1 acpi= [HW,ACPI,X86,ARM64] 2 Advanced Configuration and Power Interface 3 Format: { force | on | off | strict | noirq | rsdt | 4 copy_dsdt } 5 force -- enable ACPI if default was off 6 on -- enable ACPI but allow fallback to DT [arm64] 7 off -- disable ACPI if default was on 8 noirq -- do not use ACPI for IRQ routing 9 strict -- Be less tolerant of platforms that are not 10 strictly ACPI specification compliant. 11 rsdt -- prefer RSDT over (default) XSDT 12 copy_dsdt -- copy DSDT to memory 13 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force" 14 are available 15 16 See also Documentation/power/runtime_pm.rst, pci=noacpi 17 18 acpi_apic_instance= [ACPI, IOAPIC] 19 Format: <int> 20 2: use 2nd APIC table, if available 21 1,0: use 1st APIC table 22 default: 0 23 24 acpi_backlight= [HW,ACPI] 25 { vendor | video | native | none } 26 If set to vendor, prefer vendor-specific driver 27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead 28 of the ACPI video.ko driver. 29 If set to video, use the ACPI video.ko driver. 30 If set to native, use the device's native backlight mode. 31 If set to none, disable the ACPI backlight interface. 32 33 acpi_force_32bit_fadt_addr 34 force FADT to use 32 bit addresses rather than the 35 64 bit X_* addresses. Some firmware have broken 64 36 bit addresses for force ACPI ignore these and use 37 the older legacy 32 bit addresses. 38 39 acpica_no_return_repair [HW, ACPI] 40 Disable AML predefined validation mechanism 41 This mechanism can repair the evaluation result to make 42 the return objects more ACPI specification compliant. 43 This option is useful for developers to identify the 44 root cause of an AML interpreter issue when the issue 45 has something to do with the repair mechanism. 46 47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG] 48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG] 49 Format: <int> 50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI 51 debug output. Bits in debug_layer correspond to a 52 _COMPONENT in an ACPI source file, e.g., 53 #define _COMPONENT ACPI_PCI_COMPONENT 54 Bits in debug_level correspond to a level in 55 ACPI_DEBUG_PRINT statements, e.g., 56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ... 57 The debug_level mask defaults to "info". See 58 Documentation/firmware-guide/acpi/debug.rst for more information about 59 debug layers and levels. 60 61 Enable processor driver info messages: 62 acpi.debug_layer=0x20000000 63 Enable PCI/PCI interrupt routing info messages: 64 acpi.debug_layer=0x400000 65 Enable AML "Debug" output, i.e., stores to the Debug 66 object while interpreting AML: 67 acpi.debug_layer=0xffffffff acpi.debug_level=0x2 68 Enable all messages related to ACPI hardware: 69 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff 70 71 Some values produce so much output that the system is 72 unusable. The "log_buf_len" parameter may be useful 73 if you need to capture more output. 74 75 acpi_enforce_resources= [ACPI] 76 { strict | lax | no } 77 Check for resource conflicts between native drivers 78 and ACPI OperationRegions (SystemIO and SystemMemory 79 only). IO ports and memory declared in ACPI might be 80 used by the ACPI subsystem in arbitrary AML code and 81 can interfere with legacy drivers. 82 strict (default): access to resources claimed by ACPI 83 is denied; legacy drivers trying to access reserved 84 resources will fail to bind to device using them. 85 lax: access to resources claimed by ACPI is allowed; 86 legacy drivers trying to access reserved resources 87 will bind successfully but a warning message is logged. 88 no: ACPI OperationRegions are not marked as reserved, 89 no further checks are performed. 90 91 acpi_force_table_verification [HW,ACPI] 92 Enable table checksum verification during early stage. 93 By default, this is disabled due to x86 early mapping 94 size limitation. 95 96 acpi_irq_balance [HW,ACPI] 97 ACPI will balance active IRQs 98 default in APIC mode 99 100 acpi_irq_nobalance [HW,ACPI] 101 ACPI will not move active IRQs (default) 102 default in PIC mode 103 104 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA 105 Format: <irq>,<irq>... 106 107 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for 108 use by PCI 109 Format: <irq>,<irq>... 110 111 acpi_mask_gpe= [HW,ACPI] 112 Due to the existence of _Lxx/_Exx, some GPEs triggered 113 by unsupported hardware/firmware features can result in 114 GPE floodings that cannot be automatically disabled by 115 the GPE dispatcher. 116 This facility can be used to prevent such uncontrolled 117 GPE floodings. 118 Format: <byte> 119 120 acpi_no_auto_serialize [HW,ACPI] 121 Disable auto-serialization of AML methods 122 AML control methods that contain the opcodes to create 123 named objects will be marked as "Serialized" by the 124 auto-serialization feature. 125 This feature is enabled by default. 126 This option allows to turn off the feature. 127 128 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump 129 kernels. 130 131 acpi_no_static_ssdt [HW,ACPI] 132 Disable installation of static SSDTs at early boot time 133 By default, SSDTs contained in the RSDT/XSDT will be 134 installed automatically and they will appear under 135 /sys/firmware/acpi/tables. 136 This option turns off this feature. 137 Note that specifying this option does not affect 138 dynamic table installation which will install SSDT 139 tables to /sys/firmware/acpi/tables/dynamic. 140 141 acpi_no_watchdog [HW,ACPI,WDT] 142 Ignore the ACPI-based watchdog interface (WDAT) and let 143 a native driver control the watchdog device instead. 144 145 acpi_rsdp= [ACPI,EFI,KEXEC] 146 Pass the RSDP address to the kernel, mostly used 147 on machines running EFI runtime service to boot the 148 second kernel for kdump. 149 150 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS 151 Format: To spoof as Windows 98: ="Microsoft Windows" 152 153 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead 154 of 2 which is mandated by ACPI 6) as the supported ACPI 155 specification revision (when using this switch, it may 156 be necessary to carry out a cold reboot _twice_ in a 157 row to make it take effect on the platform firmware). 158 159 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings 160 acpi_osi="string1" # add string1 161 acpi_osi="!string2" # remove string2 162 acpi_osi=!* # remove all strings 163 acpi_osi=! # disable all built-in OS vendor 164 strings 165 acpi_osi=!! # enable all built-in OS vendor 166 strings 167 acpi_osi= # disable all strings 168 169 'acpi_osi=!' can be used in combination with single or 170 multiple 'acpi_osi="string1"' to support specific OS 171 vendor string(s). Note that such command can only 172 affect the default state of the OS vendor strings, thus 173 it cannot affect the default state of the feature group 174 strings and the current state of the OS vendor strings, 175 specifying it multiple times through kernel command line 176 is meaningless. This command is useful when one do not 177 care about the state of the feature group strings which 178 should be controlled by the OSPM. 179 Examples: 180 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent 181 to 'acpi_osi="Windows 2000" acpi_osi=!', they all 182 can make '_OSI("Windows 2000")' TRUE. 183 184 'acpi_osi=' cannot be used in combination with other 185 'acpi_osi=' command lines, the _OSI method will not 186 exist in the ACPI namespace. NOTE that such command can 187 only affect the _OSI support state, thus specifying it 188 multiple times through kernel command line is also 189 meaningless. 190 Examples: 191 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)' 192 FALSE. 193 194 'acpi_osi=!*' can be used in combination with single or 195 multiple 'acpi_osi="string1"' to support specific 196 string(s). Note that such command can affect the 197 current state of both the OS vendor strings and the 198 feature group strings, thus specifying it multiple times 199 through kernel command line is meaningful. But it may 200 still not able to affect the final state of a string if 201 there are quirks related to this string. This command 202 is useful when one want to control the state of the 203 feature group strings to debug BIOS issues related to 204 the OSPM features. 205 Examples: 206 1. 'acpi_osi="Module Device" acpi_osi=!*' can make 207 '_OSI("Module Device")' FALSE. 208 2. 'acpi_osi=!* acpi_osi="Module Device"' can make 209 '_OSI("Module Device")' TRUE. 210 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is 211 equivalent to 212 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"' 213 and 214 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!', 215 they all will make '_OSI("Windows 2000")' TRUE. 216 217 acpi_pm_good [X86] 218 Override the pmtimer bug detection: force the kernel 219 to assume that this machine's pmtimer latches its value 220 and always returns good values. 221 222 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode 223 Format: { level | edge | high | low } 224 225 acpi_skip_timer_override [HW,ACPI] 226 Recognize and ignore IRQ0/pin2 Interrupt Override. 227 For broken nForce2 BIOS resulting in XT-PIC timer. 228 229 acpi_sleep= [HW,ACPI] Sleep options 230 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig, 231 old_ordering, nonvs, sci_force_enable, nobl } 232 See Documentation/power/video.rst for information on 233 s3_bios and s3_mode. 234 s3_beep is for debugging; it makes the PC's speaker beep 235 as soon as the kernel's real-mode entry point is called. 236 s4_nohwsig prevents ACPI hardware signature from being 237 used during resume from hibernation. 238 old_ordering causes the ACPI 1.0 ordering of the _PTS 239 control method, with respect to putting devices into 240 low power states, to be enforced (the ACPI 2.0 ordering 241 of _PTS is used by default). 242 nonvs prevents the kernel from saving/restoring the 243 ACPI NVS memory during suspend/hibernation and resume. 244 sci_force_enable causes the kernel to set SCI_EN directly 245 on resume from S1/S3 (which is against the ACPI spec, 246 but some broken systems don't work without it). 247 nobl causes the internal blacklist of systems known to 248 behave incorrectly in some ways with respect to system 249 suspend and resume to be ignored (use wisely). 250 251 acpi_use_timer_override [HW,ACPI] 252 Use timer override. For some broken Nvidia NF5 boards 253 that require a timer override, but don't have HPET 254 255 add_efi_memmap [EFI; X86] Include EFI memory map in 256 kernel's map of available physical RAM. 257 258 agp= [AGP] 259 { off | try_unsupported } 260 off: disable AGP support 261 try_unsupported: try to drive unsupported chipsets 262 (may crash computer or cause data corruption) 263 264 ALSA [HW,ALSA] 265 See Documentation/sound/alsa-configuration.rst 266 267 alignment= [KNL,ARM] 268 Allow the default userspace alignment fault handler 269 behaviour to be specified. Bit 0 enables warnings, 270 bit 1 enables fixups, and bit 2 sends a segfault. 271 272 align_va_addr= [X86-64] 273 Align virtual addresses by clearing slice [14:12] when 274 allocating a VMA at process creation time. This option 275 gives you up to 3% performance improvement on AMD F15h 276 machines (where it is enabled by default) for a 277 CPU-intensive style benchmark, and it can vary highly in 278 a microbenchmark depending on workload and compiler. 279 280 32: only for 32-bit processes 281 64: only for 64-bit processes 282 on: enable for both 32- and 64-bit processes 283 off: disable for both 32- and 64-bit processes 284 285 alloc_snapshot [FTRACE] 286 Allocate the ftrace snapshot buffer on boot up when the 287 main buffer is allocated. This is handy if debugging 288 and you need to use tracing_snapshot() on boot up, and 289 do not want to use tracing_snapshot_alloc() as it needs 290 to be done where GFP_KERNEL allocations are allowed. 291 292 amd_iommu= [HW,X86-64] 293 Pass parameters to the AMD IOMMU driver in the system. 294 Possible values are: 295 fullflush - enable flushing of IO/TLB entries when 296 they are unmapped. Otherwise they are 297 flushed before they will be reused, which 298 is a lot of faster 299 off - do not initialize any AMD IOMMU found in 300 the system 301 force_isolation - Force device isolation for all 302 devices. The IOMMU driver is not 303 allowed anymore to lift isolation 304 requirements as needed. This option 305 does not override iommu=pt 306 307 amd_iommu_dump= [HW,X86-64] 308 Enable AMD IOMMU driver option to dump the ACPI table 309 for AMD IOMMU. With this option enabled, AMD IOMMU 310 driver will print ACPI tables for AMD IOMMU during 311 IOMMU initialization. 312 313 amd_iommu_intr= [HW,X86-64] 314 Specifies one of the following AMD IOMMU interrupt 315 remapping modes: 316 legacy - Use legacy interrupt remapping mode. 317 vapic - Use virtual APIC mode, which allows IOMMU 318 to inject interrupts directly into guest. 319 This mode requires kvm-amd.avic=1. 320 (Default when IOMMU HW support is present.) 321 322 amijoy.map= [HW,JOY] Amiga joystick support 323 Map of devices attached to JOY0DAT and JOY1DAT 324 Format: <a>,<b> 325 See also Documentation/input/joydev/joystick.rst 326 327 analog.map= [HW,JOY] Analog joystick and gamepad support 328 Specifies type or capabilities of an analog joystick 329 connected to one of 16 gameports 330 Format: <type1>,<type2>,..<type16> 331 332 apc= [HW,SPARC] 333 Power management functions (SPARCstation-4/5 + deriv.) 334 Format: noidle 335 Disable APC CPU standby support. SPARCstation-Fox does 336 not play well with APC CPU idle - disable it if you have 337 APC and your system crashes randomly. 338 339 apic= [APIC,X86] Advanced Programmable Interrupt Controller 340 Change the output verbosity while booting 341 Format: { quiet (default) | verbose | debug } 342 Change the amount of debugging information output 343 when initialising the APIC and IO-APIC components. 344 For X86-32, this can also be used to specify an APIC 345 driver name. 346 Format: apic=driver_name 347 Examples: apic=bigsmp 348 349 apic_extnmi= [APIC,X86] External NMI delivery setting 350 Format: { bsp (default) | all | none } 351 bsp: External NMI is delivered only to CPU 0 352 all: External NMIs are broadcast to all CPUs as a 353 backup of CPU 0 354 none: External NMI is masked for all CPUs. This is 355 useful so that a dump capture kernel won't be 356 shot down by NMI 357 358 autoconf= [IPV6] 359 See Documentation/networking/ipv6.rst. 360 361 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller 362 Limit apic dumping. The parameter defines the maximal 363 number of local apics being dumped. Also it is possible 364 to set it to "all" by meaning -- no limit here. 365 Format: { 1 (default) | 2 | ... | all }. 366 The parameter valid if only apic=debug or 367 apic=verbose is specified. 368 Example: apic=debug show_lapic=all 369 370 apm= [APM] Advanced Power Management 371 See header of arch/x86/kernel/apm_32.c. 372 373 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards 374 Format: <io>,<irq>,<nodeID> 375 376 ataflop= [HW,M68k] 377 378 atarimouse= [HW,MOUSE] Atari Mouse 379 380 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess, 381 EzKey and similar keyboards 382 383 atkbd.reset= [HW] Reset keyboard during initialization 384 385 atkbd.set= [HW] Select keyboard code set 386 Format: <int> (2 = AT (default), 3 = PS/2) 387 388 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar 389 keyboards 390 391 atkbd.softraw= [HW] Choose between synthetic and real raw mode 392 Format: <bool> (0 = real, 1 = synthetic (default)) 393 394 atkbd.softrepeat= [HW] 395 Use software keyboard repeat 396 397 audit= [KNL] Enable the audit sub-system 398 Format: { "0" | "1" | "off" | "on" } 399 0 | off - kernel audit is disabled and can not be 400 enabled until the next reboot 401 unset - kernel audit is initialized but disabled and 402 will be fully enabled by the userspace auditd. 403 1 | on - kernel audit is initialized and partially 404 enabled, storing at most audit_backlog_limit 405 messages in RAM until it is fully enabled by the 406 userspace auditd. 407 Default: unset 408 409 audit_backlog_limit= [KNL] Set the audit queue size limit. 410 Format: <int> (must be >=0) 411 Default: 64 412 413 bau= [X86_UV] Enable the BAU on SGI UV. The default 414 behavior is to disable the BAU (i.e. bau=0). 415 Format: { "0" | "1" } 416 0 - Disable the BAU. 417 1 - Enable the BAU. 418 unset - Disable the BAU. 419 420 baycom_epp= [HW,AX25] 421 Format: <io>,<mode> 422 423 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem 424 Format: <io>,<mode> 425 See header of drivers/net/hamradio/baycom_par.c. 426 427 baycom_ser_fdx= [HW,AX25] 428 BayCom Serial Port AX.25 Modem (Full Duplex Mode) 429 Format: <io>,<irq>,<mode>[,<baud>] 430 See header of drivers/net/hamradio/baycom_ser_fdx.c. 431 432 baycom_ser_hdx= [HW,AX25] 433 BayCom Serial Port AX.25 Modem (Half Duplex Mode) 434 Format: <io>,<irq>,<mode> 435 See header of drivers/net/hamradio/baycom_ser_hdx.c. 436 437 blkdevparts= Manual partition parsing of block device(s) for 438 embedded devices based on command line input. 439 See Documentation/block/cmdline-partition.rst 440 441 boot_delay= Milliseconds to delay each printk during boot. 442 Values larger than 10 seconds (10000) are changed to 443 no delay (0). 444 Format: integer 445 446 bootconfig [KNL] 447 Extended command line options can be added to an initrd 448 and this will cause the kernel to look for it. 449 450 See Documentation/admin-guide/bootconfig.rst 451 452 bert_disable [ACPI] 453 Disable BERT OS support on buggy BIOSes. 454 455 bgrt_disable [ACPI][X86] 456 Disable BGRT to avoid flickering OEM logo. 457 458 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards) 459 bttv.radio= Most important insmod options are available as 460 kernel args too. 461 bttv.pll= See Documentation/admin-guide/media/bttv.rst 462 bttv.tuner= 463 464 bulk_remove=off [PPC] This parameter disables the use of the pSeries 465 firmware feature for flushing multiple hpte entries 466 at a time. 467 468 c101= [NET] Moxa C101 synchronous serial card 469 470 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection. 471 Sometimes CPU hardware bugs make them report the cache 472 size incorrectly. The kernel will attempt work arounds 473 to fix known problems, but for some CPUs it is not 474 possible to determine what the correct size should be. 475 This option provides an override for these situations. 476 477 carrier_timeout= 478 [NET] Specifies amount of time (in seconds) that 479 the kernel should wait for a network carrier. By default 480 it waits 120 seconds. 481 482 ca_keys= [KEYS] This parameter identifies a specific key(s) on 483 the system trusted keyring to be used for certificate 484 trust validation. 485 format: { id:<keyid> | builtin } 486 487 cca= [MIPS] Override the kernel pages' cache coherency 488 algorithm. Accepted values range from 0 to 7 489 inclusive. See arch/mips/include/asm/pgtable-bits.h 490 for platform specific values (SB1, Loongson3 and 491 others). 492 493 ccw_timeout_log [S390] 494 See Documentation/s390/common_io.rst for details. 495 496 cgroup_disable= [KNL] Disable a particular controller 497 Format: {name of the controller(s) to disable} 498 The effects of cgroup_disable=foo are: 499 - foo isn't auto-mounted if you mount all cgroups in 500 a single hierarchy 501 - foo isn't visible as an individually mountable 502 subsystem 503 {Currently only "memory" controller deal with this and 504 cut the overhead, others just disable the usage. So 505 only cgroup_disable=memory is actually worthy} 506 507 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1 508 Format: { { controller | "all" | "named" } 509 [,{ controller | "all" | "named" }...] } 510 Like cgroup_disable, but only applies to cgroup v1; 511 the blacklisted controllers remain available in cgroup2. 512 "all" blacklists all controllers and "named" disables 513 named mounts. Specifying both "all" and "named" disables 514 all v1 hierarchies. 515 516 cgroup.memory= [KNL] Pass options to the cgroup memory controller. 517 Format: <string> 518 nosocket -- Disable socket memory accounting. 519 nokmem -- Disable kernel memory accounting. 520 kmem -- Enable kernel memory accounting. 521 522 checkreqprot [SELINUX] Set initial checkreqprot flag value. 523 Format: { "0" | "1" } 524 See security/selinux/Kconfig help text. 525 0 -- check protection applied by kernel (includes 526 any implied execute protection). 527 1 -- check protection requested by application. 528 Default value is set via a kernel config option. 529 Value can be changed at runtime via 530 /sys/fs/selinux/checkreqprot. 531 Setting checkreqprot to 1 is deprecated. 532 533 cio_ignore= [S390] 534 See Documentation/s390/common_io.rst for details. 535 clk_ignore_unused 536 [CLK] 537 Prevents the clock framework from automatically gating 538 clocks that have not been explicitly enabled by a Linux 539 device driver but are enabled in hardware at reset or 540 by the bootloader/firmware. Note that this does not 541 force such clocks to be always-on nor does it reserve 542 those clocks in any way. This parameter is useful for 543 debug and development, but should not be needed on a 544 platform with proper driver support. For more 545 information, see Documentation/driver-api/clk.rst. 546 547 clock= [BUGS=X86-32, HW] gettimeofday clocksource override. 548 [Deprecated] 549 Forces specified clocksource (if available) to be used 550 when calculating gettimeofday(). If specified 551 clocksource is not available, it defaults to PIT. 552 Format: { pit | tsc | cyclone | pmtmr } 553 554 clocksource= Override the default clocksource 555 Format: <string> 556 Override the default clocksource and use the clocksource 557 with the name specified. 558 Some clocksource names to choose from, depending on 559 the platform: 560 [all] jiffies (this is the base, fallback clocksource) 561 [ACPI] acpi_pm 562 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2, 563 pxa_timer,timer3,32k_counter,timer0_1 564 [X86-32] pit,hpet,tsc; 565 scx200_hrt on Geode; cyclone on IBM x440 566 [MIPS] MIPS 567 [PARISC] cr16 568 [S390] tod 569 [SH] SuperH 570 [SPARC64] tick 571 [X86-64] hpet,tsc 572 573 clocksource.arm_arch_timer.evtstrm= 574 [ARM,ARM64] 575 Format: <bool> 576 Enable/disable the eventstream feature of the ARM 577 architected timer so that code using WFE-based polling 578 loops can be debugged more effectively on production 579 systems. 580 581 clocksource.max_cswd_read_retries= [KNL] 582 Number of clocksource_watchdog() retries due to 583 external delays before the clock will be marked 584 unstable. Defaults to three retries, that is, 585 four attempts to read the clock under test. 586 587 clearcpuid=BITNUM[,BITNUM...] [X86] 588 Disable CPUID feature X for the kernel. See 589 arch/x86/include/asm/cpufeatures.h for the valid bit 590 numbers. Note the Linux specific bits are not necessarily 591 stable over kernel options, but the vendor specific 592 ones should be. 593 Also note that user programs calling CPUID directly 594 or using the feature without checking anything 595 will still see it. This just prevents it from 596 being used by the kernel or shown in /proc/cpuinfo. 597 Also note the kernel might malfunction if you disable 598 some critical bits. 599 600 cma=nn[MG]@[start[MG][-end[MG]]] 601 [KNL,CMA] 602 Sets the size of kernel global memory area for 603 contiguous memory allocations and optionally the 604 placement constraint by the physical address range of 605 memory allocations. A value of 0 disables CMA 606 altogether. For more information, see 607 kernel/dma/contiguous.c 608 609 cma_pernuma=nn[MG] 610 [ARM64,KNL] 611 Sets the size of kernel per-numa memory area for 612 contiguous memory allocations. A value of 0 disables 613 per-numa CMA altogether. And If this option is not 614 specificed, the default value is 0. 615 With per-numa CMA enabled, DMA users on node nid will 616 first try to allocate buffer from the pernuma area 617 which is located in node nid, if the allocation fails, 618 they will fallback to the global default memory area. 619 620 cmo_free_hint= [PPC] Format: { yes | no } 621 Specify whether pages are marked as being inactive 622 when they are freed. This is used in CMO environments 623 to determine OS memory pressure for page stealing by 624 a hypervisor. 625 Default: yes 626 627 coherent_pool=nn[KMG] [ARM,KNL] 628 Sets the size of memory pool for coherent, atomic dma 629 allocations, by default set to 256K. 630 631 com20020= [HW,NET] ARCnet - COM20020 chipset 632 Format: 633 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]] 634 635 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers) 636 Format: <io>[,<irq>] 637 638 com90xx= [HW,NET] 639 ARCnet - COM90xx chipset (memory-mapped buffers) 640 Format: <io>[,<irq>[,<memstart>]] 641 642 condev= [HW,S390] console device 643 conmode= 644 645 console= [KNL] Output console device and options. 646 647 tty<n> Use the virtual console device <n>. 648 649 ttyS<n>[,options] 650 ttyUSB0[,options] 651 Use the specified serial port. The options are of 652 the form "bbbbpnf", where "bbbb" is the baud rate, 653 "p" is parity ("n", "o", or "e"), "n" is number of 654 bits, and "f" is flow control ("r" for RTS or 655 omit it). Default is "9600n8". 656 657 See Documentation/admin-guide/serial-console.rst for more 658 information. See 659 Documentation/networking/netconsole.rst for an 660 alternative. 661 662 uart[8250],io,<addr>[,options] 663 uart[8250],mmio,<addr>[,options] 664 uart[8250],mmio16,<addr>[,options] 665 uart[8250],mmio32,<addr>[,options] 666 uart[8250],0x<addr>[,options] 667 Start an early, polled-mode console on the 8250/16550 668 UART at the specified I/O port or MMIO address, 669 switching to the matching ttyS device later. 670 MMIO inter-register address stride is either 8-bit 671 (mmio), 16-bit (mmio16), or 32-bit (mmio32). 672 If none of [io|mmio|mmio16|mmio32], <addr> is assumed 673 to be equivalent to 'mmio'. 'options' are specified in 674 the same format described for ttyS above; if unspecified, 675 the h/w is not re-initialized. 676 677 hvc<n> Use the hypervisor console device <n>. This is for 678 both Xen and PowerPC hypervisors. 679 680 If the device connected to the port is not a TTY but a braille 681 device, prepend "brl," before the device type, for instance 682 console=brl,ttyS0 683 For now, only VisioBraille is supported. 684 685 console_msg_format= 686 [KNL] Change console messages format 687 default 688 By default we print messages on consoles in 689 "[time stamp] text\n" format (time stamp may not be 690 printed, depending on CONFIG_PRINTK_TIME or 691 `printk_time' param). 692 syslog 693 Switch to syslog format: "<%u>[time stamp] text\n" 694 IOW, each message will have a facility and loglevel 695 prefix. The format is similar to one used by syslog() 696 syscall, or to executing "dmesg -S --raw" or to reading 697 from /proc/kmsg. 698 699 consoleblank= [KNL] The console blank (screen saver) timeout in 700 seconds. A value of 0 disables the blank timer. 701 Defaults to 0. 702 703 coredump_filter= 704 [KNL] Change the default value for 705 /proc/<pid>/coredump_filter. 706 See also Documentation/filesystems/proc.rst. 707 708 coresight_cpu_debug.enable 709 [ARM,ARM64] 710 Format: <bool> 711 Enable/disable the CPU sampling based debugging. 712 0: default value, disable debugging 713 1: enable debugging at boot time 714 715 cpuidle.off=1 [CPU_IDLE] 716 disable the cpuidle sub-system 717 718 cpuidle.governor= 719 [CPU_IDLE] Name of the cpuidle governor to use. 720 721 cpufreq.off=1 [CPU_FREQ] 722 disable the cpufreq sub-system 723 724 cpufreq.default_governor= 725 [CPU_FREQ] Name of the default cpufreq governor or 726 policy to use. This governor must be registered in the 727 kernel before the cpufreq driver probes. 728 729 cpu_init_udelay=N 730 [X86] Delay for N microsec between assert and de-assert 731 of APIC INIT to start processors. This delay occurs 732 on every CPU online, such as boot, and resume from suspend. 733 Default: 10000 734 735 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver 736 Format: 737 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>] 738 739 crashkernel=size[KMG][@offset[KMG]] 740 [KNL] Using kexec, Linux can switch to a 'crash kernel' 741 upon panic. This parameter reserves the physical 742 memory region [offset, offset + size] for that kernel 743 image. If '@offset' is omitted, then a suitable offset 744 is selected automatically. 745 [KNL, X86-64] Select a region under 4G first, and 746 fall back to reserve region above 4G when '@offset' 747 hasn't been specified. 748 See Documentation/admin-guide/kdump/kdump.rst for further details. 749 750 crashkernel=range1:size1[,range2:size2,...][@offset] 751 [KNL] Same as above, but depends on the memory 752 in the running system. The syntax of range is 753 start-[end] where start and end are both 754 a memory unit (amount[KMG]). See also 755 Documentation/admin-guide/kdump/kdump.rst for an example. 756 757 crashkernel=size[KMG],high 758 [KNL, X86-64] range could be above 4G. Allow kernel 759 to allocate physical memory region from top, so could 760 be above 4G if system have more than 4G ram installed. 761 Otherwise memory region will be allocated below 4G, if 762 available. 763 It will be ignored if crashkernel=X is specified. 764 crashkernel=size[KMG],low 765 [KNL, X86-64] range under 4G. When crashkernel=X,high 766 is passed, kernel could allocate physical memory region 767 above 4G, that cause second kernel crash on system 768 that require some amount of low memory, e.g. swiotlb 769 requires at least 64M+32K low memory, also enough extra 770 low memory is needed to make sure DMA buffers for 32-bit 771 devices won't run out. Kernel would try to allocate at 772 at least 256M below 4G automatically. 773 This one let user to specify own low range under 4G 774 for second kernel instead. 775 0: to disable low allocation. 776 It will be ignored when crashkernel=X,high is not used 777 or memory reserved is below 4G. 778 779 cryptomgr.notests 780 [KNL] Disable crypto self-tests 781 782 cs89x0_dma= [HW,NET] 783 Format: <dma> 784 785 cs89x0_media= [HW,NET] 786 Format: { rj45 | aui | bnc } 787 788 dasd= [HW,NET] 789 See header of drivers/s390/block/dasd_devmap.c. 790 791 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port 792 (one device per port) 793 Format: <port#>,<type> 794 See also Documentation/input/devices/joystick-parport.rst 795 796 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot 797 time. See 798 Documentation/admin-guide/dynamic-debug-howto.rst for 799 details. Deprecated, see dyndbg. 800 801 debug [KNL] Enable kernel debugging (events log level). 802 803 debug_boot_weak_hash 804 [KNL] Enable printing [hashed] pointers early in the 805 boot sequence. If enabled, we use a weak hash instead 806 of siphash to hash pointers. Use this option if you are 807 seeing instances of '(___ptrval___)') and need to see a 808 value (hashed pointer) instead. Cryptographically 809 insecure, please do not use on production kernels. 810 811 debug_locks_verbose= 812 [KNL] verbose self-tests 813 Format=<0|1> 814 Print debugging info while doing the locking API 815 self-tests. 816 We default to 0 (no extra messages), setting it to 817 1 will print _a lot_ more information - normally 818 only useful to kernel developers. 819 820 debug_objects [KNL] Enable object debugging 821 822 no_debug_objects 823 [KNL] Disable object debugging 824 825 debug_guardpage_minorder= 826 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this 827 parameter allows control of the order of pages that will 828 be intentionally kept free (and hence protected) by the 829 buddy allocator. Bigger value increase the probability 830 of catching random memory corruption, but reduce the 831 amount of memory for normal system use. The maximum 832 possible value is MAX_ORDER/2. Setting this parameter 833 to 1 or 2 should be enough to identify most random 834 memory corruption problems caused by bugs in kernel or 835 driver code when a CPU writes to (or reads from) a 836 random memory location. Note that there exists a class 837 of memory corruptions problems caused by buggy H/W or 838 F/W or by drivers badly programing DMA (basically when 839 memory is written at bus level and the CPU MMU is 840 bypassed) which are not detectable by 841 CONFIG_DEBUG_PAGEALLOC, hence this option will not help 842 tracking down these problems. 843 844 debug_pagealloc= 845 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter 846 enables the feature at boot time. By default, it is 847 disabled and the system will work mostly the same as a 848 kernel built without CONFIG_DEBUG_PAGEALLOC. 849 Note: to get most of debug_pagealloc error reports, it's 850 useful to also enable the page_owner functionality. 851 on: enable the feature 852 853 debugfs= [KNL] This parameter enables what is exposed to userspace 854 and debugfs internal clients. 855 Format: { on, no-mount, off } 856 on: All functions are enabled. 857 no-mount: 858 Filesystem is not registered but kernel clients can 859 access APIs and a crashkernel can be used to read 860 its content. There is nothing to mount. 861 off: Filesystem is not registered and clients 862 get a -EPERM as result when trying to register files 863 or directories within debugfs. 864 This is equivalent of the runtime functionality if 865 debugfs was not enabled in the kernel at all. 866 Default value is set in build-time with a kernel configuration. 867 868 debugpat [X86] Enable PAT debugging 869 870 decnet.addr= [HW,NET] 871 Format: <area>[,<node>] 872 See also Documentation/networking/decnet.rst. 873 874 default_hugepagesz= 875 [HW] The size of the default HugeTLB page. This is 876 the size represented by the legacy /proc/ hugepages 877 APIs. In addition, this is the default hugetlb size 878 used for shmget(), mmap() and mounting hugetlbfs 879 filesystems. If not specified, defaults to the 880 architecture's default huge page size. Huge page 881 sizes are architecture dependent. See also 882 Documentation/admin-guide/mm/hugetlbpage.rst. 883 Format: size[KMG] 884 885 deferred_probe_timeout= 886 [KNL] Debugging option to set a timeout in seconds for 887 deferred probe to give up waiting on dependencies to 888 probe. Only specific dependencies (subsystems or 889 drivers) that have opted in will be ignored. A timeout of 0 890 will timeout at the end of initcalls. This option will also 891 dump out devices still on the deferred probe list after 892 retrying. 893 894 dfltcc= [HW,S390] 895 Format: { on | off | def_only | inf_only | always } 896 on: s390 zlib hardware support for compression on 897 level 1 and decompression (default) 898 off: No s390 zlib hardware support 899 def_only: s390 zlib hardware support for deflate 900 only (compression on level 1) 901 inf_only: s390 zlib hardware support for inflate 902 only (decompression) 903 always: Same as 'on' but ignores the selected compression 904 level always using hardware support (used for debugging) 905 906 dhash_entries= [KNL] 907 Set number of hash buckets for dentry cache. 908 909 disable_1tb_segments [PPC] 910 Disables the use of 1TB hash page table segments. This 911 causes the kernel to fall back to 256MB segments which 912 can be useful when debugging issues that require an SLB 913 miss to occur. 914 915 stress_slb [PPC] 916 Limits the number of kernel SLB entries, and flushes 917 them frequently to increase the rate of SLB faults 918 on kernel addresses. 919 920 disable= [IPV6] 921 See Documentation/networking/ipv6.rst. 922 923 hardened_usercopy= 924 [KNL] Under CONFIG_HARDENED_USERCOPY, whether 925 hardening is enabled for this boot. Hardened 926 usercopy checking is used to protect the kernel 927 from reading or writing beyond known memory 928 allocation boundaries as a proactive defense 929 against bounds-checking flaws in the kernel's 930 copy_to_user()/copy_from_user() interface. 931 on Perform hardened usercopy checks (default). 932 off Disable hardened usercopy checks. 933 934 disable_radix [PPC] 935 Disable RADIX MMU mode on POWER9 936 937 radix_hcall_invalidate=on [PPC/PSERIES] 938 Disable RADIX GTSE feature and use hcall for TLB 939 invalidate. 940 941 disable_tlbie [PPC] 942 Disable TLBIE instruction. Currently does not work 943 with KVM, with HASH MMU, or with coherent accelerators. 944 945 disable_cpu_apicid= [X86,APIC,SMP] 946 Format: <int> 947 The number of initial APIC ID for the 948 corresponding CPU to be disabled at boot, 949 mostly used for the kdump 2nd kernel to 950 disable BSP to wake up multiple CPUs without 951 causing system reset or hang due to sending 952 INIT from AP to BSP. 953 954 perf_v4_pmi= [X86,INTEL] 955 Format: <bool> 956 Disable Intel PMU counter freezing feature. 957 The feature only exists starting from 958 Arch Perfmon v4 (Skylake and newer). 959 960 disable_ddw [PPC/PSERIES] 961 Disable Dynamic DMA Window support. Use this 962 to workaround buggy firmware. 963 964 disable_ipv6= [IPV6] 965 See Documentation/networking/ipv6.rst. 966 967 disable_mtrr_cleanup [X86] 968 The kernel tries to adjust MTRR layout from continuous 969 to discrete, to make X server driver able to add WB 970 entry later. This parameter disables that. 971 972 disable_mtrr_trim [X86, Intel and AMD only] 973 By default the kernel will trim any uncacheable 974 memory out of your available memory pool based on 975 MTRR settings. This parameter disables that behavior, 976 possibly causing your machine to run very slowly. 977 978 disable_timer_pin_1 [X86] 979 Disable PIN 1 of APIC timer 980 Can be useful to work around chipset bugs. 981 982 dis_ucode_ldr [X86] Disable the microcode loader. 983 984 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support, 985 this option disables the debugging code at boot. 986 987 dma_debug_entries=<number> 988 This option allows to tune the number of preallocated 989 entries for DMA-API debugging code. One entry is 990 required per DMA-API allocation. Use this if the 991 DMA-API debugging code disables itself because the 992 architectural default is too low. 993 994 dma_debug_driver=<driver_name> 995 With this option the DMA-API debugging driver 996 filter feature can be enabled at boot time. Just 997 pass the driver to filter for as the parameter. 998 The filter can be disabled or changed to another 999 driver later using sysfs. 1000 1001 driver_async_probe= [KNL] 1002 List of driver names to be probed asynchronously. 1003 Format: <driver_name1>,<driver_name2>... 1004 1005 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>] 1006 Broken monitors, graphic adapters, KVMs and EDIDless 1007 panels may send no or incorrect EDID data sets. 1008 This parameter allows to specify an EDID data sets 1009 in the /lib/firmware directory that are used instead. 1010 Generic built-in EDID data sets are used, if one of 1011 edid/1024x768.bin, edid/1280x1024.bin, 1012 edid/1680x1050.bin, or edid/1920x1080.bin is given 1013 and no file with the same name exists. Details and 1014 instructions how to build your own EDID data are 1015 available in Documentation/admin-guide/edid.rst. An EDID 1016 data set will only be used for a particular connector, 1017 if its name and a colon are prepended to the EDID 1018 name. Each connector may use a unique EDID data 1019 set by separating the files with a comma. An EDID 1020 data set with no connector name will be used for 1021 any connectors not explicitly specified. 1022 1023 dscc4.setup= [NET] 1024 1025 dt_cpu_ftrs= [PPC] 1026 Format: {"off" | "known"} 1027 Control how the dt_cpu_ftrs device-tree binding is 1028 used for CPU feature discovery and setup (if it 1029 exists). 1030 off: Do not use it, fall back to legacy cpu table. 1031 known: Do not pass through unknown features to guests 1032 or userspace, only those that the kernel is aware of. 1033 1034 dump_apple_properties [X86] 1035 Dump name and content of EFI device properties on 1036 x86 Macs. Useful for driver authors to determine 1037 what data is available or for reverse-engineering. 1038 1039 dyndbg[="val"] [KNL,DYNAMIC_DEBUG] 1040 <module>.dyndbg[="val"] 1041 Enable debug messages at boot time. See 1042 Documentation/admin-guide/dynamic-debug-howto.rst 1043 for details. 1044 1045 nopku [X86] Disable Memory Protection Keys CPU feature found 1046 in some Intel CPUs. 1047 1048 <module>.async_probe [KNL] 1049 Enable asynchronous probe on this module. 1050 1051 early_ioremap_debug [KNL] 1052 Enable debug messages in early_ioremap support. This 1053 is useful for tracking down temporary early mappings 1054 which are not unmapped. 1055 1056 earlycon= [KNL] Output early console device and options. 1057 1058 When used with no options, the early console is 1059 determined by stdout-path property in device tree's 1060 chosen node or the ACPI SPCR table if supported by 1061 the platform. 1062 1063 cdns,<addr>[,options] 1064 Start an early, polled-mode console on a Cadence 1065 (xuartps) serial port at the specified address. Only 1066 supported option is baud rate. If baud rate is not 1067 specified, the serial port must already be setup and 1068 configured. 1069 1070 uart[8250],io,<addr>[,options] 1071 uart[8250],mmio,<addr>[,options] 1072 uart[8250],mmio32,<addr>[,options] 1073 uart[8250],mmio32be,<addr>[,options] 1074 uart[8250],0x<addr>[,options] 1075 Start an early, polled-mode console on the 8250/16550 1076 UART at the specified I/O port or MMIO address. 1077 MMIO inter-register address stride is either 8-bit 1078 (mmio) or 32-bit (mmio32 or mmio32be). 1079 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed 1080 to be equivalent to 'mmio'. 'options' are specified 1081 in the same format described for "console=ttyS<n>"; if 1082 unspecified, the h/w is not initialized. 1083 1084 pl011,<addr> 1085 pl011,mmio32,<addr> 1086 Start an early, polled-mode console on a pl011 serial 1087 port at the specified address. The pl011 serial port 1088 must already be setup and configured. Options are not 1089 yet supported. If 'mmio32' is specified, then only 1090 the driver will use only 32-bit accessors to read/write 1091 the device registers. 1092 1093 meson,<addr> 1094 Start an early, polled-mode console on a meson serial 1095 port at the specified address. The serial port must 1096 already be setup and configured. Options are not yet 1097 supported. 1098 1099 msm_serial,<addr> 1100 Start an early, polled-mode console on an msm serial 1101 port at the specified address. The serial port 1102 must already be setup and configured. Options are not 1103 yet supported. 1104 1105 msm_serial_dm,<addr> 1106 Start an early, polled-mode console on an msm serial 1107 dm port at the specified address. The serial port 1108 must already be setup and configured. Options are not 1109 yet supported. 1110 1111 owl,<addr> 1112 Start an early, polled-mode console on a serial port 1113 of an Actions Semi SoC, such as S500 or S900, at the 1114 specified address. The serial port must already be 1115 setup and configured. Options are not yet supported. 1116 1117 rda,<addr> 1118 Start an early, polled-mode console on a serial port 1119 of an RDA Micro SoC, such as RDA8810PL, at the 1120 specified address. The serial port must already be 1121 setup and configured. Options are not yet supported. 1122 1123 sbi 1124 Use RISC-V SBI (Supervisor Binary Interface) for early 1125 console. 1126 1127 smh Use ARM semihosting calls for early console. 1128 1129 s3c2410,<addr> 1130 s3c2412,<addr> 1131 s3c2440,<addr> 1132 s3c6400,<addr> 1133 s5pv210,<addr> 1134 exynos4210,<addr> 1135 Use early console provided by serial driver available 1136 on Samsung SoCs, requires selecting proper type and 1137 a correct base address of the selected UART port. The 1138 serial port must already be setup and configured. 1139 Options are not yet supported. 1140 1141 lantiq,<addr> 1142 Start an early, polled-mode console on a lantiq serial 1143 (lqasc) port at the specified address. The serial port 1144 must already be setup and configured. Options are not 1145 yet supported. 1146 1147 lpuart,<addr> 1148 lpuart32,<addr> 1149 Use early console provided by Freescale LP UART driver 1150 found on Freescale Vybrid and QorIQ LS1021A processors. 1151 A valid base address must be provided, and the serial 1152 port must already be setup and configured. 1153 1154 ec_imx21,<addr> 1155 ec_imx6q,<addr> 1156 Start an early, polled-mode, output-only console on the 1157 Freescale i.MX UART at the specified address. The UART 1158 must already be setup and configured. 1159 1160 ar3700_uart,<addr> 1161 Start an early, polled-mode console on the 1162 Armada 3700 serial port at the specified 1163 address. The serial port must already be setup 1164 and configured. Options are not yet supported. 1165 1166 qcom_geni,<addr> 1167 Start an early, polled-mode console on a Qualcomm 1168 Generic Interface (GENI) based serial port at the 1169 specified address. The serial port must already be 1170 setup and configured. Options are not yet supported. 1171 1172 efifb,[options] 1173 Start an early, unaccelerated console on the EFI 1174 memory mapped framebuffer (if available). On cache 1175 coherent non-x86 systems that use system memory for 1176 the framebuffer, pass the 'ram' option so that it is 1177 mapped with the correct attributes. 1178 1179 linflex,<addr> 1180 Use early console provided by Freescale LINFlexD UART 1181 serial driver for NXP S32V234 SoCs. A valid base 1182 address must be provided, and the serial port must 1183 already be setup and configured. 1184 1185 earlyprintk= [X86,SH,ARM,M68k,S390] 1186 earlyprintk=vga 1187 earlyprintk=sclp 1188 earlyprintk=xen 1189 earlyprintk=serial[,ttySn[,baudrate]] 1190 earlyprintk=serial[,0x...[,baudrate]] 1191 earlyprintk=ttySn[,baudrate] 1192 earlyprintk=dbgp[debugController#] 1193 earlyprintk=pciserial[,force],bus:device.function[,baudrate] 1194 earlyprintk=xdbc[xhciController#] 1195 1196 earlyprintk is useful when the kernel crashes before 1197 the normal console is initialized. It is not enabled by 1198 default because it has some cosmetic problems. 1199 1200 Append ",keep" to not disable it when the real console 1201 takes over. 1202 1203 Only one of vga, efi, serial, or usb debug port can 1204 be used at a time. 1205 1206 Currently only ttyS0 and ttyS1 may be specified by 1207 name. Other I/O ports may be explicitly specified 1208 on some architectures (x86 and arm at least) by 1209 replacing ttySn with an I/O port address, like this: 1210 earlyprintk=serial,0x1008,115200 1211 You can find the port for a given device in 1212 /proc/tty/driver/serial: 1213 2: uart:ST16650V2 port:00001008 irq:18 ... 1214 1215 Interaction with the standard serial driver is not 1216 very good. 1217 1218 The VGA and EFI output is eventually overwritten by 1219 the real console. 1220 1221 The xen output can only be used by Xen PV guests. 1222 1223 The sclp output can only be used on s390. 1224 1225 The optional "force" to "pciserial" enables use of a 1226 PCI device even when its classcode is not of the 1227 UART class. 1228 1229 edac_report= [HW,EDAC] Control how to report EDAC event 1230 Format: {"on" | "off" | "force"} 1231 on: enable EDAC to report H/W event. May be overridden 1232 by other higher priority error reporting module. 1233 off: disable H/W event reporting through EDAC. 1234 force: enforce the use of EDAC to report H/W event. 1235 default: on. 1236 1237 ekgdboc= [X86,KGDB] Allow early kernel console debugging 1238 ekgdboc=kbd 1239 1240 This is designed to be used in conjunction with 1241 the boot argument: earlyprintk=vga 1242 1243 This parameter works in place of the kgdboc parameter 1244 but can only be used if the backing tty is available 1245 very early in the boot process. For early debugging 1246 via a serial port see kgdboc_earlycon instead. 1247 1248 edd= [EDD] 1249 Format: {"off" | "on" | "skip[mbr]"} 1250 1251 efi= [EFI] 1252 Format: { "debug", "disable_early_pci_dma", 1253 "nochunk", "noruntime", "nosoftreserve", 1254 "novamap", "no_disable_early_pci_dma" } 1255 debug: enable misc debug output. 1256 disable_early_pci_dma: disable the busmaster bit on all 1257 PCI bridges while in the EFI boot stub. 1258 nochunk: disable reading files in "chunks" in the EFI 1259 boot stub, as chunking can cause problems with some 1260 firmware implementations. 1261 noruntime : disable EFI runtime services support 1262 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose) 1263 attribute may cause the kernel to reserve the 1264 memory range for a memory mapping driver to 1265 claim. Specify efi=nosoftreserve to disable this 1266 reservation and treat the memory by its base type 1267 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM"). 1268 novamap: do not call SetVirtualAddressMap(). 1269 no_disable_early_pci_dma: Leave the busmaster bit set 1270 on all PCI bridges while in the EFI boot stub 1271 1272 efi_no_storage_paranoia [EFI; X86] 1273 Using this parameter you can use more than 50% of 1274 your efi variable storage. Use this parameter only if 1275 you are really sure that your UEFI does sane gc and 1276 fulfills the spec otherwise your board may brick. 1277 1278 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86] 1279 Add arbitrary attribute to specific memory range by 1280 updating original EFI memory map. 1281 Region of memory which aa attribute is added to is 1282 from ss to ss+nn. 1283 1284 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000 1285 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000) 1286 attribute is added to range 0x100000000-0x180000000 and 1287 0x10a0000000-0x1120000000. 1288 1289 If efi_fake_mem=8G@9G:0x40000 is specified, the 1290 EFI_MEMORY_SP(0x40000) attribute is added to 1291 range 0x240000000-0x43fffffff. 1292 1293 Using this parameter you can do debugging of EFI memmap 1294 related features. For example, you can do debugging of 1295 Address Range Mirroring feature even if your box 1296 doesn't support it, or mark specific memory as 1297 "soft reserved". 1298 1299 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT 1300 that is to be dynamically loaded by Linux. If there are 1301 multiple variables with the same name but with different 1302 vendor GUIDs, all of them will be loaded. See 1303 Documentation/admin-guide/acpi/ssdt-overlays.rst for details. 1304 1305 1306 eisa_irq_edge= [PARISC,HW] 1307 See header of drivers/parisc/eisa.c. 1308 1309 elanfreq= [X86-32] 1310 See comment before function elanfreq_setup() in 1311 arch/x86/kernel/cpu/cpufreq/elanfreq.c. 1312 1313 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390] 1314 Specifies physical address of start of kernel core 1315 image elf header and optionally the size. Generally 1316 kexec loader will pass this option to capture kernel. 1317 See Documentation/admin-guide/kdump/kdump.rst for details. 1318 1319 enable_mtrr_cleanup [X86] 1320 The kernel tries to adjust MTRR layout from continuous 1321 to discrete, to make X server driver able to add WB 1322 entry later. This parameter enables that. 1323 1324 enable_timer_pin_1 [X86] 1325 Enable PIN 1 of APIC timer 1326 Can be useful to work around chipset bugs 1327 (in particular on some ATI chipsets). 1328 The kernel tries to set a reasonable default. 1329 1330 enforcing [SELINUX] Set initial enforcing status. 1331 Format: {"0" | "1"} 1332 See security/selinux/Kconfig help text. 1333 0 -- permissive (log only, no denials). 1334 1 -- enforcing (deny and log). 1335 Default value is 0. 1336 Value can be changed at runtime via 1337 /sys/fs/selinux/enforce. 1338 1339 erst_disable [ACPI] 1340 Disable Error Record Serialization Table (ERST) 1341 support. 1342 1343 ether= [HW,NET] Ethernet cards parameters 1344 This option is obsoleted by the "netdev=" option, which 1345 has equivalent usage. See its documentation for details. 1346 1347 evm= [EVM] 1348 Format: { "fix" } 1349 Permit 'security.evm' to be updated regardless of 1350 current integrity status. 1351 1352 failslab= 1353 fail_usercopy= 1354 fail_page_alloc= 1355 fail_make_request=[KNL] 1356 General fault injection mechanism. 1357 Format: <interval>,<probability>,<space>,<times> 1358 See also Documentation/fault-injection/. 1359 1360 fb_tunnels= [NET] 1361 Format: { initns | none } 1362 See Documentation/admin-guide/sysctl/net.rst for 1363 fb_tunnels_only_for_init_ns 1364 1365 floppy= [HW] 1366 See Documentation/admin-guide/blockdev/floppy.rst. 1367 1368 force_pal_cache_flush 1369 [IA-64] Avoid check_sal_cache_flush which may hang on 1370 buggy SAL_CACHE_FLUSH implementations. Using this 1371 parameter will force ia64_sal_cache_flush to call 1372 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH. 1373 1374 forcepae [X86-32] 1375 Forcefully enable Physical Address Extension (PAE). 1376 Many Pentium M systems disable PAE but may have a 1377 functionally usable PAE implementation. 1378 Warning: use of this parameter will taint the kernel 1379 and may cause unknown problems. 1380 1381 ftrace=[tracer] 1382 [FTRACE] will set and start the specified tracer 1383 as early as possible in order to facilitate early 1384 boot debugging. 1385 1386 ftrace_dump_on_oops[=orig_cpu] 1387 [FTRACE] will dump the trace buffers on oops. 1388 If no parameter is passed, ftrace will dump 1389 buffers of all CPUs, but if you pass orig_cpu, it will 1390 dump only the buffer of the CPU that triggered the 1391 oops. 1392 1393 ftrace_filter=[function-list] 1394 [FTRACE] Limit the functions traced by the function 1395 tracer at boot up. function-list is a comma separated 1396 list of functions. This list can be changed at run 1397 time by the set_ftrace_filter file in the debugfs 1398 tracing directory. 1399 1400 ftrace_notrace=[function-list] 1401 [FTRACE] Do not trace the functions specified in 1402 function-list. This list can be changed at run time 1403 by the set_ftrace_notrace file in the debugfs 1404 tracing directory. 1405 1406 ftrace_graph_filter=[function-list] 1407 [FTRACE] Limit the top level callers functions traced 1408 by the function graph tracer at boot up. 1409 function-list is a comma separated list of functions 1410 that can be changed at run time by the 1411 set_graph_function file in the debugfs tracing directory. 1412 1413 ftrace_graph_notrace=[function-list] 1414 [FTRACE] Do not trace from the functions specified in 1415 function-list. This list is a comma separated list of 1416 functions that can be changed at run time by the 1417 set_graph_notrace file in the debugfs tracing directory. 1418 1419 ftrace_graph_max_depth=<uint> 1420 [FTRACE] Used with the function graph tracer. This is 1421 the max depth it will trace into a function. This value 1422 can be changed at run time by the max_graph_depth file 1423 in the tracefs tracing directory. default: 0 (no limit) 1424 1425 fw_devlink= [KNL] Create device links between consumer and supplier 1426 devices by scanning the firmware to infer the 1427 consumer/supplier relationships. This feature is 1428 especially useful when drivers are loaded as modules as 1429 it ensures proper ordering of tasks like device probing 1430 (suppliers first, then consumers), supplier boot state 1431 clean up (only after all consumers have probed), 1432 suspend/resume & runtime PM (consumers first, then 1433 suppliers). 1434 Format: { off | permissive | on | rpm } 1435 off -- Don't create device links from firmware info. 1436 permissive -- Create device links from firmware info 1437 but use it only for ordering boot state clean 1438 up (sync_state() calls). 1439 on -- Create device links from firmware info and use it 1440 to enforce probe and suspend/resume ordering. 1441 rpm -- Like "on", but also use to order runtime PM. 1442 1443 gamecon.map[2|3]= 1444 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad 1445 support via parallel port (up to 5 devices per port) 1446 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5> 1447 See also Documentation/input/devices/joystick-parport.rst 1448 1449 gamma= [HW,DRM] 1450 1451 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART 1452 Format: off | on 1453 default: on 1454 1455 gcov_persist= [GCOV] When non-zero (default), profiling data for 1456 kernel modules is saved and remains accessible via 1457 debugfs, even when the module is unloaded/reloaded. 1458 When zero, profiling data is discarded and associated 1459 debugfs files are removed at module unload time. 1460 1461 goldfish [X86] Enable the goldfish android emulator platform. 1462 Don't use this when you are not running on the 1463 android emulator 1464 1465 gpt [EFI] Forces disk with valid GPT signature but 1466 invalid Protective MBR to be treated as GPT. If the 1467 primary GPT is corrupted, it enables the backup/alternate 1468 GPT to be used instead. 1469 1470 grcan.enable0= [HW] Configuration of physical interface 0. Determines 1471 the "Enable 0" bit of the configuration register. 1472 Format: 0 | 1 1473 Default: 0 1474 grcan.enable1= [HW] Configuration of physical interface 1. Determines 1475 the "Enable 0" bit of the configuration register. 1476 Format: 0 | 1 1477 Default: 0 1478 grcan.select= [HW] Select which physical interface to use. 1479 Format: 0 | 1 1480 Default: 0 1481 grcan.txsize= [HW] Sets the size of the tx buffer. 1482 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0. 1483 Default: 1024 1484 grcan.rxsize= [HW] Sets the size of the rx buffer. 1485 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0. 1486 Default: 1024 1487 1488 gpio-mockup.gpio_mockup_ranges 1489 [HW] Sets the ranges of gpiochip of for this device. 1490 Format: <start1>,<end1>,<start2>,<end2>... 1491 1492 hardlockup_all_cpu_backtrace= 1493 [KNL] Should the hard-lockup detector generate 1494 backtraces on all cpus. 1495 Format: 0 | 1 1496 1497 hashdist= [KNL,NUMA] Large hashes allocated during boot 1498 are distributed across NUMA nodes. Defaults on 1499 for 64-bit NUMA, off otherwise. 1500 Format: 0 | 1 (for off | on) 1501 1502 hcl= [IA-64] SGI's Hardware Graph compatibility layer 1503 1504 hd= [EIDE] (E)IDE hard drive subsystem geometry 1505 Format: <cyl>,<head>,<sect> 1506 1507 hest_disable [ACPI] 1508 Disable Hardware Error Source Table (HEST) support; 1509 corresponding firmware-first mode error processing 1510 logic will be disabled. 1511 1512 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact 1513 size of <nn>. This works even on boxes that have no 1514 highmem otherwise. This also works to reduce highmem 1515 size on bigger boxes. 1516 1517 highres= [KNL] Enable/disable high resolution timer mode. 1518 Valid parameters: "on", "off" 1519 Default: "on" 1520 1521 hlt [BUGS=ARM,SH] 1522 1523 hpet= [X86-32,HPET] option to control HPET usage 1524 Format: { enable (default) | disable | force | 1525 verbose } 1526 disable: disable HPET and use PIT instead 1527 force: allow force enabled of undocumented chips (ICH4, 1528 VIA, nVidia) 1529 verbose: show contents of HPET registers during setup 1530 1531 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET 1532 registers. Default set by CONFIG_HPET_MMAP_DEFAULT. 1533 1534 hugetlb_cma= [HW] The size of a cma area used for allocation 1535 of gigantic hugepages. 1536 Format: nn[KMGTPE] 1537 1538 Reserve a cma area of given size and allocate gigantic 1539 hugepages using the cma allocator. If enabled, the 1540 boot-time allocation of gigantic hugepages is skipped. 1541 1542 hugepages= [HW] Number of HugeTLB pages to allocate at boot. 1543 If this follows hugepagesz (below), it specifies 1544 the number of pages of hugepagesz to be allocated. 1545 If this is the first HugeTLB parameter on the command 1546 line, it specifies the number of pages to allocate for 1547 the default huge page size. See also 1548 Documentation/admin-guide/mm/hugetlbpage.rst. 1549 Format: <integer> 1550 1551 hugepagesz= 1552 [HW] The size of the HugeTLB pages. This is used in 1553 conjunction with hugepages (above) to allocate huge 1554 pages of a specific size at boot. The pair 1555 hugepagesz=X hugepages=Y can be specified once for 1556 each supported huge page size. Huge page sizes are 1557 architecture dependent. See also 1558 Documentation/admin-guide/mm/hugetlbpage.rst. 1559 Format: size[KMG] 1560 1561 hung_task_panic= 1562 [KNL] Should the hung task detector generate panics. 1563 Format: 0 | 1 1564 1565 A value of 1 instructs the kernel to panic when a 1566 hung task is detected. The default value is controlled 1567 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time 1568 option. The value selected by this boot parameter can 1569 be changed later by the kernel.hung_task_panic sysctl. 1570 1571 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC) 1572 terminal devices. Valid values: 0..8 1573 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs. 1574 If specified, z/VM IUCV HVC accepts connections 1575 from listed z/VM user IDs only. 1576 1577 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations 1578 which allow the hypervisor to 'idle' the 1579 guest on lock contention. 1580 1581 keep_bootcon [KNL] 1582 Do not unregister boot console at start. This is only 1583 useful for debugging when something happens in the window 1584 between unregistering the boot console and initializing 1585 the real console. 1586 1587 i2c_bus= [HW] Override the default board specific I2C bus speed 1588 or register an additional I2C bus that is not 1589 registered from board initialization code. 1590 Format: 1591 <bus_id>,<clkrate> 1592 1593 i8042.debug [HW] Toggle i8042 debug mode 1594 i8042.unmask_kbd_data 1595 [HW] Enable printing of interrupt data from the KBD port 1596 (disabled by default, and as a pre-condition 1597 requires that i8042.debug=1 be enabled) 1598 i8042.direct [HW] Put keyboard port into non-translated mode 1599 i8042.dumbkbd [HW] Pretend that controller can only read data from 1600 keyboard and cannot control its state 1601 (Don't attempt to blink the leds) 1602 i8042.noaux [HW] Don't check for auxiliary (== mouse) port 1603 i8042.nokbd [HW] Don't check/create keyboard port 1604 i8042.noloop [HW] Disable the AUX Loopback command while probing 1605 for the AUX port 1606 i8042.nomux [HW] Don't check presence of an active multiplexing 1607 controller 1608 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX 1609 controllers 1610 i8042.notimeout [HW] Ignore timeout condition signalled by controller 1611 i8042.reset [HW] Reset the controller during init, cleanup and 1612 suspend-to-ram transitions, only during s2r 1613 transitions, or never reset 1614 Format: { 1 | Y | y | 0 | N | n } 1615 1, Y, y: always reset controller 1616 0, N, n: don't ever reset controller 1617 Default: only on s2r transitions on x86; most other 1618 architectures force reset to be always executed 1619 i8042.unlock [HW] Unlock (ignore) the keylock 1620 i8042.kbdreset [HW] Reset device connected to KBD port 1621 1622 i810= [HW,DRM] 1623 1624 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data 1625 indicates that the driver is running on unsupported 1626 hardware. 1627 i8k.force [HW] Activate i8k driver even if SMM BIOS signature 1628 does not match list of supported models. 1629 i8k.power_status 1630 [HW] Report power status in /proc/i8k 1631 (disabled by default) 1632 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN 1633 capability is set. 1634 1635 i915.invert_brightness= 1636 [DRM] Invert the sense of the variable that is used to 1637 set the brightness of the panel backlight. Normally a 1638 brightness value of 0 indicates backlight switched off, 1639 and the maximum of the brightness value sets the backlight 1640 to maximum brightness. If this parameter is set to 0 1641 (default) and the machine requires it, or this parameter 1642 is set to 1, a brightness value of 0 sets the backlight 1643 to maximum brightness, and the maximum of the brightness 1644 value switches the backlight off. 1645 -1 -- never invert brightness 1646 0 -- machine default 1647 1 -- force brightness inversion 1648 1649 icn= [HW,ISDN] 1650 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]] 1651 1652 ide-core.nodma= [HW] (E)IDE subsystem 1653 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc 1654 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr 1655 .cdrom .chs .ignore_cable are additional options 1656 See Documentation/ide/ide.rst. 1657 1658 ide-generic.probe-mask= [HW] (E)IDE subsystem 1659 Format: <int> 1660 Probe mask for legacy ISA IDE ports. Depending on 1661 platform up to 6 ports are supported, enabled by 1662 setting corresponding bits in the mask to 1. The 1663 default value is 0x0, which has a special meaning. 1664 On systems that have PCI, it triggers scanning the 1665 PCI bus for the first and the second port, which 1666 are then probed. On systems without PCI the value 1667 of 0x0 enables probing the two first ports as if it 1668 was 0x3. 1669 1670 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem 1671 Claim all unknown PCI IDE storage controllers. 1672 1673 idle= [X86] 1674 Format: idle=poll, idle=halt, idle=nomwait 1675 Poll forces a polling idle loop that can slightly 1676 improve the performance of waking up a idle CPU, but 1677 will use a lot of power and make the system run hot. 1678 Not recommended. 1679 idle=halt: Halt is forced to be used for CPU idle. 1680 In such case C2/C3 won't be used again. 1681 idle=nomwait: Disable mwait for CPU C-states 1682 1683 ieee754= [MIPS] Select IEEE Std 754 conformance mode 1684 Format: { strict | legacy | 2008 | relaxed } 1685 Default: strict 1686 1687 Choose which programs will be accepted for execution 1688 based on the IEEE 754 NaN encoding(s) supported by 1689 the FPU and the NaN encoding requested with the value 1690 of an ELF file header flag individually set by each 1691 binary. Hardware implementations are permitted to 1692 support either or both of the legacy and the 2008 NaN 1693 encoding mode. 1694 1695 Available settings are as follows: 1696 strict accept binaries that request a NaN encoding 1697 supported by the FPU 1698 legacy only accept legacy-NaN binaries, if supported 1699 by the FPU 1700 2008 only accept 2008-NaN binaries, if supported 1701 by the FPU 1702 relaxed accept any binaries regardless of whether 1703 supported by the FPU 1704 1705 The FPU emulator is always able to support both NaN 1706 encodings, so if no FPU hardware is present or it has 1707 been disabled with 'nofpu', then the settings of 1708 'legacy' and '2008' strap the emulator accordingly, 1709 'relaxed' straps the emulator for both legacy-NaN and 1710 2008-NaN, whereas 'strict' enables legacy-NaN only on 1711 legacy processors and both NaN encodings on MIPS32 or 1712 MIPS64 CPUs. 1713 1714 The setting for ABS.fmt/NEG.fmt instruction execution 1715 mode generally follows that for the NaN encoding, 1716 except where unsupported by hardware. 1717 1718 ignore_loglevel [KNL] 1719 Ignore loglevel setting - this will print /all/ 1720 kernel messages to the console. Useful for debugging. 1721 We also add it as printk module parameter, so users 1722 could change it dynamically, usually by 1723 /sys/module/printk/parameters/ignore_loglevel. 1724 1725 ignore_rlimit_data 1726 Ignore RLIMIT_DATA setting for data mappings, 1727 print warning at first misuse. Can be changed via 1728 /sys/module/kernel/parameters/ignore_rlimit_data. 1729 1730 ihash_entries= [KNL] 1731 Set number of hash buckets for inode cache. 1732 1733 ima_appraise= [IMA] appraise integrity measurements 1734 Format: { "off" | "enforce" | "fix" | "log" } 1735 default: "enforce" 1736 1737 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead. 1738 The builtin appraise policy appraises all files 1739 owned by uid=0. 1740 1741 ima_canonical_fmt [IMA] 1742 Use the canonical format for the binary runtime 1743 measurements, instead of host native format. 1744 1745 ima_hash= [IMA] 1746 Format: { md5 | sha1 | rmd160 | sha256 | sha384 1747 | sha512 | ... } 1748 default: "sha1" 1749 1750 The list of supported hash algorithms is defined 1751 in crypto/hash_info.h. 1752 1753 ima_policy= [IMA] 1754 The builtin policies to load during IMA setup. 1755 Format: "tcb | appraise_tcb | secure_boot | 1756 fail_securely" 1757 1758 The "tcb" policy measures all programs exec'd, files 1759 mmap'd for exec, and all files opened with the read 1760 mode bit set by either the effective uid (euid=0) or 1761 uid=0. 1762 1763 The "appraise_tcb" policy appraises the integrity of 1764 all files owned by root. 1765 1766 The "secure_boot" policy appraises the integrity 1767 of files (eg. kexec kernel image, kernel modules, 1768 firmware, policy, etc) based on file signatures. 1769 1770 The "fail_securely" policy forces file signature 1771 verification failure also on privileged mounted 1772 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE 1773 flag. 1774 1775 ima_tcb [IMA] Deprecated. Use ima_policy= instead. 1776 Load a policy which meets the needs of the Trusted 1777 Computing Base. This means IMA will measure all 1778 programs exec'd, files mmap'd for exec, and all files 1779 opened for read by uid=0. 1780 1781 ima_template= [IMA] 1782 Select one of defined IMA measurements template formats. 1783 Formats: { "ima" | "ima-ng" | "ima-sig" } 1784 Default: "ima-ng" 1785 1786 ima_template_fmt= 1787 [IMA] Define a custom template format. 1788 Format: { "field1|...|fieldN" } 1789 1790 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage 1791 Format: <min_file_size> 1792 Set the minimal file size for using asynchronous hash. 1793 If left unspecified, ahash usage is disabled. 1794 1795 ahash performance varies for different data sizes on 1796 different crypto accelerators. This option can be used 1797 to achieve the best performance for a particular HW. 1798 1799 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size 1800 Format: <bufsize> 1801 Set hashing buffer size. Default: 4k. 1802 1803 ahash performance varies for different chunk sizes on 1804 different crypto accelerators. This option can be used 1805 to achieve best performance for particular HW. 1806 1807 init= [KNL] 1808 Format: <full_path> 1809 Run specified binary instead of /sbin/init as init 1810 process. 1811 1812 initcall_debug [KNL] Trace initcalls as they are executed. Useful 1813 for working out where the kernel is dying during 1814 startup. 1815 1816 initcall_blacklist= [KNL] Do not execute a comma-separated list of 1817 initcall functions. Useful for debugging built-in 1818 modules and initcalls. 1819 1820 initrd= [BOOT] Specify the location of the initial ramdisk 1821 1822 initrdmem= [KNL] Specify a physical address and size from which to 1823 load the initrd. If an initrd is compiled in or 1824 specified in the bootparams, it takes priority over this 1825 setting. 1826 Format: ss[KMG],nn[KMG] 1827 Default is 0, 0 1828 1829 init_on_alloc= [MM] Fill newly allocated pages and heap objects with 1830 zeroes. 1831 Format: 0 | 1 1832 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON. 1833 1834 init_on_free= [MM] Fill freed pages and heap objects with zeroes. 1835 Format: 0 | 1 1836 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON. 1837 1838 init_pkru= [X86] Specify the default memory protection keys rights 1839 register contents for all processes. 0x55555554 by 1840 default (disallow access to all but pkey 0). Can 1841 override in debugfs after boot. 1842 1843 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver 1844 Format: <irq> 1845 1846 int_pln_enable [X86] Enable power limit notification interrupt 1847 1848 integrity_audit=[IMA] 1849 Format: { "0" | "1" } 1850 0 -- basic integrity auditing messages. (Default) 1851 1 -- additional integrity auditing messages. 1852 1853 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option 1854 on 1855 Enable intel iommu driver. 1856 off 1857 Disable intel iommu driver. 1858 igfx_off [Default Off] 1859 By default, gfx is mapped as normal device. If a gfx 1860 device has a dedicated DMAR unit, the DMAR unit is 1861 bypassed by not enabling DMAR with this option. In 1862 this case, gfx device will use physical address for 1863 DMA. 1864 forcedac [X86-64] 1865 With this option iommu will not optimize to look 1866 for io virtual address below 32-bit forcing dual 1867 address cycle on pci bus for cards supporting greater 1868 than 32-bit addressing. The default is to look 1869 for translation below 32-bit and if not available 1870 then look in the higher range. 1871 strict [Default Off] 1872 With this option on every unmap_single operation will 1873 result in a hardware IOTLB flush operation as opposed 1874 to batching them for performance. 1875 sp_off [Default Off] 1876 By default, super page will be supported if Intel IOMMU 1877 has the capability. With this option, super page will 1878 not be supported. 1879 sm_on [Default Off] 1880 By default, scalable mode will be disabled even if the 1881 hardware advertises that it has support for the scalable 1882 mode translation. With this option set, scalable mode 1883 will be used on hardware which claims to support it. 1884 tboot_noforce [Default Off] 1885 Do not force the Intel IOMMU enabled under tboot. 1886 By default, tboot will force Intel IOMMU on, which 1887 could harm performance of some high-throughput 1888 devices like 40GBit network cards, even if identity 1889 mapping is enabled. 1890 Note that using this option lowers the security 1891 provided by tboot because it makes the system 1892 vulnerable to DMA attacks. 1893 nobounce [Default off] 1894 Disable bounce buffer for untrusted devices such as 1895 the Thunderbolt devices. This will treat the untrusted 1896 devices as the trusted ones, hence might expose security 1897 risks of DMA attacks. 1898 1899 intel_idle.max_cstate= [KNL,HW,ACPI,X86] 1900 0 disables intel_idle and fall back on acpi_idle. 1901 1 to 9 specify maximum depth of C-state. 1902 1903 intel_pstate= [X86] 1904 disable 1905 Do not enable intel_pstate as the default 1906 scaling driver for the supported processors 1907 passive 1908 Use intel_pstate as a scaling driver, but configure it 1909 to work with generic cpufreq governors (instead of 1910 enabling its internal governor). This mode cannot be 1911 used along with the hardware-managed P-states (HWP) 1912 feature. 1913 force 1914 Enable intel_pstate on systems that prohibit it by default 1915 in favor of acpi-cpufreq. Forcing the intel_pstate driver 1916 instead of acpi-cpufreq may disable platform features, such 1917 as thermal controls and power capping, that rely on ACPI 1918 P-States information being indicated to OSPM and therefore 1919 should be used with caution. This option does not work with 1920 processors that aren't supported by the intel_pstate driver 1921 or on platforms that use pcc-cpufreq instead of acpi-cpufreq. 1922 no_hwp 1923 Do not enable hardware P state control (HWP) 1924 if available. 1925 hwp_only 1926 Only load intel_pstate on systems which support 1927 hardware P state control (HWP) if available. 1928 support_acpi_ppc 1929 Enforce ACPI _PPC performance limits. If the Fixed ACPI 1930 Description Table, specifies preferred power management 1931 profile as "Enterprise Server" or "Performance Server", 1932 then this feature is turned on by default. 1933 per_cpu_perf_limits 1934 Allow per-logical-CPU P-State performance control limits using 1935 cpufreq sysfs interface 1936 1937 intremap= [X86-64, Intel-IOMMU] 1938 on enable Interrupt Remapping (default) 1939 off disable Interrupt Remapping 1940 nosid disable Source ID checking 1941 no_x2apic_optout 1942 BIOS x2APIC opt-out request will be ignored 1943 nopost disable Interrupt Posting 1944 1945 iomem= Disable strict checking of access to MMIO memory 1946 strict regions from userspace. 1947 relaxed 1948 1949 iommu= [X86] 1950 off 1951 force 1952 noforce 1953 biomerge 1954 panic 1955 nopanic 1956 merge 1957 nomerge 1958 soft 1959 pt [X86] 1960 nopt [X86] 1961 nobypass [PPC/POWERNV] 1962 Disable IOMMU bypass, using IOMMU for PCI devices. 1963 1964 iommu.strict= [ARM64] Configure TLB invalidation behaviour 1965 Format: { "0" | "1" } 1966 0 - Lazy mode. 1967 Request that DMA unmap operations use deferred 1968 invalidation of hardware TLBs, for increased 1969 throughput at the cost of reduced device isolation. 1970 Will fall back to strict mode if not supported by 1971 the relevant IOMMU driver. 1972 1 - Strict mode (default). 1973 DMA unmap operations invalidate IOMMU hardware TLBs 1974 synchronously. 1975 1976 iommu.passthrough= 1977 [ARM64, X86] Configure DMA to bypass the IOMMU by default. 1978 Format: { "0" | "1" } 1979 0 - Use IOMMU translation for DMA. 1980 1 - Bypass the IOMMU for DMA. 1981 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH. 1982 1983 io7= [HW] IO7 for Marvel-based Alpha systems 1984 See comment before marvel_specify_io7 in 1985 arch/alpha/kernel/core_marvel.c. 1986 1987 io_delay= [X86] I/O delay method 1988 0x80 1989 Standard port 0x80 based delay 1990 0xed 1991 Alternate port 0xed based delay (needed on some systems) 1992 udelay 1993 Simple two microseconds delay 1994 none 1995 No delay 1996 1997 ip= [IP_PNP] 1998 See Documentation/admin-guide/nfs/nfsroot.rst. 1999 2000 ipcmni_extend [KNL] Extend the maximum number of unique System V 2001 IPC identifiers from 32,768 to 16,777,216. 2002 2003 irqaffinity= [SMP] Set the default irq affinity mask 2004 The argument is a cpu list, as described above. 2005 2006 irqchip.gicv2_force_probe= 2007 [ARM, ARM64] 2008 Format: <bool> 2009 Force the kernel to look for the second 4kB page 2010 of a GICv2 controller even if the memory range 2011 exposed by the device tree is too small. 2012 2013 irqchip.gicv3_nolpi= 2014 [ARM, ARM64] 2015 Force the kernel to ignore the availability of 2016 LPIs (and by consequence ITSs). Intended for system 2017 that use the kernel as a bootloader, and thus want 2018 to let secondary kernels in charge of setting up 2019 LPIs. 2020 2021 irqchip.gicv3_pseudo_nmi= [ARM64] 2022 Enables support for pseudo-NMIs in the kernel. This 2023 requires the kernel to be built with 2024 CONFIG_ARM64_PSEUDO_NMI. 2025 2026 irqfixup [HW] 2027 When an interrupt is not handled search all handlers 2028 for it. Intended to get systems with badly broken 2029 firmware running. 2030 2031 irqpoll [HW] 2032 When an interrupt is not handled search all handlers 2033 for it. Also check all handlers each timer 2034 interrupt. Intended to get systems with badly broken 2035 firmware running. 2036 2037 isapnp= [ISAPNP] 2038 Format: <RDP>,<reset>,<pci_scan>,<verbosity> 2039 2040 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance. 2041 [Deprecated - use cpusets instead] 2042 Format: [flag-list,]<cpu-list> 2043 2044 Specify one or more CPUs to isolate from disturbances 2045 specified in the flag list (default: domain): 2046 2047 nohz 2048 Disable the tick when a single task runs. 2049 2050 A residual 1Hz tick is offloaded to workqueues, which you 2051 need to affine to housekeeping through the global 2052 workqueue's affinity configured via the 2053 /sys/devices/virtual/workqueue/cpumask sysfs file, or 2054 by using the 'domain' flag described below. 2055 2056 NOTE: by default the global workqueue runs on all CPUs, 2057 so to protect individual CPUs the 'cpumask' file has to 2058 be configured manually after bootup. 2059 2060 domain 2061 Isolate from the general SMP balancing and scheduling 2062 algorithms. Note that performing domain isolation this way 2063 is irreversible: it's not possible to bring back a CPU to 2064 the domains once isolated through isolcpus. It's strongly 2065 advised to use cpusets instead to disable scheduler load 2066 balancing through the "cpuset.sched_load_balance" file. 2067 It offers a much more flexible interface where CPUs can 2068 move in and out of an isolated set anytime. 2069 2070 You can move a process onto or off an "isolated" CPU via 2071 the CPU affinity syscalls or cpuset. 2072 <cpu number> begins at 0 and the maximum value is 2073 "number of CPUs in system - 1". 2074 2075 managed_irq 2076 2077 Isolate from being targeted by managed interrupts 2078 which have an interrupt mask containing isolated 2079 CPUs. The affinity of managed interrupts is 2080 handled by the kernel and cannot be changed via 2081 the /proc/irq/* interfaces. 2082 2083 This isolation is best effort and only effective 2084 if the automatically assigned interrupt mask of a 2085 device queue contains isolated and housekeeping 2086 CPUs. If housekeeping CPUs are online then such 2087 interrupts are directed to the housekeeping CPU 2088 so that IO submitted on the housekeeping CPU 2089 cannot disturb the isolated CPU. 2090 2091 If a queue's affinity mask contains only isolated 2092 CPUs then this parameter has no effect on the 2093 interrupt routing decision, though interrupts are 2094 only delivered when tasks running on those 2095 isolated CPUs submit IO. IO submitted on 2096 housekeeping CPUs has no influence on those 2097 queues. 2098 2099 The format of <cpu-list> is described above. 2100 2101 iucv= [HW,NET] 2102 2103 ivrs_ioapic [HW,X86-64] 2104 Provide an override to the IOAPIC-ID<->DEVICE-ID 2105 mapping provided in the IVRS ACPI table. For 2106 example, to map IOAPIC-ID decimal 10 to 2107 PCI device 00:14.0 write the parameter as: 2108 ivrs_ioapic[10]=00:14.0 2109 2110 ivrs_hpet [HW,X86-64] 2111 Provide an override to the HPET-ID<->DEVICE-ID 2112 mapping provided in the IVRS ACPI table. For 2113 example, to map HPET-ID decimal 0 to 2114 PCI device 00:14.0 write the parameter as: 2115 ivrs_hpet[0]=00:14.0 2116 2117 ivrs_acpihid [HW,X86-64] 2118 Provide an override to the ACPI-HID:UID<->DEVICE-ID 2119 mapping provided in the IVRS ACPI table. For 2120 example, to map UART-HID:UID AMD0020:0 to 2121 PCI device 00:14.5 write the parameter as: 2122 ivrs_acpihid[00:14.5]=AMD0020:0 2123 2124 js= [HW,JOY] Analog joystick 2125 See Documentation/input/joydev/joystick.rst. 2126 2127 nokaslr [KNL] 2128 When CONFIG_RANDOMIZE_BASE is set, this disables 2129 kernel and module base offset ASLR (Address Space 2130 Layout Randomization). 2131 2132 kasan_multi_shot 2133 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print 2134 report on every invalid memory access. Without this 2135 parameter KASAN will print report only for the first 2136 invalid access. 2137 2138 keepinitrd [HW,ARM] 2139 2140 kernelcore= [KNL,X86,IA-64,PPC] 2141 Format: nn[KMGTPE] | nn% | "mirror" 2142 This parameter specifies the amount of memory usable by 2143 the kernel for non-movable allocations. The requested 2144 amount is spread evenly throughout all nodes in the 2145 system as ZONE_NORMAL. The remaining memory is used for 2146 movable memory in its own zone, ZONE_MOVABLE. In the 2147 event, a node is too small to have both ZONE_NORMAL and 2148 ZONE_MOVABLE, kernelcore memory will take priority and 2149 other nodes will have a larger ZONE_MOVABLE. 2150 2151 ZONE_MOVABLE is used for the allocation of pages that 2152 may be reclaimed or moved by the page migration 2153 subsystem. Note that allocations like PTEs-from-HighMem 2154 still use the HighMem zone if it exists, and the Normal 2155 zone if it does not. 2156 2157 It is possible to specify the exact amount of memory in 2158 the form of "nn[KMGTPE]", a percentage of total system 2159 memory in the form of "nn%", or "mirror". If "mirror" 2160 option is specified, mirrored (reliable) memory is used 2161 for non-movable allocations and remaining memory is used 2162 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror" 2163 are exclusive, so you cannot specify multiple forms. 2164 2165 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port. 2166 Format: <Controller#>[,poll interval] 2167 The controller # is the number of the ehci usb debug 2168 port as it is probed via PCI. The poll interval is 2169 optional and is the number seconds in between 2170 each poll cycle to the debug port in case you need 2171 the functionality for interrupting the kernel with 2172 gdb or control-c on the dbgp connection. When 2173 not using this parameter you use sysrq-g to break into 2174 the kernel debugger. 2175 2176 kgdboc= [KGDB,HW] kgdb over consoles. 2177 Requires a tty driver that supports console polling, 2178 or a supported polling keyboard driver (non-usb). 2179 Serial only format: <serial_device>[,baud] 2180 keyboard only format: kbd 2181 keyboard and serial format: kbd,<serial_device>[,baud] 2182 Optional Kernel mode setting: 2183 kms, kbd format: kms,kbd 2184 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud] 2185 2186 kgdboc_earlycon= [KGDB,HW] 2187 If the boot console provides the ability to read 2188 characters and can work in polling mode, you can use 2189 this parameter to tell kgdb to use it as a backend 2190 until the normal console is registered. Intended to 2191 be used together with the kgdboc parameter which 2192 specifies the normal console to transition to. 2193 2194 The name of the early console should be specified 2195 as the value of this parameter. Note that the name of 2196 the early console might be different than the tty 2197 name passed to kgdboc. It's OK to leave the value 2198 blank and the first boot console that implements 2199 read() will be picked. 2200 2201 kgdbwait [KGDB] Stop kernel execution and enter the 2202 kernel debugger at the earliest opportunity. 2203 2204 kmac= [MIPS] Korina ethernet MAC address. 2205 Configure the RouterBoard 532 series on-chip 2206 Ethernet adapter MAC address. 2207 2208 kmemleak= [KNL] Boot-time kmemleak enable/disable 2209 Valid arguments: on, off 2210 Default: on 2211 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y, 2212 the default is off. 2213 2214 kprobe_event=[probe-list] 2215 [FTRACE] Add kprobe events and enable at boot time. 2216 The probe-list is a semicolon delimited list of probe 2217 definitions. Each definition is same as kprobe_events 2218 interface, but the parameters are comma delimited. 2219 For example, to add a kprobe event on vfs_read with 2220 arg1 and arg2, add to the command line; 2221 2222 kprobe_event=p,vfs_read,$arg1,$arg2 2223 2224 See also Documentation/trace/kprobetrace.rst "Kernel 2225 Boot Parameter" section. 2226 2227 kpti= [ARM64] Control page table isolation of user 2228 and kernel address spaces. 2229 Default: enabled on cores which need mitigation. 2230 0: force disabled 2231 1: force enabled 2232 2233 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 2234 Default is 0 (don't ignore, but inject #GP) 2235 2236 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface. 2237 Default is false (don't support). 2238 2239 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit 2240 KVM MMU at runtime. 2241 Default is 0 (off) 2242 2243 kvm.nx_huge_pages= 2244 [KVM] Controls the software workaround for the 2245 X86_BUG_ITLB_MULTIHIT bug. 2246 force : Always deploy workaround. 2247 off : Never deploy workaround. 2248 auto : Deploy workaround based on the presence of 2249 X86_BUG_ITLB_MULTIHIT. 2250 2251 Default is 'auto'. 2252 2253 If the software workaround is enabled for the host, 2254 guests do need not to enable it for nested guests. 2255 2256 kvm.nx_huge_pages_recovery_ratio= 2257 [KVM] Controls how many 4KiB pages are periodically zapped 2258 back to huge pages. 0 disables the recovery, otherwise if 2259 the value is N KVM will zap 1/Nth of the 4KiB pages every 2260 minute. The default is 60. 2261 2262 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM. 2263 Default is 1 (enabled) 2264 2265 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU) 2266 for all guests. 2267 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode. 2268 2269 kvm-arm.vgic_v3_group0_trap= 2270 [KVM,ARM] Trap guest accesses to GICv3 group-0 2271 system registers 2272 2273 kvm-arm.vgic_v3_group1_trap= 2274 [KVM,ARM] Trap guest accesses to GICv3 group-1 2275 system registers 2276 2277 kvm-arm.vgic_v3_common_trap= 2278 [KVM,ARM] Trap guest accesses to GICv3 common 2279 system registers 2280 2281 kvm-arm.vgic_v4_enable= 2282 [KVM,ARM] Allow use of GICv4 for direct injection of 2283 LPIs. 2284 2285 kvm_cma_resv_ratio=n [PPC] 2286 Reserves given percentage from system memory area for 2287 contiguous memory allocation for KVM hash pagetable 2288 allocation. 2289 By default it reserves 5% of total system memory. 2290 Format: <integer> 2291 Default: 5 2292 2293 kvm-intel.ept= [KVM,Intel] Disable extended page tables 2294 (virtualized MMU) support on capable Intel chips. 2295 Default is 1 (enabled) 2296 2297 kvm-intel.emulate_invalid_guest_state= 2298 [KVM,Intel] Enable emulation of invalid guest states 2299 Default is 0 (disabled) 2300 2301 kvm-intel.flexpriority= 2302 [KVM,Intel] Disable FlexPriority feature (TPR shadow). 2303 Default is 1 (enabled) 2304 2305 kvm-intel.nested= 2306 [KVM,Intel] Enable VMX nesting (nVMX). 2307 Default is 0 (disabled) 2308 2309 kvm-intel.unrestricted_guest= 2310 [KVM,Intel] Disable unrestricted guest feature 2311 (virtualized real and unpaged mode) on capable 2312 Intel chips. Default is 1 (enabled) 2313 2314 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault 2315 CVE-2018-3620. 2316 2317 Valid arguments: never, cond, always 2318 2319 always: L1D cache flush on every VMENTER. 2320 cond: Flush L1D on VMENTER only when the code between 2321 VMEXIT and VMENTER can leak host memory. 2322 never: Disables the mitigation 2323 2324 Default is cond (do L1 cache flush in specific instances) 2325 2326 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification 2327 feature (tagged TLBs) on capable Intel chips. 2328 Default is 1 (enabled) 2329 2330 l1tf= [X86] Control mitigation of the L1TF vulnerability on 2331 affected CPUs 2332 2333 The kernel PTE inversion protection is unconditionally 2334 enabled and cannot be disabled. 2335 2336 full 2337 Provides all available mitigations for the 2338 L1TF vulnerability. Disables SMT and 2339 enables all mitigations in the 2340 hypervisors, i.e. unconditional L1D flush. 2341 2342 SMT control and L1D flush control via the 2343 sysfs interface is still possible after 2344 boot. Hypervisors will issue a warning 2345 when the first VM is started in a 2346 potentially insecure configuration, 2347 i.e. SMT enabled or L1D flush disabled. 2348 2349 full,force 2350 Same as 'full', but disables SMT and L1D 2351 flush runtime control. Implies the 2352 'nosmt=force' command line option. 2353 (i.e. sysfs control of SMT is disabled.) 2354 2355 flush 2356 Leaves SMT enabled and enables the default 2357 hypervisor mitigation, i.e. conditional 2358 L1D flush. 2359 2360 SMT control and L1D flush control via the 2361 sysfs interface is still possible after 2362 boot. Hypervisors will issue a warning 2363 when the first VM is started in a 2364 potentially insecure configuration, 2365 i.e. SMT enabled or L1D flush disabled. 2366 2367 flush,nosmt 2368 2369 Disables SMT and enables the default 2370 hypervisor mitigation. 2371 2372 SMT control and L1D flush control via the 2373 sysfs interface is still possible after 2374 boot. Hypervisors will issue a warning 2375 when the first VM is started in a 2376 potentially insecure configuration, 2377 i.e. SMT enabled or L1D flush disabled. 2378 2379 flush,nowarn 2380 Same as 'flush', but hypervisors will not 2381 warn when a VM is started in a potentially 2382 insecure configuration. 2383 2384 off 2385 Disables hypervisor mitigations and doesn't 2386 emit any warnings. 2387 It also drops the swap size and available 2388 RAM limit restriction on both hypervisor and 2389 bare metal. 2390 2391 Default is 'flush'. 2392 2393 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst 2394 2395 l2cr= [PPC] 2396 2397 l3cr= [PPC] 2398 2399 lapic [X86-32,APIC] Enable the local APIC even if BIOS 2400 disabled it. 2401 2402 lapic= [X86,APIC] Do not use TSC deadline 2403 value for LAPIC timer one-shot implementation. Default 2404 back to the programmable timer unit in the LAPIC. 2405 Format: notscdeadline 2406 2407 lapic_timer_c2_ok [X86,APIC] trust the local apic timer 2408 in C2 power state. 2409 2410 libata.dma= [LIBATA] DMA control 2411 libata.dma=0 Disable all PATA and SATA DMA 2412 libata.dma=1 PATA and SATA Disk DMA only 2413 libata.dma=2 ATAPI (CDROM) DMA only 2414 libata.dma=4 Compact Flash DMA only 2415 Combinations also work, so libata.dma=3 enables DMA 2416 for disks and CDROMs, but not CFs. 2417 2418 libata.ignore_hpa= [LIBATA] Ignore HPA limit 2419 libata.ignore_hpa=0 keep BIOS limits (default) 2420 libata.ignore_hpa=1 ignore limits, using full disk 2421 2422 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume 2423 when set. 2424 Format: <int> 2425 2426 libata.force= [LIBATA] Force configurations. The format is comma 2427 separated list of "[ID:]VAL" where ID is 2428 PORT[.DEVICE]. PORT and DEVICE are decimal numbers 2429 matching port, link or device. Basically, it matches 2430 the ATA ID string printed on console by libata. If 2431 the whole ID part is omitted, the last PORT and DEVICE 2432 values are used. If ID hasn't been specified yet, the 2433 configuration applies to all ports, links and devices. 2434 2435 If only DEVICE is omitted, the parameter applies to 2436 the port and all links and devices behind it. DEVICE 2437 number of 0 either selects the first device or the 2438 first fan-out link behind PMP device. It does not 2439 select the host link. DEVICE number of 15 selects the 2440 host link and device attached to it. 2441 2442 The VAL specifies the configuration to force. As long 2443 as there's no ambiguity shortcut notation is allowed. 2444 For example, both 1.5 and 1.5G would work for 1.5Gbps. 2445 The following configurations can be forced. 2446 2447 * Cable type: 40c, 80c, short40c, unk, ign or sata. 2448 Any ID with matching PORT is used. 2449 2450 * SATA link speed limit: 1.5Gbps or 3.0Gbps. 2451 2452 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7]. 2453 udma[/][16,25,33,44,66,100,133] notation is also 2454 allowed. 2455 2456 * [no]ncq: Turn on or off NCQ. 2457 2458 * [no]ncqtrim: Turn off queued DSM TRIM. 2459 2460 * nohrst, nosrst, norst: suppress hard, soft 2461 and both resets. 2462 2463 * rstonce: only attempt one reset during 2464 hot-unplug link recovery 2465 2466 * dump_id: dump IDENTIFY data. 2467 2468 * atapi_dmadir: Enable ATAPI DMADIR bridge support 2469 2470 * disable: Disable this device. 2471 2472 If there are multiple matching configurations changing 2473 the same attribute, the last one is used. 2474 2475 memblock=debug [KNL] Enable memblock debug messages. 2476 2477 load_ramdisk= [RAM] [Deprecated] 2478 2479 lockd.nlm_grace_period=P [NFS] Assign grace period. 2480 Format: <integer> 2481 2482 lockd.nlm_tcpport=N [NFS] Assign TCP port. 2483 Format: <integer> 2484 2485 lockd.nlm_timeout=T [NFS] Assign timeout value. 2486 Format: <integer> 2487 2488 lockd.nlm_udpport=M [NFS] Assign UDP port. 2489 Format: <integer> 2490 2491 lockdown= [SECURITY] 2492 { integrity | confidentiality } 2493 Enable the kernel lockdown feature. If set to 2494 integrity, kernel features that allow userland to 2495 modify the running kernel are disabled. If set to 2496 confidentiality, kernel features that allow userland 2497 to extract confidential information from the kernel 2498 are also disabled. 2499 2500 locktorture.nreaders_stress= [KNL] 2501 Set the number of locking read-acquisition kthreads. 2502 Defaults to being automatically set based on the 2503 number of online CPUs. 2504 2505 locktorture.nwriters_stress= [KNL] 2506 Set the number of locking write-acquisition kthreads. 2507 2508 locktorture.onoff_holdoff= [KNL] 2509 Set time (s) after boot for CPU-hotplug testing. 2510 2511 locktorture.onoff_interval= [KNL] 2512 Set time (s) between CPU-hotplug operations, or 2513 zero to disable CPU-hotplug testing. 2514 2515 locktorture.shuffle_interval= [KNL] 2516 Set task-shuffle interval (jiffies). Shuffling 2517 tasks allows some CPUs to go into dyntick-idle 2518 mode during the locktorture test. 2519 2520 locktorture.shutdown_secs= [KNL] 2521 Set time (s) after boot system shutdown. This 2522 is useful for hands-off automated testing. 2523 2524 locktorture.stat_interval= [KNL] 2525 Time (s) between statistics printk()s. 2526 2527 locktorture.stutter= [KNL] 2528 Time (s) to stutter testing, for example, 2529 specifying five seconds causes the test to run for 2530 five seconds, wait for five seconds, and so on. 2531 This tests the locking primitive's ability to 2532 transition abruptly to and from idle. 2533 2534 locktorture.torture_type= [KNL] 2535 Specify the locking implementation to test. 2536 2537 locktorture.verbose= [KNL] 2538 Enable additional printk() statements. 2539 2540 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver 2541 Format: <irq> 2542 2543 loglevel= All Kernel Messages with a loglevel smaller than the 2544 console loglevel will be printed to the console. It can 2545 also be changed with klogd or other programs. The 2546 loglevels are defined as follows: 2547 2548 0 (KERN_EMERG) system is unusable 2549 1 (KERN_ALERT) action must be taken immediately 2550 2 (KERN_CRIT) critical conditions 2551 3 (KERN_ERR) error conditions 2552 4 (KERN_WARNING) warning conditions 2553 5 (KERN_NOTICE) normal but significant condition 2554 6 (KERN_INFO) informational 2555 7 (KERN_DEBUG) debug-level messages 2556 2557 log_buf_len=n[KMG] Sets the size of the printk ring buffer, 2558 in bytes. n must be a power of two and greater 2559 than the minimal size. The minimal size is defined 2560 by LOG_BUF_SHIFT kernel config parameter. There is 2561 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter 2562 that allows to increase the default size depending on 2563 the number of CPUs. See init/Kconfig for more details. 2564 2565 logo.nologo [FB] Disables display of the built-in Linux logo. 2566 This may be used to provide more screen space for 2567 kernel log messages and is useful when debugging 2568 kernel boot problems. 2569 2570 lp=0 [LP] Specify parallel ports to use, e.g, 2571 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses 2572 lp=reset first parallel port). 'lp=0' disables the 2573 lp=auto printer driver. 'lp=reset' (which can be 2574 specified in addition to the ports) causes 2575 attached printers to be reset. Using 2576 lp=port1,port2,... specifies the parallel ports 2577 to associate lp devices with, starting with 2578 lp0. A port specification may be 'none' to skip 2579 that lp device, or a parport name such as 2580 'parport0'. Specifying 'lp=auto' instead of a 2581 port specification list means that device IDs 2582 from each port should be examined, to see if 2583 an IEEE 1284-compliant printer is attached; if 2584 so, the driver will manage that printer. 2585 See also header of drivers/char/lp.c. 2586 2587 lpj=n [KNL] 2588 Sets loops_per_jiffy to given constant, thus avoiding 2589 time-consuming boot-time autodetection (up to 250 ms per 2590 CPU). 0 enables autodetection (default). To determine 2591 the correct value for your kernel, boot with normal 2592 autodetection and see what value is printed. Note that 2593 on SMP systems the preset will be applied to all CPUs, 2594 which is likely to cause problems if your CPUs need 2595 significantly divergent settings. An incorrect value 2596 will cause delays in the kernel to be wrong, leading to 2597 unpredictable I/O errors and other breakage. Although 2598 unlikely, in the extreme case this might damage your 2599 hardware. 2600 2601 ltpc= [NET] 2602 Format: <io>,<irq>,<dma> 2603 2604 lsm.debug [SECURITY] Enable LSM initialization debugging output. 2605 2606 lsm=lsm1,...,lsmN 2607 [SECURITY] Choose order of LSM initialization. This 2608 overrides CONFIG_LSM, and the "security=" parameter. 2609 2610 machvec= [IA-64] Force the use of a particular machine-vector 2611 (machvec) in a generic kernel. 2612 Example: machvec=hpzx1 2613 2614 machtype= [Loongson] Share the same kernel image file between 2615 different yeeloong laptops. 2616 Example: machtype=lemote-yeeloong-2f-7inch 2617 2618 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater 2619 than or equal to this physical address is ignored. 2620 2621 maxcpus= [SMP] Maximum number of processors that an SMP kernel 2622 will bring up during bootup. maxcpus=n : n >= 0 limits 2623 the kernel to bring up 'n' processors. Surely after 2624 bootup you can bring up the other plugged cpu by executing 2625 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus 2626 only takes effect during system bootup. 2627 While n=0 is a special case, it is equivalent to "nosmp", 2628 which also disables the IO APIC. 2629 2630 max_loop= [LOOP] The number of loop block devices that get 2631 (loop.max_loop) unconditionally pre-created at init time. The default 2632 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead 2633 of statically allocating a predefined number, loop 2634 devices can be requested on-demand with the 2635 /dev/loop-control interface. 2636 2637 mce [X86-32] Machine Check Exception 2638 2639 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst 2640 2641 md= [HW] RAID subsystems devices and level 2642 See Documentation/admin-guide/md.rst. 2643 2644 mdacon= [MDA] 2645 Format: <first>,<last> 2646 Specifies range of consoles to be captured by the MDA. 2647 2648 mds= [X86,INTEL] 2649 Control mitigation for the Micro-architectural Data 2650 Sampling (MDS) vulnerability. 2651 2652 Certain CPUs are vulnerable to an exploit against CPU 2653 internal buffers which can forward information to a 2654 disclosure gadget under certain conditions. 2655 2656 In vulnerable processors, the speculatively 2657 forwarded data can be used in a cache side channel 2658 attack, to access data to which the attacker does 2659 not have direct access. 2660 2661 This parameter controls the MDS mitigation. The 2662 options are: 2663 2664 full - Enable MDS mitigation on vulnerable CPUs 2665 full,nosmt - Enable MDS mitigation and disable 2666 SMT on vulnerable CPUs 2667 off - Unconditionally disable MDS mitigation 2668 2669 On TAA-affected machines, mds=off can be prevented by 2670 an active TAA mitigation as both vulnerabilities are 2671 mitigated with the same mechanism so in order to disable 2672 this mitigation, you need to specify tsx_async_abort=off 2673 too. 2674 2675 Not specifying this option is equivalent to 2676 mds=full. 2677 2678 For details see: Documentation/admin-guide/hw-vuln/mds.rst 2679 2680 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory 2681 Amount of memory to be used in cases as follows: 2682 2683 1 for test; 2684 2 when the kernel is not able to see the whole system memory; 2685 3 memory that lies after 'mem=' boundary is excluded from 2686 the hypervisor, then assigned to KVM guests. 2687 2688 [X86] Work as limiting max address. Use together 2689 with memmap= to avoid physical address space collisions. 2690 Without memmap= PCI devices could be placed at addresses 2691 belonging to unused RAM. 2692 2693 Note that this only takes effects during boot time since 2694 in above case 3, memory may need be hot added after boot 2695 if system memory of hypervisor is not sufficient. 2696 2697 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel 2698 memory. 2699 2700 memchunk=nn[KMG] 2701 [KNL,SH] Allow user to override the default size for 2702 per-device physically contiguous DMA buffers. 2703 2704 memhp_default_state=online/offline 2705 [KNL] Set the initial state for the memory hotplug 2706 onlining policy. If not specified, the default value is 2707 set according to the 2708 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config 2709 option. 2710 See Documentation/admin-guide/mm/memory-hotplug.rst. 2711 2712 memmap=exactmap [KNL,X86] Enable setting of an exact 2713 E820 memory map, as specified by the user. 2714 Such memmap=exactmap lines can be constructed based on 2715 BIOS output or other requirements. See the memmap=nn@ss 2716 option description. 2717 2718 memmap=nn[KMG]@ss[KMG] 2719 [KNL] Force usage of a specific region of memory. 2720 Region of memory to be used is from ss to ss+nn. 2721 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG], 2722 which limits max address to nn[KMG]. 2723 Multiple different regions can be specified, 2724 comma delimited. 2725 Example: 2726 memmap=100M@2G,100M#3G,1G!1024G 2727 2728 memmap=nn[KMG]#ss[KMG] 2729 [KNL,ACPI] Mark specific memory as ACPI data. 2730 Region of memory to be marked is from ss to ss+nn. 2731 2732 memmap=nn[KMG]$ss[KMG] 2733 [KNL,ACPI] Mark specific memory as reserved. 2734 Region of memory to be reserved is from ss to ss+nn. 2735 Example: Exclude memory from 0x18690000-0x1869ffff 2736 memmap=64K$0x18690000 2737 or 2738 memmap=0x10000$0x18690000 2739 Some bootloaders may need an escape character before '$', 2740 like Grub2, otherwise '$' and the following number 2741 will be eaten. 2742 2743 memmap=nn[KMG]!ss[KMG] 2744 [KNL,X86] Mark specific memory as protected. 2745 Region of memory to be used, from ss to ss+nn. 2746 The memory region may be marked as e820 type 12 (0xc) 2747 and is NVDIMM or ADR memory. 2748 2749 memmap=<size>%<offset>-<oldtype>+<newtype> 2750 [KNL,ACPI] Convert memory within the specified region 2751 from <oldtype> to <newtype>. If "-<oldtype>" is left 2752 out, the whole region will be marked as <newtype>, 2753 even if previously unavailable. If "+<newtype>" is left 2754 out, matching memory will be removed. Types are 2755 specified as e820 types, e.g., 1 = RAM, 2 = reserved, 2756 3 = ACPI, 12 = PRAM. 2757 2758 memory_corruption_check=0/1 [X86] 2759 Some BIOSes seem to corrupt the first 64k of 2760 memory when doing things like suspend/resume. 2761 Setting this option will scan the memory 2762 looking for corruption. Enabling this will 2763 both detect corruption and prevent the kernel 2764 from using the memory being corrupted. 2765 However, its intended as a diagnostic tool; if 2766 repeatable BIOS-originated corruption always 2767 affects the same memory, you can use memmap= 2768 to prevent the kernel from using that memory. 2769 2770 memory_corruption_check_size=size [X86] 2771 By default it checks for corruption in the low 2772 64k, making this memory unavailable for normal 2773 use. Use this parameter to scan for 2774 corruption in more or less memory. 2775 2776 memory_corruption_check_period=seconds [X86] 2777 By default it checks for corruption every 60 2778 seconds. Use this parameter to check at some 2779 other rate. 0 disables periodic checking. 2780 2781 memtest= [KNL,X86,ARM,PPC] Enable memtest 2782 Format: <integer> 2783 default : 0 <disable> 2784 Specifies the number of memtest passes to be 2785 performed. Each pass selects another test 2786 pattern from a given set of patterns. Memtest 2787 fills the memory with this pattern, validates 2788 memory contents and reserves bad memory 2789 regions that are detected. 2790 2791 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control 2792 Valid arguments: on, off 2793 Default (depends on kernel configuration option): 2794 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y) 2795 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n) 2796 mem_encrypt=on: Activate SME 2797 mem_encrypt=off: Do not activate SME 2798 2799 Refer to Documentation/virt/kvm/amd-memory-encryption.rst 2800 for details on when memory encryption can be activated. 2801 2802 mem_sleep_default= [SUSPEND] Default system suspend mode: 2803 s2idle - Suspend-To-Idle 2804 shallow - Power-On Suspend or equivalent (if supported) 2805 deep - Suspend-To-RAM or equivalent (if supported) 2806 See Documentation/admin-guide/pm/sleep-states.rst. 2807 2808 meye.*= [HW] Set MotionEye Camera parameters 2809 See Documentation/admin-guide/media/meye.rst. 2810 2811 mfgpt_irq= [IA-32] Specify the IRQ to use for the 2812 Multi-Function General Purpose Timers on AMD Geode 2813 platforms. 2814 2815 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when 2816 the BIOS has incorrectly applied a workaround. TinyBIOS 2817 version 0.98 is known to be affected, 0.99 fixes the 2818 problem by letting the user disable the workaround. 2819 2820 mga= [HW,DRM] 2821 2822 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this 2823 physical address is ignored. 2824 2825 mini2440= [ARM,HW,KNL] 2826 Format:[0..2][b][c][t] 2827 Default: "0tb" 2828 MINI2440 configuration specification: 2829 0 - The attached screen is the 3.5" TFT 2830 1 - The attached screen is the 7" TFT 2831 2 - The VGA Shield is attached (1024x768) 2832 Leaving out the screen size parameter will not load 2833 the TFT driver, and the framebuffer will be left 2834 unconfigured. 2835 b - Enable backlight. The TFT backlight pin will be 2836 linked to the kernel VESA blanking code and a GPIO 2837 LED. This parameter is not necessary when using the 2838 VGA shield. 2839 c - Enable the s3c camera interface. 2840 t - Reserved for enabling touchscreen support. The 2841 touchscreen support is not enabled in the mainstream 2842 kernel as of 2.6.30, a preliminary port can be found 2843 in the "bleeding edge" mini2440 support kernel at 2844 https://repo.or.cz/w/linux-2.6/mini2440.git 2845 2846 mitigations= 2847 [X86,PPC,S390,ARM64] Control optional mitigations for 2848 CPU vulnerabilities. This is a set of curated, 2849 arch-independent options, each of which is an 2850 aggregation of existing arch-specific options. 2851 2852 off 2853 Disable all optional CPU mitigations. This 2854 improves system performance, but it may also 2855 expose users to several CPU vulnerabilities. 2856 Equivalent to: nopti [X86,PPC] 2857 kpti=0 [ARM64] 2858 nospectre_v1 [X86,PPC] 2859 nobp=0 [S390] 2860 nospectre_v2 [X86,PPC,S390,ARM64] 2861 spectre_v2_user=off [X86] 2862 spec_store_bypass_disable=off [X86,PPC] 2863 ssbd=force-off [ARM64] 2864 l1tf=off [X86] 2865 mds=off [X86] 2866 tsx_async_abort=off [X86] 2867 kvm.nx_huge_pages=off [X86] 2868 no_entry_flush [PPC] 2869 no_uaccess_flush [PPC] 2870 mmio_stale_data=off [X86] 2871 2872 Exceptions: 2873 This does not have any effect on 2874 kvm.nx_huge_pages when 2875 kvm.nx_huge_pages=force. 2876 2877 auto (default) 2878 Mitigate all CPU vulnerabilities, but leave SMT 2879 enabled, even if it's vulnerable. This is for 2880 users who don't want to be surprised by SMT 2881 getting disabled across kernel upgrades, or who 2882 have other ways of avoiding SMT-based attacks. 2883 Equivalent to: (default behavior) 2884 2885 auto,nosmt 2886 Mitigate all CPU vulnerabilities, disabling SMT 2887 if needed. This is for users who always want to 2888 be fully mitigated, even if it means losing SMT. 2889 Equivalent to: l1tf=flush,nosmt [X86] 2890 mds=full,nosmt [X86] 2891 tsx_async_abort=full,nosmt [X86] 2892 mmio_stale_data=full,nosmt [X86] 2893 2894 mminit_loglevel= 2895 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this 2896 parameter allows control of the logging verbosity for 2897 the additional memory initialisation checks. A value 2898 of 0 disables mminit logging and a level of 4 will 2899 log everything. Information is printed at KERN_DEBUG 2900 so loglevel=8 may also need to be specified. 2901 2902 mmio_stale_data= 2903 [X86,INTEL] Control mitigation for the Processor 2904 MMIO Stale Data vulnerabilities. 2905 2906 Processor MMIO Stale Data is a class of 2907 vulnerabilities that may expose data after an MMIO 2908 operation. Exposed data could originate or end in 2909 the same CPU buffers as affected by MDS and TAA. 2910 Therefore, similar to MDS and TAA, the mitigation 2911 is to clear the affected CPU buffers. 2912 2913 This parameter controls the mitigation. The 2914 options are: 2915 2916 full - Enable mitigation on vulnerable CPUs 2917 2918 full,nosmt - Enable mitigation and disable SMT on 2919 vulnerable CPUs. 2920 2921 off - Unconditionally disable mitigation 2922 2923 On MDS or TAA affected machines, 2924 mmio_stale_data=off can be prevented by an active 2925 MDS or TAA mitigation as these vulnerabilities are 2926 mitigated with the same mechanism so in order to 2927 disable this mitigation, you need to specify 2928 mds=off and tsx_async_abort=off too. 2929 2930 Not specifying this option is equivalent to 2931 mmio_stale_data=full. 2932 2933 For details see: 2934 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst 2935 2936 module.sig_enforce 2937 [KNL] When CONFIG_MODULE_SIG is set, this means that 2938 modules without (valid) signatures will fail to load. 2939 Note that if CONFIG_MODULE_SIG_FORCE is set, that 2940 is always true, so this option does nothing. 2941 2942 module_blacklist= [KNL] Do not load a comma-separated list of 2943 modules. Useful for debugging problem modules. 2944 2945 mousedev.tap_time= 2946 [MOUSE] Maximum time between finger touching and 2947 leaving touchpad surface for touch to be considered 2948 a tap and be reported as a left button click (for 2949 touchpads working in absolute mode only). 2950 Format: <msecs> 2951 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices 2952 reporting absolute coordinates, such as tablets 2953 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices 2954 reporting absolute coordinates, such as tablets 2955 2956 movablecore= [KNL,X86,IA-64,PPC] 2957 Format: nn[KMGTPE] | nn% 2958 This parameter is the complement to kernelcore=, it 2959 specifies the amount of memory used for migratable 2960 allocations. If both kernelcore and movablecore is 2961 specified, then kernelcore will be at *least* the 2962 specified value but may be more. If movablecore on its 2963 own is specified, the administrator must be careful 2964 that the amount of memory usable for all allocations 2965 is not too small. 2966 2967 movable_node [KNL] Boot-time switch to make hotplugable memory 2968 NUMA nodes to be movable. This means that the memory 2969 of such nodes will be usable only for movable 2970 allocations which rules out almost all kernel 2971 allocations. Use with caution! 2972 2973 MTD_Partition= [MTD] 2974 Format: <name>,<region-number>,<size>,<offset> 2975 2976 MTD_Region= [MTD] Format: 2977 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>] 2978 2979 mtdparts= [MTD] 2980 See drivers/mtd/parsers/cmdlinepart.c 2981 2982 multitce=off [PPC] This parameter disables the use of the pSeries 2983 firmware feature for updating multiple TCE entries 2984 at a time. 2985 2986 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration 2987 2988 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock] 2989 2990 boundary - index of last SLC block on Flex-OneNAND. 2991 The remaining blocks are configured as MLC blocks. 2992 lock - Configure if Flex-OneNAND boundary should be locked. 2993 Once locked, the boundary cannot be changed. 2994 1 indicates lock status, 0 indicates unlock status. 2995 2996 mtdset= [ARM] 2997 ARM/S3C2412 JIVE boot control 2998 2999 See arch/arm/mach-s3c2412/mach-jive.c 3000 3001 mtouchusb.raw_coordinates= 3002 [HW] Make the MicroTouch USB driver use raw coordinates 3003 ('y', default) or cooked coordinates ('n') 3004 3005 mtrr_chunk_size=nn[KMG] [X86] 3006 used for mtrr cleanup. It is largest continuous chunk 3007 that could hold holes aka. UC entries. 3008 3009 mtrr_gran_size=nn[KMG] [X86] 3010 Used for mtrr cleanup. It is granularity of mtrr block. 3011 Default is 1. 3012 Large value could prevent small alignment from 3013 using up MTRRs. 3014 3015 mtrr_spare_reg_nr=n [X86] 3016 Format: <integer> 3017 Range: 0,7 : spare reg number 3018 Default : 1 3019 Used for mtrr cleanup. It is spare mtrr entries number. 3020 Set to 2 or more if your graphical card needs more. 3021 3022 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card 3023 3024 netdev= [NET] Network devices parameters 3025 Format: <irq>,<io>,<mem_start>,<mem_end>,<name> 3026 Note that mem_start is often overloaded to mean 3027 something different and driver-specific. 3028 This usage is only documented in each driver source 3029 file if at all. 3030 3031 nf_conntrack.acct= 3032 [NETFILTER] Enable connection tracking flow accounting 3033 0 to disable accounting 3034 1 to enable accounting 3035 Default value is 0. 3036 3037 nfsaddrs= [NFS] Deprecated. Use ip= instead. 3038 See Documentation/admin-guide/nfs/nfsroot.rst. 3039 3040 nfsroot= [NFS] nfs root filesystem for disk-less boxes. 3041 See Documentation/admin-guide/nfs/nfsroot.rst. 3042 3043 nfsrootdebug [NFS] enable nfsroot debugging messages. 3044 See Documentation/admin-guide/nfs/nfsroot.rst. 3045 3046 nfs.callback_nr_threads= 3047 [NFSv4] set the total number of threads that the 3048 NFS client will assign to service NFSv4 callback 3049 requests. 3050 3051 nfs.callback_tcpport= 3052 [NFS] set the TCP port on which the NFSv4 callback 3053 channel should listen. 3054 3055 nfs.cache_getent= 3056 [NFS] sets the pathname to the program which is used 3057 to update the NFS client cache entries. 3058 3059 nfs.cache_getent_timeout= 3060 [NFS] sets the timeout after which an attempt to 3061 update a cache entry is deemed to have failed. 3062 3063 nfs.idmap_cache_timeout= 3064 [NFS] set the maximum lifetime for idmapper cache 3065 entries. 3066 3067 nfs.enable_ino64= 3068 [NFS] enable 64-bit inode numbers. 3069 If zero, the NFS client will fake up a 32-bit inode 3070 number for the readdir() and stat() syscalls instead 3071 of returning the full 64-bit number. 3072 The default is to return 64-bit inode numbers. 3073 3074 nfs.max_session_cb_slots= 3075 [NFSv4.1] Sets the maximum number of session 3076 slots the client will assign to the callback 3077 channel. This determines the maximum number of 3078 callbacks the client will process in parallel for 3079 a particular server. 3080 3081 nfs.max_session_slots= 3082 [NFSv4.1] Sets the maximum number of session slots 3083 the client will attempt to negotiate with the server. 3084 This limits the number of simultaneous RPC requests 3085 that the client can send to the NFSv4.1 server. 3086 Note that there is little point in setting this 3087 value higher than the max_tcp_slot_table_limit. 3088 3089 nfs.nfs4_disable_idmapping= 3090 [NFSv4] When set to the default of '1', this option 3091 ensures that both the RPC level authentication 3092 scheme and the NFS level operations agree to use 3093 numeric uids/gids if the mount is using the 3094 'sec=sys' security flavour. In effect it is 3095 disabling idmapping, which can make migration from 3096 legacy NFSv2/v3 systems to NFSv4 easier. 3097 Servers that do not support this mode of operation 3098 will be autodetected by the client, and it will fall 3099 back to using the idmapper. 3100 To turn off this behaviour, set the value to '0'. 3101 nfs.nfs4_unique_id= 3102 [NFS4] Specify an additional fixed unique ident- 3103 ification string that NFSv4 clients can insert into 3104 their nfs_client_id4 string. This is typically a 3105 UUID that is generated at system install time. 3106 3107 nfs.send_implementation_id = 3108 [NFSv4.1] Send client implementation identification 3109 information in exchange_id requests. 3110 If zero, no implementation identification information 3111 will be sent. 3112 The default is to send the implementation identification 3113 information. 3114 3115 nfs.recover_lost_locks = 3116 [NFSv4] Attempt to recover locks that were lost due 3117 to a lease timeout on the server. Please note that 3118 doing this risks data corruption, since there are 3119 no guarantees that the file will remain unchanged 3120 after the locks are lost. 3121 If you want to enable the kernel legacy behaviour of 3122 attempting to recover these locks, then set this 3123 parameter to '1'. 3124 The default parameter value of '0' causes the kernel 3125 not to attempt recovery of lost locks. 3126 3127 nfs4.layoutstats_timer = 3128 [NFSv4.2] Change the rate at which the kernel sends 3129 layoutstats to the pNFS metadata server. 3130 3131 Setting this to value to 0 causes the kernel to use 3132 whatever value is the default set by the layout 3133 driver. A non-zero value sets the minimum interval 3134 in seconds between layoutstats transmissions. 3135 3136 nfsd.nfs4_disable_idmapping= 3137 [NFSv4] When set to the default of '1', the NFSv4 3138 server will return only numeric uids and gids to 3139 clients using auth_sys, and will accept numeric uids 3140 and gids from such clients. This is intended to ease 3141 migration from NFSv2/v3. 3142 3143 nmi_backtrace.backtrace_idle [KNL] 3144 Dump stacks even of idle CPUs in response to an 3145 NMI stack-backtrace request. 3146 3147 nmi_debug= [KNL,SH] Specify one or more actions to take 3148 when a NMI is triggered. 3149 Format: [state][,regs][,debounce][,die] 3150 3151 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels 3152 Format: [panic,][nopanic,][num] 3153 Valid num: 0 or 1 3154 0 - turn hardlockup detector in nmi_watchdog off 3155 1 - turn hardlockup detector in nmi_watchdog on 3156 When panic is specified, panic when an NMI watchdog 3157 timeout occurs (or 'nopanic' to not panic on an NMI 3158 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set) 3159 To disable both hard and soft lockup detectors, 3160 please see 'nowatchdog'. 3161 This is useful when you use a panic=... timeout and 3162 need the box quickly up again. 3163 3164 These settings can be accessed at runtime via 3165 the nmi_watchdog and hardlockup_panic sysctls. 3166 3167 netpoll.carrier_timeout= 3168 [NET] Specifies amount of time (in seconds) that 3169 netpoll should wait for a carrier. By default netpoll 3170 waits 4 seconds. 3171 3172 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths 3173 emulation library even if a 387 maths coprocessor 3174 is present. 3175 3176 no5lvl [X86-64] Disable 5-level paging mode. Forces 3177 kernel to use 4-level paging instead. 3178 3179 nofsgsbase [X86] Disables FSGSBASE instructions. 3180 3181 no_console_suspend 3182 [HW] Never suspend the console 3183 Disable suspending of consoles during suspend and 3184 hibernate operations. Once disabled, debugging 3185 messages can reach various consoles while the rest 3186 of the system is being put to sleep (ie, while 3187 debugging driver suspend/resume hooks). This may 3188 not work reliably with all consoles, but is known 3189 to work with serial and VGA consoles. 3190 To facilitate more flexible debugging, we also add 3191 console_suspend, a printk module parameter to control 3192 it. Users could use console_suspend (usually 3193 /sys/module/printk/parameters/console_suspend) to 3194 turn on/off it dynamically. 3195 3196 novmcoredd [KNL,KDUMP] 3197 Disable device dump. Device dump allows drivers to 3198 append dump data to vmcore so you can collect driver 3199 specified debug info. Drivers can append the data 3200 without any limit and this data is stored in memory, 3201 so this may cause significant memory stress. Disabling 3202 device dump can help save memory but the driver debug 3203 data will be no longer available. This parameter 3204 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP 3205 is set. 3206 3207 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien 3208 caches in the slab allocator. Saves per-node memory, 3209 but will impact performance. 3210 3211 noalign [KNL,ARM] 3212 3213 noaltinstr [S390] Disables alternative instructions patching 3214 (CPU alternatives feature). 3215 3216 noapic [SMP,APIC] Tells the kernel to not make use of any 3217 IOAPICs that may be present in the system. 3218 3219 noautogroup Disable scheduler automatic task group creation. 3220 3221 nobats [PPC] Do not use BATs for mapping kernel lowmem 3222 on "Classic" PPC cores. 3223 3224 nocache [ARM] 3225 3226 noclflush [BUGS=X86] Don't use the CLFLUSH instruction 3227 3228 nodelayacct [KNL] Disable per-task delay accounting 3229 3230 nodsp [SH] Disable hardware DSP at boot time. 3231 3232 noefi Disable EFI runtime services support. 3233 3234 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel. 3235 3236 noexec [IA-64] 3237 3238 noexec [X86] 3239 On X86-32 available only on PAE configured kernels. 3240 noexec=on: enable non-executable mappings (default) 3241 noexec=off: disable non-executable mappings 3242 3243 nosmap [X86,PPC] 3244 Disable SMAP (Supervisor Mode Access Prevention) 3245 even if it is supported by processor. 3246 3247 nosmep [X86,PPC] 3248 Disable SMEP (Supervisor Mode Execution Prevention) 3249 even if it is supported by processor. 3250 3251 noexec32 [X86-64] 3252 This affects only 32-bit executables. 3253 noexec32=on: enable non-executable mappings (default) 3254 read doesn't imply executable mappings 3255 noexec32=off: disable non-executable mappings 3256 read implies executable mappings 3257 3258 nofpu [MIPS,SH] Disable hardware FPU at boot time. 3259 3260 nofxsr [BUGS=X86-32] Disables x86 floating point extended 3261 register save and restore. The kernel will only save 3262 legacy floating-point registers on task switch. 3263 3264 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings. 3265 3266 nosmt [KNL,S390] Disable symmetric multithreading (SMT). 3267 Equivalent to smt=1. 3268 3269 [KNL,X86] Disable symmetric multithreading (SMT). 3270 nosmt=force: Force disable SMT, cannot be undone 3271 via the sysfs control file. 3272 3273 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1 3274 (bounds check bypass). With this option data leaks are 3275 possible in the system. 3276 3277 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for 3278 the Spectre variant 2 (indirect branch prediction) 3279 vulnerability. System may allow data leaks with this 3280 option. 3281 3282 nospec_store_bypass_disable 3283 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability 3284 3285 no_uaccess_flush 3286 [PPC] Don't flush the L1-D cache after accessing user data. 3287 3288 noxsave [BUGS=X86] Disables x86 extended register state save 3289 and restore using xsave. The kernel will fallback to 3290 enabling legacy floating-point and sse state. 3291 3292 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended 3293 register states. The kernel will fall back to use 3294 xsave to save the states. By using this parameter, 3295 performance of saving the states is degraded because 3296 xsave doesn't support modified optimization while 3297 xsaveopt supports it on xsaveopt enabled systems. 3298 3299 noxsaves [X86] Disables xsaves and xrstors used in saving and 3300 restoring x86 extended register state in compacted 3301 form of xsave area. The kernel will fall back to use 3302 xsaveopt and xrstor to save and restore the states 3303 in standard form of xsave area. By using this 3304 parameter, xsave area per process might occupy more 3305 memory on xsaves enabled systems. 3306 3307 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or 3308 wfi(ARM) instruction doesn't work correctly and not to 3309 use it. This is also useful when using JTAG debugger. 3310 3311 no_file_caps Tells the kernel not to honor file capabilities. The 3312 only way then for a file to be executed with privilege 3313 is to be setuid root or executed by root. 3314 3315 nohalt [IA-64] Tells the kernel not to use the power saving 3316 function PAL_HALT_LIGHT when idle. This increases 3317 power-consumption. On the positive side, it reduces 3318 interrupt wake-up latency, which may improve performance 3319 in certain environments such as networked servers or 3320 real-time systems. 3321 3322 nohibernate [HIBERNATION] Disable hibernation and resume. 3323 3324 nohz= [KNL] Boottime enable/disable dynamic ticks 3325 Valid arguments: on, off 3326 Default: on 3327 3328 nohz_full= [KNL,BOOT,SMP,ISOL] 3329 The argument is a cpu list, as described above. 3330 In kernels built with CONFIG_NO_HZ_FULL=y, set 3331 the specified list of CPUs whose tick will be stopped 3332 whenever possible. The boot CPU will be forced outside 3333 the range to maintain the timekeeping. Any CPUs 3334 in this list will have their RCU callbacks offloaded, 3335 just as if they had also been called out in the 3336 rcu_nocbs= boot parameter. 3337 3338 noiotrap [SH] Disables trapped I/O port accesses. 3339 3340 noirqdebug [X86-32] Disables the code which attempts to detect and 3341 disable unhandled interrupt sources. 3342 3343 no_timer_check [X86,APIC] Disables the code which tests for 3344 broken timer IRQ sources. 3345 3346 noisapnp [ISAPNP] Disables ISA PnP code. 3347 3348 noinitrd [RAM] Tells the kernel not to load any configured 3349 initial RAM disk. 3350 3351 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt 3352 remapping. 3353 [Deprecated - use intremap=off] 3354 3355 nointroute [IA-64] 3356 3357 noinvpcid [X86] Disable the INVPCID cpu feature. 3358 3359 nojitter [IA-64] Disables jitter checking for ITC timers. 3360 3361 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver 3362 3363 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page 3364 fault handling. 3365 3366 no-vmw-sched-clock 3367 [X86,PV_OPS] Disable paravirtualized VMware scheduler 3368 clock and use the default one. 3369 3370 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time 3371 accounting. steal time is computed, but won't 3372 influence scheduler behaviour 3373 3374 nolapic [X86-32,APIC] Do not enable or use the local APIC. 3375 3376 nolapic_timer [X86-32,APIC] Do not use the local APIC timer. 3377 3378 noltlbs [PPC] Do not use large page/tlb entries for kernel 3379 lowmem mapping on PPC40x and PPC8xx 3380 3381 nomca [IA-64] Disable machine check abort handling 3382 3383 nomce [X86-32] Disable Machine Check Exception 3384 3385 nomfgpt [X86-32] Disable Multi-Function General Purpose 3386 Timer usage (for AMD Geode machines). 3387 3388 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to 3389 shutdown the other cpus. Instead use the REBOOT_VECTOR 3390 irq. 3391 3392 nomodule Disable module load 3393 3394 nopat [X86] Disable PAT (page attribute table extension of 3395 pagetables) support. 3396 3397 nopcid [X86-64] Disable the PCID cpu feature. 3398 3399 norandmaps Don't use address space randomization. Equivalent to 3400 echo 0 > /proc/sys/kernel/randomize_va_space 3401 3402 noreplace-smp [X86-32,SMP] Don't replace SMP instructions 3403 with UP alternatives 3404 3405 nordrand [X86] Disable kernel use of the RDRAND and 3406 RDSEED instructions even if they are supported 3407 by the processor. RDRAND and RDSEED are still 3408 available to user space applications. 3409 3410 noresume [SWSUSP] Disables resume and restores original swap 3411 space. 3412 3413 no-scroll [VGA] Disables scrollback. 3414 This is required for the Braillex ib80-piezo Braille 3415 reader made by F.H. Papenmeier (Germany). 3416 3417 nosbagart [IA-64] 3418 3419 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support. 3420 3421 nosmp [SMP] Tells an SMP kernel to act as a UP kernel, 3422 and disable the IO APIC. legacy for "maxcpus=0". 3423 3424 nosoftlockup [KNL] Disable the soft-lockup detector. 3425 3426 nosync [HW,M68K] Disables sync negotiation for all devices. 3427 3428 nowatchdog [KNL] Disable both lockup detectors, i.e. 3429 soft-lockup and NMI watchdog (hard-lockup). 3430 3431 nowb [ARM] 3432 3433 nox2apic [X86-64,APIC] Do not enable x2APIC mode. 3434 3435 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when 3436 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off. 3437 Some features depend on CPU0. Known dependencies are: 3438 1. Resume from suspend/hibernate depends on CPU0. 3439 Suspend/hibernate will fail if CPU0 is offline and you 3440 need to online CPU0 before suspend/hibernate. 3441 2. PIC interrupts also depend on CPU0. CPU0 can't be 3442 removed if a PIC interrupt is detected. 3443 It's said poweroff/reboot may depend on CPU0 on some 3444 machines although I haven't seen such issues so far 3445 after CPU0 is offline on a few tested machines. 3446 If the dependencies are under your control, you can 3447 turn on cpu0_hotplug. 3448 3449 nps_mtm_hs_ctr= [KNL,ARC] 3450 This parameter sets the maximum duration, in 3451 cycles, each HW thread of the CTOP can run 3452 without interruptions, before HW switches it. 3453 The actual maximum duration is 16 times this 3454 parameter's value. 3455 Format: integer between 1 and 255 3456 Default: 255 3457 3458 nptcg= [IA-64] Override max number of concurrent global TLB 3459 purges which is reported from either PAL_VM_SUMMARY or 3460 SAL PALO. 3461 3462 nr_cpus= [SMP] Maximum number of processors that an SMP kernel 3463 could support. nr_cpus=n : n >= 1 limits the kernel to 3464 support 'n' processors. It could be larger than the 3465 number of already plugged CPU during bootup, later in 3466 runtime you can physically add extra cpu until it reaches 3467 n. So during boot up some boot time memory for per-cpu 3468 variables need be pre-allocated for later physical cpu 3469 hot plugging. 3470 3471 nr_uarts= [SERIAL] maximum number of UARTs to be registered. 3472 3473 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing. 3474 Allowed values are enable and disable 3475 3476 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA. 3477 'node', 'default' can be specified 3478 This can be set from sysctl after boot. 3479 See Documentation/admin-guide/sysctl/vm.rst for details. 3480 3481 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver. 3482 See Documentation/core-api/debugging-via-ohci1394.rst for more 3483 info. 3484 3485 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands 3486 Rather than timing out after 20 ms if an EC 3487 command is not properly ACKed, override the length 3488 of the timeout. We have interrupts disabled while 3489 waiting for the ACK, so if this is set too high 3490 interrupts *may* be lost! 3491 3492 omap_mux= [OMAP] Override bootloader pin multiplexing. 3493 Format: <mux_mode0.mode_name=value>... 3494 For example, to override I2C bus2: 3495 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100 3496 3497 oprofile.timer= [HW] 3498 Use timer interrupt instead of performance counters 3499 3500 oprofile.cpu_type= Force an oprofile cpu type 3501 This might be useful if you have an older oprofile 3502 userland or if you want common events. 3503 Format: { arch_perfmon } 3504 arch_perfmon: [X86] Force use of architectural 3505 perfmon on Intel CPUs instead of the 3506 CPU specific event set. 3507 timer: [X86] Force use of architectural NMI 3508 timer mode (see also oprofile.timer 3509 for generic hr timer mode) 3510 3511 oops=panic Always panic on oopses. Default is to just kill the 3512 process, but there is a small probability of 3513 deadlocking the machine. 3514 This will also cause panics on machine check exceptions. 3515 Useful together with panic=30 to trigger a reboot. 3516 3517 page_alloc.shuffle= 3518 [KNL] Boolean flag to control whether the page allocator 3519 should randomize its free lists. The randomization may 3520 be automatically enabled if the kernel detects it is 3521 running on a platform with a direct-mapped memory-side 3522 cache, and this parameter can be used to 3523 override/disable that behavior. The state of the flag 3524 can be read from sysfs at: 3525 /sys/module/page_alloc/parameters/shuffle. 3526 3527 page_owner= [KNL] Boot-time page_owner enabling option. 3528 Storage of the information about who allocated 3529 each page is disabled in default. With this switch, 3530 we can turn it on. 3531 on: enable the feature 3532 3533 page_poison= [KNL] Boot-time parameter changing the state of 3534 poisoning on the buddy allocator, available with 3535 CONFIG_PAGE_POISONING=y. 3536 off: turn off poisoning (default) 3537 on: turn on poisoning 3538 3539 panic= [KNL] Kernel behaviour on panic: delay <timeout> 3540 timeout > 0: seconds before rebooting 3541 timeout = 0: wait forever 3542 timeout < 0: reboot immediately 3543 Format: <timeout> 3544 3545 panic_print= Bitmask for printing system info when panic happens. 3546 User can chose combination of the following bits: 3547 bit 0: print all tasks info 3548 bit 1: print system memory info 3549 bit 2: print timer info 3550 bit 3: print locks info if CONFIG_LOCKDEP is on 3551 bit 4: print ftrace buffer 3552 bit 5: print all printk messages in buffer 3553 3554 panic_on_taint= Bitmask for conditionally calling panic() in add_taint() 3555 Format: <hex>[,nousertaint] 3556 Hexadecimal bitmask representing the set of TAINT flags 3557 that will cause the kernel to panic when add_taint() is 3558 called with any of the flags in this set. 3559 The optional switch "nousertaint" can be utilized to 3560 prevent userspace forced crashes by writing to sysctl 3561 /proc/sys/kernel/tainted any flagset matching with the 3562 bitmask set on panic_on_taint. 3563 See Documentation/admin-guide/tainted-kernels.rst for 3564 extra details on the taint flags that users can pick 3565 to compose the bitmask to assign to panic_on_taint. 3566 3567 panic_on_warn panic() instead of WARN(). Useful to cause kdump 3568 on a WARN(). 3569 3570 crash_kexec_post_notifiers 3571 Run kdump after running panic-notifiers and dumping 3572 kmsg. This only for the users who doubt kdump always 3573 succeeds in any situation. 3574 Note that this also increases risks of kdump failure, 3575 because some panic notifiers can make the crashed 3576 kernel more unstable. 3577 3578 parkbd.port= [HW] Parallel port number the keyboard adapter is 3579 connected to, default is 0. 3580 Format: <parport#> 3581 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation, 3582 0 for XT, 1 for AT (default is AT). 3583 Format: <mode> 3584 3585 parport= [HW,PPT] Specify parallel ports. 0 disables. 3586 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] } 3587 Use 'auto' to force the driver to use any 3588 IRQ/DMA settings detected (the default is to 3589 ignore detected IRQ/DMA settings because of 3590 possible conflicts). You can specify the base 3591 address, IRQ, and DMA settings; IRQ and DMA 3592 should be numbers, or 'auto' (for using detected 3593 settings on that particular port), or 'nofifo' 3594 (to avoid using a FIFO even if it is detected). 3595 Parallel ports are assigned in the order they 3596 are specified on the command line, starting 3597 with parport0. 3598 3599 parport_init_mode= [HW,PPT] 3600 Configure VIA parallel port to operate in 3601 a specific mode. This is necessary on Pegasos 3602 computer where firmware has no options for setting 3603 up parallel port mode and sets it to spp. 3604 Currently this function knows 686a and 8231 chips. 3605 Format: [spp|ps2|epp|ecp|ecpepp] 3606 3607 pause_on_oops= 3608 Halt all CPUs after the first oops has been printed for 3609 the specified number of seconds. This is to be used if 3610 your oopses keep scrolling off the screen. 3611 3612 pcbit= [HW,ISDN] 3613 3614 pcd. [PARIDE] 3615 See header of drivers/block/paride/pcd.c. 3616 See also Documentation/admin-guide/blockdev/paride.rst. 3617 3618 pci=option[,option...] [PCI] various PCI subsystem options. 3619 3620 Some options herein operate on a specific device 3621 or a set of devices (<pci_dev>). These are 3622 specified in one of the following formats: 3623 3624 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]* 3625 pci:<vendor>:<device>[:<subvendor>:<subdevice>] 3626 3627 Note: the first format specifies a PCI 3628 bus/device/function address which may change 3629 if new hardware is inserted, if motherboard 3630 firmware changes, or due to changes caused 3631 by other kernel parameters. If the 3632 domain is left unspecified, it is 3633 taken to be zero. Optionally, a path 3634 to a device through multiple device/function 3635 addresses can be specified after the base 3636 address (this is more robust against 3637 renumbering issues). The second format 3638 selects devices using IDs from the 3639 configuration space which may match multiple 3640 devices in the system. 3641 3642 earlydump dump PCI config space before the kernel 3643 changes anything 3644 off [X86] don't probe for the PCI bus 3645 bios [X86-32] force use of PCI BIOS, don't access 3646 the hardware directly. Use this if your machine 3647 has a non-standard PCI host bridge. 3648 nobios [X86-32] disallow use of PCI BIOS, only direct 3649 hardware access methods are allowed. Use this 3650 if you experience crashes upon bootup and you 3651 suspect they are caused by the BIOS. 3652 conf1 [X86] Force use of PCI Configuration Access 3653 Mechanism 1 (config address in IO port 0xCF8, 3654 data in IO port 0xCFC, both 32-bit). 3655 conf2 [X86] Force use of PCI Configuration Access 3656 Mechanism 2 (IO port 0xCF8 is an 8-bit port for 3657 the function, IO port 0xCFA, also 8-bit, sets 3658 bus number. The config space is then accessed 3659 through ports 0xC000-0xCFFF). 3660 See http://wiki.osdev.org/PCI for more info 3661 on the configuration access mechanisms. 3662 noaer [PCIE] If the PCIEAER kernel config parameter is 3663 enabled, this kernel boot option can be used to 3664 disable the use of PCIE advanced error reporting. 3665 nodomains [PCI] Disable support for multiple PCI 3666 root domains (aka PCI segments, in ACPI-speak). 3667 nommconf [X86] Disable use of MMCONFIG for PCI 3668 Configuration 3669 check_enable_amd_mmconf [X86] check for and enable 3670 properly configured MMIO access to PCI 3671 config space on AMD family 10h CPU 3672 nomsi [MSI] If the PCI_MSI kernel config parameter is 3673 enabled, this kernel boot option can be used to 3674 disable the use of MSI interrupts system-wide. 3675 noioapicquirk [APIC] Disable all boot interrupt quirks. 3676 Safety option to keep boot IRQs enabled. This 3677 should never be necessary. 3678 ioapicreroute [APIC] Enable rerouting of boot IRQs to the 3679 primary IO-APIC for bridges that cannot disable 3680 boot IRQs. This fixes a source of spurious IRQs 3681 when the system masks IRQs. 3682 noioapicreroute [APIC] Disable workaround that uses the 3683 boot IRQ equivalent of an IRQ that connects to 3684 a chipset where boot IRQs cannot be disabled. 3685 The opposite of ioapicreroute. 3686 biosirq [X86-32] Use PCI BIOS calls to get the interrupt 3687 routing table. These calls are known to be buggy 3688 on several machines and they hang the machine 3689 when used, but on other computers it's the only 3690 way to get the interrupt routing table. Try 3691 this option if the kernel is unable to allocate 3692 IRQs or discover secondary PCI buses on your 3693 motherboard. 3694 rom [X86] Assign address space to expansion ROMs. 3695 Use with caution as certain devices share 3696 address decoders between ROMs and other 3697 resources. 3698 norom [X86] Do not assign address space to 3699 expansion ROMs that do not already have 3700 BIOS assigned address ranges. 3701 nobar [X86] Do not assign address space to the 3702 BARs that weren't assigned by the BIOS. 3703 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be 3704 assigned automatically to PCI devices. You can 3705 make the kernel exclude IRQs of your ISA cards 3706 this way. 3707 pirqaddr=0xAAAAA [X86] Specify the physical address 3708 of the PIRQ table (normally generated 3709 by the BIOS) if it is outside the 3710 F0000h-100000h range. 3711 lastbus=N [X86] Scan all buses thru bus #N. Can be 3712 useful if the kernel is unable to find your 3713 secondary buses and you want to tell it 3714 explicitly which ones they are. 3715 assign-busses [X86] Always assign all PCI bus 3716 numbers ourselves, overriding 3717 whatever the firmware may have done. 3718 usepirqmask [X86] Honor the possible IRQ mask stored 3719 in the BIOS $PIR table. This is needed on 3720 some systems with broken BIOSes, notably 3721 some HP Pavilion N5400 and Omnibook XE3 3722 notebooks. This will have no effect if ACPI 3723 IRQ routing is enabled. 3724 noacpi [X86] Do not use ACPI for IRQ routing 3725 or for PCI scanning. 3726 use_crs [X86] Use PCI host bridge window information 3727 from ACPI. On BIOSes from 2008 or later, this 3728 is enabled by default. If you need to use this, 3729 please report a bug. 3730 nocrs [X86] Ignore PCI host bridge windows from ACPI. 3731 If you need to use this, please report a bug. 3732 routeirq Do IRQ routing for all PCI devices. 3733 This is normally done in pci_enable_device(), 3734 so this option is a temporary workaround 3735 for broken drivers that don't call it. 3736 skip_isa_align [X86] do not align io start addr, so can 3737 handle more pci cards 3738 noearly [X86] Don't do any early type 1 scanning. 3739 This might help on some broken boards which 3740 machine check when some devices' config space 3741 is read. But various workarounds are disabled 3742 and some IOMMU drivers will not work. 3743 bfsort Sort PCI devices into breadth-first order. 3744 This sorting is done to get a device 3745 order compatible with older (<= 2.4) kernels. 3746 nobfsort Don't sort PCI devices into breadth-first order. 3747 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size) 3748 tuning and use the BIOS-configured MPS defaults. 3749 pcie_bus_safe Set every device's MPS to the largest value 3750 supported by all devices below the root complex. 3751 pcie_bus_perf Set device MPS to the largest allowable MPS 3752 based on its parent bus. Also set MRRS (Max 3753 Read Request Size) to the largest supported 3754 value (no larger than the MPS that the device 3755 or bus can support) for best performance. 3756 pcie_bus_peer2peer Set every device's MPS to 128B, which 3757 every device is guaranteed to support. This 3758 configuration allows peer-to-peer DMA between 3759 any pair of devices, possibly at the cost of 3760 reduced performance. This also guarantees 3761 that hot-added devices will work. 3762 cbiosize=nn[KMG] The fixed amount of bus space which is 3763 reserved for the CardBus bridge's IO window. 3764 The default value is 256 bytes. 3765 cbmemsize=nn[KMG] The fixed amount of bus space which is 3766 reserved for the CardBus bridge's memory 3767 window. The default value is 64 megabytes. 3768 resource_alignment= 3769 Format: 3770 [<order of align>@]<pci_dev>[; ...] 3771 Specifies alignment and device to reassign 3772 aligned memory resources. How to 3773 specify the device is described above. 3774 If <order of align> is not specified, 3775 PAGE_SIZE is used as alignment. 3776 A PCI-PCI bridge can be specified if resource 3777 windows need to be expanded. 3778 To specify the alignment for several 3779 instances of a device, the PCI vendor, 3780 device, subvendor, and subdevice may be 3781 specified, e.g., 12@pci:8086:9c22:103c:198f 3782 for 4096-byte alignment. 3783 ecrc= Enable/disable PCIe ECRC (transaction layer 3784 end-to-end CRC checking). 3785 bios: Use BIOS/firmware settings. This is the 3786 the default. 3787 off: Turn ECRC off 3788 on: Turn ECRC on. 3789 hpiosize=nn[KMG] The fixed amount of bus space which is 3790 reserved for hotplug bridge's IO window. 3791 Default size is 256 bytes. 3792 hpmmiosize=nn[KMG] The fixed amount of bus space which is 3793 reserved for hotplug bridge's MMIO window. 3794 Default size is 2 megabytes. 3795 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is 3796 reserved for hotplug bridge's MMIO_PREF window. 3797 Default size is 2 megabytes. 3798 hpmemsize=nn[KMG] The fixed amount of bus space which is 3799 reserved for hotplug bridge's MMIO and 3800 MMIO_PREF window. 3801 Default size is 2 megabytes. 3802 hpbussize=nn The minimum amount of additional bus numbers 3803 reserved for buses below a hotplug bridge. 3804 Default is 1. 3805 realloc= Enable/disable reallocating PCI bridge resources 3806 if allocations done by BIOS are too small to 3807 accommodate resources required by all child 3808 devices. 3809 off: Turn realloc off 3810 on: Turn realloc on 3811 realloc same as realloc=on 3812 noari do not use PCIe ARI. 3813 noats [PCIE, Intel-IOMMU, AMD-IOMMU] 3814 do not use PCIe ATS (and IOMMU device IOTLB). 3815 pcie_scan_all Scan all possible PCIe devices. Otherwise we 3816 only look for one device below a PCIe downstream 3817 port. 3818 big_root_window Try to add a big 64bit memory window to the PCIe 3819 root complex on AMD CPUs. Some GFX hardware 3820 can resize a BAR to allow access to all VRAM. 3821 Adding the window is slightly risky (it may 3822 conflict with unreported devices), so this 3823 taints the kernel. 3824 disable_acs_redir=<pci_dev>[; ...] 3825 Specify one or more PCI devices (in the format 3826 specified above) separated by semicolons. 3827 Each device specified will have the PCI ACS 3828 redirect capabilities forced off which will 3829 allow P2P traffic between devices through 3830 bridges without forcing it upstream. Note: 3831 this removes isolation between devices and 3832 may put more devices in an IOMMU group. 3833 force_floating [S390] Force usage of floating interrupts. 3834 nomio [S390] Do not use MIO instructions. 3835 norid [S390] ignore the RID field and force use of 3836 one PCI domain per PCI function 3837 3838 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power 3839 Management. 3840 off Disable ASPM. 3841 force Enable ASPM even on devices that claim not to support it. 3842 WARNING: Forcing ASPM on may cause system lockups. 3843 3844 pcie_ports= [PCIE] PCIe port services handling: 3845 native Use native PCIe services (PME, AER, DPC, PCIe hotplug) 3846 even if the platform doesn't give the OS permission to 3847 use them. This may cause conflicts if the platform 3848 also tries to use these services. 3849 dpc-native Use native PCIe service for DPC only. May 3850 cause conflicts if firmware uses AER or DPC. 3851 compat Disable native PCIe services (PME, AER, DPC, PCIe 3852 hotplug). 3853 3854 pcie_port_pm= [PCIE] PCIe port power management handling: 3855 off Disable power management of all PCIe ports 3856 force Forcibly enable power management of all PCIe ports 3857 3858 pcie_pme= [PCIE,PM] Native PCIe PME signaling options: 3859 nomsi Do not use MSI for native PCIe PME signaling (this makes 3860 all PCIe root ports use INTx for all services). 3861 3862 pcmv= [HW,PCMCIA] BadgePAD 4 3863 3864 pd_ignore_unused 3865 [PM] 3866 Keep all power-domains already enabled by bootloader on, 3867 even if no driver has claimed them. This is useful 3868 for debug and development, but should not be 3869 needed on a platform with proper driver support. 3870 3871 pd. [PARIDE] 3872 See Documentation/admin-guide/blockdev/paride.rst. 3873 3874 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at 3875 boot time. 3876 Format: { 0 | 1 } 3877 See arch/parisc/kernel/pdc_chassis.c 3878 3879 percpu_alloc= Select which percpu first chunk allocator to use. 3880 Currently supported values are "embed" and "page". 3881 Archs may support subset or none of the selections. 3882 See comments in mm/percpu.c for details on each 3883 allocator. This parameter is primarily for debugging 3884 and performance comparison. 3885 3886 pf. [PARIDE] 3887 See Documentation/admin-guide/blockdev/paride.rst. 3888 3889 pg. [PARIDE] 3890 See Documentation/admin-guide/blockdev/paride.rst. 3891 3892 pirq= [SMP,APIC] Manual mp-table setup 3893 See Documentation/x86/i386/IO-APIC.rst. 3894 3895 plip= [PPT,NET] Parallel port network link 3896 Format: { parport<nr> | timid | 0 } 3897 See also Documentation/admin-guide/parport.rst. 3898 3899 pmtmr= [X86] Manual setup of pmtmr I/O Port. 3900 Override pmtimer IOPort with a hex value. 3901 e.g. pmtmr=0x508 3902 3903 pm_debug_messages [SUSPEND,KNL] 3904 Enable suspend/resume debug messages during boot up. 3905 3906 pnp.debug=1 [PNP] 3907 Enable PNP debug messages (depends on the 3908 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time 3909 via /sys/module/pnp/parameters/debug. We always show 3910 current resource usage; turning this on also shows 3911 possible settings and some assignment information. 3912 3913 pnpacpi= [ACPI] 3914 { off } 3915 3916 pnpbios= [ISAPNP] 3917 { on | off | curr | res | no-curr | no-res } 3918 3919 pnp_reserve_irq= 3920 [ISAPNP] Exclude IRQs for the autoconfiguration 3921 3922 pnp_reserve_dma= 3923 [ISAPNP] Exclude DMAs for the autoconfiguration 3924 3925 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration 3926 Ranges are in pairs (I/O port base and size). 3927 3928 pnp_reserve_mem= 3929 [ISAPNP] Exclude memory regions for the 3930 autoconfiguration. 3931 Ranges are in pairs (memory base and size). 3932 3933 ports= [IP_VS_FTP] IPVS ftp helper module 3934 Default is 21. 3935 Up to 8 (IP_VS_APP_MAX_PORTS) ports 3936 may be specified. 3937 Format: <port>,<port>.... 3938 3939 powersave=off [PPC] This option disables power saving features. 3940 It specifically disables cpuidle and sets the 3941 platform machine description specific power_save 3942 function to NULL. On Idle the CPU just reduces 3943 execution priority. 3944 3945 ppc_strict_facility_enable 3946 [PPC] This option catches any kernel floating point, 3947 Altivec, VSX and SPE outside of regions specifically 3948 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()). 3949 There is some performance impact when enabling this. 3950 3951 ppc_tm= [PPC] 3952 Format: {"off"} 3953 Disable Hardware Transactional Memory 3954 3955 print-fatal-signals= 3956 [KNL] debug: print fatal signals 3957 3958 If enabled, warn about various signal handling 3959 related application anomalies: too many signals, 3960 too many POSIX.1 timers, fatal signals causing a 3961 coredump - etc. 3962 3963 If you hit the warning due to signal overflow, 3964 you might want to try "ulimit -i unlimited". 3965 3966 default: off. 3967 3968 printk.always_kmsg_dump= 3969 Trigger kmsg_dump for cases other than kernel oops or 3970 panics 3971 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 3972 default: disabled 3973 3974 printk.devkmsg={on,off,ratelimit} 3975 Control writing to /dev/kmsg. 3976 on - unlimited logging to /dev/kmsg from userspace 3977 off - logging to /dev/kmsg disabled 3978 ratelimit - ratelimit the logging 3979 Default: ratelimit 3980 3981 printk.time= Show timing data prefixed to each printk message line 3982 Format: <bool> (1/Y/y=enable, 0/N/n=disable) 3983 3984 processor.max_cstate= [HW,ACPI] 3985 Limit processor to maximum C-state 3986 max_cstate=9 overrides any DMI blacklist limit. 3987 3988 processor.nocst [HW,ACPI] 3989 Ignore the _CST method to determine C-states, 3990 instead using the legacy FADT method 3991 3992 profile= [KNL] Enable kernel profiling via /proc/profile 3993 Format: [<profiletype>,]<number> 3994 Param: <profiletype>: "schedule", "sleep", or "kvm" 3995 [defaults to kernel profiling] 3996 Param: "schedule" - profile schedule points. 3997 Param: "sleep" - profile D-state sleeping (millisecs). 3998 Requires CONFIG_SCHEDSTATS 3999 Param: "kvm" - profile VM exits. 4000 Param: <number> - step/bucket size as a power of 2 for 4001 statistical time based profiling. 4002 4003 prompt_ramdisk= [RAM] [Deprecated] 4004 4005 prot_virt= [S390] enable hosting protected virtual machines 4006 isolated from the hypervisor (if hardware supports 4007 that). 4008 Format: <bool> 4009 4010 psi= [KNL] Enable or disable pressure stall information 4011 tracking. 4012 Format: <bool> 4013 4014 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to 4015 probe for; one of (bare|imps|exps|lifebook|any). 4016 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports 4017 per second. 4018 psmouse.resetafter= [HW,MOUSE] 4019 Try to reset the device after so many bad packets 4020 (0 = never). 4021 psmouse.resolution= 4022 [HW,MOUSE] Set desired mouse resolution, in dpi. 4023 psmouse.smartscroll= 4024 [HW,MOUSE] Controls Logitech smartscroll autorepeat. 4025 0 = disabled, 1 = enabled (default). 4026 4027 pstore.backend= Specify the name of the pstore backend to use 4028 4029 pt. [PARIDE] 4030 See Documentation/admin-guide/blockdev/paride.rst. 4031 4032 pti= [X86-64] Control Page Table Isolation of user and 4033 kernel address spaces. Disabling this feature 4034 removes hardening, but improves performance of 4035 system calls and interrupts. 4036 4037 on - unconditionally enable 4038 off - unconditionally disable 4039 auto - kernel detects whether your CPU model is 4040 vulnerable to issues that PTI mitigates 4041 4042 Not specifying this option is equivalent to pti=auto. 4043 4044 nopti [X86-64] 4045 Equivalent to pti=off 4046 4047 pty.legacy_count= 4048 [KNL] Number of legacy pty's. Overwrites compiled-in 4049 default number. 4050 4051 quiet [KNL] Disable most log messages 4052 4053 r128= [HW,DRM] 4054 4055 raid= [HW,RAID] 4056 See Documentation/admin-guide/md.rst. 4057 4058 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes 4059 See Documentation/admin-guide/blockdev/ramdisk.rst. 4060 4061 ramdisk_start= [RAM] RAM disk image start address 4062 4063 random.trust_cpu={on,off} 4064 [KNL] Enable or disable trusting the use of the 4065 CPU's random number generator (if available) to 4066 fully seed the kernel's CRNG. Default is controlled 4067 by CONFIG_RANDOM_TRUST_CPU. 4068 4069 ras=option[,option,...] [KNL] RAS-specific options 4070 4071 cec_disable [X86] 4072 Disable the Correctable Errors Collector, 4073 see CONFIG_RAS_CEC help text. 4074 4075 rcu_nocbs= [KNL] 4076 The argument is a cpu list, as described above, 4077 except that the string "all" can be used to 4078 specify every CPU on the system. 4079 4080 In kernels built with CONFIG_RCU_NOCB_CPU=y, set 4081 the specified list of CPUs to be no-callback CPUs. 4082 Invocation of these CPUs' RCU callbacks will be 4083 offloaded to "rcuox/N" kthreads created for that 4084 purpose, where "x" is "p" for RCU-preempt, and 4085 "s" for RCU-sched, and "N" is the CPU number. 4086 This reduces OS jitter on the offloaded CPUs, 4087 which can be useful for HPC and real-time 4088 workloads. It can also improve energy efficiency 4089 for asymmetric multiprocessors. 4090 4091 rcu_nocb_poll [KNL] 4092 Rather than requiring that offloaded CPUs 4093 (specified by rcu_nocbs= above) explicitly 4094 awaken the corresponding "rcuoN" kthreads, 4095 make these kthreads poll for callbacks. 4096 This improves the real-time response for the 4097 offloaded CPUs by relieving them of the need to 4098 wake up the corresponding kthread, but degrades 4099 energy efficiency by requiring that the kthreads 4100 periodically wake up to do the polling. 4101 4102 rcutree.blimit= [KNL] 4103 Set maximum number of finished RCU callbacks to 4104 process in one batch. 4105 4106 rcutree.dump_tree= [KNL] 4107 Dump the structure of the rcu_node combining tree 4108 out at early boot. This is used for diagnostic 4109 purposes, to verify correct tree setup. 4110 4111 rcutree.gp_cleanup_delay= [KNL] 4112 Set the number of jiffies to delay each step of 4113 RCU grace-period cleanup. 4114 4115 rcutree.gp_init_delay= [KNL] 4116 Set the number of jiffies to delay each step of 4117 RCU grace-period initialization. 4118 4119 rcutree.gp_preinit_delay= [KNL] 4120 Set the number of jiffies to delay each step of 4121 RCU grace-period pre-initialization, that is, 4122 the propagation of recent CPU-hotplug changes up 4123 the rcu_node combining tree. 4124 4125 rcutree.use_softirq= [KNL] 4126 If set to zero, move all RCU_SOFTIRQ processing to 4127 per-CPU rcuc kthreads. Defaults to a non-zero 4128 value, meaning that RCU_SOFTIRQ is used by default. 4129 Specify rcutree.use_softirq=0 to use rcuc kthreads. 4130 4131 rcutree.rcu_fanout_exact= [KNL] 4132 Disable autobalancing of the rcu_node combining 4133 tree. This is used by rcutorture, and might 4134 possibly be useful for architectures having high 4135 cache-to-cache transfer latencies. 4136 4137 rcutree.rcu_fanout_leaf= [KNL] 4138 Change the number of CPUs assigned to each 4139 leaf rcu_node structure. Useful for very 4140 large systems, which will choose the value 64, 4141 and for NUMA systems with large remote-access 4142 latencies, which will choose a value aligned 4143 with the appropriate hardware boundaries. 4144 4145 rcutree.rcu_min_cached_objs= [KNL] 4146 Minimum number of objects which are cached and 4147 maintained per one CPU. Object size is equal 4148 to PAGE_SIZE. The cache allows to reduce the 4149 pressure to page allocator, also it makes the 4150 whole algorithm to behave better in low memory 4151 condition. 4152 4153 rcutree.jiffies_till_first_fqs= [KNL] 4154 Set delay from grace-period initialization to 4155 first attempt to force quiescent states. 4156 Units are jiffies, minimum value is zero, 4157 and maximum value is HZ. 4158 4159 rcutree.jiffies_till_next_fqs= [KNL] 4160 Set delay between subsequent attempts to force 4161 quiescent states. Units are jiffies, minimum 4162 value is one, and maximum value is HZ. 4163 4164 rcutree.jiffies_till_sched_qs= [KNL] 4165 Set required age in jiffies for a 4166 given grace period before RCU starts 4167 soliciting quiescent-state help from 4168 rcu_note_context_switch() and cond_resched(). 4169 If not specified, the kernel will calculate 4170 a value based on the most recent settings 4171 of rcutree.jiffies_till_first_fqs 4172 and rcutree.jiffies_till_next_fqs. 4173 This calculated value may be viewed in 4174 rcutree.jiffies_to_sched_qs. Any attempt to set 4175 rcutree.jiffies_to_sched_qs will be cheerfully 4176 overwritten. 4177 4178 rcutree.kthread_prio= [KNL,BOOT] 4179 Set the SCHED_FIFO priority of the RCU per-CPU 4180 kthreads (rcuc/N). This value is also used for 4181 the priority of the RCU boost threads (rcub/N) 4182 and for the RCU grace-period kthreads (rcu_bh, 4183 rcu_preempt, and rcu_sched). If RCU_BOOST is 4184 set, valid values are 1-99 and the default is 1 4185 (the least-favored priority). Otherwise, when 4186 RCU_BOOST is not set, valid values are 0-99 and 4187 the default is zero (non-realtime operation). 4188 4189 rcutree.rcu_nocb_gp_stride= [KNL] 4190 Set the number of NOCB callback kthreads in 4191 each group, which defaults to the square root 4192 of the number of CPUs. Larger numbers reduce 4193 the wakeup overhead on the global grace-period 4194 kthread, but increases that same overhead on 4195 each group's NOCB grace-period kthread. 4196 4197 rcutree.qhimark= [KNL] 4198 Set threshold of queued RCU callbacks beyond which 4199 batch limiting is disabled. 4200 4201 rcutree.qlowmark= [KNL] 4202 Set threshold of queued RCU callbacks below which 4203 batch limiting is re-enabled. 4204 4205 rcutree.qovld= [KNL] 4206 Set threshold of queued RCU callbacks beyond which 4207 RCU's force-quiescent-state scan will aggressively 4208 enlist help from cond_resched() and sched IPIs to 4209 help CPUs more quickly reach quiescent states. 4210 Set to less than zero to make this be set based 4211 on rcutree.qhimark at boot time and to zero to 4212 disable more aggressive help enlistment. 4213 4214 rcutree.rcu_idle_gp_delay= [KNL] 4215 Set wakeup interval for idle CPUs that have 4216 RCU callbacks (RCU_FAST_NO_HZ=y). 4217 4218 rcutree.rcu_idle_lazy_gp_delay= [KNL] 4219 Set wakeup interval for idle CPUs that have 4220 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y). 4221 Lazy RCU callbacks are those which RCU can 4222 prove do nothing more than free memory. 4223 4224 rcutree.rcu_kick_kthreads= [KNL] 4225 Cause the grace-period kthread to get an extra 4226 wake_up() if it sleeps three times longer than 4227 it should at force-quiescent-state time. 4228 This wake_up() will be accompanied by a 4229 WARN_ONCE() splat and an ftrace_dump(). 4230 4231 rcutree.rcu_unlock_delay= [KNL] 4232 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels, 4233 this specifies an rcu_read_unlock()-time delay 4234 in microseconds. This defaults to zero. 4235 Larger delays increase the probability of 4236 catching RCU pointer leaks, that is, buggy use 4237 of RCU-protected pointers after the relevant 4238 rcu_read_unlock() has completed. 4239 4240 rcutree.sysrq_rcu= [KNL] 4241 Commandeer a sysrq key to dump out Tree RCU's 4242 rcu_node tree with an eye towards determining 4243 why a new grace period has not yet started. 4244 4245 rcuscale.gp_async= [KNL] 4246 Measure performance of asynchronous 4247 grace-period primitives such as call_rcu(). 4248 4249 rcuscale.gp_async_max= [KNL] 4250 Specify the maximum number of outstanding 4251 callbacks per writer thread. When a writer 4252 thread exceeds this limit, it invokes the 4253 corresponding flavor of rcu_barrier() to allow 4254 previously posted callbacks to drain. 4255 4256 rcuscale.gp_exp= [KNL] 4257 Measure performance of expedited synchronous 4258 grace-period primitives. 4259 4260 rcuscale.holdoff= [KNL] 4261 Set test-start holdoff period. The purpose of 4262 this parameter is to delay the start of the 4263 test until boot completes in order to avoid 4264 interference. 4265 4266 rcuscale.kfree_rcu_test= [KNL] 4267 Set to measure performance of kfree_rcu() flooding. 4268 4269 rcuscale.kfree_nthreads= [KNL] 4270 The number of threads running loops of kfree_rcu(). 4271 4272 rcuscale.kfree_alloc_num= [KNL] 4273 Number of allocations and frees done in an iteration. 4274 4275 rcuscale.kfree_loops= [KNL] 4276 Number of loops doing rcuscale.kfree_alloc_num number 4277 of allocations and frees. 4278 4279 rcuscale.nreaders= [KNL] 4280 Set number of RCU readers. The value -1 selects 4281 N, where N is the number of CPUs. A value 4282 "n" less than -1 selects N-n+1, where N is again 4283 the number of CPUs. For example, -2 selects N 4284 (the number of CPUs), -3 selects N+1, and so on. 4285 A value of "n" less than or equal to -N selects 4286 a single reader. 4287 4288 rcuscale.nwriters= [KNL] 4289 Set number of RCU writers. The values operate 4290 the same as for rcuscale.nreaders. 4291 N, where N is the number of CPUs 4292 4293 rcuscale.perf_type= [KNL] 4294 Specify the RCU implementation to test. 4295 4296 rcuscale.shutdown= [KNL] 4297 Shut the system down after performance tests 4298 complete. This is useful for hands-off automated 4299 testing. 4300 4301 rcuscale.verbose= [KNL] 4302 Enable additional printk() statements. 4303 4304 rcuscale.writer_holdoff= [KNL] 4305 Write-side holdoff between grace periods, 4306 in microseconds. The default of zero says 4307 no holdoff. 4308 4309 rcutorture.fqs_duration= [KNL] 4310 Set duration of force_quiescent_state bursts 4311 in microseconds. 4312 4313 rcutorture.fqs_holdoff= [KNL] 4314 Set holdoff time within force_quiescent_state bursts 4315 in microseconds. 4316 4317 rcutorture.fqs_stutter= [KNL] 4318 Set wait time between force_quiescent_state bursts 4319 in seconds. 4320 4321 rcutorture.fwd_progress= [KNL] 4322 Enable RCU grace-period forward-progress testing 4323 for the types of RCU supporting this notion. 4324 4325 rcutorture.fwd_progress_div= [KNL] 4326 Specify the fraction of a CPU-stall-warning 4327 period to do tight-loop forward-progress testing. 4328 4329 rcutorture.fwd_progress_holdoff= [KNL] 4330 Number of seconds to wait between successive 4331 forward-progress tests. 4332 4333 rcutorture.fwd_progress_need_resched= [KNL] 4334 Enclose cond_resched() calls within checks for 4335 need_resched() during tight-loop forward-progress 4336 testing. 4337 4338 rcutorture.gp_cond= [KNL] 4339 Use conditional/asynchronous update-side 4340 primitives, if available. 4341 4342 rcutorture.gp_exp= [KNL] 4343 Use expedited update-side primitives, if available. 4344 4345 rcutorture.gp_normal= [KNL] 4346 Use normal (non-expedited) asynchronous 4347 update-side primitives, if available. 4348 4349 rcutorture.gp_sync= [KNL] 4350 Use normal (non-expedited) synchronous 4351 update-side primitives, if available. If all 4352 of rcutorture.gp_cond=, rcutorture.gp_exp=, 4353 rcutorture.gp_normal=, and rcutorture.gp_sync= 4354 are zero, rcutorture acts as if is interpreted 4355 they are all non-zero. 4356 4357 rcutorture.irqreader= [KNL] 4358 Run RCU readers from irq handlers, or, more 4359 accurately, from a timer handler. Not all RCU 4360 flavors take kindly to this sort of thing. 4361 4362 rcutorture.leakpointer= [KNL] 4363 Leak an RCU-protected pointer out of the reader. 4364 This can of course result in splats, and is 4365 intended to test the ability of things like 4366 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect 4367 such leaks. 4368 4369 rcutorture.n_barrier_cbs= [KNL] 4370 Set callbacks/threads for rcu_barrier() testing. 4371 4372 rcutorture.nfakewriters= [KNL] 4373 Set number of concurrent RCU writers. These just 4374 stress RCU, they don't participate in the actual 4375 test, hence the "fake". 4376 4377 rcutorture.nreaders= [KNL] 4378 Set number of RCU readers. The value -1 selects 4379 N-1, where N is the number of CPUs. A value 4380 "n" less than -1 selects N-n-2, where N is again 4381 the number of CPUs. For example, -2 selects N 4382 (the number of CPUs), -3 selects N+1, and so on. 4383 4384 rcutorture.object_debug= [KNL] 4385 Enable debug-object double-call_rcu() testing. 4386 4387 rcutorture.onoff_holdoff= [KNL] 4388 Set time (s) after boot for CPU-hotplug testing. 4389 4390 rcutorture.onoff_interval= [KNL] 4391 Set time (jiffies) between CPU-hotplug operations, 4392 or zero to disable CPU-hotplug testing. 4393 4394 rcutorture.read_exit= [KNL] 4395 Set the number of read-then-exit kthreads used 4396 to test the interaction of RCU updaters and 4397 task-exit processing. 4398 4399 rcutorture.read_exit_burst= [KNL] 4400 The number of times in a given read-then-exit 4401 episode that a set of read-then-exit kthreads 4402 is spawned. 4403 4404 rcutorture.read_exit_delay= [KNL] 4405 The delay, in seconds, between successive 4406 read-then-exit testing episodes. 4407 4408 rcutorture.shuffle_interval= [KNL] 4409 Set task-shuffle interval (s). Shuffling tasks 4410 allows some CPUs to go into dyntick-idle mode 4411 during the rcutorture test. 4412 4413 rcutorture.shutdown_secs= [KNL] 4414 Set time (s) after boot system shutdown. This 4415 is useful for hands-off automated testing. 4416 4417 rcutorture.stall_cpu= [KNL] 4418 Duration of CPU stall (s) to test RCU CPU stall 4419 warnings, zero to disable. 4420 4421 rcutorture.stall_cpu_block= [KNL] 4422 Sleep while stalling if set. This will result 4423 in warnings from preemptible RCU in addition 4424 to any other stall-related activity. 4425 4426 rcutorture.stall_cpu_holdoff= [KNL] 4427 Time to wait (s) after boot before inducing stall. 4428 4429 rcutorture.stall_cpu_irqsoff= [KNL] 4430 Disable interrupts while stalling if set. 4431 4432 rcutorture.stall_gp_kthread= [KNL] 4433 Duration (s) of forced sleep within RCU 4434 grace-period kthread to test RCU CPU stall 4435 warnings, zero to disable. If both stall_cpu 4436 and stall_gp_kthread are specified, the 4437 kthread is starved first, then the CPU. 4438 4439 rcutorture.stat_interval= [KNL] 4440 Time (s) between statistics printk()s. 4441 4442 rcutorture.stutter= [KNL] 4443 Time (s) to stutter testing, for example, specifying 4444 five seconds causes the test to run for five seconds, 4445 wait for five seconds, and so on. This tests RCU's 4446 ability to transition abruptly to and from idle. 4447 4448 rcutorture.test_boost= [KNL] 4449 Test RCU priority boosting? 0=no, 1=maybe, 2=yes. 4450 "Maybe" means test if the RCU implementation 4451 under test support RCU priority boosting. 4452 4453 rcutorture.test_boost_duration= [KNL] 4454 Duration (s) of each individual boost test. 4455 4456 rcutorture.test_boost_interval= [KNL] 4457 Interval (s) between each boost test. 4458 4459 rcutorture.test_no_idle_hz= [KNL] 4460 Test RCU's dyntick-idle handling. See also the 4461 rcutorture.shuffle_interval parameter. 4462 4463 rcutorture.torture_type= [KNL] 4464 Specify the RCU implementation to test. 4465 4466 rcutorture.verbose= [KNL] 4467 Enable additional printk() statements. 4468 4469 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL] 4470 Dump ftrace buffer after reporting RCU CPU 4471 stall warning. 4472 4473 rcupdate.rcu_cpu_stall_suppress= [KNL] 4474 Suppress RCU CPU stall warning messages. 4475 4476 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL] 4477 Suppress RCU CPU stall warning messages and 4478 rcutorture writer stall warnings that occur 4479 during early boot, that is, during the time 4480 before the init task is spawned. 4481 4482 rcupdate.rcu_cpu_stall_timeout= [KNL] 4483 Set timeout for RCU CPU stall warning messages. 4484 4485 rcupdate.rcu_expedited= [KNL] 4486 Use expedited grace-period primitives, for 4487 example, synchronize_rcu_expedited() instead 4488 of synchronize_rcu(). This reduces latency, 4489 but can increase CPU utilization, degrade 4490 real-time latency, and degrade energy efficiency. 4491 No effect on CONFIG_TINY_RCU kernels. 4492 4493 rcupdate.rcu_normal= [KNL] 4494 Use only normal grace-period primitives, 4495 for example, synchronize_rcu() instead of 4496 synchronize_rcu_expedited(). This improves 4497 real-time latency, CPU utilization, and 4498 energy efficiency, but can expose users to 4499 increased grace-period latency. This parameter 4500 overrides rcupdate.rcu_expedited. No effect on 4501 CONFIG_TINY_RCU kernels. 4502 4503 rcupdate.rcu_normal_after_boot= [KNL] 4504 Once boot has completed (that is, after 4505 rcu_end_inkernel_boot() has been invoked), use 4506 only normal grace-period primitives. No effect 4507 on CONFIG_TINY_RCU kernels. 4508 4509 rcupdate.rcu_task_ipi_delay= [KNL] 4510 Set time in jiffies during which RCU tasks will 4511 avoid sending IPIs, starting with the beginning 4512 of a given grace period. Setting a large 4513 number avoids disturbing real-time workloads, 4514 but lengthens grace periods. 4515 4516 rcupdate.rcu_task_stall_timeout= [KNL] 4517 Set timeout in jiffies for RCU task stall warning 4518 messages. Disable with a value less than or equal 4519 to zero. 4520 4521 rcupdate.rcu_self_test= [KNL] 4522 Run the RCU early boot self tests 4523 4524 rdinit= [KNL] 4525 Format: <full_path> 4526 Run specified binary instead of /init from the ramdisk, 4527 used for early userspace startup. See initrd. 4528 4529 rdrand= [X86] 4530 force - Override the decision by the kernel to hide the 4531 advertisement of RDRAND support (this affects 4532 certain AMD processors because of buggy BIOS 4533 support, specifically around the suspend/resume 4534 path). 4535 4536 rdt= [HW,X86,RDT] 4537 Turn on/off individual RDT features. List is: 4538 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp, 4539 mba. 4540 E.g. to turn on cmt and turn off mba use: 4541 rdt=cmt,!mba 4542 4543 reboot= [KNL] 4544 Format (x86 or x86_64): 4545 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \ 4546 [[,]s[mp]#### \ 4547 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \ 4548 [[,]f[orce] 4549 Where reboot_mode is one of warm (soft) or cold (hard) or gpio 4550 (prefix with 'panic_' to set mode for panic 4551 reboot only), 4552 reboot_type is one of bios, acpi, kbd, triple, efi, or pci, 4553 reboot_force is either force or not specified, 4554 reboot_cpu is s[mp]#### with #### being the processor 4555 to be used for rebooting. 4556 4557 refscale.holdoff= [KNL] 4558 Set test-start holdoff period. The purpose of 4559 this parameter is to delay the start of the 4560 test until boot completes in order to avoid 4561 interference. 4562 4563 refscale.loops= [KNL] 4564 Set the number of loops over the synchronization 4565 primitive under test. Increasing this number 4566 reduces noise due to loop start/end overhead, 4567 but the default has already reduced the per-pass 4568 noise to a handful of picoseconds on ca. 2020 4569 x86 laptops. 4570 4571 refscale.nreaders= [KNL] 4572 Set number of readers. The default value of -1 4573 selects N, where N is roughly 75% of the number 4574 of CPUs. A value of zero is an interesting choice. 4575 4576 refscale.nruns= [KNL] 4577 Set number of runs, each of which is dumped onto 4578 the console log. 4579 4580 refscale.readdelay= [KNL] 4581 Set the read-side critical-section duration, 4582 measured in microseconds. 4583 4584 refscale.scale_type= [KNL] 4585 Specify the read-protection implementation to test. 4586 4587 refscale.shutdown= [KNL] 4588 Shut down the system at the end of the performance 4589 test. This defaults to 1 (shut it down) when 4590 refscale is built into the kernel and to 0 (leave 4591 it running) when refscale is built as a module. 4592 4593 refscale.verbose= [KNL] 4594 Enable additional printk() statements. 4595 4596 relax_domain_level= 4597 [KNL, SMP] Set scheduler's default relax_domain_level. 4598 See Documentation/admin-guide/cgroup-v1/cpusets.rst. 4599 4600 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory 4601 Format: <base1>,<size1>[,<base2>,<size2>,...] 4602 Reserve I/O ports or memory so the kernel won't use 4603 them. If <base> is less than 0x10000, the region 4604 is assumed to be I/O ports; otherwise it is memory. 4605 4606 reservetop= [X86-32] 4607 Format: nn[KMG] 4608 Reserves a hole at the top of the kernel virtual 4609 address space. 4610 4611 reservelow= [X86] 4612 Format: nn[K] 4613 Set the amount of memory to reserve for BIOS at 4614 the bottom of the address space. 4615 4616 reset_devices [KNL] Force drivers to reset the underlying device 4617 during initialization. 4618 4619 resume= [SWSUSP] 4620 Specify the partition device for software suspend 4621 Format: 4622 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>} 4623 4624 resume_offset= [SWSUSP] 4625 Specify the offset from the beginning of the partition 4626 given by "resume=" at which the swap header is located, 4627 in <PAGE_SIZE> units (needed only for swap files). 4628 See Documentation/power/swsusp-and-swap-files.rst 4629 4630 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to 4631 read the resume files 4632 4633 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up. 4634 Useful for devices that are detected asynchronously 4635 (e.g. USB and MMC devices). 4636 4637 hibernate= [HIBERNATION] 4638 noresume Don't check if there's a hibernation image 4639 present during boot. 4640 nocompress Don't compress/decompress hibernation images. 4641 no Disable hibernation and resume. 4642 protect_image Turn on image protection during restoration 4643 (that will set all pages holding image data 4644 during restoration read-only). 4645 4646 retain_initrd [RAM] Keep initrd memory after extraction 4647 4648 retbleed= [X86] Control mitigation of RETBleed (Arbitrary 4649 Speculative Code Execution with Return Instructions) 4650 vulnerability. 4651 4652 off - no mitigation 4653 auto - automatically select a migitation 4654 auto,nosmt - automatically select a mitigation, 4655 disabling SMT if necessary for 4656 the full mitigation (only on Zen1 4657 and older without STIBP). 4658 ibpb - mitigate short speculation windows on 4659 basic block boundaries too. Safe, highest 4660 perf impact. 4661 unret - force enable untrained return thunks, 4662 only effective on AMD f15h-f17h 4663 based systems. 4664 unret,nosmt - like unret, will disable SMT when STIBP 4665 is not available. 4666 4667 Selecting 'auto' will choose a mitigation method at run 4668 time according to the CPU. 4669 4670 Not specifying this option is equivalent to retbleed=auto. 4671 4672 rfkill.default_state= 4673 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm, 4674 etc. communication is blocked by default. 4675 1 Unblocked. 4676 4677 rfkill.master_switch_mode= 4678 0 The "airplane mode" button does nothing. 4679 1 The "airplane mode" button toggles between everything 4680 blocked and the previous configuration. 4681 2 The "airplane mode" button toggles between everything 4682 blocked and everything unblocked. 4683 4684 rhash_entries= [KNL,NET] 4685 Set number of hash buckets for route cache 4686 4687 ring3mwait=disable 4688 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported 4689 CPUs. 4690 4691 ro [KNL] Mount root device read-only on boot 4692 4693 rodata= [KNL] 4694 on Mark read-only kernel memory as read-only (default). 4695 off Leave read-only kernel memory writable for debugging. 4696 4697 rockchip.usb_uart 4698 Enable the uart passthrough on the designated usb port 4699 on Rockchip SoCs. When active, the signals of the 4700 debug-uart get routed to the D+ and D- pins of the usb 4701 port and the regular usb controller gets disabled. 4702 4703 root= [KNL] Root filesystem 4704 See name_to_dev_t comment in init/do_mounts.c. 4705 4706 rootdelay= [KNL] Delay (in seconds) to pause before attempting to 4707 mount the root filesystem 4708 4709 rootflags= [KNL] Set root filesystem mount option string 4710 4711 rootfstype= [KNL] Set root filesystem type 4712 4713 rootwait [KNL] Wait (indefinitely) for root device to show up. 4714 Useful for devices that are detected asynchronously 4715 (e.g. USB and MMC devices). 4716 4717 rproc_mem=nn[KMG][@address] 4718 [KNL,ARM,CMA] Remoteproc physical memory block. 4719 Memory area to be used by remote processor image, 4720 managed by CMA. 4721 4722 rw [KNL] Mount root device read-write on boot 4723 4724 S [KNL] Run init in single mode 4725 4726 s390_iommu= [HW,S390] 4727 Set s390 IOTLB flushing mode 4728 strict 4729 With strict flushing every unmap operation will result in 4730 an IOTLB flush. Default is lazy flushing before reuse, 4731 which is faster. 4732 4733 sa1100ir [NET] 4734 See drivers/net/irda/sa1100_ir.c. 4735 4736 sbni= [NET] Granch SBNI12 leased line adapter 4737 4738 sched_debug [KNL] Enables verbose scheduler debug messages. 4739 4740 schedstats= [KNL,X86] Enable or disable scheduled statistics. 4741 Allowed values are enable and disable. This feature 4742 incurs a small amount of overhead in the scheduler 4743 but is useful for debugging and performance tuning. 4744 4745 sched_thermal_decay_shift= 4746 [KNL, SMP] Set a decay shift for scheduler thermal 4747 pressure signal. Thermal pressure signal follows the 4748 default decay period of other scheduler pelt 4749 signals(usually 32 ms but configurable). Setting 4750 sched_thermal_decay_shift will left shift the decay 4751 period for the thermal pressure signal by the shift 4752 value. 4753 i.e. with the default pelt decay period of 32 ms 4754 sched_thermal_decay_shift thermal pressure decay pr 4755 1 64 ms 4756 2 128 ms 4757 and so on. 4758 Format: integer between 0 and 10 4759 Default is 0. 4760 4761 scftorture.holdoff= [KNL] 4762 Number of seconds to hold off before starting 4763 test. Defaults to zero for module insertion and 4764 to 10 seconds for built-in smp_call_function() 4765 tests. 4766 4767 scftorture.longwait= [KNL] 4768 Request ridiculously long waits randomly selected 4769 up to the chosen limit in seconds. Zero (the 4770 default) disables this feature. Please note 4771 that requesting even small non-zero numbers of 4772 seconds can result in RCU CPU stall warnings, 4773 softlockup complaints, and so on. 4774 4775 scftorture.nthreads= [KNL] 4776 Number of kthreads to spawn to invoke the 4777 smp_call_function() family of functions. 4778 The default of -1 specifies a number of kthreads 4779 equal to the number of CPUs. 4780 4781 scftorture.onoff_holdoff= [KNL] 4782 Number seconds to wait after the start of the 4783 test before initiating CPU-hotplug operations. 4784 4785 scftorture.onoff_interval= [KNL] 4786 Number seconds to wait between successive 4787 CPU-hotplug operations. Specifying zero (which 4788 is the default) disables CPU-hotplug operations. 4789 4790 scftorture.shutdown_secs= [KNL] 4791 The number of seconds following the start of the 4792 test after which to shut down the system. The 4793 default of zero avoids shutting down the system. 4794 Non-zero values are useful for automated tests. 4795 4796 scftorture.stat_interval= [KNL] 4797 The number of seconds between outputting the 4798 current test statistics to the console. A value 4799 of zero disables statistics output. 4800 4801 scftorture.stutter_cpus= [KNL] 4802 The number of jiffies to wait between each change 4803 to the set of CPUs under test. 4804 4805 scftorture.use_cpus_read_lock= [KNL] 4806 Use use_cpus_read_lock() instead of the default 4807 preempt_disable() to disable CPU hotplug 4808 while invoking one of the smp_call_function*() 4809 functions. 4810 4811 scftorture.verbose= [KNL] 4812 Enable additional printk() statements. 4813 4814 scftorture.weight_single= [KNL] 4815 The probability weighting to use for the 4816 smp_call_function_single() function with a zero 4817 "wait" parameter. A value of -1 selects the 4818 default if all other weights are -1. However, 4819 if at least one weight has some other value, a 4820 value of -1 will instead select a weight of zero. 4821 4822 scftorture.weight_single_wait= [KNL] 4823 The probability weighting to use for the 4824 smp_call_function_single() function with a 4825 non-zero "wait" parameter. See weight_single. 4826 4827 scftorture.weight_many= [KNL] 4828 The probability weighting to use for the 4829 smp_call_function_many() function with a zero 4830 "wait" parameter. See weight_single. 4831 Note well that setting a high probability for 4832 this weighting can place serious IPI load 4833 on the system. 4834 4835 scftorture.weight_many_wait= [KNL] 4836 The probability weighting to use for the 4837 smp_call_function_many() function with a 4838 non-zero "wait" parameter. See weight_single 4839 and weight_many. 4840 4841 scftorture.weight_all= [KNL] 4842 The probability weighting to use for the 4843 smp_call_function_all() function with a zero 4844 "wait" parameter. See weight_single and 4845 weight_many. 4846 4847 scftorture.weight_all_wait= [KNL] 4848 The probability weighting to use for the 4849 smp_call_function_all() function with a 4850 non-zero "wait" parameter. See weight_single 4851 and weight_many. 4852 4853 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate 4854 xtime_lock contention on larger systems, and/or RCU lock 4855 contention on all systems with CONFIG_MAXSMP set. 4856 Format: { "0" | "1" } 4857 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1" 4858 1 -- enable. 4859 Note: increases power consumption, thus should only be 4860 enabled if running jitter sensitive (HPC/RT) workloads. 4861 4862 security= [SECURITY] Choose a legacy "major" security module to 4863 enable at boot. This has been deprecated by the 4864 "lsm=" parameter. 4865 4866 selinux= [SELINUX] Disable or enable SELinux at boot time. 4867 Format: { "0" | "1" } 4868 See security/selinux/Kconfig help text. 4869 0 -- disable. 4870 1 -- enable. 4871 Default value is 1. 4872 4873 apparmor= [APPARMOR] Disable or enable AppArmor at boot time 4874 Format: { "0" | "1" } 4875 See security/apparmor/Kconfig help text 4876 0 -- disable. 4877 1 -- enable. 4878 Default value is set via kernel config option. 4879 4880 serialnumber [BUGS=X86-32] 4881 4882 shapers= [NET] 4883 Maximal number of shapers. 4884 4885 simeth= [IA-64] 4886 simscsi= 4887 4888 slram= [HW,MTD] 4889 4890 slab_nomerge [MM] 4891 Disable merging of slabs with similar size. May be 4892 necessary if there is some reason to distinguish 4893 allocs to different slabs, especially in hardened 4894 environments where the risk of heap overflows and 4895 layout control by attackers can usually be 4896 frustrated by disabling merging. This will reduce 4897 most of the exposure of a heap attack to a single 4898 cache (risks via metadata attacks are mostly 4899 unchanged). Debug options disable merging on their 4900 own. 4901 For more information see Documentation/vm/slub.rst. 4902 4903 slab_max_order= [MM, SLAB] 4904 Determines the maximum allowed order for slabs. 4905 A high setting may cause OOMs due to memory 4906 fragmentation. Defaults to 1 for systems with 4907 more than 32MB of RAM, 0 otherwise. 4908 4909 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB] 4910 Enabling slub_debug allows one to determine the 4911 culprit if slab objects become corrupted. Enabling 4912 slub_debug can create guard zones around objects and 4913 may poison objects when not in use. Also tracks the 4914 last alloc / free. For more information see 4915 Documentation/vm/slub.rst. 4916 4917 slub_memcg_sysfs= [MM, SLUB] 4918 Determines whether to enable sysfs directories for 4919 memory cgroup sub-caches. 1 to enable, 0 to disable. 4920 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON. 4921 Enabling this can lead to a very high number of debug 4922 directories and files being created under 4923 /sys/kernel/slub. 4924 4925 slub_max_order= [MM, SLUB] 4926 Determines the maximum allowed order for slabs. 4927 A high setting may cause OOMs due to memory 4928 fragmentation. For more information see 4929 Documentation/vm/slub.rst. 4930 4931 slub_min_objects= [MM, SLUB] 4932 The minimum number of objects per slab. SLUB will 4933 increase the slab order up to slub_max_order to 4934 generate a sufficiently large slab able to contain 4935 the number of objects indicated. The higher the number 4936 of objects the smaller the overhead of tracking slabs 4937 and the less frequently locks need to be acquired. 4938 For more information see Documentation/vm/slub.rst. 4939 4940 slub_min_order= [MM, SLUB] 4941 Determines the minimum page order for slabs. Must be 4942 lower than slub_max_order. 4943 For more information see Documentation/vm/slub.rst. 4944 4945 slub_nomerge [MM, SLUB] 4946 Same with slab_nomerge. This is supported for legacy. 4947 See slab_nomerge for more information. 4948 4949 smart2= [HW] 4950 Format: <io1>[,<io2>[,...,<io8>]] 4951 4952 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices 4953 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port 4954 smsc-ircc2.ircc_sir= [HW] SIR base I/O port 4955 smsc-ircc2.ircc_fir= [HW] FIR base I/O port 4956 smsc-ircc2.ircc_irq= [HW] IRQ line 4957 smsc-ircc2.ircc_dma= [HW] DMA channel 4958 smsc-ircc2.ircc_transceiver= [HW] Transceiver type: 4959 0: Toshiba Satellite 1800 (GP data pin select) 4960 1: Fast pin select (default) 4961 2: ATC IRMode 4962 4963 smt [KNL,S390] Set the maximum number of threads (logical 4964 CPUs) to use per physical CPU on systems capable of 4965 symmetric multithreading (SMT). Will be capped to the 4966 actual hardware limit. 4967 Format: <integer> 4968 Default: -1 (no limit) 4969 4970 softlockup_panic= 4971 [KNL] Should the soft-lockup detector generate panics. 4972 Format: 0 | 1 4973 4974 A value of 1 instructs the soft-lockup detector 4975 to panic the machine when a soft-lockup occurs. It is 4976 also controlled by the kernel.softlockup_panic sysctl 4977 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the 4978 respective build-time switch to that functionality. 4979 4980 softlockup_all_cpu_backtrace= 4981 [KNL] Should the soft-lockup detector generate 4982 backtraces on all cpus. 4983 Format: 0 | 1 4984 4985 sonypi.*= [HW] Sony Programmable I/O Control Device driver 4986 See Documentation/admin-guide/laptops/sonypi.rst 4987 4988 spectre_v2= [X86] Control mitigation of Spectre variant 2 4989 (indirect branch speculation) vulnerability. 4990 The default operation protects the kernel from 4991 user space attacks. 4992 4993 on - unconditionally enable, implies 4994 spectre_v2_user=on 4995 off - unconditionally disable, implies 4996 spectre_v2_user=off 4997 auto - kernel detects whether your CPU model is 4998 vulnerable 4999 5000 Selecting 'on' will, and 'auto' may, choose a 5001 mitigation method at run time according to the 5002 CPU, the available microcode, the setting of the 5003 CONFIG_RETPOLINE configuration option, and the 5004 compiler with which the kernel was built. 5005 5006 Selecting 'on' will also enable the mitigation 5007 against user space to user space task attacks. 5008 5009 Selecting 'off' will disable both the kernel and 5010 the user space protections. 5011 5012 Specific mitigations can also be selected manually: 5013 5014 retpoline - replace indirect branches 5015 retpoline,generic - Retpolines 5016 retpoline,lfence - LFENCE; indirect branch 5017 retpoline,amd - alias for retpoline,lfence 5018 eibrs - enhanced IBRS 5019 eibrs,retpoline - enhanced IBRS + Retpolines 5020 eibrs,lfence - enhanced IBRS + LFENCE 5021 ibrs - use IBRS to protect kernel 5022 5023 Not specifying this option is equivalent to 5024 spectre_v2=auto. 5025 5026 spectre_v2_user= 5027 [X86] Control mitigation of Spectre variant 2 5028 (indirect branch speculation) vulnerability between 5029 user space tasks 5030 5031 on - Unconditionally enable mitigations. Is 5032 enforced by spectre_v2=on 5033 5034 off - Unconditionally disable mitigations. Is 5035 enforced by spectre_v2=off 5036 5037 prctl - Indirect branch speculation is enabled, 5038 but mitigation can be enabled via prctl 5039 per thread. The mitigation control state 5040 is inherited on fork. 5041 5042 prctl,ibpb 5043 - Like "prctl" above, but only STIBP is 5044 controlled per thread. IBPB is issued 5045 always when switching between different user 5046 space processes. 5047 5048 seccomp 5049 - Same as "prctl" above, but all seccomp 5050 threads will enable the mitigation unless 5051 they explicitly opt out. 5052 5053 seccomp,ibpb 5054 - Like "seccomp" above, but only STIBP is 5055 controlled per thread. IBPB is issued 5056 always when switching between different 5057 user space processes. 5058 5059 auto - Kernel selects the mitigation depending on 5060 the available CPU features and vulnerability. 5061 5062 Default mitigation: 5063 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl" 5064 5065 Not specifying this option is equivalent to 5066 spectre_v2_user=auto. 5067 5068 spec_store_bypass_disable= 5069 [HW] Control Speculative Store Bypass (SSB) Disable mitigation 5070 (Speculative Store Bypass vulnerability) 5071 5072 Certain CPUs are vulnerable to an exploit against a 5073 a common industry wide performance optimization known 5074 as "Speculative Store Bypass" in which recent stores 5075 to the same memory location may not be observed by 5076 later loads during speculative execution. The idea 5077 is that such stores are unlikely and that they can 5078 be detected prior to instruction retirement at the 5079 end of a particular speculation execution window. 5080 5081 In vulnerable processors, the speculatively forwarded 5082 store can be used in a cache side channel attack, for 5083 example to read memory to which the attacker does not 5084 directly have access (e.g. inside sandboxed code). 5085 5086 This parameter controls whether the Speculative Store 5087 Bypass optimization is used. 5088 5089 On x86 the options are: 5090 5091 on - Unconditionally disable Speculative Store Bypass 5092 off - Unconditionally enable Speculative Store Bypass 5093 auto - Kernel detects whether the CPU model contains an 5094 implementation of Speculative Store Bypass and 5095 picks the most appropriate mitigation. If the 5096 CPU is not vulnerable, "off" is selected. If the 5097 CPU is vulnerable the default mitigation is 5098 architecture and Kconfig dependent. See below. 5099 prctl - Control Speculative Store Bypass per thread 5100 via prctl. Speculative Store Bypass is enabled 5101 for a process by default. The state of the control 5102 is inherited on fork. 5103 seccomp - Same as "prctl" above, but all seccomp threads 5104 will disable SSB unless they explicitly opt out. 5105 5106 Default mitigations: 5107 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl" 5108 5109 On powerpc the options are: 5110 5111 on,auto - On Power8 and Power9 insert a store-forwarding 5112 barrier on kernel entry and exit. On Power7 5113 perform a software flush on kernel entry and 5114 exit. 5115 off - No action. 5116 5117 Not specifying this option is equivalent to 5118 spec_store_bypass_disable=auto. 5119 5120 spia_io_base= [HW,MTD] 5121 spia_fio_base= 5122 spia_pedr= 5123 spia_peddr= 5124 5125 split_lock_detect= 5126 [X86] Enable split lock detection 5127 5128 When enabled (and if hardware support is present), atomic 5129 instructions that access data across cache line 5130 boundaries will result in an alignment check exception. 5131 5132 off - not enabled 5133 5134 warn - the kernel will emit rate limited warnings 5135 about applications triggering the #AC 5136 exception. This mode is the default on CPUs 5137 that supports split lock detection. 5138 5139 fatal - the kernel will send SIGBUS to applications 5140 that trigger the #AC exception. 5141 5142 If an #AC exception is hit in the kernel or in 5143 firmware (i.e. not while executing in user mode) 5144 the kernel will oops in either "warn" or "fatal" 5145 mode. 5146 5147 srbds= [X86,INTEL] 5148 Control the Special Register Buffer Data Sampling 5149 (SRBDS) mitigation. 5150 5151 Certain CPUs are vulnerable to an MDS-like 5152 exploit which can leak bits from the random 5153 number generator. 5154 5155 By default, this issue is mitigated by 5156 microcode. However, the microcode fix can cause 5157 the RDRAND and RDSEED instructions to become 5158 much slower. Among other effects, this will 5159 result in reduced throughput from /dev/urandom. 5160 5161 The microcode mitigation can be disabled with 5162 the following option: 5163 5164 off: Disable mitigation and remove 5165 performance impact to RDRAND and RDSEED 5166 5167 srcutree.counter_wrap_check [KNL] 5168 Specifies how frequently to check for 5169 grace-period sequence counter wrap for the 5170 srcu_data structure's ->srcu_gp_seq_needed field. 5171 The greater the number of bits set in this kernel 5172 parameter, the less frequently counter wrap will 5173 be checked for. Note that the bottom two bits 5174 are ignored. 5175 5176 srcutree.exp_holdoff [KNL] 5177 Specifies how many nanoseconds must elapse 5178 since the end of the last SRCU grace period for 5179 a given srcu_struct until the next normal SRCU 5180 grace period will be considered for automatic 5181 expediting. Set to zero to disable automatic 5182 expediting. 5183 5184 ssbd= [ARM64,HW] 5185 Speculative Store Bypass Disable control 5186 5187 On CPUs that are vulnerable to the Speculative 5188 Store Bypass vulnerability and offer a 5189 firmware based mitigation, this parameter 5190 indicates how the mitigation should be used: 5191 5192 force-on: Unconditionally enable mitigation for 5193 for both kernel and userspace 5194 force-off: Unconditionally disable mitigation for 5195 for both kernel and userspace 5196 kernel: Always enable mitigation in the 5197 kernel, and offer a prctl interface 5198 to allow userspace to register its 5199 interest in being mitigated too. 5200 5201 stack_guard_gap= [MM] 5202 override the default stack gap protection. The value 5203 is in page units and it defines how many pages prior 5204 to (for stacks growing down) resp. after (for stacks 5205 growing up) the main stack are reserved for no other 5206 mapping. Default value is 256 pages. 5207 5208 stacktrace [FTRACE] 5209 Enabled the stack tracer on boot up. 5210 5211 stacktrace_filter=[function-list] 5212 [FTRACE] Limit the functions that the stack tracer 5213 will trace at boot up. function-list is a comma separated 5214 list of functions. This list can be changed at run 5215 time by the stack_trace_filter file in the debugfs 5216 tracing directory. Note, this enables stack tracing 5217 and the stacktrace above is not needed. 5218 5219 sti= [PARISC,HW] 5220 Format: <num> 5221 Set the STI (builtin display/keyboard on the HP-PARISC 5222 machines) console (graphic card) which should be used 5223 as the initial boot-console. 5224 See also comment in drivers/video/console/sticore.c. 5225 5226 sti_font= [HW] 5227 See comment in drivers/video/console/sticore.c. 5228 5229 stifb= [HW] 5230 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]] 5231 5232 sunrpc.min_resvport= 5233 sunrpc.max_resvport= 5234 [NFS,SUNRPC] 5235 SunRPC servers often require that client requests 5236 originate from a privileged port (i.e. a port in the 5237 range 0 < portnr < 1024). 5238 An administrator who wishes to reserve some of these 5239 ports for other uses may adjust the range that the 5240 kernel's sunrpc client considers to be privileged 5241 using these two parameters to set the minimum and 5242 maximum port values. 5243 5244 sunrpc.svc_rpc_per_connection_limit= 5245 [NFS,SUNRPC] 5246 Limit the number of requests that the server will 5247 process in parallel from a single connection. 5248 The default value is 0 (no limit). 5249 5250 sunrpc.pool_mode= 5251 [NFS] 5252 Control how the NFS server code allocates CPUs to 5253 service thread pools. Depending on how many NICs 5254 you have and where their interrupts are bound, this 5255 option will affect which CPUs will do NFS serving. 5256 Note: this parameter cannot be changed while the 5257 NFS server is running. 5258 5259 auto the server chooses an appropriate mode 5260 automatically using heuristics 5261 global a single global pool contains all CPUs 5262 percpu one pool for each CPU 5263 pernode one pool for each NUMA node (equivalent 5264 to global on non-NUMA machines) 5265 5266 sunrpc.tcp_slot_table_entries= 5267 sunrpc.udp_slot_table_entries= 5268 [NFS,SUNRPC] 5269 Sets the upper limit on the number of simultaneous 5270 RPC calls that can be sent from the client to a 5271 server. Increasing these values may allow you to 5272 improve throughput, but will also increase the 5273 amount of memory reserved for use by the client. 5274 5275 suspend.pm_test_delay= 5276 [SUSPEND] 5277 Sets the number of seconds to remain in a suspend test 5278 mode before resuming the system (see 5279 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG 5280 is set. Default value is 5. 5281 5282 svm= [PPC] 5283 Format: { on | off | y | n | 1 | 0 } 5284 This parameter controls use of the Protected 5285 Execution Facility on pSeries. 5286 5287 swapaccount=[0|1] 5288 [KNL] Enable accounting of swap in memory resource 5289 controller if no parameter or 1 is given or disable 5290 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst) 5291 5292 swiotlb= [ARM,IA-64,PPC,MIPS,X86] 5293 Format: { <int> | force | noforce } 5294 <int> -- Number of I/O TLB slabs 5295 force -- force using of bounce buffers even if they 5296 wouldn't be automatically used by the kernel 5297 noforce -- Never use bounce buffers (for debugging) 5298 5299 switches= [HW,M68k] 5300 5301 sysctl.*= [KNL] 5302 Set a sysctl parameter, right before loading the init 5303 process, as if the value was written to the respective 5304 /proc/sys/... file. Both '.' and '/' are recognized as 5305 separators. Unrecognized parameters and invalid values 5306 are reported in the kernel log. Sysctls registered 5307 later by a loaded module cannot be set this way. 5308 Example: sysctl.vm.swappiness=40 5309 5310 sysfs.deprecated=0|1 [KNL] 5311 Enable/disable old style sysfs layout for old udev 5312 on older distributions. When this option is enabled 5313 very new udev will not work anymore. When this option 5314 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled) 5315 in older udev will not work anymore. 5316 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in 5317 the kernel configuration. 5318 5319 sysrq_always_enabled 5320 [KNL] 5321 Ignore sysrq setting - this boot parameter will 5322 neutralize any effect of /proc/sys/kernel/sysrq. 5323 Useful for debugging. 5324 5325 tcpmhash_entries= [KNL,NET] 5326 Set the number of tcp_metrics_hash slots. 5327 Default value is 8192 or 16384 depending on total 5328 ram pages. This is used to specify the TCP metrics 5329 cache size. See Documentation/networking/ip-sysctl.rst 5330 "tcp_no_metrics_save" section for more details. 5331 5332 tdfx= [HW,DRM] 5333 5334 test_suspend= [SUSPEND][,N] 5335 Specify "mem" (for Suspend-to-RAM) or "standby" (for 5336 standby suspend) or "freeze" (for suspend type freeze) 5337 as the system sleep state during system startup with 5338 the optional capability to repeat N number of times. 5339 The system is woken from this state using a 5340 wakeup-capable RTC alarm. 5341 5342 thash_entries= [KNL,NET] 5343 Set number of hash buckets for TCP connection 5344 5345 thermal.act= [HW,ACPI] 5346 -1: disable all active trip points in all thermal zones 5347 <degrees C>: override all lowest active trip points 5348 5349 thermal.crt= [HW,ACPI] 5350 -1: disable all critical trip points in all thermal zones 5351 <degrees C>: override all critical trip points 5352 5353 thermal.nocrt= [HW,ACPI] 5354 Set to disable actions on ACPI thermal zone 5355 critical and hot trip points. 5356 5357 thermal.off= [HW,ACPI] 5358 1: disable ACPI thermal control 5359 5360 thermal.psv= [HW,ACPI] 5361 -1: disable all passive trip points 5362 <degrees C>: override all passive trip points to this 5363 value 5364 5365 thermal.tzp= [HW,ACPI] 5366 Specify global default ACPI thermal zone polling rate 5367 <deci-seconds>: poll all this frequency 5368 0: no polling (default) 5369 5370 threadirqs [KNL] 5371 Force threading of all interrupt handlers except those 5372 marked explicitly IRQF_NO_THREAD. 5373 5374 topology= [S390] 5375 Format: {off | on} 5376 Specify if the kernel should make use of the cpu 5377 topology information if the hardware supports this. 5378 The scheduler will make use of this information and 5379 e.g. base its process migration decisions on it. 5380 Default is on. 5381 5382 topology_updates= [KNL, PPC, NUMA] 5383 Format: {off} 5384 Specify if the kernel should ignore (off) 5385 topology updates sent by the hypervisor to this 5386 LPAR. 5387 5388 torture.disable_onoff_at_boot= [KNL] 5389 Prevent the CPU-hotplug component of torturing 5390 until after init has spawned. 5391 5392 torture.ftrace_dump_at_shutdown= [KNL] 5393 Dump the ftrace buffer at torture-test shutdown, 5394 even if there were no errors. This can be a 5395 very costly operation when many torture tests 5396 are running concurrently, especially on systems 5397 with rotating-rust storage. 5398 5399 tp720= [HW,PS2] 5400 5401 tpm_suspend_pcr=[HW,TPM] 5402 Format: integer pcr id 5403 Specify that at suspend time, the tpm driver 5404 should extend the specified pcr with zeros, 5405 as a workaround for some chips which fail to 5406 flush the last written pcr on TPM_SaveState. 5407 This will guarantee that all the other pcrs 5408 are saved. 5409 5410 trace_buf_size=nn[KMG] 5411 [FTRACE] will set tracing buffer size on each cpu. 5412 5413 trace_event=[event-list] 5414 [FTRACE] Set and start specified trace events in order 5415 to facilitate early boot debugging. The event-list is a 5416 comma separated list of trace events to enable. See 5417 also Documentation/trace/events.rst 5418 5419 trace_options=[option-list] 5420 [FTRACE] Enable or disable tracer options at boot. 5421 The option-list is a comma delimited list of options 5422 that can be enabled or disabled just as if you were 5423 to echo the option name into 5424 5425 /sys/kernel/debug/tracing/trace_options 5426 5427 For example, to enable stacktrace option (to dump the 5428 stack trace of each event), add to the command line: 5429 5430 trace_options=stacktrace 5431 5432 See also Documentation/trace/ftrace.rst "trace options" 5433 section. 5434 5435 tp_printk[FTRACE] 5436 Have the tracepoints sent to printk as well as the 5437 tracing ring buffer. This is useful for early boot up 5438 where the system hangs or reboots and does not give the 5439 option for reading the tracing buffer or performing a 5440 ftrace_dump_on_oops. 5441 5442 To turn off having tracepoints sent to printk, 5443 echo 0 > /proc/sys/kernel/tracepoint_printk 5444 Note, echoing 1 into this file without the 5445 tracepoint_printk kernel cmdline option has no effect. 5446 5447 ** CAUTION ** 5448 5449 Having tracepoints sent to printk() and activating high 5450 frequency tracepoints such as irq or sched, can cause 5451 the system to live lock. 5452 5453 traceoff_on_warning 5454 [FTRACE] enable this option to disable tracing when a 5455 warning is hit. This turns off "tracing_on". Tracing can 5456 be enabled again by echoing '1' into the "tracing_on" 5457 file located in /sys/kernel/debug/tracing/ 5458 5459 This option is useful, as it disables the trace before 5460 the WARNING dump is called, which prevents the trace to 5461 be filled with content caused by the warning output. 5462 5463 This option can also be set at run time via the sysctl 5464 option: kernel/traceoff_on_warning 5465 5466 transparent_hugepage= 5467 [KNL] 5468 Format: [always|madvise|never] 5469 Can be used to control the default behavior of the system 5470 with respect to transparent hugepages. 5471 See Documentation/admin-guide/mm/transhuge.rst 5472 for more details. 5473 5474 tsc= Disable clocksource stability checks for TSC. 5475 Format: <string> 5476 [x86] reliable: mark tsc clocksource as reliable, this 5477 disables clocksource verification at runtime, as well 5478 as the stability checks done at bootup. Used to enable 5479 high-resolution timer mode on older hardware, and in 5480 virtualized environment. 5481 [x86] noirqtime: Do not use TSC to do irq accounting. 5482 Used to run time disable IRQ_TIME_ACCOUNTING on any 5483 platforms where RDTSC is slow and this accounting 5484 can add overhead. 5485 [x86] unstable: mark the TSC clocksource as unstable, this 5486 marks the TSC unconditionally unstable at bootup and 5487 avoids any further wobbles once the TSC watchdog notices. 5488 [x86] nowatchdog: disable clocksource watchdog. Used 5489 in situations with strict latency requirements (where 5490 interruptions from clocksource watchdog are not 5491 acceptable). 5492 5493 tsc_early_khz= [X86] Skip early TSC calibration and use the given 5494 value instead. Useful when the early TSC frequency discovery 5495 procedure is not reliable, such as on overclocked systems 5496 with CPUID.16h support and partial CPUID.15h support. 5497 Format: <unsigned int> 5498 5499 tsx= [X86] Control Transactional Synchronization 5500 Extensions (TSX) feature in Intel processors that 5501 support TSX control. 5502 5503 This parameter controls the TSX feature. The options are: 5504 5505 on - Enable TSX on the system. Although there are 5506 mitigations for all known security vulnerabilities, 5507 TSX has been known to be an accelerator for 5508 several previous speculation-related CVEs, and 5509 so there may be unknown security risks associated 5510 with leaving it enabled. 5511 5512 off - Disable TSX on the system. (Note that this 5513 option takes effect only on newer CPUs which are 5514 not vulnerable to MDS, i.e., have 5515 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get 5516 the new IA32_TSX_CTRL MSR through a microcode 5517 update. This new MSR allows for the reliable 5518 deactivation of the TSX functionality.) 5519 5520 auto - Disable TSX if X86_BUG_TAA is present, 5521 otherwise enable TSX on the system. 5522 5523 Not specifying this option is equivalent to tsx=off. 5524 5525 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 5526 for more details. 5527 5528 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async 5529 Abort (TAA) vulnerability. 5530 5531 Similar to Micro-architectural Data Sampling (MDS) 5532 certain CPUs that support Transactional 5533 Synchronization Extensions (TSX) are vulnerable to an 5534 exploit against CPU internal buffers which can forward 5535 information to a disclosure gadget under certain 5536 conditions. 5537 5538 In vulnerable processors, the speculatively forwarded 5539 data can be used in a cache side channel attack, to 5540 access data to which the attacker does not have direct 5541 access. 5542 5543 This parameter controls the TAA mitigation. The 5544 options are: 5545 5546 full - Enable TAA mitigation on vulnerable CPUs 5547 if TSX is enabled. 5548 5549 full,nosmt - Enable TAA mitigation and disable SMT on 5550 vulnerable CPUs. If TSX is disabled, SMT 5551 is not disabled because CPU is not 5552 vulnerable to cross-thread TAA attacks. 5553 off - Unconditionally disable TAA mitigation 5554 5555 On MDS-affected machines, tsx_async_abort=off can be 5556 prevented by an active MDS mitigation as both vulnerabilities 5557 are mitigated with the same mechanism so in order to disable 5558 this mitigation, you need to specify mds=off too. 5559 5560 Not specifying this option is equivalent to 5561 tsx_async_abort=full. On CPUs which are MDS affected 5562 and deploy MDS mitigation, TAA mitigation is not 5563 required and doesn't provide any additional 5564 mitigation. 5565 5566 For details see: 5567 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst 5568 5569 turbografx.map[2|3]= [HW,JOY] 5570 TurboGraFX parallel port interface 5571 Format: 5572 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7> 5573 See also Documentation/input/devices/joystick-parport.rst 5574 5575 udbg-immortal [PPC] When debugging early kernel crashes that 5576 happen after console_init() and before a proper 5577 console driver takes over, this boot options might 5578 help "seeing" what's going on. 5579 5580 uhash_entries= [KNL,NET] 5581 Set number of hash buckets for UDP/UDP-Lite connections 5582 5583 uhci-hcd.ignore_oc= 5584 [USB] Ignore overcurrent events (default N). 5585 Some badly-designed motherboards generate lots of 5586 bogus events, for ports that aren't wired to 5587 anything. Set this parameter to avoid log spamming. 5588 Note that genuine overcurrent events won't be 5589 reported either. 5590 5591 unknown_nmi_panic 5592 [X86] Cause panic on unknown NMI. 5593 5594 usbcore.authorized_default= 5595 [USB] Default USB device authorization: 5596 (default -1 = authorized except for wireless USB, 5597 0 = not authorized, 1 = authorized, 2 = authorized 5598 if device connected to internal port) 5599 5600 usbcore.autosuspend= 5601 [USB] The autosuspend time delay (in seconds) used 5602 for newly-detected USB devices (default 2). This 5603 is the time required before an idle device will be 5604 autosuspended. Devices for which the delay is set 5605 to a negative value won't be autosuspended at all. 5606 5607 usbcore.usbfs_snoop= 5608 [USB] Set to log all usbfs traffic (default 0 = off). 5609 5610 usbcore.usbfs_snoop_max= 5611 [USB] Maximum number of bytes to snoop in each URB 5612 (default = 65536). 5613 5614 usbcore.blinkenlights= 5615 [USB] Set to cycle leds on hubs (default 0 = off). 5616 5617 usbcore.old_scheme_first= 5618 [USB] Start with the old device initialization 5619 scheme (default 0 = off). 5620 5621 usbcore.usbfs_memory_mb= 5622 [USB] Memory limit (in MB) for buffers allocated by 5623 usbfs (default = 16, 0 = max = 2047). 5624 5625 usbcore.use_both_schemes= 5626 [USB] Try the other device initialization scheme 5627 if the first one fails (default 1 = enabled). 5628 5629 usbcore.initial_descriptor_timeout= 5630 [USB] Specifies timeout for the initial 64-byte 5631 USB_REQ_GET_DESCRIPTOR request in milliseconds 5632 (default 5000 = 5.0 seconds). 5633 5634 usbcore.nousb [USB] Disable the USB subsystem 5635 5636 usbcore.quirks= 5637 [USB] A list of quirk entries to augment the built-in 5638 usb core quirk list. List entries are separated by 5639 commas. Each entry has the form 5640 VendorID:ProductID:Flags. The IDs are 4-digit hex 5641 numbers and Flags is a set of letters. Each letter 5642 will change the built-in quirk; setting it if it is 5643 clear and clearing it if it is set. The letters have 5644 the following meanings: 5645 a = USB_QUIRK_STRING_FETCH_255 (string 5646 descriptors must not be fetched using 5647 a 255-byte read); 5648 b = USB_QUIRK_RESET_RESUME (device can't resume 5649 correctly so reset it instead); 5650 c = USB_QUIRK_NO_SET_INTF (device can't handle 5651 Set-Interface requests); 5652 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't 5653 handle its Configuration or Interface 5654 strings); 5655 e = USB_QUIRK_RESET (device can't be reset 5656 (e.g morph devices), don't use reset); 5657 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has 5658 more interface descriptions than the 5659 bNumInterfaces count, and can't handle 5660 talking to these interfaces); 5661 g = USB_QUIRK_DELAY_INIT (device needs a pause 5662 during initialization, after we read 5663 the device descriptor); 5664 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For 5665 high speed and super speed interrupt 5666 endpoints, the USB 2.0 and USB 3.0 spec 5667 require the interval in microframes (1 5668 microframe = 125 microseconds) to be 5669 calculated as interval = 2 ^ 5670 (bInterval-1). 5671 Devices with this quirk report their 5672 bInterval as the result of this 5673 calculation instead of the exponent 5674 variable used in the calculation); 5675 i = USB_QUIRK_DEVICE_QUALIFIER (device can't 5676 handle device_qualifier descriptor 5677 requests); 5678 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device 5679 generates spurious wakeup, ignore 5680 remote wakeup capability); 5681 k = USB_QUIRK_NO_LPM (device can't handle Link 5682 Power Management); 5683 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL 5684 (Device reports its bInterval as linear 5685 frames instead of the USB 2.0 5686 calculation); 5687 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs 5688 to be disconnected before suspend to 5689 prevent spurious wakeup); 5690 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a 5691 pause after every control message); 5692 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra 5693 delay after resetting its port); 5694 Example: quirks=0781:5580:bk,0a5c:5834:gij 5695 5696 usbhid.mousepoll= 5697 [USBHID] The interval which mice are to be polled at. 5698 5699 usbhid.jspoll= 5700 [USBHID] The interval which joysticks are to be polled at. 5701 5702 usbhid.kbpoll= 5703 [USBHID] The interval which keyboards are to be polled at. 5704 5705 usb-storage.delay_use= 5706 [UMS] The delay in seconds before a new device is 5707 scanned for Logical Units (default 1). 5708 5709 usb-storage.quirks= 5710 [UMS] A list of quirks entries to supplement or 5711 override the built-in unusual_devs list. List 5712 entries are separated by commas. Each entry has 5713 the form VID:PID:Flags where VID and PID are Vendor 5714 and Product ID values (4-digit hex numbers) and 5715 Flags is a set of characters, each corresponding 5716 to a common usb-storage quirk flag as follows: 5717 a = SANE_SENSE (collect more than 18 bytes 5718 of sense data, not on uas); 5719 b = BAD_SENSE (don't collect more than 18 5720 bytes of sense data, not on uas); 5721 c = FIX_CAPACITY (decrease the reported 5722 device capacity by one sector); 5723 d = NO_READ_DISC_INFO (don't use 5724 READ_DISC_INFO command, not on uas); 5725 e = NO_READ_CAPACITY_16 (don't use 5726 READ_CAPACITY_16 command); 5727 f = NO_REPORT_OPCODES (don't use report opcodes 5728 command, uas only); 5729 g = MAX_SECTORS_240 (don't transfer more than 5730 240 sectors at a time, uas only); 5731 h = CAPACITY_HEURISTICS (decrease the 5732 reported device capacity by one 5733 sector if the number is odd); 5734 i = IGNORE_DEVICE (don't bind to this 5735 device); 5736 j = NO_REPORT_LUNS (don't use report luns 5737 command, uas only); 5738 k = NO_SAME (do not use WRITE_SAME, uas only) 5739 l = NOT_LOCKABLE (don't try to lock and 5740 unlock ejectable media, not on uas); 5741 m = MAX_SECTORS_64 (don't transfer more 5742 than 64 sectors = 32 KB at a time, 5743 not on uas); 5744 n = INITIAL_READ10 (force a retry of the 5745 initial READ(10) command, not on uas); 5746 o = CAPACITY_OK (accept the capacity 5747 reported by the device, not on uas); 5748 p = WRITE_CACHE (the device cache is ON 5749 by default, not on uas); 5750 r = IGNORE_RESIDUE (the device reports 5751 bogus residue values, not on uas); 5752 s = SINGLE_LUN (the device has only one 5753 Logical Unit); 5754 t = NO_ATA_1X (don't allow ATA(12) and ATA(16) 5755 commands, uas only); 5756 u = IGNORE_UAS (don't bind to the uas driver); 5757 w = NO_WP_DETECT (don't test whether the 5758 medium is write-protected). 5759 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE 5760 even if the device claims no cache, 5761 not on uas) 5762 Example: quirks=0419:aaf5:rl,0421:0433:rc 5763 5764 user_debug= [KNL,ARM] 5765 Format: <int> 5766 See arch/arm/Kconfig.debug help text. 5767 1 - undefined instruction events 5768 2 - system calls 5769 4 - invalid data aborts 5770 8 - SIGSEGV faults 5771 16 - SIGBUS faults 5772 Example: user_debug=31 5773 5774 userpte= 5775 [X86] Flags controlling user PTE allocations. 5776 5777 nohigh = do not allocate PTE pages in 5778 HIGHMEM regardless of setting 5779 of CONFIG_HIGHPTE. 5780 5781 vdso= [X86,SH] 5782 On X86_32, this is an alias for vdso32=. Otherwise: 5783 5784 vdso=1: enable VDSO (the default) 5785 vdso=0: disable VDSO mapping 5786 5787 vdso32= [X86] Control the 32-bit vDSO 5788 vdso32=1: enable 32-bit VDSO 5789 vdso32=0 or vdso32=2: disable 32-bit VDSO 5790 5791 See the help text for CONFIG_COMPAT_VDSO for more 5792 details. If CONFIG_COMPAT_VDSO is set, the default is 5793 vdso32=0; otherwise, the default is vdso32=1. 5794 5795 For compatibility with older kernels, vdso32=2 is an 5796 alias for vdso32=0. 5797 5798 Try vdso32=0 if you encounter an error that says: 5799 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed! 5800 5801 vector= [IA-64,SMP] 5802 vector=percpu: enable percpu vector domain 5803 5804 video= [FB] Frame buffer configuration 5805 See Documentation/fb/modedb.rst. 5806 5807 video.brightness_switch_enabled= [0,1] 5808 If set to 1, on receiving an ACPI notify event 5809 generated by hotkey, video driver will adjust brightness 5810 level and then send out the event to user space through 5811 the allocated input device; If set to 0, video driver 5812 will only send out the event without touching backlight 5813 brightness level. 5814 default: 1 5815 5816 virtio_mmio.device= 5817 [VMMIO] Memory mapped virtio (platform) device. 5818 5819 <size>@<baseaddr>:<irq>[:<id>] 5820 where: 5821 <size> := size (can use standard suffixes 5822 like K, M and G) 5823 <baseaddr> := physical base address 5824 <irq> := interrupt number (as passed to 5825 request_irq()) 5826 <id> := (optional) platform device id 5827 example: 5828 virtio_mmio.device=1K@0x100b0000:48:7 5829 5830 Can be used multiple times for multiple devices. 5831 5832 vga= [BOOT,X86-32] Select a particular video mode 5833 See Documentation/x86/boot.rst and 5834 Documentation/admin-guide/svga.rst. 5835 Use vga=ask for menu. 5836 This is actually a boot loader parameter; the value is 5837 passed to the kernel using a special protocol. 5838 5839 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y. 5840 May slow down system boot speed, especially when 5841 enabled on systems with a large amount of memory. 5842 All options are enabled by default, and this 5843 interface is meant to allow for selectively 5844 enabling or disabling specific virtual memory 5845 debugging features. 5846 5847 Available options are: 5848 P Enable page structure init time poisoning 5849 - Disable all of the above options 5850 5851 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact 5852 size of <nn>. This can be used to increase the 5853 minimum size (128MB on x86). It can also be used to 5854 decrease the size and leave more room for directly 5855 mapped kernel RAM. 5856 5857 vmcp_cma=nn[MG] [KNL,S390] 5858 Sets the memory size reserved for contiguous memory 5859 allocations for the vmcp device driver. 5860 5861 vmhalt= [KNL,S390] Perform z/VM CP command after system halt. 5862 Format: <command> 5863 5864 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic. 5865 Format: <command> 5866 5867 vmpoff= [KNL,S390] Perform z/VM CP command after power off. 5868 Format: <command> 5869 5870 vsyscall= [X86-64] 5871 Controls the behavior of vsyscalls (i.e. calls to 5872 fixed addresses of 0xffffffffff600x00 from legacy 5873 code). Most statically-linked binaries and older 5874 versions of glibc use these calls. Because these 5875 functions are at fixed addresses, they make nice 5876 targets for exploits that can control RIP. 5877 5878 emulate [default] Vsyscalls turn into traps and are 5879 emulated reasonably safely. The vsyscall 5880 page is readable. 5881 5882 xonly Vsyscalls turn into traps and are 5883 emulated reasonably safely. The vsyscall 5884 page is not readable. 5885 5886 none Vsyscalls don't work at all. This makes 5887 them quite hard to use for exploits but 5888 might break your system. 5889 5890 vt.color= [VT] Default text color. 5891 Format: 0xYX, X = foreground, Y = background. 5892 Default: 0x07 = light gray on black. 5893 5894 vt.cur_default= [VT] Default cursor shape. 5895 Format: 0xCCBBAA, where AA, BB, and CC are the same as 5896 the parameters of the <Esc>[?A;B;Cc escape sequence; 5897 see VGA-softcursor.txt. Default: 2 = underline. 5898 5899 vt.default_blu= [VT] 5900 Format: <blue0>,<blue1>,<blue2>,...,<blue15> 5901 Change the default blue palette of the console. 5902 This is a 16-member array composed of values 5903 ranging from 0-255. 5904 5905 vt.default_grn= [VT] 5906 Format: <green0>,<green1>,<green2>,...,<green15> 5907 Change the default green palette of the console. 5908 This is a 16-member array composed of values 5909 ranging from 0-255. 5910 5911 vt.default_red= [VT] 5912 Format: <red0>,<red1>,<red2>,...,<red15> 5913 Change the default red palette of the console. 5914 This is a 16-member array composed of values 5915 ranging from 0-255. 5916 5917 vt.default_utf8= 5918 [VT] 5919 Format=<0|1> 5920 Set system-wide default UTF-8 mode for all tty's. 5921 Default is 1, i.e. UTF-8 mode is enabled for all 5922 newly opened terminals. 5923 5924 vt.global_cursor_default= 5925 [VT] 5926 Format=<-1|0|1> 5927 Set system-wide default for whether a cursor 5928 is shown on new VTs. Default is -1, 5929 i.e. cursors will be created by default unless 5930 overridden by individual drivers. 0 will hide 5931 cursors, 1 will display them. 5932 5933 vt.italic= [VT] Default color for italic text; 0-15. 5934 Default: 2 = green. 5935 5936 vt.underline= [VT] Default color for underlined text; 0-15. 5937 Default: 3 = cyan. 5938 5939 watchdog timers [HW,WDT] For information on watchdog timers, 5940 see Documentation/watchdog/watchdog-parameters.rst 5941 or other driver-specific files in the 5942 Documentation/watchdog/ directory. 5943 5944 watchdog_thresh= 5945 [KNL] 5946 Set the hard lockup detector stall duration 5947 threshold in seconds. The soft lockup detector 5948 threshold is set to twice the value. A value of 0 5949 disables both lockup detectors. Default is 10 5950 seconds. 5951 5952 workqueue.watchdog_thresh= 5953 If CONFIG_WQ_WATCHDOG is configured, workqueue can 5954 warn stall conditions and dump internal state to 5955 help debugging. 0 disables workqueue stall 5956 detection; otherwise, it's the stall threshold 5957 duration in seconds. The default value is 30 and 5958 it can be updated at runtime by writing to the 5959 corresponding sysfs file. 5960 5961 workqueue.disable_numa 5962 By default, all work items queued to unbound 5963 workqueues are affine to the NUMA nodes they're 5964 issued on, which results in better behavior in 5965 general. If NUMA affinity needs to be disabled for 5966 whatever reason, this option can be used. Note 5967 that this also can be controlled per-workqueue for 5968 workqueues visible under /sys/bus/workqueue/. 5969 5970 workqueue.power_efficient 5971 Per-cpu workqueues are generally preferred because 5972 they show better performance thanks to cache 5973 locality; unfortunately, per-cpu workqueues tend to 5974 be more power hungry than unbound workqueues. 5975 5976 Enabling this makes the per-cpu workqueues which 5977 were observed to contribute significantly to power 5978 consumption unbound, leading to measurably lower 5979 power usage at the cost of small performance 5980 overhead. 5981 5982 The default value of this parameter is determined by 5983 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. 5984 5985 workqueue.debug_force_rr_cpu 5986 Workqueue used to implicitly guarantee that work 5987 items queued without explicit CPU specified are put 5988 on the local CPU. This guarantee is no longer true 5989 and while local CPU is still preferred work items 5990 may be put on foreign CPUs. This debug option 5991 forces round-robin CPU selection to flush out 5992 usages which depend on the now broken guarantee. 5993 When enabled, memory and cache locality will be 5994 impacted. 5995 5996 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of 5997 default x2apic cluster mode on platforms 5998 supporting x2apic. 5999 6000 x86_intel_mid_timer= [X86-32,APBT] 6001 Choose timer option for x86 Intel MID platform. 6002 Two valid options are apbt timer only and lapic timer 6003 plus one apbt timer for broadcast timer. 6004 x86_intel_mid_timer=apbt_only | lapic_and_apbt 6005 6006 xen_512gb_limit [KNL,X86-64,XEN] 6007 Restricts the kernel running paravirtualized under Xen 6008 to use only up to 512 GB of RAM. The reason to do so is 6009 crash analysis tools and Xen tools for doing domain 6010 save/restore/migration must be enabled to handle larger 6011 domains. 6012 6013 xen_emul_unplug= [HW,X86,XEN] 6014 Unplug Xen emulated devices 6015 Format: [unplug0,][unplug1] 6016 ide-disks -- unplug primary master IDE devices 6017 aux-ide-disks -- unplug non-primary-master IDE devices 6018 nics -- unplug network devices 6019 all -- unplug all emulated devices (NICs and IDE disks) 6020 unnecessary -- unplugging emulated devices is 6021 unnecessary even if the host did not respond to 6022 the unplug protocol 6023 never -- do not unplug even if version check succeeds 6024 6025 xen_legacy_crash [X86,XEN] 6026 Crash from Xen panic notifier, without executing late 6027 panic() code such as dumping handler. 6028 6029 xen_nopvspin [X86,XEN] 6030 Disables the qspinlock slowpath using Xen PV optimizations. 6031 This parameter is obsoleted by "nopvspin" parameter, which 6032 has equivalent effect for XEN platform. 6033 6034 xen_nopv [X86] 6035 Disables the PV optimizations forcing the HVM guest to 6036 run as generic HVM guest with no PV drivers. 6037 This option is obsoleted by the "nopv" option, which 6038 has equivalent effect for XEN platform. 6039 6040 xen_no_vector_callback 6041 [KNL,X86,XEN] Disable the vector callback for Xen 6042 event channel interrupts. 6043 6044 xen_scrub_pages= [XEN] 6045 Boolean option to control scrubbing pages before giving them back 6046 to Xen, for use by other domains. Can be also changed at runtime 6047 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages. 6048 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT. 6049 6050 xen_timer_slop= [X86-64,XEN] 6051 Set the timer slop (in nanoseconds) for the virtual Xen 6052 timers (default is 100000). This adjusts the minimum 6053 delta of virtualized Xen timers, where lower values 6054 improve timer resolution at the expense of processing 6055 more timer interrupts. 6056 6057 xen.event_eoi_delay= [XEN] 6058 How long to delay EOI handling in case of event 6059 storms (jiffies). Default is 10. 6060 6061 xen.event_loop_timeout= [XEN] 6062 After which time (jiffies) the event handling loop 6063 should start to delay EOI handling. Default is 2. 6064 6065 xen.fifo_events= [XEN] 6066 Boolean parameter to disable using fifo event handling 6067 even if available. Normally fifo event handling is 6068 preferred over the 2-level event handling, as it is 6069 fairer and the number of possible event channels is 6070 much higher. Default is on (use fifo events). 6071 6072 nopv= [X86,XEN,KVM,HYPER_V,VMWARE] 6073 Disables the PV optimizations forcing the guest to run 6074 as generic guest with no PV drivers. Currently support 6075 XEN HVM, KVM, HYPER_V and VMWARE guest. 6076 6077 nopvspin [X86,XEN,KVM] 6078 Disables the qspinlock slow path using PV optimizations 6079 which allow the hypervisor to 'idle' the guest on lock 6080 contention. 6081 6082 xirc2ps_cs= [NET,PCMCIA] 6083 Format: 6084 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] 6085 6086 xive= [PPC] 6087 By default on POWER9 and above, the kernel will 6088 natively use the XIVE interrupt controller. This option 6089 allows the fallback firmware mode to be used: 6090 6091 off Fallback to firmware control of XIVE interrupt 6092 controller on both pseries and powernv 6093 platforms. Only useful on POWER9 and above. 6094 6095 xhci-hcd.quirks [USB,KNL] 6096 A hex value specifying bitmask with supplemental xhci 6097 host controller quirks. Meaning of each bit can be 6098 consulted in header drivers/usb/host/xhci.h. 6099 6100 xmon [PPC] 6101 Format: { early | on | rw | ro | off } 6102 Controls if xmon debugger is enabled. Default is off. 6103 Passing only "xmon" is equivalent to "xmon=early". 6104 early Call xmon as early as possible on boot; xmon 6105 debugger is called from setup_arch(). 6106 on xmon debugger hooks will be installed so xmon 6107 is only called on a kernel crash. Default mode, 6108 i.e. either "ro" or "rw" mode, is controlled 6109 with CONFIG_XMON_DEFAULT_RO_MODE. 6110 rw xmon debugger hooks will be installed so xmon 6111 is called only on a kernel crash, mode is write, 6112 meaning SPR registers, memory and, other data 6113 can be written using xmon commands. 6114 ro same as "rw" option above but SPR registers, 6115 memory, and other data can't be written using 6116 xmon commands. 6117 off xmon is disabled. 6118