1menu "Kernel hacking" 2 3source "lib/Kconfig.debug" 4 5config DEBUG_STACKOVERFLOW 6 bool "Check for stack overflows" 7 depends on DEBUG_KERNEL 8 help 9 This option will cause messages to be printed if free stack space 10 drops below a certain limit. 11 12config DEBUG_VERBOSE 13 bool "Verbose fault messages" 14 default y 15 select PRINTK 16 help 17 When a program crashes due to an exception, or the kernel detects 18 an internal error, the kernel can print a not so brief message 19 explaining what the problem was. This debugging information is 20 useful to developers and kernel hackers when tracking down problems, 21 but mostly meaningless to other people. This is always helpful for 22 debugging but serves no purpose on a production system. 23 Most people should say N here. 24 25config DEBUG_MMRS 26 tristate "Generate Blackfin MMR tree" 27 select DEBUG_FS 28 help 29 Create a tree of Blackfin MMRs via the debugfs tree. If 30 you enable this, you will find all MMRs laid out in the 31 /sys/kernel/debug/blackfin/ directory where you can read/write 32 MMRs directly from userspace. This is obviously just a debug 33 feature. 34 35config DEBUG_HWERR 36 bool "Hardware error interrupt debugging" 37 depends on DEBUG_KERNEL 38 help 39 When enabled, the hardware error interrupt is never disabled, and 40 will happen immediately when an error condition occurs. This comes 41 at a slight cost in code size, but is necessary if you are getting 42 hardware error interrupts and need to know where they are coming 43 from. 44 45config EXACT_HWERR 46 bool "Try to make Hardware errors exact" 47 depends on DEBUG_HWERR 48 help 49 By default, the Blackfin hardware errors are not exact - the error 50 be reported multiple cycles after the error happens. This delay 51 can cause the wrong application, or even the kernel to receive a 52 signal to be killed. If you are getting HW errors in your system, 53 try turning this on to ensure they are at least coming from the 54 proper thread. 55 56 On production systems, it is safe (and a small optimization) to say N. 57 58config DEBUG_DOUBLEFAULT 59 bool "Debug Double Faults" 60 default n 61 help 62 If an exception is caused while executing code within the exception 63 handler, the NMI handler, the reset vector, or in emulator mode, 64 a double fault occurs. On the Blackfin, this is a unrecoverable 65 event. You have two options: 66 - RESET exactly when double fault occurs. The excepting 67 instruction address is stored in RETX, where the next kernel 68 boot will print it out. 69 - Print debug message. This is much more error prone, although 70 easier to handle. It is error prone since: 71 - The excepting instruction is not committed. 72 - All writebacks from the instruction are prevented. 73 - The generated exception is not taken. 74 - The EXCAUSE field is updated with an unrecoverable event 75 The only way to check this is to see if EXCAUSE contains the 76 unrecoverable event value at every exception return. By selecting 77 this option, you are skipping over the faulting instruction, and 78 hoping things stay together enough to print out a debug message. 79 80 This does add a little kernel code, but is the only method to debug 81 double faults - if unsure say "Y" 82 83choice 84 prompt "Double Fault Failure Method" 85 default DEBUG_DOUBLEFAULT_PRINT 86 depends on DEBUG_DOUBLEFAULT 87 88config DEBUG_DOUBLEFAULT_PRINT 89 bool "Print" 90 91config DEBUG_DOUBLEFAULT_RESET 92 bool "Reset" 93 94endchoice 95 96config DEBUG_HUNT_FOR_ZERO 97 bool "Catch NULL pointer reads/writes" 98 default y 99 help 100 Say Y here to catch reads/writes to anywhere in the memory range 101 from 0x0000 - 0x0FFF (the first 4k) of memory. This is useful in 102 catching common programming errors such as NULL pointer dereferences. 103 104 Misbehaving applications will be killed (generate a SEGV) while the 105 kernel will trigger a panic. 106 107 Enabling this option will take up an extra entry in CPLB table. 108 Otherwise, there is no extra overhead. 109 110config DEBUG_BFIN_HWTRACE_ON 111 bool "Turn on Blackfin's Hardware Trace" 112 default y 113 help 114 All Blackfins include a Trace Unit which stores a history of the last 115 16 changes in program flow taken by the program sequencer. The history 116 allows the user to recreate the program sequencer’s recent path. This 117 can be handy when an application dies - we print out the execution 118 path of how it got to the offending instruction. 119 120 By turning this off, you may save a tiny amount of power. 121 122choice 123 prompt "Omit loop Tracing" 124 default DEBUG_BFIN_HWTRACE_COMPRESSION_OFF 125 depends on DEBUG_BFIN_HWTRACE_ON 126 help 127 The trace buffer can be configured to omit recording of changes in 128 program flow that match either the last entry or one of the last 129 two entries. Omitting one of these entries from the record prevents 130 the trace buffer from overflowing because of any sort of loop (for, do 131 while, etc) in the program. 132 133 Because zero-overhead Hardware loops are not recorded in the trace buffer, 134 this feature can be used to prevent trace overflow from loops that 135 are nested four deep. 136 137config DEBUG_BFIN_HWTRACE_COMPRESSION_OFF 138 bool "Trace all Loops" 139 help 140 The trace buffer records all changes of flow 141 142config DEBUG_BFIN_HWTRACE_COMPRESSION_ONE 143 bool "Compress single-level loops" 144 help 145 The trace buffer does not record single loops - helpful if trace 146 is spinning on a while or do loop. 147 148config DEBUG_BFIN_HWTRACE_COMPRESSION_TWO 149 bool "Compress two-level loops" 150 help 151 The trace buffer does not record loops two levels deep. Helpful if 152 the trace is spinning in a nested loop 153 154endchoice 155 156config DEBUG_BFIN_HWTRACE_COMPRESSION 157 int 158 depends on DEBUG_BFIN_HWTRACE_ON 159 default 0 if DEBUG_BFIN_HWTRACE_COMPRESSION_OFF 160 default 1 if DEBUG_BFIN_HWTRACE_COMPRESSION_ONE 161 default 2 if DEBUG_BFIN_HWTRACE_COMPRESSION_TWO 162 163 164config DEBUG_BFIN_HWTRACE_EXPAND 165 bool "Expand Trace Buffer greater than 16 entries" 166 depends on DEBUG_BFIN_HWTRACE_ON 167 default n 168 help 169 By selecting this option, every time the 16 hardware entries in 170 the Blackfin's HW Trace buffer are full, the kernel will move them 171 into a software buffer, for dumping when there is an issue. This 172 has a great impact on performance, (an interrupt every 16 change of 173 flows) and should normally be turned off, except in those nasty 174 debugging sessions 175 176config DEBUG_BFIN_HWTRACE_EXPAND_LEN 177 int "Size of Trace buffer (in power of 2k)" 178 range 0 4 179 depends on DEBUG_BFIN_HWTRACE_EXPAND 180 default 1 181 help 182 This sets the size of the software buffer that the trace information 183 is kept in. 184 0 for (2^0) 1k, or 256 entries, 185 1 for (2^1) 2k, or 512 entries, 186 2 for (2^2) 4k, or 1024 entries, 187 3 for (2^3) 8k, or 2048 entries, 188 4 for (2^4) 16k, or 4096 entries 189 190config DEBUG_BFIN_NO_KERN_HWTRACE 191 bool "Turn off hwtrace in CPLB handlers" 192 depends on DEBUG_BFIN_HWTRACE_ON 193 default y 194 help 195 The CPLB error handler contains a lot of flow changes which can 196 quickly fill up the hardware trace buffer. When debugging crashes, 197 the hardware trace may indicate that the problem lies in kernel 198 space when in reality an application is buggy. 199 200 Say Y here to disable hardware tracing in some known "jumpy" pieces 201 of code so that the trace buffer will extend further back. 202 203config EARLY_PRINTK 204 bool "Early printk" 205 default n 206 select SERIAL_CORE_CONSOLE 207 help 208 This option enables special console drivers which allow the kernel 209 to print messages very early in the bootup process. 210 211 This is useful for kernel debugging when your machine crashes very 212 early before the console code is initialized. After enabling this 213 feature, you must add "earlyprintk=serial,uart0,57600" to the 214 command line (bootargs). It is safe to say Y here in all cases, as 215 all of this lives in the init section and is thrown away after the 216 kernel boots completely. 217 218config NMI_WATCHDOG 219 bool "Enable NMI watchdog to help debugging lockup on SMP" 220 default n 221 depends on SMP 222 help 223 If any CPU in the system does not execute the period local timer 224 interrupt for more than 5 seconds, then the NMI handler dumps debug 225 information. This information can be used to debug the lockup. 226 227config CPLB_INFO 228 bool "Display the CPLB information" 229 help 230 Display the CPLB information via /proc/cplbinfo. 231 232config ACCESS_CHECK 233 bool "Check the user pointer address" 234 default y 235 help 236 Usually the pointer transfer from user space is checked to see if its 237 address is in the kernel space. 238 239 Say N here to disable that check to improve the performance. 240 241config BFIN_ISRAM_SELF_TEST 242 bool "isram boot self tests" 243 default n 244 help 245 Run some self tests of the isram driver code at boot. 246 247config BFIN_PSEUDODBG_INSNS 248 bool "Support pseudo debug instructions" 249 default n 250 help 251 This option allows the kernel to emulate some pseudo instructions which 252 allow simulator test cases to be run under Linux with no changes. 253 254 Most people should say N here. 255 256endmenu 257