module/pnp/parameters/debug. We always show 4814 current resource usage; turning this on also shows 4815 possible settings and some assignment information. 4816 4817 pnpacpi= [ACPI] 4818 { off } 4819 4820 pnpbios= [ISAPNP] 4821 { on | off | curr | res | no-curr | no-res } 4822 4823 pnp_reserve_irq= 4824 [ISAPNP] Exclude IRQs for the autoconfiguration 4825 4826 pnp_reserve_dma= 4827 [ISAPNP] Exclude DMAs for the autoconfiguration 4828 4829 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 4830 Ranges are in pairs (I/O port base and size). 4831 4832 pnp_reserve_mem= 4833 [ISAPNP] Exclude memory regions for the 4834 autoconfiguration. 4835 Ranges are in pairs (memory base and size). 4836 4837 ports= [IP_VS_FTP] IPVS ftp helper module 4838 Default is 21. 4839 Up to 8 (IP_VS_APP_MAX_PORTS) ports 4840 may be specified. 4841 Format: <port>,<port>.... 4842 4843 possible_cpus= [SMP,S390,X86] 4844 Format: <unsigned int> 4845 Set the number of possible CPUs, overriding the 4846 regular discovery mechanisms (such as ACPI/FW, etc). 4847 4848 powersave=off [PPC] This option disables power saving features. 4849 It specifically disables cpuidle and sets the 4850 platform machine description specific power_save 4851 function to NULL. On Idle the CPU just reduces 4852 execution priority. 4853 4854 ppc_strict_facility_enable 4855 [PPC,ENABLE] This option catches any kernel floating point, 4856 Altivec, VSX and SPE outside of regions specifically 4857 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 4858 There is some performance impact when enabling this. 4859 4860 ppc_tm= [PPC,EARLY] 4861 Format: {"off"} 4862 Disable Hardware Transactional Memory 4863 4864 preempt= [KNL] 4865 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC 4866 none - Limited to cond_resched() calls 4867 voluntary - Limited to cond_resched() and might_sleep() calls 4868 full - Any section that isn't explicitly preempt disabled 4869 can be preempted anytime. Tasks will also yield 4870 contended spinlocks (if the critical section isn't 4871 explicitly preempt disabled beyond the lock itself). 4872 4873 print-fatal-signals= 4874 [KNL] debug: print fatal signals 4875 4876 If enabled, warn about various signal handling 4877 related application anomalies: too many signals, 4878 too many POSIX.1 timers, fatal signals causing a 4879 coredump - etc. 4880 4881 If you hit the warning due to signal overflow, 4882 you might want to try "ulimit -i unlimited". 4883 4884 default: off. 4885 4886 printk.always_kmsg_dump= 4887 Trigger kmsg_dump for cases other than kernel oops or 4888 panics 4889 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 4890 default: disabled 4891 4892 printk.console_no_auto_verbose= 4893 Disable console loglevel raise on oops, panic 4894 or lockdep-detected issues (only if lock debug is on). 4895 With an exception to setups with low baudrate on 4896 serial console, keeping this 0 is a good choice 4897 in order to provide more debug information. 4898 Format: <bool> 4899 default: 0 (auto_verbose is enabled) 4900 4901 printk.debug_non_panic_cpus= 4902 Allows storing messages from non-panic CPUs into 4903 the printk log buffer during panic(). They are 4904 flushed to consoles by the panic-CPU on 4905 a best-effort basis. 4906 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 4907 Default: disabled 4908 4909 printk.devkmsg={on,off,ratelimit} 4910 Control writing to /dev/kmsg. 4911 on - unlimited logging to /dev/kmsg from userspace 4912 off - logging to /dev/kmsg disabled 4913 ratelimit - ratelimit the logging 4914 Default: ratelimit 4915 4916 printk.time= Show timing data prefixed to each printk message line 4917 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 4918 4919 proc_mem.force_override= [KNL] 4920 Format: {always | ptrace | never} 4921 Traditionally /proc/pid/mem allows memory permissions to be 4922 overridden without restrictions. This option may be set to 4923 restrict that. Can be one of: 4924 - 'always': traditional behavior always allows mem overrides. 4925 - 'ptrace': only allow mem overrides for active ptracers. 4926 - 'never': never allow mem overrides. 4927 If not specified, default is the CONFIG_PROC_MEM_* choice. 4928 4929 processor.max_cstate= [HW,ACPI] 4930 Limit processor to maximum C-state 4931 max_cstate=9 overrides any DMI blacklist limit. 4932 4933 processor.nocst [HW,ACPI] 4934 Ignore the _CST method to determine C-states, 4935 instead using the legacy FADT method 4936 4937 profile= [KNL] Enable kernel profiling via /proc/profile 4938 Format: [<profiletype>,]<number> 4939 Param: <profiletype>: "schedule" or "kvm" 4940 [defaults to kernel profiling] 4941 Param: "schedule" - profile schedule points. 4942 Param: "kvm" - profile VM exits. 4943 Param: <number> - step/bucket size as a power of 2 for 4944 statistical time based profiling. 4945 4946 prompt_ramdisk= [RAM] [Deprecated] 4947 4948 prot_virt= [S390] enable hosting protected virtual machines 4949 isolated from the hypervisor (if hardware supports 4950 that). If enabled, the default kernel base address 4951 might be overridden even when Kernel Address Space 4952 Layout Randomization is disabled. 4953 Format: <bool> 4954 4955 psi= [KNL] Enable or disable pressure stall information 4956 tracking. 4957 Format: <bool> 4958 4959 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 4960 probe for; one of (bare|imps|exps|lifebook|any). 4961 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 4962 per second. 4963 psmouse.resetafter= [HW,MOUSE] 4964 Try to reset the device after so many bad packets 4965 (0 = never). 4966 psmouse.resolution= 4967 [HW,MOUSE] Set desired mouse resolution, in dpi. 4968 psmouse.smartscroll= 4969 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 4970 0 = disabled, 1 = enabled (default). 4971 4972 pstore.backend= Specify the name of the pstore backend to use 4973 4974 pti= [X86-64] Control Page Table Isolation of user and 4975 kernel address spaces. Disabling this feature 4976 removes hardening, but improves performance of 4977 system calls and interrupts. 4978 4979 on - unconditionally enable 4980 off - unconditionally disable 4981 auto - kernel detects whether your CPU model is 4982 vulnerable to issues that PTI mitigates 4983 4984 Not specifying this option is equivalent to pti=auto. 4985 4986 pty.legacy_count= 4987 [KNL] Number of legacy pty's. Overwrites compiled-in 4988 default number. 4989 4990 quiet [KNL,EARLY] Disable most log messages 4991 4992 r128= [HW,DRM] 4993 4994 radix_hcall_invalidate=on [PPC/PSERIES] 4995 Disable RADIX GTSE feature and use hcall for TLB 4996 invalidate. 4997 4998 raid= [HW,RAID] 4999 See Documentation/admin-guide/md.rst. 5000 5001 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 5002 See Documentation/admin-guide/blockdev/ramdisk.rst. 5003 5004 ramdisk_start= [RAM] RAM disk image start address 5005 5006 random.trust_cpu=off 5007 [KNL,EARLY] Disable trusting the use of the CPU's 5008 random number generator (if available) to 5009 initialize the kernel's RNG. 5010 5011 random.trust_bootloader=off 5012 [KNL,EARLY] Disable trusting the use of the a seed 5013 passed by the bootloader (if available) to 5014 initialize the kernel's RNG. 5015 5016 randomize_kstack_offset= 5017 [KNL,EARLY] Enable or disable kernel stack offset 5018 randomization, which provides roughly 5 bits of 5019 entropy, frustrating memory corruption attacks 5020 that depend on stack address determinism or 5021 cross-syscall address exposures. This is only 5022 available on architectures that have defined 5023 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET. 5024 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 5025 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT. 5026 5027 ras=option[,option,...] [KNL] RAS-specific options 5028 5029 cec_disable [X86] 5030 Disable the Correctable Errors Collector, 5031 see CONFIG_RAS_CEC help text. 5032 5033 rcu_nocbs[=cpu-list] 5034 [KNL] The optional argument is a cpu list, 5035 as described above. 5036 5037 In kernels built with CONFIG_RCU_NOCB_CPU=y, 5038 enable the no-callback CPU mode, which prevents 5039 such CPUs' callbacks from being invoked in 5040 softirq context. Invocation of such CPUs' RCU 5041 callbacks will instead be offloaded to "rcuox/N" 5042 kthreads created for that purpose, where "x" is 5043 "p" for RCU-preempt, "s" for RCU-sched, and "g" 5044 for the kthreads that mediate grace periods; and 5045 "N" is the CPU number. This reduces OS jitter on 5046 the offloaded CPUs, which can be useful for HPC 5047 and real-time workloads. It can also improve 5048 energy efficiency for asymmetric multiprocessors. 5049 5050 If a cpulist is passed as an argument, the specified 5051 list of CPUs is set to no-callback mode from boot. 5052 5053 Otherwise, if the '=' sign and the cpulist 5054 arguments are omitted, no CPU will be set to 5055 no-callback mode from boot but the mode may be 5056 toggled at runtime via cpusets. 5057 5058 Note that this argument takes precedence over 5059 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option. 5060 5061 rcu_nocb_poll [KNL] 5062 Rather than requiring that offloaded CPUs 5063 (specified by rcu_nocbs= above) explicitly 5064 awaken the corresponding "rcuoN" kthreads, 5065 make these kthreads poll for callbacks. 5066 This improves the real-time response for the 5067 offloaded CPUs by relieving them of the need to 5068 wake up the corresponding kthread, but degrades 5069 energy efficiency by requiring that the kthreads 5070 periodically wake up to do the polling. 5071 5072 rcutree.blimit= [KNL] 5073 Set maximum number of finished RCU callbacks to 5074 process in one batch. 5075 5076 rcutree.csd_lock_suppress_rcu_stall= [KNL] 5077 Do only a one-line RCU CPU stall warning when 5078 there is an ongoing too-long CSD-lock wait. 5079 5080 rcutree.do_rcu_barrier= [KNL] 5081 Request a call to rcu_barrier(). This is 5082 throttled so that userspace tests can safely 5083 hammer on the sysfs variable if they so choose. 5084 If triggered before the RCU grace-period machinery 5085 is fully active, this will error out with EAGAIN. 5086 5087 rcutree.dump_tree= [KNL] 5088 Dump the structure of the rcu_node combining tree 5089 out at early boot. This is used for diagnostic 5090 purposes, to verify correct tree setup. 5091 5092 rcutree.gp_cleanup_delay= [KNL] 5093 Set the number of jiffies to delay each step of 5094 RCU grace-period cleanup. 5095 5096 rcutree.gp_init_delay= [KNL] 5097 Set the number of jiffies to delay each step of 5098 RCU grace-period initialization. 5099 5100 rcutree.gp_preinit_delay= [KNL] 5101 Set the number of jiffies to delay each step of 5102 RCU grace-period pre-initialization, that is, 5103 the propagation of recent CPU-hotplug changes up 5104 the rcu_node combining tree. 5105 5106 rcutree.jiffies_till_first_fqs= [KNL] 5107 Set delay from grace-period initialization to 5108 first attempt to force quiescent states. 5109 Units are jiffies, minimum value is zero, 5110 and maximum value is HZ. 5111 5112 rcutree.jiffies_till_next_fqs= [KNL] 5113 Set delay between subsequent attempts to force 5114 quiescent states. Units are jiffies, minimum 5115 value is one, and maximum value is HZ. 5116 5117 rcutree.jiffies_till_sched_qs= [KNL] 5118 Set required age in jiffies for a 5119 given grace period before RCU starts 5120 soliciting quiescent-state help from 5121 rcu_note_context_switch() and cond_resched(). 5122 If not specified, the kernel will calculate 5123 a value based on the most recent settings 5124 of rcutree.jiffies_till_first_fqs 5125 and rcutree.jiffies_till_next_fqs. 5126 This calculated value may be viewed in 5127 rcutree.jiffies_to_sched_qs. Any attempt to set 5128 rcutree.jiffies_to_sched_qs will be cheerfully 5129 overwritten. 5130 5131 rcutree.kthread_prio= [KNL,BOOT] 5132 Set the SCHED_FIFO priority of the RCU per-CPU 5133 kthreads (rcuc/N). This value is also used for 5134 the priority of the RCU boost threads (rcub/N) 5135 and for the RCU grace-period kthreads (rcu_bh, 5136 rcu_preempt, and rcu_sched). If RCU_BOOST is 5137 set, valid values are 1-99 and the default is 1 5138 (the least-favored priority). Otherwise, when 5139 RCU_BOOST is not set, valid values are 0-99 and 5140 the default is zero (non-realtime operation). 5141 When RCU_NOCB_CPU is set, also adjust the 5142 priority of NOCB callback kthreads. 5143 5144 rcutree.nocb_nobypass_lim_per_jiffy= [KNL] 5145 On callback-offloaded (rcu_nocbs) CPUs, 5146 RCU reduces the lock contention that would 5147 otherwise be caused by callback floods through 5148 use of the ->nocb_bypass list. However, in the 5149 common non-flooded case, RCU queues directly to 5150 the main ->cblist in order to avoid the extra 5151 overhead of the ->nocb_bypass list and its lock. 5152 But if there are too many callbacks queued during 5153 a single jiffy, RCU pre-queues the callbacks into 5154 the ->nocb_bypass queue. The definition of "too 5155 many" is supplied by this kernel boot parameter. 5156 5157 rcutree.nohz_full_patience_delay= [KNL] 5158 On callback-offloaded (rcu_nocbs) CPUs, avoid 5159 disturbing RCU unless the grace period has 5160 reached the specified age in milliseconds. 5161 Defaults to zero. Large values will be capped 5162 at five seconds. All values will be rounded down 5163 to the nearest value representable by jiffies. 5164 5165 rcutree.qhimark= [KNL] 5166 Set threshold of queued RCU callbacks beyond which 5167 batch limiting is disabled. 5168 5169 rcutree.qlowmark= [KNL] 5170 Set threshold of queued RCU callbacks below which 5171 batch limiting is re-enabled. 5172 5173 rcutree.qovld= [KNL] 5174 Set threshold of queued RCU callbacks beyond which 5175 RCU's force-quiescent-state scan will aggressively 5176 enlist help from cond_resched() and sched IPIs to 5177 help CPUs more quickly reach quiescent states. 5178 Set to less than zero to make this be set based 5179 on rcutree.qhimark at boot time and to zero to 5180 disable more aggressive help enlistment. 5181 5182 rcutree.rcu_delay_page_cache_fill_msec= [KNL] 5183 Set the page-cache refill delay (in milliseconds) 5184 in response to low-memory conditions. The range 5185 of permitted values is in the range 0:100000. 5186 5187 rcutree.rcu_divisor= [KNL] 5188 Set the shift-right count to use to compute 5189 the callback-invocation batch limit bl from 5190 the number of callbacks queued on this CPU. 5191 The result will be bounded below by the value of 5192 the rcutree.blimit kernel parameter. Every bl 5193 callbacks, the softirq handler will exit in 5194 order to allow the CPU to do other work. 5195 5196 Please note that this callback-invocation batch 5197 limit applies only to non-offloaded callback 5198 invocation. Offloaded callbacks are instead 5199 invoked in the context of an rcuoc kthread, which 5200 scheduler will preempt as it does any other task. 5201 5202 rcutree.rcu_fanout_exact= [KNL] 5203 Disable autobalancing of the rcu_node combining 5204 tree. This is used by rcutorture, and might 5205 possibly be useful for architectures having high 5206 cache-to-cache transfer latencies. 5207 5208 rcutree.rcu_fanout_leaf= [KNL] 5209 Change the number of CPUs assigned to each 5210 leaf rcu_node structure. Useful for very 5211 large systems, which will choose the value 64, 5212 and for NUMA systems with large remote-access 5213 latencies, which will choose a value aligned 5214 with the appropriate hardware boundaries. 5215 5216 rcutree.rcu_min_cached_objs= [KNL] 5217 Minimum number of objects which are cached and 5218 maintained per one CPU. Object size is equal 5219 to PAGE_SIZE. The cache allows to reduce the 5220 pressure to page allocator, also it makes the 5221 whole algorithm to behave better in low memory 5222 condition. 5223 5224 rcutree.rcu_nocb_gp_stride= [KNL] 5225 Set the number of NOCB callback kthreads in 5226 each group, which defaults to the square root 5227 of the number of CPUs. Larger numbers reduce 5228 the wakeup overhead on the global grace-period 5229 kthread, but increases that same overhead on 5230 each group's NOCB grace-period kthread. 5231 5232 rcutree.rcu_kick_kthreads= [KNL] 5233 Cause the grace-period kthread to get an extra 5234 wake_up() if it sleeps three times longer than 5235 it should at force-quiescent-state time. 5236 This wake_up() will be accompanied by a 5237 WARN_ONCE() splat and an ftrace_dump(). 5238 5239 rcutree.rcu_resched_ns= [KNL] 5240 Limit the time spend invoking a batch of RCU 5241 callbacks to the specified number of nanoseconds. 5242 By default, this limit is checked only once 5243 every 32 callbacks in order to limit the pain 5244 inflicted by local_clock() overhead. 5245 5246 rcutree.rcu_unlock_delay= [KNL] 5247 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 5248 this specifies an rcu_read_unlock()-time delay 5249 in microseconds. This defaults to zero. 5250 Larger delays increase the probability of 5251 catching RCU pointer leaks, that is, buggy use 5252 of RCU-protected pointers after the relevant 5253 rcu_read_unlock() has completed. 5254 5255 rcutree.sysrq_rcu= [KNL] 5256 Commandeer a sysrq key to dump out Tree RCU's 5257 rcu_node tree with an eye towards determining 5258 why a new grace period has not yet started. 5259 5260 rcutree.use_softirq= [KNL] 5261 If set to zero, move all RCU_SOFTIRQ processing to 5262 per-CPU rcuc kthreads. Defaults to a non-zero 5263 value, meaning that RCU_SOFTIRQ is used by default. 5264 Specify rcutree.use_softirq=0 to use rcuc kthreads. 5265 5266 But note that CONFIG_PREEMPT_RT=y kernels disable 5267 this kernel boot parameter, forcibly setting it 5268 to zero. 5269 5270 rcutree.enable_rcu_lazy= [KNL] 5271 To save power, batch RCU callbacks and flush after 5272 delay, memory pressure or callback list growing too 5273 big. 5274 5275 rcutree.rcu_normal_wake_from_gp= [KNL] 5276 Reduces a latency of synchronize_rcu() call. This approach 5277 maintains its own track of synchronize_rcu() callers, so it 5278 does not interact with regular callbacks because it does not 5279 use a call_rcu[_hurry]() path. Please note, this is for a 5280 normal grace period. 5281 5282 How to enable it: 5283 5284 echo 1 > /sys/module/rcutree/parameters/rcu_normal_wake_from_gp 5285 or pass a boot parameter "rcutree.rcu_normal_wake_from_gp=1" 5286 5287 Default is 0. 5288 5289 rcuscale.gp_async= [KNL] 5290 Measure performance of asynchronous 5291 grace-period primitives such as call_rcu(). 5292 5293 rcuscale.gp_async_max= [KNL] 5294 Specify the maximum number of outstanding 5295 callbacks per writer thread. When a writer 5296 thread exceeds this limit, it invokes the 5297 corresponding flavor of rcu_barrier() to allow 5298 previously posted callbacks to drain. 5299 5300 rcuscale.gp_exp= [KNL] 5301 Measure performance of expedited synchronous 5302 grace-period primitives. 5303 5304 rcuscale.holdoff= [KNL] 5305 Set test-start holdoff period. The purpose of 5306 this parameter is to delay the start of the 5307 test until boot completes in order to avoid 5308 interference. 5309 5310 rcuscale.kfree_by_call_rcu= [KNL] 5311 In kernels built with CONFIG_RCU_LAZY=y, test 5312 call_rcu() instead of kfree_rcu(). 5313 5314 rcuscale.kfree_mult= [KNL] 5315 Instead of allocating an object of size kfree_obj, 5316 allocate one of kfree_mult * sizeof(kfree_obj). 5317 Defaults to 1. 5318 5319 rcuscale.kfree_rcu_test= [KNL] 5320 Set to measure performance of kfree_rcu() flooding. 5321 5322 rcuscale.kfree_rcu_test_double= [KNL] 5323 Test the double-argument variant of kfree_rcu(). 5324 If this parameter has the same value as 5325 rcuscale.kfree_rcu_test_single, both the single- 5326 and double-argument variants are tested. 5327 5328 rcuscale.kfree_rcu_test_single= [KNL] 5329 Test the single-argument variant of kfree_rcu(). 5330 If this parameter has the same value as 5331 rcuscale.kfree_rcu_test_double, both the single- 5332 and double-argument variants are tested. 5333 5334 rcuscale.kfree_nthreads= [KNL] 5335 The number of threads running loops of kfree_rcu(). 5336 5337 rcuscale.kfree_alloc_num= [KNL] 5338 Number of allocations and frees done in an iteration. 5339 5340 rcuscale.kfree_loops= [KNL] 5341 Number of loops doing rcuscale.kfree_alloc_num number 5342 of allocations and frees. 5343 5344 rcuscale.minruntime= [KNL] 5345 Set the minimum test run time in seconds. This 5346 does not affect the data-collection interval, 5347 but instead allows better measurement of things 5348 like CPU consumption. 5349 5350 rcuscale.nreaders= [KNL] 5351 Set number of RCU readers. The value -1 selects 5352 N, where N is the number of CPUs. A value 5353 "n" less than -1 selects N-n+1, where N is again 5354 the number of CPUs. For example, -2 selects N 5355 (the number of CPUs), -3 selects N+1, and so on. 5356 A value of "n" less than or equal to -N selects 5357 a single reader. 5358 5359 rcuscale.nwriters= [KNL] 5360 Set number of RCU writers. The values operate 5361 the same as for rcuscale.nreaders. 5362 N, where N is the number of CPUs 5363 5364 rcuscale.scale_type= [KNL] 5365 Specify the RCU implementation to test. 5366 5367 rcuscale.shutdown= [KNL] 5368 Shut the system down after performance tests 5369 complete. This is useful for hands-off automated 5370 testing. 5371 5372 rcuscale.verbose= [KNL] 5373 Enable additional printk() statements. 5374 5375 rcuscale.writer_holdoff= [KNL] 5376 Write-side holdoff between grace periods, 5377 in microseconds. The default of zero says 5378 no holdoff. 5379 5380 rcuscale.writer_holdoff_jiffies= [KNL] 5381 Additional write-side holdoff between grace 5382 periods, but in jiffies. The default of zero 5383 says no holdoff. 5384 5385 rcutorture.fqs_duration= [KNL] 5386 Set duration of force_quiescent_state bursts 5387 in microseconds. 5388 5389 rcutorture.fqs_holdoff= [KNL] 5390 Set holdoff time within force_quiescent_state bursts 5391 in microseconds. 5392 5393 rcutorture.fqs_stutter= [KNL] 5394 Set wait time between force_quiescent_state bursts 5395 in seconds. 5396 5397 rcutorture.fwd_progress= [KNL] 5398 Specifies the number of kthreads to be used 5399 for RCU grace-period forward-progress testing 5400 for the types of RCU supporting this notion. 5401 Defaults to 1 kthread, values less than zero or 5402 greater than the number of CPUs cause the number 5403 of CPUs to be used. 5404 5405 rcutorture.fwd_progress_div= [KNL] 5406 Specify the fraction of a CPU-stall-warning 5407 period to do tight-loop forward-progress testing. 5408 5409 rcutorture.fwd_progress_holdoff= [KNL] 5410 Number of seconds to wait between successive 5411 forward-progress tests. 5412 5413 rcutorture.fwd_progress_need_resched= [KNL] 5414 Enclose cond_resched() calls within checks for 5415 need_resched() during tight-loop forward-progress 5416 testing. 5417 5418 rcutorture.gp_cond= [KNL] 5419 Use conditional/asynchronous update-side 5420 primitives, if available. 5421 5422 rcutorture.gp_exp= [KNL] 5423 Use expedited update-side primitives, if available. 5424 5425 rcutorture.gp_normal= [KNL] 5426 Use normal (non-expedited) asynchronous 5427 update-side primitives, if available. 5428 5429 rcutorture.gp_sync= [KNL] 5430 Use normal (non-expedited) synchronous 5431 update-side primitives, if available. If all 5432 of rcutorture.gp_cond=, rcutorture.gp_exp=, 5433 rcutorture.gp_normal=, and rcutorture.gp_sync= 5434 are zero, rcutorture acts as if is interpreted 5435 they are all non-zero. 5436 5437 rcutorture.irqreader= [KNL] 5438 Run RCU readers from irq handlers, or, more 5439 accurately, from a timer handler. Not all RCU 5440 flavors take kindly to this sort of thing. 5441 5442 rcutorture.leakpointer= [KNL] 5443 Leak an RCU-protected pointer out of the reader. 5444 This can of course result in splats, and is 5445 intended to test the ability of things like 5446 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 5447 such leaks. 5448 5449 rcutorture.n_barrier_cbs= [KNL] 5450 Set callbacks/threads for rcu_barrier() testing. 5451 5452 rcutorture.nfakewriters= [KNL] 5453 Set number of concurrent RCU writers. These just 5454 stress RCU, they don't participate in the actual 5455 test, hence the "fake". 5456 5457 rcutorture.nocbs_nthreads= [KNL] 5458 Set number of RCU callback-offload togglers. 5459 Zero (the default) disables toggling. 5460 5461 rcutorture.nocbs_toggle= [KNL] 5462 Set the delay in milliseconds between successive 5463 callback-offload toggling attempts. 5464 5465 rcutorture.nreaders= [KNL] 5466 Set number of RCU readers. The value -1 selects 5467 N-1, where N is the number of CPUs. A value 5468 "n" less than -1 selects N-n-2, where N is again 5469 the number of CPUs. For example, -2 selects N 5470 (the number of CPUs), -3 selects N+1, and so on. 5471 5472 rcutorture.object_debug= [KNL] 5473 Enable debug-object double-call_rcu() testing. 5474 5475 rcutorture.onoff_holdoff= [KNL] 5476 Set time (s) after boot for CPU-hotplug testing. 5477 5478 rcutorture.onoff_interval= [KNL] 5479 Set time (jiffies) between CPU-hotplug operations, 5480 or zero to disable CPU-hotplug testing. 5481 5482 rcutorture.read_exit= [KNL] 5483 Set the number of read-then-exit kthreads used 5484 to test the interaction of RCU updaters and 5485 task-exit processing. 5486 5487 rcutorture.read_exit_burst= [KNL] 5488 The number of times in a given read-then-exit 5489 episode that a set of read-then-exit kthreads 5490 is spawned. 5491 5492 rcutorture.read_exit_delay= [KNL] 5493 The delay, in seconds, between successive 5494 read-then-exit testing episodes. 5495 5496 rcutorture.shuffle_interval= [KNL] 5497 Set task-shuffle interval (s). Shuffling tasks 5498 allows some CPUs to go into dyntick-idle mode 5499 during the rcutorture test. 5500 5501 rcutorture.shutdown_secs= [KNL] 5502 Set time (s) after boot system shutdown. This 5503 is useful for hands-off automated testing. 5504 5505 rcutorture.stall_cpu= [KNL] 5506 Duration of CPU stall (s) to test RCU CPU stall 5507 warnings, zero to disable. 5508 5509 rcutorture.stall_cpu_block= [KNL] 5510 Sleep while stalling if set. This will result 5511 in warnings from preemptible RCU in addition to 5512 any other stall-related activity. Note that 5513 in kernels built with CONFIG_PREEMPTION=n and 5514 CONFIG_PREEMPT_COUNT=y, this parameter will 5515 cause the CPU to pass through a quiescent state. 5516 Given CONFIG_PREEMPTION=n, this will suppress 5517 RCU CPU stall warnings, but will instead result 5518 in scheduling-while-atomic splats. 5519 5520 Use of this module parameter results in splats. 5521 5522 5523 rcutorture.stall_cpu_holdoff= [KNL] 5524 Time to wait (s) after boot before inducing stall. 5525 5526 rcutorture.stall_cpu_irqsoff= [KNL] 5527 Disable interrupts while stalling if set, but only 5528 on the first stall in the set. 5529 5530 rcutorture.stall_cpu_repeat= [KNL] 5531 Number of times to repeat the stall sequence, 5532 so that rcutorture.stall_cpu_repeat=3 will result 5533 in four stall sequences. 5534 5535 rcutorture.stall_gp_kthread= [KNL] 5536 Duration (s) of forced sleep within RCU 5537 grace-period kthread to test RCU CPU stall 5538 warnings, zero to disable. If both stall_cpu 5539 and stall_gp_kthread are specified, the 5540 kthread is starved first, then the CPU. 5541 5542 rcutorture.stat_interval= [KNL] 5543 Time (s) between statistics printk()s. 5544 5545 rcutorture.stutter= [KNL] 5546 Time (s) to stutter testing, for example, specifying 5547 five seconds causes the test to run for five seconds, 5548 wait for five seconds, and so on. This tests RCU's 5549 ability to transition abruptly to and from idle. 5550 5551 rcutorture.test_boost= [KNL] 5552 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 5553 "Maybe" means test if the RCU implementation 5554 under test support RCU priority boosting. 5555 5556 rcutorture.test_boost_duration= [KNL] 5557 Duration (s) of each individual boost test. 5558 5559 rcutorture.test_boost_interval= [KNL] 5560 Interval (s) between each boost test. 5561 5562 rcutorture.test_no_idle_hz= [KNL] 5563 Test RCU's dyntick-idle handling. See also the 5564 rcutorture.shuffle_interval parameter. 5565 5566 rcutorture.torture_type= [KNL] 5567 Specify the RCU implementation to test. 5568 5569 rcutorture.verbose= [KNL] 5570 Enable additional printk() statements. 5571 5572 rcupdate.rcu_boot_end_delay= [KNL] 5573 Minimum time in milliseconds from the start of boot 5574 that must elapse before the boot sequence can be marked 5575 complete from RCU's perspective, after which RCU's 5576 behavior becomes more relaxed. The default value is also 5577 configurable via CONFIG_RCU_BOOT_END_DELAY. 5578 Userspace can also mark the boot as completed 5579 sooner by writing the time in milliseconds, say once 5580 userspace considers the system as booted, to: 5581 /sys/module/rcupdate/parameters/rcu_boot_end_delay 5582 Or even just writing a value of 0 to this sysfs node. 5583 The sysfs node can also be used to extend the delay 5584 to be larger than the default, assuming the marking 5585 of boot complete has not yet occurred. 5586 5587 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 5588 Dump ftrace buffer after reporting RCU CPU 5589 stall warning. 5590 5591 rcupdate.rcu_cpu_stall_notifiers= [KNL] 5592 Provide RCU CPU stall notifiers, but see the 5593 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig 5594 option's help text. TL;DR: You almost certainly 5595 do not want rcupdate.rcu_cpu_stall_notifiers. 5596 5597 rcupdate.rcu_cpu_stall_suppress= [KNL] 5598 Suppress RCU CPU stall warning messages. 5599 5600 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 5601 Suppress RCU CPU stall warning messages and 5602 rcutorture writer stall warnings that occur 5603 during early boot, that is, during the time 5604 before the init task is spawned. 5605 5606 rcupdate.rcu_cpu_stall_timeout= [KNL] 5607 Set timeout for RCU CPU stall warning messages. 5608 The value is in seconds and the maximum allowed 5609 value is 300 seconds. 5610 5611 rcupdate.rcu_exp_cpu_stall_timeout= [KNL] 5612 Set timeout for expedited RCU CPU stall warning 5613 messages. The value is in milliseconds 5614 and the maximum allowed value is 21000 5615 milliseconds. Please note that this value is 5616 adjusted to an arch timer tick resolution. 5617 Setting this to zero causes the value from 5618 rcupdate.rcu_cpu_stall_timeout to be used (after 5619 conversion from seconds to milliseconds). 5620 5621 rcupdate.rcu_cpu_stall_cputime= [KNL] 5622 Provide statistics on the cputime and count of 5623 interrupts and tasks during the sampling period. For 5624 multiple continuous RCU stalls, all sampling periods 5625 begin at half of the first RCU stall timeout. 5626 5627 rcupdate.rcu_exp_stall_task_details= [KNL] 5628 Print stack dumps of any tasks blocking the 5629 current expedited RCU grace period during an 5630 expedited RCU CPU stall warning. 5631 5632 rcupdate.rcu_expedited= [KNL] 5633 Use expedited grace-period primitives, for 5634 example, synchronize_rcu_expedited() instead 5635 of synchronize_rcu(). This reduces latency, 5636 but can increase CPU utilization, degrade 5637 real-time latency, and degrade energy efficiency. 5638 No effect on CONFIG_TINY_RCU kernels. 5639 5640 rcupdate.rcu_normal= [KNL] 5641 Use only normal grace-period primitives, 5642 for example, synchronize_rcu() instead of 5643 synchronize_rcu_expedited(). This improves 5644 real-time latency, CPU utilization, and 5645 energy efficiency, but can expose users to 5646 increased grace-period latency. This parameter 5647 overrides rcupdate.rcu_expedited. No effect on 5648 CONFIG_TINY_RCU kernels. 5649 5650 rcupdate.rcu_normal_after_boot= [KNL] 5651 Once boot has completed (that is, after 5652 rcu_end_inkernel_boot() has been invoked), use 5653 only normal grace-period primitives. No effect 5654 on CONFIG_TINY_RCU kernels. 5655 5656 But note that CONFIG_PREEMPT_RT=y kernels enables 5657 this kernel boot parameter, forcibly setting 5658 it to the value one, that is, converting any 5659 post-boot attempt at an expedited RCU grace 5660 period to instead use normal non-expedited 5661 grace-period processing. 5662 5663 rcupdate.rcu_task_collapse_lim= [KNL] 5664 Set the maximum number of callbacks present 5665 at the beginning of a grace period that allows 5666 the RCU Tasks flavors to collapse back to using 5667 a single callback queue. This switching only 5668 occurs when rcupdate.rcu_task_enqueue_lim is 5669 set to the default value of -1. 5670 5671 rcupdate.rcu_task_contend_lim= [KNL] 5672 Set the minimum number of callback-queuing-time 5673 lock-contention events per jiffy required to 5674 cause the RCU Tasks flavors to switch to per-CPU 5675 callback queuing. This switching only occurs 5676 when rcupdate.rcu_task_enqueue_lim is set to 5677 the default value of -1. 5678 5679 rcupdate.rcu_task_enqueue_lim= [KNL] 5680 Set the number of callback queues to use for the 5681 RCU Tasks family of RCU flavors. The default 5682 of -1 allows this to be automatically (and 5683 dynamically) adjusted. This parameter is intended 5684 for use in testing. 5685 5686 rcupdate.rcu_task_ipi_delay= [KNL] 5687 Set time in jiffies during which RCU tasks will 5688 avoid sending IPIs, starting with the beginning 5689 of a given grace period. Setting a large 5690 number avoids disturbing real-time workloads, 5691 but lengthens grace periods. 5692 5693 rcupdate.rcu_task_lazy_lim= [KNL] 5694 Number of callbacks on a given CPU that will 5695 cancel laziness on that CPU. Use -1 to disable 5696 cancellation of laziness, but be advised that 5697 doing so increases the danger of OOM due to 5698 callback flooding. 5699 5700 rcupdate.rcu_task_stall_info= [KNL] 5701 Set initial timeout in jiffies for RCU task stall 5702 informational messages, which give some indication 5703 of the problem for those not patient enough to 5704 wait for ten minutes. Informational messages are 5705 only printed prior to the stall-warning message 5706 for a given grace period. Disable with a value 5707 less than or equal to zero. Defaults to ten 5708 seconds. A change in value does not take effect 5709 until the beginning of the next grace period. 5710 5711 rcupdate.rcu_task_stall_info_mult= [KNL] 5712 Multiplier for time interval between successive 5713 RCU task stall informational messages for a given 5714 RCU tasks grace period. This value is clamped 5715 to one through ten, inclusive. It defaults to 5716 the value three, so that the first informational 5717 message is printed 10 seconds into the grace 5718 period, the second at 40 seconds, the third at 5719 160 seconds, and then the stall warning at 600 5720 seconds would prevent a fourth at 640 seconds. 5721 5722 rcupdate.rcu_task_stall_timeout= [KNL] 5723 Set timeout in jiffies for RCU task stall 5724 warning messages. Disable with a value less 5725 than or equal to zero. Defaults to ten minutes. 5726 A change in value does not take effect until 5727 the beginning of the next grace period. 5728 5729 rcupdate.rcu_tasks_lazy_ms= [KNL] 5730 Set timeout in milliseconds RCU Tasks asynchronous 5731 callback batching for call_rcu_tasks(). 5732 A negative value will take the default. A value 5733 of zero will disable batching. Batching is 5734 always disabled for synchronize_rcu_tasks(). 5735 5736 rcupdate.rcu_tasks_trace_lazy_ms= [KNL] 5737 Set timeout in milliseconds RCU Tasks 5738 Trace asynchronous callback batching for 5739 call_rcu_tasks_trace(). A negative value 5740 will take the default. A value of zero will 5741 disable batching. Batching is always disabled 5742 for synchronize_rcu_tasks_trace(). 5743 5744 rcupdate.rcu_self_test= [KNL] 5745 Run the RCU early boot self tests 5746 5747 rdinit= [KNL] 5748 Format: <full_path> 5749 Run specified binary instead of /init from the ramdisk, 5750 used for early userspace startup. See initrd. 5751 5752 rdrand= [X86,EARLY] 5753 force - Override the decision by the kernel to hide the 5754 advertisement of RDRAND support (this affects 5755 certain AMD processors because of buggy BIOS 5756 support, specifically around the suspend/resume 5757 path). 5758 5759 rdt= [HW,X86,RDT] 5760 Turn on/off individual RDT features. List is: 5761 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 5762 mba, smba, bmec. 5763 E.g. to turn on cmt and turn off mba use: 5764 rdt=cmt,!mba 5765 5766 reboot= [KNL] 5767 Format (x86 or x86_64): 5768 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \ 5769 [[,]s[mp]#### \ 5770 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 5771 [[,]f[orce] 5772 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 5773 (prefix with 'panic_' to set mode for panic 5774 reboot only), 5775 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 5776 reboot_force is either force or not specified, 5777 reboot_cpu is s[mp]#### with #### being the processor 5778 to be used for rebooting. 5779 5780 refscale.holdoff= [KNL] 5781 Set test-start holdoff period. The purpose of 5782 this parameter is to delay the start of the 5783 test until boot completes in order to avoid 5784 interference. 5785 5786 refscale.lookup_instances= [KNL] 5787 Number of data elements to use for the forms of 5788 SLAB_TYPESAFE_BY_RCU testing. A negative number 5789 is negated and multiplied by nr_cpu_ids, while 5790 zero specifies nr_cpu_ids. 5791 5792 refscale.loops= [KNL] 5793 Set the number of loops over the synchronization 5794 primitive under test. Increasing this number 5795 reduces noise due to loop start/end overhead, 5796 but the default has already reduced the per-pass 5797 noise to a handful of picoseconds on ca. 2020 5798 x86 laptops. 5799 5800 refscale.nreaders= [KNL] 5801 Set number of readers. The default value of -1 5802 selects N, where N is roughly 75% of the number 5803 of CPUs. A value of zero is an interesting choice. 5804 5805 refscale.nruns= [KNL] 5806 Set number of runs, each of which is dumped onto 5807 the console log. 5808 5809 refscale.readdelay= [KNL] 5810 Set the read-side critical-section duration, 5811 measured in microseconds. 5812 5813 refscale.scale_type= [KNL] 5814 Specify the read-protection implementation to test. 5815 5816 refscale.shutdown= [KNL] 5817 Shut down the system at the end of the performance 5818 test. This defaults to 1 (shut it down) when 5819 refscale is built into the kernel and to 0 (leave 5820 it running) when refscale is built as a module. 5821 5822 refscale.verbose= [KNL] 5823 Enable additional printk() statements. 5824 5825 refscale.verbose_batched= [KNL] 5826 Batch the additional printk() statements. If zero 5827 (the default) or negative, print everything. Otherwise, 5828 print every Nth verbose statement, where N is the value 5829 specified. 5830 5831 regulator_ignore_unused 5832 [REGULATOR] 5833 Prevents regulator framework from disabling regulators 5834 that are unused, due no driver claiming them. This may 5835 be useful for debug and development, but should not be 5836 needed on a platform with proper driver support. 5837 5838 relax_domain_level= 5839 [KNL, SMP] Set scheduler's default relax_domain_level. 5840 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 5841 5842 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 5843 Format: <base1>,<size1>[,<base2>,<size2>,...] 5844 Reserve I/O ports or memory so the kernel won't use 5845 them. If <base> is less than 0x10000, the region 5846 is assumed to be I/O ports; otherwise it is memory. 5847 5848 reserve_mem= [RAM] 5849 Format: nn[KNG]:<align>:<label> 5850 Reserve physical memory and label it with a name that 5851 other subsystems can use to access it. This is typically 5852 used for systems that do not wipe the RAM, and this command 5853 line will try to reserve the same physical memory on 5854 soft reboots. Note, it is not guaranteed to be the same 5855 location. For example, if anything about the system changes 5856 or if booting a different kernel. It can also fail if KASLR 5857 places the kernel at the location of where the RAM reservation 5858 was from a previous boot, the new reservation will be at a 5859 different location. 5860 Any subsystem using this feature must add a way to verify 5861 that the contents of the physical memory is from a previous 5862 boot, as there may be cases where the memory will not be 5863 located at the same location. 5864 5865 The format is size:align:label for example, to request 5866 12 megabytes of 4096 alignment for ramoops: 5867 5868 reserve_mem=12M:4096:oops ramoops.mem_name=oops 5869 5870 reservetop= [X86-32,EARLY] 5871 Format: nn[KMG] 5872 Reserves a hole at the top of the kernel virtual 5873 address space. 5874 5875 reset_devices [KNL] Force drivers to reset the underlying device 5876 during initialization. 5877 5878 resume= [SWSUSP] 5879 Specify the partition device for software suspend 5880 Format: 5881 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 5882 5883 resume_offset= [SWSUSP] 5884 Specify the offset from the beginning of the partition 5885 given by "resume=" at which the swap header is located, 5886 in <PAGE_SIZE> units (needed only for swap files). 5887 See Documentation/power/swsusp-and-swap-files.rst 5888 5889 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 5890 read the resume files 5891 5892 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 5893 Useful for devices that are detected asynchronously 5894 (e.g. USB and MMC devices). 5895 5896 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will 5897 be accessible via /sys/firmware/initrd. 5898 5899 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 5900 Speculative Code Execution with Return Instructions) 5901 vulnerability. 5902 5903 AMD-based UNRET and IBPB mitigations alone do not stop 5904 sibling threads from influencing the predictions of other 5905 sibling threads. For that reason, STIBP is used on pro- 5906 cessors that support it, and mitigate SMT on processors 5907 that don't. 5908 5909 off - no mitigation 5910 auto - automatically select a migitation 5911 auto,nosmt - automatically select a mitigation, 5912 disabling SMT if necessary for 5913 the full mitigation (only on Zen1 5914 and older without STIBP). 5915 ibpb - On AMD, mitigate short speculation 5916 windows on basic block boundaries too. 5917 Safe, highest perf impact. It also 5918 enables STIBP if present. Not suitable 5919 on Intel. 5920 ibpb,nosmt - Like "ibpb" above but will disable SMT 5921 when STIBP is not available. This is 5922 the alternative for systems which do not 5923 have STIBP. 5924 unret - Force enable untrained return thunks, 5925 only effective on AMD f15h-f17h based 5926 systems. 5927 unret,nosmt - Like unret, but will disable SMT when STIBP 5928 is not available. This is the alternative for 5929 systems which do not have STIBP. 5930 5931 Selecting 'auto' will choose a mitigation method at run 5932 time according to the CPU. 5933 5934 Not specifying this option is equivalent to retbleed=auto. 5935 5936 rfkill.default_state= 5937 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 5938 etc. communication is blocked by default. 5939 1 Unblocked. 5940 5941 rfkill.master_switch_mode= 5942 0 The "airplane mode" button does nothing. 5943 1 The "airplane mode" button toggles between everything 5944 blocked and the previous configuration. 5945 2 The "airplane mode" button toggles between everything 5946 blocked and everything unblocked. 5947 5948 ring3mwait=disable 5949 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 5950 CPUs. 5951 5952 riscv_isa_fallback [RISCV,EARLY] 5953 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit 5954 falling back to detecting extension support by parsing 5955 "riscv,isa" property on devicetree systems when the 5956 replacement properties are not found. See the Kconfig 5957 entry for RISCV_ISA_FALLBACK. 5958 5959 ro [KNL] Mount root device read-only on boot 5960 5961 rodata= [KNL,EARLY] 5962 on Mark read-only kernel memory as read-only (default). 5963 off Leave read-only kernel memory writable for debugging. 5964 full Mark read-only kernel memory and aliases as read-only 5965 [arm64] 5966 5967 rockchip.usb_uart 5968 [EARLY] 5969 Enable the uart passthrough on the designated usb port 5970 on Rockchip SoCs. When active, the signals of the 5971 debug-uart get routed to the D+ and D- pins of the usb 5972 port and the regular usb controller gets disabled. 5973 5974 root= [KNL] Root filesystem 5975 Usually this a a block device specifier of some kind, 5976 see the early_lookup_bdev comment in 5977 block/early-lookup.c for details. 5978 Alternatively this can be "ram" for the legacy initial 5979 ramdisk, "nfs" and "cifs" for root on a network file 5980 system, or "mtd" and "ubi" for mounting from raw flash. 5981 5982 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 5983 mount the root filesystem 5984 5985 rootflags= [KNL] Set root filesystem mount option string 5986 5987 rootfstype= [KNL] Set root filesystem type 5988 5989 rootwait [KNL] Wait (indefinitely) for root device to show up. 5990 Useful for devices that are detected asynchronously 5991 (e.g. USB and MMC devices). 5992 5993 rootwait= [KNL] Maximum time (in seconds) to wait for root device 5994 to show up before attempting to mount the root 5995 filesystem. 5996 5997 rproc_mem=nn[KMG][@address] 5998 [KNL,ARM,CMA] Remoteproc physical memory block. 5999 Memory area to be used by remote processor image, 6000 managed by CMA. 6001 6002 rw [KNL] Mount root device read-write on boot 6003 6004 S [KNL] Run init in single mode 6005 6006 s390_iommu= [HW,S390] 6007 Set s390 IOTLB flushing mode 6008 strict 6009 With strict flushing every unmap operation will result 6010 in an IOTLB flush. Default is lazy flushing before 6011 reuse, which is faster. Deprecated, equivalent to 6012 iommu.strict=1. 6013 6014 s390_iommu_aperture= [KNL,S390] 6015 Specifies the size of the per device DMA address space 6016 accessible through the DMA and IOMMU APIs as a decimal 6017 factor of the size of main memory. 6018 The default is 1 meaning that one can concurrently use 6019 as many DMA addresses as physical memory is installed, 6020 if supported by hardware, and thus map all of memory 6021 once. With a value of 2 one can map all of memory twice 6022 and so on. As a special case a factor of 0 imposes no 6023 restrictions other than those given by hardware at the 6024 cost of significant additional memory use for tables. 6025 6026 sa1100ir [NET] 6027 See drivers/net/irda/sa1100_ir.c. 6028 6029 sched_proxy_exec= [KNL] 6030 Enables or disables "proxy execution" style 6031 solution to mutex-based priority inversion. 6032 Format: <bool> 6033 6034 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages. 6035 6036 schedstats= [KNL,X86] Enable or disable scheduled statistics. 6037 Allowed values are enable and disable. This feature 6038 incurs a small amount of overhead in the scheduler 6039 but is useful for debugging and performance tuning. 6040 6041 sched_thermal_decay_shift= 6042 [Deprecated] 6043 [KNL, SMP] Set a decay shift for scheduler thermal 6044 pressure signal. Thermal pressure signal follows the 6045 default decay period of other scheduler pelt 6046 signals(usually 32 ms but configurable). Setting 6047 sched_thermal_decay_shift will left shift the decay 6048 period for the thermal pressure signal by the shift 6049 value. 6050 i.e. with the default pelt decay period of 32 ms 6051 sched_thermal_decay_shift thermal pressure decay pr 6052 1 64 ms 6053 2 128 ms 6054 and so on. 6055 Format: integer between 0 and 10 6056 Default is 0. 6057 6058 scftorture.holdoff= [KNL] 6059 Number of seconds to hold off before starting 6060 test. Defaults to zero for module insertion and 6061 to 10 seconds for built-in smp_call_function() 6062 tests. 6063 6064 scftorture.longwait= [KNL] 6065 Request ridiculously long waits randomly selected 6066 up to the chosen limit in seconds. Zero (the 6067 default) disables this feature. Please note 6068 that requesting even small non-zero numbers of 6069 seconds can result in RCU CPU stall warnings, 6070 softlockup complaints, and so on. 6071 6072 scftorture.nthreads= [KNL] 6073 Number of kthreads to spawn to invoke the 6074 smp_call_function() family of functions. 6075 The default of -1 specifies a number of kthreads 6076 equal to the number of CPUs. 6077 6078 scftorture.onoff_holdoff= [KNL] 6079 Number seconds to wait after the start of the 6080 test before initiating CPU-hotplug operations. 6081 6082 scftorture.onoff_interval= [KNL] 6083 Number seconds to wait between successive 6084 CPU-hotplug operations. Specifying zero (which 6085 is the default) disables CPU-hotplug operations. 6086 6087 scftorture.shutdown_secs= [KNL] 6088 The number of seconds following the start of the 6089 test after which to shut down the system. The 6090 default of zero avoids shutting down the system. 6091 Non-zero values are useful for automated tests. 6092 6093 scftorture.stat_interval= [KNL] 6094 The number of seconds between outputting the 6095 current test statistics to the console. A value 6096 of zero disables statistics output. 6097 6098 scftorture.stutter_cpus= [KNL] 6099 The number of jiffies to wait between each change 6100 to the set of CPUs under test. 6101 6102 scftorture.use_cpus_read_lock= [KNL] 6103 Use use_cpus_read_lock() instead of the default 6104 preempt_disable() to disable CPU hotplug 6105 while invoking one of the smp_call_function*() 6106 functions. 6107 6108 scftorture.verbose= [KNL] 6109 Enable additional printk() statements. 6110 6111 scftorture.weight_single= [KNL] 6112 The probability weighting to use for the 6113 smp_call_function_single() function with a zero 6114 "wait" parameter. A value of -1 selects the 6115 default if all other weights are -1. However, 6116 if at least one weight has some other value, a 6117 value of -1 will instead select a weight of zero. 6118 6119 scftorture.weight_single_wait= [KNL] 6120 The probability weighting to use for the 6121 smp_call_function_single() function with a 6122 non-zero "wait" parameter. See weight_single. 6123 6124 scftorture.weight_many= [KNL] 6125 The probability weighting to use for the 6126 smp_call_function_many() function with a zero 6127 "wait" parameter. See weight_single. 6128 Note well that setting a high probability for 6129 this weighting can place serious IPI load 6130 on the system. 6131 6132 scftorture.weight_many_wait= [KNL] 6133 The probability weighting to use for the 6134 smp_call_function_many() function with a 6135 non-zero "wait" parameter. See weight_single 6136 and weight_many. 6137 6138 scftorture.weight_all= [KNL] 6139 The probability weighting to use for the 6140 smp_call_function_all() function with a zero 6141 "wait" parameter. See weight_single and 6142 weight_many. 6143 6144 scftorture.weight_all_wait= [KNL] 6145 The probability weighting to use for the 6146 smp_call_function_all() function with a 6147 non-zero "wait" parameter. See weight_single 6148 and weight_many. 6149 6150 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate 6151 xtime_lock contention on larger systems, and/or RCU lock 6152 contention on all systems with CONFIG_MAXSMP set. 6153 Format: { "0" | "1" } 6154 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" 6155 1 -- enable. 6156 Note: increases power consumption, thus should only be 6157 enabled if running jitter sensitive (HPC/RT) workloads. 6158 6159 security= [SECURITY] Choose a legacy "major" security module to 6160 enable at boot. This has been deprecated by the 6161 "lsm=" parameter. 6162 6163 selinux= [SELINUX] Disable or enable SELinux at boot time. 6164 Format: { "0" | "1" } 6165 See security/selinux/Kconfig help text. 6166 0 -- disable. 6167 1 -- enable. 6168 Default value is 1. 6169 6170 serialnumber [BUGS=X86-32] 6171 6172 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst 6173 6174 shapers= [NET] 6175 Maximal number of shapers. 6176 6177 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 6178 Limit apic dumping. The parameter defines the maximal 6179 number of local apics being dumped. Also it is possible 6180 to set it to "all" by meaning -- no limit here. 6181 Format: { 1 (default) | 2 | ... | all }. 6182 The parameter valid if only apic=debug or 6183 apic=verbose is specified. 6184 Example: apic=debug show_lapic=all 6185 6186 slab_debug[=options[,slabs][;[options[,slabs]]...] [MM] 6187 Enabling slab_debug allows one to determine the 6188 culprit if slab objects become corrupted. Enabling 6189 slab_debug can create guard zones around objects and 6190 may poison objects when not in use. Also tracks the 6191 last alloc / free. For more information see 6192 Documentation/mm/slub.rst. 6193 (slub_debug legacy name also accepted for now) 6194 6195 slab_max_order= [MM] 6196 Determines the maximum allowed order for slabs. 6197 A high setting may cause OOMs due to memory 6198 fragmentation. For more information see 6199 Documentation/mm/slub.rst. 6200 (slub_max_order legacy name also accepted for now) 6201 6202 slab_merge [MM] 6203 Enable merging of slabs with similar size when the 6204 kernel is built without CONFIG_SLAB_MERGE_DEFAULT. 6205 (slub_merge legacy name also accepted for now) 6206 6207 slab_min_objects= [MM] 6208 The minimum number of objects per slab. SLUB will 6209 increase the slab order up to slab_max_order to 6210 generate a sufficiently large slab able to contain 6211 the number of objects indicated. The higher the number 6212 of objects the smaller the overhead of tracking slabs 6213 and the less frequently locks need to be acquired. 6214 For more information see Documentation/mm/slub.rst. 6215 (slub_min_objects legacy name also accepted for now) 6216 6217 slab_min_order= [MM] 6218 Determines the minimum page order for slabs. Must be 6219 lower or equal to slab_max_order. For more information see 6220 Documentation/mm/slub.rst. 6221 (slub_min_order legacy name also accepted for now) 6222 6223 slab_nomerge [MM] 6224 Disable merging of slabs with similar size. May be 6225 necessary if there is some reason to distinguish 6226 allocs to different slabs, especially in hardened 6227 environments where the risk of heap overflows and 6228 layout control by attackers can usually be 6229 frustrated by disabling merging. This will reduce 6230 most of the exposure of a heap attack to a single 6231 cache (risks via metadata attacks are mostly 6232 unchanged). Debug options disable merging on their 6233 own. 6234 For more information see Documentation/mm/slub.rst. 6235 (slub_nomerge legacy name also accepted for now) 6236 6237 slram= [HW,MTD] 6238 6239 smart2= [HW] 6240 Format: <io1>[,<io2>[,...,<io8>]] 6241 6242 smp.csd_lock_timeout= [KNL] 6243 Specify the period of time in milliseconds 6244 that smp_call_function() and friends will wait 6245 for a CPU to release the CSD lock. This is 6246 useful when diagnosing bugs involving CPUs 6247 disabling interrupts for extended periods 6248 of time. Defaults to 5,000 milliseconds, and 6249 setting a value of zero disables this feature. 6250 This feature may be more efficiently disabled 6251 using the csdlock_debug- kernel parameter. 6252 6253 smp.panic_on_ipistall= [KNL] 6254 If a csd_lock_timeout extends for more than 6255 the specified number of milliseconds, panic the 6256 system. By default, let CSD-lock acquisition 6257 take as long as they take. Specifying 300,000 6258 for this value provides a 5-minute timeout. 6259 6260 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices 6261 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port 6262 smsc-ircc2.ircc_sir= [HW] SIR base I/O port 6263 smsc-ircc2.ircc_fir= [HW] FIR base I/O port 6264 smsc-ircc2.ircc_irq= [HW] IRQ line 6265 smsc-ircc2.ircc_dma= [HW] DMA channel 6266 smsc-ircc2.ircc_transceiver= [HW] Transceiver type: 6267 0: Toshiba Satellite 1800 (GP data pin select) 6268 1: Fast pin select (default) 6269 2: ATC IRMode 6270 6271 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads 6272 (logical CPUs) to use per physical CPU on systems 6273 capable of symmetric multithreading (SMT). Will 6274 be capped to the actual hardware limit. 6275 Format: <integer> 6276 Default: -1 (no limit) 6277 6278 softlockup_panic= 6279 [KNL] Should the soft-lockup detector generate panics. 6280 Format: 0 | 1 6281 6282 A value of 1 instructs the soft-lockup detector 6283 to panic the machine when a soft-lockup occurs. It is 6284 also controlled by the kernel.softlockup_panic sysctl 6285 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 6286 respective build-time switch to that functionality. 6287 6288 softlockup_all_cpu_backtrace= 6289 [KNL] Should the soft-lockup detector generate 6290 backtraces on all cpus. 6291 Format: 0 | 1 6292 6293 sonypi.*= [HW] Sony Programmable I/O Control Device driver 6294 See Documentation/admin-guide/laptops/sonypi.rst 6295 6296 spectre_bhi= [X86] Control mitigation of Branch History Injection 6297 (BHI) vulnerability. This setting affects the 6298 deployment of the HW BHI control and the SW BHB 6299 clearing sequence. 6300 6301 on - (default) Enable the HW or SW mitigation as 6302 needed. This protects the kernel from 6303 both syscalls and VMs. 6304 vmexit - On systems which don't have the HW mitigation 6305 available, enable the SW mitigation on vmexit 6306 ONLY. On such systems, the host kernel is 6307 protected from VM-originated BHI attacks, but 6308 may still be vulnerable to syscall attacks. 6309 off - Disable the mitigation. 6310 6311 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2 6312 (indirect branch speculation) vulnerability. 6313 The default operation protects the kernel from 6314 user space attacks. 6315 6316 on - unconditionally enable, implies 6317 spectre_v2_user=on 6318 off - unconditionally disable, implies 6319 spectre_v2_user=off 6320 auto - kernel detects whether your CPU model is 6321 vulnerable 6322 6323 Selecting 'on' will, and 'auto' may, choose a 6324 mitigation method at run time according to the 6325 CPU, the available microcode, the setting of the 6326 CONFIG_MITIGATION_RETPOLINE configuration option, 6327 and the compiler with which the kernel was built. 6328 6329 Selecting 'on' will also enable the mitigation 6330 against user space to user space task attacks. 6331 Selecting specific mitigation does not force enable 6332 user mitigations. 6333 6334 Selecting 'off' will disable both the kernel and 6335 the user space protections. 6336 6337 Specific mitigations can also be selected manually: 6338 6339 retpoline - replace indirect branches 6340 retpoline,generic - Retpolines 6341 retpoline,lfence - LFENCE; indirect branch 6342 retpoline,amd - alias for retpoline,lfence 6343 eibrs - Enhanced/Auto IBRS 6344 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines 6345 eibrs,lfence - Enhanced/Auto IBRS + LFENCE 6346 ibrs - use IBRS to protect kernel 6347 6348 Not specifying this option is equivalent to 6349 spectre_v2=auto. 6350 6351 spectre_v2_user= 6352 [X86] Control mitigation of Spectre variant 2 6353 (indirect branch speculation) vulnerability between 6354 user space tasks 6355 6356 on - Unconditionally enable mitigations. Is 6357 enforced by spectre_v2=on 6358 6359 off - Unconditionally disable mitigations. Is 6360 enforced by spectre_v2=off 6361 6362 prctl - Indirect branch speculation is enabled, 6363 but mitigation can be enabled via prctl 6364 per thread. The mitigation control state 6365 is inherited on fork. 6366 6367 prctl,ibpb 6368 - Like "prctl" above, but only STIBP is 6369 controlled per thread. IBPB is issued 6370 always when switching between different user 6371 space processes. 6372 6373 seccomp 6374 - Same as "prctl" above, but all seccomp 6375 threads will enable the mitigation unless 6376 they explicitly opt out. 6377 6378 seccomp,ibpb 6379 - Like "seccomp" above, but only STIBP is 6380 controlled per thread. IBPB is issued 6381 always when switching between different 6382 user space processes. 6383 6384 auto - Kernel selects the mitigation depending on 6385 the available CPU features and vulnerability. 6386 6387 Default mitigation: "prctl" 6388 6389 Not specifying this option is equivalent to 6390 spectre_v2_user=auto. 6391 6392 spec_rstack_overflow= 6393 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs 6394 6395 off - Disable mitigation 6396 microcode - Enable microcode mitigation only 6397 safe-ret - Enable sw-only safe RET mitigation (default) 6398 ibpb - Enable mitigation by issuing IBPB on 6399 kernel entry 6400 ibpb-vmexit - Issue IBPB only on VMEXIT 6401 (cloud-specific mitigation) 6402 6403 spec_store_bypass_disable= 6404 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation 6405 (Speculative Store Bypass vulnerability) 6406 6407 Certain CPUs are vulnerable to an exploit against a 6408 a common industry wide performance optimization known 6409 as "Speculative Store Bypass" in which recent stores 6410 to the same memory location may not be observed by 6411 later loads during speculative execution. The idea 6412 is that such stores are unlikely and that they can 6413 be detected prior to instruction retirement at the 6414 end of a particular speculation execution window. 6415 6416 In vulnerable processors, the speculatively forwarded 6417 store can be used in a cache side channel attack, for 6418 example to read memory to which the attacker does not 6419 directly have access (e.g. inside sandboxed code). 6420 6421 This parameter controls whether the Speculative Store 6422 Bypass optimization is used. 6423 6424 On x86 the options are: 6425 6426 on - Unconditionally disable Speculative Store Bypass 6427 off - Unconditionally enable Speculative Store Bypass 6428 auto - Kernel detects whether the CPU model contains an 6429 implementation of Speculative Store Bypass and 6430 picks the most appropriate mitigation. If the 6431 CPU is not vulnerable, "off" is selected. If the 6432 CPU is vulnerable the default mitigation is 6433 architecture and Kconfig dependent. See below. 6434 prctl - Control Speculative Store Bypass per thread 6435 via prctl. Speculative Store Bypass is enabled 6436 for a process by default. The state of the control 6437 is inherited on fork. 6438 seccomp - Same as "prctl" above, but all seccomp threads 6439 will disable SSB unless they explicitly opt out. 6440 6441 Default mitigations: 6442 X86: "prctl" 6443 6444 On powerpc the options are: 6445 6446 on,auto - On Power8 and Power9 insert a store-forwarding 6447 barrier on kernel entry and exit. On Power7 6448 perform a software flush on kernel entry and 6449 exit. 6450 off - No action. 6451 6452 Not specifying this option is equivalent to 6453 spec_store_bypass_disable=auto. 6454 6455 split_lock_detect= 6456 [X86] Enable split lock detection or bus lock detection 6457 6458 When enabled (and if hardware support is present), atomic 6459 instructions that access data across cache line 6460 boundaries will result in an alignment check exception 6461 for split lock detection or a debug exception for 6462 bus lock detection. 6463 6464 off - not enabled 6465 6466 warn - the kernel will emit rate-limited warnings 6467 about applications triggering the #AC 6468 exception or the #DB exception. This mode is 6469 the default on CPUs that support split lock 6470 detection or bus lock detection. Default 6471 behavior is by #AC if both features are 6472 enabled in hardware. 6473 6474 fatal - the kernel will send SIGBUS to applications 6475 that trigger the #AC exception or the #DB 6476 exception. Default behavior is by #AC if 6477 both features are enabled in hardware. 6478 6479 ratelimit:N - 6480 Set system wide rate limit to N bus locks 6481 per second for bus lock detection. 6482 0 < N <= 1000. 6483 6484 N/A for split lock detection. 6485 6486 6487 If an #AC exception is hit in the kernel or in 6488 firmware (i.e. not while executing in user mode) 6489 the kernel will oops in either "warn" or "fatal" 6490 mode. 6491 6492 #DB exception for bus lock is triggered only when 6493 CPL > 0. 6494 6495 srbds= [X86,INTEL,EARLY] 6496 Control the Special Register Buffer Data Sampling 6497 (SRBDS) mitigation. 6498 6499 Certain CPUs are vulnerable to an MDS-like 6500 exploit which can leak bits from the random 6501 number generator. 6502 6503 By default, this issue is mitigated by 6504 microcode. However, the microcode fix can cause 6505 the RDRAND and RDSEED instructions to become 6506 much slower. Among other effects, this will 6507 result in reduced throughput from /dev/urandom. 6508 6509 The microcode mitigation can be disabled with 6510 the following option: 6511 6512 off: Disable mitigation and remove 6513 performance impact to RDRAND and RDSEED 6514 6515 srcutree.big_cpu_lim [KNL] 6516 Specifies the number of CPUs constituting a 6517 large system, such that srcu_struct structures 6518 should immediately allocate an srcu_node array. 6519 This kernel-boot parameter defaults to 128, 6520 but takes effect only when the low-order four 6521 bits of srcutree.convert_to_big is equal to 3 6522 (decide at boot). 6523 6524 srcutree.convert_to_big [KNL] 6525 Specifies under what conditions an SRCU tree 6526 srcu_struct structure will be converted to big 6527 form, that is, with an rcu_node tree: 6528 6529 0: Never. 6530 1: At init_srcu_struct() time. 6531 2: When rcutorture decides to. 6532 3: Decide at boot time (default). 6533 0x1X: Above plus if high contention. 6534 6535 Either way, the srcu_node tree will be sized based 6536 on the actual runtime number of CPUs (nr_cpu_ids) 6537 instead of the compile-time CONFIG_NR_CPUS. 6538 6539 srcutree.counter_wrap_check [KNL] 6540 Specifies how frequently to check for 6541 grace-period sequence counter wrap for the 6542 srcu_data structure's ->srcu_gp_seq_needed field. 6543 The greater the number of bits set in this kernel 6544 parameter, the less frequently counter wrap will 6545 be checked for. Note that the bottom two bits 6546 are ignored. 6547 6548 srcutree.exp_holdoff [KNL] 6549 Specifies how many nanoseconds must elapse 6550 since the end of the last SRCU grace period for 6551 a given srcu_struct until the next normal SRCU 6552 grace period will be considered for automatic 6553 expediting. Set to zero to disable automatic 6554 expediting. 6555 6556 srcutree.srcu_max_nodelay [KNL] 6557 Specifies the number of no-delay instances 6558 per jiffy for which the SRCU grace period 6559 worker thread will be rescheduled with zero 6560 delay. Beyond this limit, worker thread will 6561 be rescheduled with a sleep delay of one jiffy. 6562 6563 srcutree.srcu_max_nodelay_phase [KNL] 6564 Specifies the per-grace-period phase, number of 6565 non-sleeping polls of readers. Beyond this limit, 6566 grace period worker thread will be rescheduled 6567 with a sleep delay of one jiffy, between each 6568 rescan of the readers, for a grace period phase. 6569 6570 srcutree.srcu_retry_check_delay [KNL] 6571 Specifies number of microseconds of non-sleeping 6572 delay between each non-sleeping poll of readers. 6573 6574 srcutree.small_contention_lim [KNL] 6575 Specifies the number of update-side contention 6576 events per jiffy will be tolerated before 6577 initiating a conversion of an srcu_struct 6578 structure to big form. Note that the value of 6579 srcutree.convert_to_big must have the 0x10 bit 6580 set for contention-based conversions to occur. 6581 6582 ssbd= [ARM64,HW,EARLY] 6583 Speculative Store Bypass Disable control 6584 6585 On CPUs that are vulnerable to the Speculative 6586 Store Bypass vulnerability and offer a 6587 firmware based mitigation, this parameter 6588 indicates how the mitigation should be used: 6589 6590 force-on: Unconditionally enable mitigation for 6591 for both kernel and userspace 6592 force-off: Unconditionally disable mitigation for 6593 for both kernel and userspace 6594 kernel: Always enable mitigation in the 6595 kernel, and offer a prctl interface 6596 to allow userspace to register its 6597 interest in being mitigated too. 6598 6599 stack_guard_gap= [MM] 6600 override the default stack gap protection. The value 6601 is in page units and it defines how many pages prior 6602 to (for stacks growing down) resp. after (for stacks 6603 growing up) the main stack are reserved for no other 6604 mapping. Default value is 256 pages. 6605 6606 stack_depot_disable= [KNL,EARLY] 6607 Setting this to true through kernel command line will 6608 disable the stack depot thereby saving the static memory 6609 consumed by the stack hash table. By default this is set 6610 to false. 6611 6612 stacktrace [FTRACE] 6613 Enabled the stack tracer on boot up. 6614 6615 stacktrace_filter=[function-list] 6616 [FTRACE] Limit the functions that the stack tracer 6617 will trace at boot up. function-list is a comma-separated 6618 list of functions. This list can be changed at run 6619 time by the stack_trace_filter file in the debugfs 6620 tracing directory. Note, this enables stack tracing 6621 and the stacktrace above is not needed. 6622 6623 sti= [PARISC,HW] 6624 Format: <num> 6625 Set the STI (builtin display/keyboard on the HP-PARISC 6626 machines) console (graphic card) which should be used 6627 as the initial boot-console. 6628 See also comment in drivers/video/console/sticore.c. 6629 6630 sti_font= [HW] 6631 See comment in drivers/video/console/sticore.c. 6632 6633 stifb= [HW] 6634 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 6635 6636 strict_sas_size= 6637 [X86] 6638 Format: <bool> 6639 Enable or disable strict sigaltstack size checks 6640 against the required signal frame size which 6641 depends on the supported FPU features. This can 6642 be used to filter out binaries which have 6643 not yet been made aware of AT_MINSIGSTKSZ. 6644 6645 stress_hpt [PPC,EARLY] 6646 Limits the number of kernel HPT entries in the hash 6647 page table to increase the rate of hash page table 6648 faults on kernel addresses. 6649 6650 stress_slb [PPC,EARLY] 6651 Limits the number of kernel SLB entries, and flushes 6652 them frequently to increase the rate of SLB faults 6653 on kernel addresses. 6654 6655 sunrpc.min_resvport= 6656 sunrpc.max_resvport= 6657 [NFS,SUNRPC] 6658 SunRPC servers often require that client requests 6659 originate from a privileged port (i.e. a port in the 6660 range 0 < portnr < 1024). 6661 An administrator who wishes to reserve some of these 6662 ports for other uses may adjust the range that the 6663 kernel's sunrpc client considers to be privileged 6664 using these two parameters to set the minimum and 6665 maximum port values. 6666 6667 sunrpc.svc_rpc_per_connection_limit= 6668 [NFS,SUNRPC] 6669 Limit the number of requests that the server will 6670 process in parallel from a single connection. 6671 The default value is 0 (no limit). 6672 6673 sunrpc.pool_mode= 6674 [NFS] 6675 Control how the NFS server code allocates CPUs to 6676 service thread pools. Depending on how many NICs 6677 you have and where their interrupts are bound, this 6678 option will affect which CPUs will do NFS serving. 6679 Note: this parameter cannot be changed while the 6680 NFS server is running. 6681 6682 auto the server chooses an appropriate mode 6683 automatically using heuristics 6684 global a single global pool contains all CPUs 6685 percpu one pool for each CPU 6686 pernode one pool for each NUMA node (equivalent 6687 to global on non-NUMA machines) 6688 6689 sunrpc.tcp_slot_table_entries= 6690 sunrpc.udp_slot_table_entries= 6691 [NFS,SUNRPC] 6692 Sets the upper limit on the number of simultaneous 6693 RPC calls that can be sent from the client to a 6694 server. Increasing these values may allow you to 6695 improve throughput, but will also increase the 6696 amount of memory reserved for use by the client. 6697 6698 suspend.pm_test_delay= 6699 [SUSPEND] 6700 Sets the number of seconds to remain in a suspend test 6701 mode before resuming the system (see 6702 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 6703 is set. Default value is 5. 6704 6705 svm= [PPC] 6706 Format: { on | off | y | n | 1 | 0 } 6707 This parameter controls use of the Protected 6708 Execution Facility on pSeries. 6709 6710 swiotlb= [ARM,PPC,MIPS,X86,S390,EARLY] 6711 Format: { <int> [,<int>] | force | noforce } 6712 <int> -- Number of I/O TLB slabs 6713 <int> -- Second integer after comma. Number of swiotlb 6714 areas with their own lock. Will be rounded up 6715 to a power of 2. 6716 force -- force using of bounce buffers even if they 6717 wouldn't be automatically used by the kernel 6718 noforce -- Never use bounce buffers (for debugging) 6719 6720 switches= [HW,M68k,EARLY] 6721 6722 sysctl.*= [KNL] 6723 Set a sysctl parameter, right before loading the init 6724 process, as if the value was written to the respective 6725 /proc/sys/... file. Both '.' and '/' are recognized as 6726 separators. Unrecognized parameters and invalid values 6727 are reported in the kernel log. Sysctls registered 6728 later by a loaded module cannot be set this way. 6729 Example: sysctl.vm.swappiness=40 6730 6731 sysrq_always_enabled 6732 [KNL] 6733 Ignore sysrq setting - this boot parameter will 6734 neutralize any effect of /proc/sys/kernel/sysrq. 6735 Useful for debugging. 6736 6737 tcpmhash_entries= [KNL,NET] 6738 Set the number of tcp_metrics_hash slots. 6739 Default value is 8192 or 16384 depending on total 6740 ram pages. This is used to specify the TCP metrics 6741 cache size. See Documentation/networking/ip-sysctl.rst 6742 "tcp_no_metrics_save" section for more details. 6743 6744 tdfx= [HW,DRM] 6745 6746 test_suspend= [SUSPEND] 6747 Format: { "mem" | "standby" | "freeze" }[,N] 6748 Specify "mem" (for Suspend-to-RAM) or "standby" (for 6749 standby suspend) or "freeze" (for suspend type freeze) 6750 as the system sleep state during system startup with 6751 the optional capability to repeat N number of times. 6752 The system is woken from this state using a 6753 wakeup-capable RTC alarm. 6754 6755 thash_entries= [KNL,NET] 6756 Set number of hash buckets for TCP connection 6757 6758 thermal.act= [HW,ACPI] 6759 -1: disable all active trip points in all thermal zones 6760 <degrees C>: override all lowest active trip points 6761 6762 thermal.crt= [HW,ACPI] 6763 -1: disable all critical trip points in all thermal zones 6764 <degrees C>: override all critical trip points 6765 6766 thermal.off= [HW,ACPI] 6767 1: disable ACPI thermal control 6768 6769 thermal.psv= [HW,ACPI] 6770 -1: disable all passive trip points 6771 <degrees C>: override all passive trip points to this 6772 value 6773 6774 thermal.tzp= [HW,ACPI] 6775 Specify global default ACPI thermal zone polling rate 6776 <deci-seconds>: poll all this frequency 6777 0: no polling (default) 6778 6779 thp_anon= [KNL] 6780 Format: <size>[KMG],<size>[KMG]:<state>;<size>[KMG]-<size>[KMG]:<state> 6781 state is one of "always", "madvise", "never" or "inherit". 6782 Control the default behavior of the system with respect 6783 to anonymous transparent hugepages. 6784 Can be used multiple times for multiple anon THP sizes. 6785 See Documentation/admin-guide/mm/transhuge.rst for more 6786 details. 6787 6788 threadirqs [KNL,EARLY] 6789 Force threading of all interrupt handlers except those 6790 marked explicitly IRQF_NO_THREAD. 6791 6792 topology= [S390,EARLY] 6793 Format: {off | on} 6794 Specify if the kernel should make use of the cpu 6795 topology information if the hardware supports this. 6796 The scheduler will make use of this information and 6797 e.g. base its process migration decisions on it. 6798 Default is on. 6799 6800 torture.disable_onoff_at_boot= [KNL] 6801 Prevent the CPU-hotplug component of torturing 6802 until after init has spawned. 6803 6804 torture.ftrace_dump_at_shutdown= [KNL] 6805 Dump the ftrace buffer at torture-test shutdown, 6806 even if there were no errors. This can be a 6807 very costly operation when many torture tests 6808 are running concurrently, especially on systems 6809 with rotating-rust storage. 6810 6811 torture.verbose_sleep_frequency= [KNL] 6812 Specifies how many verbose printk()s should be 6813 emitted between each sleep. The default of zero 6814 disables verbose-printk() sleeping. 6815 6816 torture.verbose_sleep_duration= [KNL] 6817 Duration of each verbose-printk() sleep in jiffies. 6818 6819 tpm.disable_pcr_integrity= [HW,TPM] 6820 Do not protect PCR registers from unintended physical 6821 access, or interposers in the bus by the means of 6822 having an integrity protected session wrapped around 6823 TPM2_PCR_Extend command. Consider this in a situation 6824 where TPM is heavily utilized by IMA, thus protection 6825 causing a major performance hit, and the space where 6826 machines are deployed is by other means guarded. 6827 6828 tpm_suspend_pcr=[HW,TPM] 6829 Format: integer pcr id 6830 Specify that at suspend time, the tpm driver 6831 should extend the specified pcr with zeros, 6832 as a workaround for some chips which fail to 6833 flush the last written pcr on TPM_SaveState. 6834 This will guarantee that all the other pcrs 6835 are saved. 6836 6837 tpm_tis.interrupts= [HW,TPM] 6838 Enable interrupts for the MMIO based physical layer 6839 for the FIFO interface. By default it is set to false 6840 (0). For more information about TPM hardware interfaces 6841 defined by Trusted Computing Group (TCG) see 6842 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/ 6843 6844 tp_printk [FTRACE] 6845 Have the tracepoints sent to printk as well as the 6846 tracing ring buffer. This is useful for early boot up 6847 where the system hangs or reboots and does not give the 6848 option for reading the tracing buffer or performing a 6849 ftrace_dump_on_oops. 6850 6851 To turn off having tracepoints sent to printk, 6852 echo 0 > /proc/sys/kernel/tracepoint_printk 6853 Note, echoing 1 into this file without the 6854 tp_printk kernel cmdline option has no effect. 6855 6856 The tp_printk_stop_on_boot (see below) can also be used 6857 to stop the printing of events to console at 6858 late_initcall_sync. 6859 6860 ** CAUTION ** 6861 6862 Having tracepoints sent to printk() and activating high 6863 frequency tracepoints such as irq or sched, can cause 6864 the system to live lock. 6865 6866 tp_printk_stop_on_boot [FTRACE] 6867 When tp_printk (above) is set, it can cause a lot of noise 6868 on the console. It may be useful to only include the 6869 printing of events during boot up, as user space may 6870 make the system inoperable. 6871 6872 This command line option will stop the printing of events 6873 to console at the late_initcall_sync() time frame. 6874 6875 trace_buf_size=nn[KMG] 6876 [FTRACE] will set tracing buffer size on each cpu. 6877 6878 trace_clock= [FTRACE] Set the clock used for tracing events 6879 at boot up. 6880 local - Use the per CPU time stamp counter 6881 (converted into nanoseconds). Fast, but 6882 depending on the architecture, may not be 6883 in sync between CPUs. 6884 global - Event time stamps are synchronize across 6885 CPUs. May be slower than the local clock, 6886 but better for some race conditions. 6887 counter - Simple counting of events (1, 2, ..) 6888 note, some counts may be skipped due to the 6889 infrastructure grabbing the clock more than 6890 once per event. 6891 uptime - Use jiffies as the time stamp. 6892 perf - Use the same clock that perf uses. 6893 mono - Use ktime_get_mono_fast_ns() for time stamps. 6894 mono_raw - Use ktime_get_raw_fast_ns() for time 6895 stamps. 6896 boot - Use ktime_get_boot_fast_ns() for time stamps. 6897 Architectures may add more clocks. See 6898 Documentation/trace/ftrace.rst for more details. 6899 6900 trace_event=[event-list] 6901 [FTRACE] Set and start specified trace events in order 6902 to facilitate early boot debugging. The event-list is a 6903 comma-separated list of trace events to enable. See 6904 also Documentation/trace/events.rst 6905 6906 trace_instance=[instance-info] 6907 [FTRACE] Create a ring buffer instance early in boot up. 6908 This will be listed in: 6909 6910 /sys/kernel/tracing/instances 6911 6912 Events can be enabled at the time the instance is created 6913 via: 6914 6915 trace_instance=<name>,<system1>:<event1>,<system2>:<event2> 6916 6917 Note, the "<system*>:" portion is optional if the event is 6918 unique. 6919 6920 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall 6921 6922 will enable the "sched_switch" event (note, the "sched:" is optional, and 6923 the same thing would happen if it was left off). The irq_handler_entry 6924 event, and all events under the "initcall" system. 6925 6926 Flags can be added to the instance to modify its behavior when it is 6927 created. The flags are separated by '^'. 6928 6929 The available flags are: 6930 6931 traceoff - Have the tracing instance tracing disabled after it is created. 6932 traceprintk - Have trace_printk() write into this trace instance 6933 (note, "printk" and "trace_printk" can also be used) 6934 6935 trace_instance=foo^traceoff^traceprintk,sched,irq 6936 6937 The flags must come before the defined events. 6938 6939 If memory has been reserved (see memmap for x86), the instance 6940 can use that memory: 6941 6942 memmap=12M$0x284500000 trace_instance=boot_map@0x284500000:12M 6943 6944 The above will create a "boot_map" instance that uses the physical 6945 memory at 0x284500000 that is 12Megs. The per CPU buffers of that 6946 instance will be split up accordingly. 6947 6948 Alternatively, the memory can be reserved by the reserve_mem option: 6949 6950 reserve_mem=12M:4096:trace trace_instance=boot_map@trace 6951 6952 This will reserve 12 megabytes at boot up with a 4096 byte alignment 6953 and place the ring buffer in this memory. Note that due to KASLR, the 6954 memory may not be the same location each time, which will not preserve 6955 the buffer content. 6956 6957 Also note that the layout of the ring buffer data may change between 6958 kernel versions where the validator will fail and reset the ring buffer 6959 if the layout is not the same as the previous kernel. 6960 6961 If the ring buffer is used for persistent bootups and has events enabled, 6962 it is recommend to disable tracing so that events from a previous boot do not 6963 mix with events of the current boot (unless you are debugging a random crash 6964 at boot up). 6965 6966 reserve_mem=12M:4096:trace trace_instance=boot_map^traceoff^traceprintk@trace,sched,irq 6967 6968 See also Documentation/trace/debugging.rst 6969 6970 6971 trace_options=[option-list] 6972 [FTRACE] Enable or disable tracer options at boot. 6973 The option-list is a comma delimited list of options 6974 that can be enabled or disabled just as if you were 6975 to echo the option name into 6976 6977 /sys/kernel/tracing/trace_options 6978 6979 For example, to enable stacktrace option (to dump the 6980 stack trace of each event), add to the command line: 6981 6982 trace_options=stacktrace 6983 6984 See also Documentation/trace/ftrace.rst "trace options" 6985 section. 6986 6987 trace_trigger=[trigger-list] 6988 [FTRACE] Add a event trigger on specific events. 6989 Set a trigger on top of a specific event, with an optional 6990 filter. 6991 6992 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..." 6993 Where more than one trigger may be specified that are comma deliminated. 6994 6995 For example: 6996 6997 trace_trigger="sched_switch.stacktrace if prev_state == 2" 6998 6999 The above will enable the "stacktrace" trigger on the "sched_switch" 7000 event but only trigger it if the "prev_state" of the "sched_switch" 7001 event is "2" (TASK_UNINTERUPTIBLE). 7002 7003 See also "Event triggers" in Documentation/trace/events.rst 7004 7005 7006 traceoff_on_warning 7007 [FTRACE] enable this option to disable tracing when a 7008 warning is hit. This turns off "tracing_on". Tracing can 7009 be enabled again by echoing '1' into the "tracing_on" 7010 file located in /sys/kernel/tracing/ 7011 7012 This option is useful, as it disables the trace before 7013 the WARNING dump is called, which prevents the trace to 7014 be filled with content caused by the warning output. 7015 7016 This option can also be set at run time via the sysctl 7017 option: kernel/traceoff_on_warning 7018 7019 transparent_hugepage= 7020 [KNL] 7021 Format: [always|madvise|never] 7022 Can be used to control the default behavior of the system 7023 with respect to transparent hugepages. 7024 See Documentation/admin-guide/mm/transhuge.rst 7025 for more details. 7026 7027 trusted.source= [KEYS] 7028 Format: <string> 7029 This parameter identifies the trust source as a backend 7030 for trusted keys implementation. Supported trust 7031 sources: 7032 - "tpm" 7033 - "tee" 7034 - "caam" 7035 - "dcp" 7036 If not specified then it defaults to iterating through 7037 the trust source list starting with TPM and assigns the 7038 first trust source as a backend which is initialized 7039 successfully during iteration. 7040 7041 trusted.rng= [KEYS] 7042 Format: <string> 7043 The RNG used to generate key material for trusted keys. 7044 Can be one of: 7045 - "kernel" 7046 - the same value as trusted.source: "tpm" or "tee" 7047 - "default" 7048 If not specified, "default" is used. In this case, 7049 the RNG's choice is left to each individual trust source. 7050 7051 trusted.dcp_use_otp_key 7052 This is intended to be used in combination with 7053 trusted.source=dcp and will select the DCP OTP key 7054 instead of the DCP UNIQUE key blob encryption. 7055 7056 trusted.dcp_skip_zk_test 7057 This is intended to be used in combination with 7058 trusted.source=dcp and will disable the check if the 7059 blob key is all zeros. This is helpful for situations where 7060 having this key zero'ed is acceptable. E.g. in testing 7061 scenarios. 7062 7063 tsa= [X86] Control mitigation for Transient Scheduler 7064 Attacks on AMD CPUs. Search the following in your 7065 favourite search engine for more details: 7066 7067 "Technical guidance for mitigating transient scheduler 7068 attacks". 7069 7070 off - disable the mitigation 7071 on - enable the mitigation (default) 7072 user - mitigate only user/kernel transitions 7073 vm - mitigate only guest/host transitions 7074 7075 7076 tsc= Disable clocksource stability checks for TSC. 7077 Format: <string> 7078 [x86] reliable: mark tsc clocksource as reliable, this 7079 disables clocksource verification at runtime, as well 7080 as the stability checks done at bootup. Used to enable 7081 high-resolution timer mode on older hardware, and in 7082 virtualized environment. 7083 [x86] noirqtime: Do not use TSC to do irq accounting. 7084 Used to run time disable IRQ_TIME_ACCOUNTING on any 7085 platforms where RDTSC is slow and this accounting 7086 can add overhead. 7087 [x86] unstable: mark the TSC clocksource as unstable, this 7088 marks the TSC unconditionally unstable at bootup and 7089 avoids any further wobbles once the TSC watchdog notices. 7090 [x86] nowatchdog: disable clocksource watchdog. Used 7091 in situations with strict latency requirements (where 7092 interruptions from clocksource watchdog are not 7093 acceptable). 7094 [x86] recalibrate: force recalibration against a HW timer 7095 (HPET or PM timer) on systems whose TSC frequency was 7096 obtained from HW or FW using either an MSR or CPUID(0x15). 7097 Warn if the difference is more than 500 ppm. 7098 [x86] watchdog: Use TSC as the watchdog clocksource with 7099 which to check other HW timers (HPET or PM timer), but 7100 only on systems where TSC has been deemed trustworthy. 7101 This will be suppressed by an earlier tsc=nowatchdog and 7102 can be overridden by a later tsc=nowatchdog. A console 7103 message will flag any such suppression or overriding. 7104 7105 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given 7106 value instead. Useful when the early TSC frequency discovery 7107 procedure is not reliable, such as on overclocked systems 7108 with CPUID.16h support and partial CPUID.15h support. 7109 Format: <unsigned int> 7110 7111 tsx= [X86] Control Transactional Synchronization 7112 Extensions (TSX) feature in Intel processors that 7113 support TSX control. 7114 7115 This parameter controls the TSX feature. The options are: 7116 7117 on - Enable TSX on the system. Although there are 7118 mitigations for all known security vulnerabilities, 7119 TSX has been known to be an accelerator for 7120 several previous speculation-related CVEs, and 7121 so there may be unknown security risks associated 7122 with leaving it enabled. 7123 7124 off - Disable TSX on the system. (Note that this 7125 option takes effect only on newer CPUs which are 7126 not vulnerable to MDS, i.e., have 7127 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 7128 the new IA32_TSX_CTRL MSR through a microcode 7129 update. This new MSR allows for the reliable 7130 deactivation of the TSX functionality.) 7131 7132 auto - Disable TSX if X86_BUG_TAA is present, 7133 otherwise enable TSX on the system. 7134 7135 Not specifying this option is equivalent to tsx=off. 7136 7137 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7138 for more details. 7139 7140 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async 7141 Abort (TAA) vulnerability. 7142 7143 Similar to Micro-architectural Data Sampling (MDS) 7144 certain CPUs that support Transactional 7145 Synchronization Extensions (TSX) are vulnerable to an 7146 exploit against CPU internal buffers which can forward 7147 information to a disclosure gadget under certain 7148 conditions. 7149 7150 In vulnerable processors, the speculatively forwarded 7151 data can be used in a cache side channel attack, to 7152 access data to which the attacker does not have direct 7153 access. 7154 7155 This parameter controls the TAA mitigation. The 7156 options are: 7157 7158 full - Enable TAA mitigation on vulnerable CPUs 7159 if TSX is enabled. 7160 7161 full,nosmt - Enable TAA mitigation and disable SMT on 7162 vulnerable CPUs. If TSX is disabled, SMT 7163 is not disabled because CPU is not 7164 vulnerable to cross-thread TAA attacks. 7165 off - Unconditionally disable TAA mitigation 7166 7167 On MDS-affected machines, tsx_async_abort=off can be 7168 prevented by an active MDS mitigation as both vulnerabilities 7169 are mitigated with the same mechanism so in order to disable 7170 this mitigation, you need to specify mds=off too. 7171 7172 Not specifying this option is equivalent to 7173 tsx_async_abort=full. On CPUs which are MDS affected 7174 and deploy MDS mitigation, TAA mitigation is not 7175 required and doesn't provide any additional 7176 mitigation. 7177 7178 For details see: 7179 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 7180 7181 turbografx.map[2|3]= [HW,JOY] 7182 TurboGraFX parallel port interface 7183 Format: 7184 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 7185 See also Documentation/input/devices/joystick-parport.rst 7186 7187 udbg-immortal [PPC] When debugging early kernel crashes that 7188 happen after console_init() and before a proper 7189 console driver takes over, this boot options might 7190 help "seeing" what's going on. 7191 7192 uhash_entries= [KNL,NET] 7193 Set number of hash buckets for UDP/UDP-Lite connections 7194 7195 uhci-hcd.ignore_oc= 7196 [USB] Ignore overcurrent events (default N). 7197 Some badly-designed motherboards generate lots of 7198 bogus events, for ports that aren't wired to 7199 anything. Set this parameter to avoid log spamming. 7200 Note that genuine overcurrent events won't be 7201 reported either. 7202 7203 unknown_nmi_panic 7204 [X86] Cause panic on unknown NMI. 7205 7206 unwind_debug [X86-64,EARLY] 7207 Enable unwinder debug output. This can be 7208 useful for debugging certain unwinder error 7209 conditions, including corrupt stacks and 7210 bad/missing unwinder metadata. 7211 7212 usbcore.authorized_default= 7213 [USB] Default USB device authorization: 7214 (default -1 = authorized (same as 1), 7215 0 = not authorized, 1 = authorized, 2 = authorized 7216 if device connected to internal port) 7217 7218 usbcore.autosuspend= 7219 [USB] The autosuspend time delay (in seconds) used 7220 for newly-detected USB devices (default 2). This 7221 is the time required before an idle device will be 7222 autosuspended. Devices for which the delay is set 7223 to a negative value won't be autosuspended at all. 7224 7225 usbcore.usbfs_snoop= 7226 [USB] Set to log all usbfs traffic (default 0 = off). 7227 7228 usbcore.usbfs_snoop_max= 7229 [USB] Maximum number of bytes to snoop in each URB 7230 (default = 65536). 7231 7232 usbcore.blinkenlights= 7233 [USB] Set to cycle leds on hubs (default 0 = off). 7234 7235 usbcore.old_scheme_first= 7236 [USB] Start with the old device initialization 7237 scheme (default 0 = off). 7238 7239 usbcore.usbfs_memory_mb= 7240 [USB] Memory limit (in MB) for buffers allocated by 7241 usbfs (default = 16, 0 = max = 2047). 7242 7243 usbcore.use_both_schemes= 7244 [USB] Try the other device initialization scheme 7245 if the first one fails (default 1 = enabled). 7246 7247 usbcore.initial_descriptor_timeout= 7248 [USB] Specifies timeout for the initial 64-byte 7249 USB_REQ_GET_DESCRIPTOR request in milliseconds 7250 (default 5000 = 5.0 seconds). 7251 7252 usbcore.nousb [USB] Disable the USB subsystem 7253 7254 usbcore.quirks= 7255 [USB] A list of quirk entries to augment the built-in 7256 usb core quirk list. List entries are separated by 7257 commas. Each entry has the form 7258 VendorID:ProductID:Flags. The IDs are 4-digit hex 7259 numbers and Flags is a set of letters. Each letter 7260 will change the built-in quirk; setting it if it is 7261 clear and clearing it if it is set. The letters have 7262 the following meanings: 7263 a = USB_QUIRK_STRING_FETCH_255 (string 7264 descriptors must not be fetched using 7265 a 255-byte read); 7266 b = USB_QUIRK_RESET_RESUME (device can't resume 7267 correctly so reset it instead); 7268 c = USB_QUIRK_NO_SET_INTF (device can't handle 7269 Set-Interface requests); 7270 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 7271 handle its Configuration or Interface 7272 strings); 7273 e = USB_QUIRK_RESET (device can't be reset 7274 (e.g morph devices), don't use reset); 7275 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 7276 more interface descriptions than the 7277 bNumInterfaces count, and can't handle 7278 talking to these interfaces); 7279 g = USB_QUIRK_DELAY_INIT (device needs a pause 7280 during initialization, after we read 7281 the device descriptor); 7282 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 7283 high speed and super speed interrupt 7284 endpoints, the USB 2.0 and USB 3.0 spec 7285 require the interval in microframes (1 7286 microframe = 125 microseconds) to be 7287 calculated as interval = 2 ^ 7288 (bInterval-1). 7289 Devices with this quirk report their 7290 bInterval as the result of this 7291 calculation instead of the exponent 7292 variable used in the calculation); 7293 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 7294 handle device_qualifier descriptor 7295 requests); 7296 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 7297 generates spurious wakeup, ignore 7298 remote wakeup capability); 7299 k = USB_QUIRK_NO_LPM (device can't handle Link 7300 Power Management); 7301 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 7302 (Device reports its bInterval as linear 7303 frames instead of the USB 2.0 7304 calculation); 7305 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 7306 to be disconnected before suspend to 7307 prevent spurious wakeup); 7308 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 7309 pause after every control message); 7310 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 7311 delay after resetting its port); 7312 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT 7313 (Reduce timeout of the SET_ADDRESS 7314 request from 5000 ms to 500 ms); 7315 Example: quirks=0781:5580:bk,0a5c:5834:gij 7316 7317 usbhid.mousepoll= 7318 [USBHID] The interval which mice are to be polled at. 7319 7320 usbhid.jspoll= 7321 [USBHID] The interval which joysticks are to be polled at. 7322 7323 usbhid.kbpoll= 7324 [USBHID] The interval which keyboards are to be polled at. 7325 7326 usb-storage.delay_use= 7327 [UMS] The delay in seconds before a new device is 7328 scanned for Logical Units (default 1). 7329 Optionally the delay in milliseconds if the value has 7330 suffix with "ms". 7331 Example: delay_use=2567ms 7332 7333 usb-storage.quirks= 7334 [UMS] A list of quirks entries to supplement or 7335 override the built-in unusual_devs list. List 7336 entries are separated by commas. Each entry has 7337 the form VID:PID:Flags where VID and PID are Vendor 7338 and Product ID values (4-digit hex numbers) and 7339 Flags is a set of characters, each corresponding 7340 to a common usb-storage quirk flag as follows: 7341 a = SANE_SENSE (collect more than 18 bytes 7342 of sense data, not on uas); 7343 b = BAD_SENSE (don't collect more than 18 7344 bytes of sense data, not on uas); 7345 c = FIX_CAPACITY (decrease the reported 7346 device capacity by one sector); 7347 d = NO_READ_DISC_INFO (don't use 7348 READ_DISC_INFO command, not on uas); 7349 e = NO_READ_CAPACITY_16 (don't use 7350 READ_CAPACITY_16 command); 7351 f = NO_REPORT_OPCODES (don't use report opcodes 7352 command, uas only); 7353 g = MAX_SECTORS_240 (don't transfer more than 7354 240 sectors at a time, uas only); 7355 h = CAPACITY_HEURISTICS (decrease the 7356 reported device capacity by one 7357 sector if the number is odd); 7358 i = IGNORE_DEVICE (don't bind to this 7359 device); 7360 j = NO_REPORT_LUNS (don't use report luns 7361 command, uas only); 7362 k = NO_SAME (do not use WRITE_SAME, uas only) 7363 l = NOT_LOCKABLE (don't try to lock and 7364 unlock ejectable media, not on uas); 7365 m = MAX_SECTORS_64 (don't transfer more 7366 than 64 sectors = 32 KB at a time, 7367 not on uas); 7368 n = INITIAL_READ10 (force a retry of the 7369 initial READ(10) command, not on uas); 7370 o = CAPACITY_OK (accept the capacity 7371 reported by the device, not on uas); 7372 p = WRITE_CACHE (the device cache is ON 7373 by default, not on uas); 7374 r = IGNORE_RESIDUE (the device reports 7375 bogus residue values, not on uas); 7376 s = SINGLE_LUN (the device has only one 7377 Logical Unit); 7378 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 7379 commands, uas only); 7380 u = IGNORE_UAS (don't bind to the uas driver); 7381 w = NO_WP_DETECT (don't test whether the 7382 medium is write-protected). 7383 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 7384 even if the device claims no cache, 7385 not on uas) 7386 Example: quirks=0419:aaf5:rl,0421:0433:rc 7387 7388 user_debug= [KNL,ARM] 7389 Format: <int> 7390 See arch/arm/Kconfig.debug help text. 7391 1 - undefined instruction events 7392 2 - system calls 7393 4 - invalid data aborts 7394 8 - SIGSEGV faults 7395 16 - SIGBUS faults 7396 Example: user_debug=31 7397 7398 userpte= 7399 [X86,EARLY] Flags controlling user PTE allocations. 7400 7401 nohigh = do not allocate PTE pages in 7402 HIGHMEM regardless of setting 7403 of CONFIG_HIGHPTE. 7404 7405 vdso= [X86,SH,SPARC] 7406 On X86_32, this is an alias for vdso32=. Otherwise: 7407 7408 vdso=1: enable VDSO (the default) 7409 vdso=0: disable VDSO mapping 7410 7411 vdso32= [X86] Control the 32-bit vDSO 7412 vdso32=1: enable 32-bit VDSO 7413 vdso32=0 or vdso32=2: disable 32-bit VDSO 7414 7415 See the help text for CONFIG_COMPAT_VDSO for more 7416 details. If CONFIG_COMPAT_VDSO is set, the default is 7417 vdso32=0; otherwise, the default is vdso32=1. 7418 7419 For compatibility with older kernels, vdso32=2 is an 7420 alias for vdso32=0. 7421 7422 Try vdso32=0 if you encounter an error that says: 7423 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 7424 7425 video= [FB,EARLY] Frame buffer configuration 7426 See Documentation/fb/modedb.rst. 7427 7428 video.brightness_switch_enabled= [ACPI] 7429 Format: [0|1] 7430 If set to 1, on receiving an ACPI notify event 7431 generated by hotkey, video driver will adjust brightness 7432 level and then send out the event to user space through 7433 the allocated input device. If set to 0, video driver 7434 will only send out the event without touching backlight 7435 brightness level. 7436 default: 1 7437 7438 virtio_mmio.device= 7439 [VMMIO] Memory mapped virtio (platform) device. 7440 7441 <size>@<baseaddr>:<irq>[:<id>] 7442 where: 7443 <size> := size (can use standard suffixes 7444 like K, M and G) 7445 <baseaddr> := physical base address 7446 <irq> := interrupt number (as passed to 7447 request_irq()) 7448 <id> := (optional) platform device id 7449 example: 7450 virtio_mmio.device=1K@0x100b0000:48:7 7451 7452 Can be used multiple times for multiple devices. 7453 7454 vga= [BOOT,X86-32] Select a particular video mode 7455 See Documentation/arch/x86/boot.rst and 7456 Documentation/admin-guide/svga.rst. 7457 Use vga=ask for menu. 7458 This is actually a boot loader parameter; the value is 7459 passed to the kernel using a special protocol. 7460 7461 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 7462 May slow down system boot speed, especially when 7463 enabled on systems with a large amount of memory. 7464 All options are enabled by default, and this 7465 interface is meant to allow for selectively 7466 enabling or disabling specific virtual memory 7467 debugging features. 7468 7469 Available options are: 7470 P Enable page structure init time poisoning 7471 - Disable all of the above options 7472 7473 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an 7474 exact size of <nn>. This can be used to increase 7475 the minimum size (128MB on x86, arm32 platforms). 7476 It can also be used to decrease the size and leave more room 7477 for directly mapped kernel RAM. Note that this parameter does 7478 not exist on many other platforms (including arm64, alpha, 7479 loongarch, arc, csky, hexagon, microblaze, mips, nios2, openrisc, 7480 parisc, m64k, powerpc, riscv, sh, um, xtensa, s390, sparc). 7481 7482 vmcp_cma=nn[MG] [KNL,S390,EARLY] 7483 Sets the memory size reserved for contiguous memory 7484 allocations for the vmcp device driver. 7485 7486 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 7487 Format: <command> 7488 7489 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 7490 Format: <command> 7491 7492 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 7493 Format: <command> 7494 7495 vmscape= [X86] Controls mitigation for VMscape attacks. 7496 VMscape attacks can leak information from a userspace 7497 hypervisor to a guest via speculative side-channels. 7498 7499 off - disable the mitigation 7500 ibpb - use Indirect Branch Prediction Barrier 7501 (IBPB) mitigation (default) 7502 force - force vulnerability detection even on 7503 unaffected processors 7504 7505 vsyscall= [X86-64,EARLY] 7506 Controls the behavior of vsyscalls (i.e. calls to 7507 fixed addresses of 0xffffffffff600x00 from legacy 7508 code). Most statically-linked binaries and older 7509 versions of glibc use these calls. Because these 7510 functions are at fixed addresses, they make nice 7511 targets for exploits that can control RIP. 7512 7513 emulate Vsyscalls turn into traps and are emulated 7514 reasonably safely. The vsyscall page is 7515 readable. 7516 7517 xonly [default] Vsyscalls turn into traps and are 7518 emulated reasonably safely. The vsyscall 7519 page is not readable. 7520 7521 none Vsyscalls don't work at all. This makes 7522 them quite hard to use for exploits but 7523 might break your system. 7524 7525 vt.color= [VT] Default text color. 7526 Format: 0xYX, X = foreground, Y = background. 7527 Default: 0x07 = light gray on black. 7528 7529 vt.cur_default= [VT] Default cursor shape. 7530 Format: 0xCCBBAA, where AA, BB, and CC are the same as 7531 the parameters of the <Esc>[?A;B;Cc escape sequence; 7532 see VGA-softcursor.txt. Default: 2 = underline. 7533 7534 vt.default_blu= [VT] 7535 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 7536 Change the default blue palette of the console. 7537 This is a 16-member array composed of values 7538 ranging from 0-255. 7539 7540 vt.default_grn= [VT] 7541 Format: <green0>,<green1>,<green2>,...,<green15> 7542 Change the default green palette of the console. 7543 This is a 16-member array composed of values 7544 ranging from 0-255. 7545 7546 vt.default_red= [VT] 7547 Format: <red0>,<red1>,<red2>,...,<red15> 7548 Change the default red palette of the console. 7549 This is a 16-member array composed of values 7550 ranging from 0-255. 7551 7552 vt.default_utf8= 7553 [VT] 7554 Format=<0|1> 7555 Set system-wide default UTF-8 mode for all tty's. 7556 Default is 1, i.e. UTF-8 mode is enabled for all 7557 newly opened terminals. 7558 7559 vt.global_cursor_default= 7560 [VT] 7561 Format=<-1|0|1> 7562 Set system-wide default for whether a cursor 7563 is shown on new VTs. Default is -1, 7564 i.e. cursors will be created by default unless 7565 overridden by individual drivers. 0 will hide 7566 cursors, 1 will display them. 7567 7568 vt.italic= [VT] Default color for italic text; 0-15. 7569 Default: 2 = green. 7570 7571 vt.underline= [VT] Default color for underlined text; 0-15. 7572 Default: 3 = cyan. 7573 7574 watchdog timers [HW,WDT] For information on watchdog timers, 7575 see Documentation/watchdog/watchdog-parameters.rst 7576 or other driver-specific files in the 7577 Documentation/watchdog/ directory. 7578 7579 watchdog_thresh= 7580 [KNL] 7581 Set the hard lockup detector stall duration 7582 threshold in seconds. The soft lockup detector 7583 threshold is set to twice the value. A value of 0 7584 disables both lockup detectors. Default is 10 7585 seconds. 7586 7587 workqueue.unbound_cpus= 7588 [KNL,SMP] Specify to constrain one or some CPUs 7589 to use in unbound workqueues. 7590 Format: <cpu-list> 7591 By default, all online CPUs are available for 7592 unbound workqueues. 7593 7594 workqueue.watchdog_thresh= 7595 If CONFIG_WQ_WATCHDOG is configured, workqueue can 7596 warn stall conditions and dump internal state to 7597 help debugging. 0 disables workqueue stall 7598 detection; otherwise, it's the stall threshold 7599 duration in seconds. The default value is 30 and 7600 it can be updated at runtime by writing to the 7601 corresponding sysfs file. 7602 7603 workqueue.panic_on_stall=<uint> 7604 Panic when workqueue stall is detected by 7605 CONFIG_WQ_WATCHDOG. It sets the number times of the 7606 stall to trigger panic. 7607 7608 The default is 0, which disables the panic on stall. 7609 7610 workqueue.cpu_intensive_thresh_us= 7611 Per-cpu work items which run for longer than this 7612 threshold are automatically considered CPU intensive 7613 and excluded from concurrency management to prevent 7614 them from noticeably delaying other per-cpu work 7615 items. Default is 10000 (10ms). 7616 7617 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 7618 will report the work functions which violate this 7619 threshold repeatedly. They are likely good 7620 candidates for using WQ_UNBOUND workqueues instead. 7621 7622 workqueue.cpu_intensive_warning_thresh=<uint> 7623 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel 7624 will report the work functions which violate the 7625 intensive_threshold_us repeatedly. In order to prevent 7626 spurious warnings, start printing only after a work 7627 function has violated this threshold number of times. 7628 7629 The default is 4 times. 0 disables the warning. 7630 7631 workqueue.power_efficient 7632 Per-cpu workqueues are generally preferred because 7633 they show better performance thanks to cache 7634 locality; unfortunately, per-cpu workqueues tend to 7635 be more power hungry than unbound workqueues. 7636 7637 Enabling this makes the per-cpu workqueues which 7638 were observed to contribute significantly to power 7639 consumption unbound, leading to measurably lower 7640 power usage at the cost of small performance 7641 overhead. 7642 7643 The default value of this parameter is determined by 7644 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 7645 7646 workqueue.default_affinity_scope= 7647 Select the default affinity scope to use for unbound 7648 workqueues. Can be one of "cpu", "smt", "cache", 7649 "numa" and "system". Default is "cache". For more 7650 information, see the Affinity Scopes section in 7651 Documentation/core-api/workqueue.rst. 7652 7653 This can be changed after boot by writing to the 7654 matching /sys/module/workqueue/parameters file. All 7655 workqueues with the "default" affinity scope will be 7656 updated accordingly. 7657 7658 workqueue.debug_force_rr_cpu 7659 Workqueue used to implicitly guarantee that work 7660 items queued without explicit CPU specified are put 7661 on the local CPU. This guarantee is no longer true 7662 and while local CPU is still preferred work items 7663 may be put on foreign CPUs. This debug option 7664 forces round-robin CPU selection to flush out 7665 usages which depend on the now broken guarantee. 7666 When enabled, memory and cache locality will be 7667 impacted. 7668 7669 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access 7670 Type) of ioremap_wc(). 7671 7672 on - Enable writecombine, use WUC for ioremap_wc() 7673 off - Disable writecombine, use SUC for ioremap_wc() 7674 7675 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of 7676 default x2apic cluster mode on platforms 7677 supporting x2apic. 7678 7679 xen_512gb_limit [KNL,X86-64,XEN] 7680 Restricts the kernel running paravirtualized under Xen 7681 to use only up to 512 GB of RAM. The reason to do so is 7682 crash analysis tools and Xen tools for doing domain 7683 save/restore/migration must be enabled to handle larger 7684 domains. 7685 7686 xen_emul_unplug= [HW,X86,XEN,EARLY] 7687 Unplug Xen emulated devices 7688 Format: [unplug0,][unplug1] 7689 ide-disks -- unplug primary master IDE devices 7690 aux-ide-disks -- unplug non-primary-master IDE devices 7691 nics -- unplug network devices 7692 all -- unplug all emulated devices (NICs and IDE disks) 7693 unnecessary -- unplugging emulated devices is 7694 unnecessary even if the host did not respond to 7695 the unplug protocol 7696 never -- do not unplug even if version check succeeds 7697 7698 xen_legacy_crash [X86,XEN,EARLY] 7699 Crash from Xen panic notifier, without executing late 7700 panic() code such as dumping handler. 7701 7702 xen_mc_debug [X86,XEN,EARLY] 7703 Enable multicall debugging when running as a Xen PV guest. 7704 Enabling this feature will reduce performance a little 7705 bit, so it should only be enabled for obtaining extended 7706 debug data in case of multicall errors. 7707 7708 xen_msr_safe= [X86,XEN,EARLY] 7709 Format: <bool> 7710 Select whether to always use non-faulting (safe) MSR 7711 access functions when running as Xen PV guest. The 7712 default value is controlled by CONFIG_XEN_PV_MSR_SAFE. 7713 7714 xen_nopv [X86] 7715 Disables the PV optimizations forcing the HVM guest to 7716 run as generic HVM guest with no PV drivers. 7717 This option is obsoleted by the "nopv" option, which 7718 has equivalent effect for XEN platform. 7719 7720 xen_no_vector_callback 7721 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen 7722 event channel interrupts. 7723 7724 xen_scrub_pages= [XEN] 7725 Boolean option to control scrubbing pages before giving them back 7726 to Xen, for use by other domains. Can be also changed at runtime 7727 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 7728 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 7729 7730 xen_timer_slop= [X86-64,XEN,EARLY] 7731 Set the timer slop (in nanoseconds) for the virtual Xen 7732 timers (default is 100000). This adjusts the minimum 7733 delta of virtualized Xen timers, where lower values 7734 improve timer resolution at the expense of processing 7735 more timer interrupts. 7736 7737 xen.balloon_boot_timeout= [XEN] 7738 The time (in seconds) to wait before giving up to boot 7739 in case initial ballooning fails to free enough memory. 7740 Applies only when running as HVM or PVH guest and 7741 started with less memory configured than allowed at 7742 max. Default is 180. 7743 7744 xen.event_eoi_delay= [XEN] 7745 How long to delay EOI handling in case of event 7746 storms (jiffies). Default is 10. 7747 7748 xen.event_loop_timeout= [XEN] 7749 After which time (jiffies) the event handling loop 7750 should start to delay EOI handling. Default is 2. 7751 7752 xen.fifo_events= [XEN] 7753 Boolean parameter to disable using fifo event handling 7754 even if available. Normally fifo event handling is 7755 preferred over the 2-level event handling, as it is 7756 fairer and the number of possible event channels is 7757 much higher. Default is on (use fifo events). 7758 7759 xirc2ps_cs= [NET,PCMCIA] 7760 Format: 7761 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 7762 7763 xive= [PPC] 7764 By default on POWER9 and above, the kernel will 7765 natively use the XIVE interrupt controller. This option 7766 allows the fallback firmware mode to be used: 7767 7768 off Fallback to firmware control of XIVE interrupt 7769 controller on both pseries and powernv 7770 platforms. Only useful on POWER9 and above. 7771 7772 xive.store-eoi=off [PPC] 7773 By default on POWER10 and above, the kernel will use 7774 stores for EOI handling when the XIVE interrupt mode 7775 is active. This option allows the XIVE driver to use 7776 loads instead, as on POWER9. 7777 7778 xhci-hcd.quirks [USB,KNL] 7779 A hex value specifying bitmask with supplemental xhci 7780 host controller quirks. Meaning of each bit can be 7781 consulted in header drivers/usb/host/xhci.h. 7782 7783 xmon [PPC,EARLY] 7784 Format: { early | on | rw | ro | off } 7785 Controls if xmon debugger is enabled. Default is off. 7786 Passing only "xmon" is equivalent to "xmon=early". 7787 early Call xmon as early as possible on boot; xmon 7788 debugger is called from setup_arch(). 7789 on xmon debugger hooks will be installed so xmon 7790 is only called on a kernel crash. Default mode, 7791 i.e. either "ro" or "rw" mode, is controlled 7792 with CONFIG_XMON_DEFAULT_RO_MODE. 7793 rw xmon debugger hooks will be installed so xmon 7794 is called only on a kernel crash, mode is write, 7795 meaning SPR registers, memory and, other data 7796 can be written using xmon commands. 7797 ro same as "rw" option above but SPR registers, 7798 memory, and other data can't be written using 7799 xmon commands. 7800 off xmon is disabled. 7801