Kconfig.debug 8.1 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284
  1. menu "Kernel hacking"
  2. config TRACE_IRQFLAGS_SUPPORT
  3. def_bool y
  4. source "lib/Kconfig.debug"
  5. config NONPROMISC_DEVMEM
  6. bool "Filter access to /dev/mem"
  7. help
  8. If this option is left off, you allow userspace access to all
  9. of memory, including kernel and userspace memory. Accidental
  10. access to this is obviously disastrous, but specific access can
  11. be used by people debugging the kernel.
  12. If this option is switched on, the /dev/mem file only allows
  13. userspace access to PCI space and the BIOS code and data regions.
  14. This is sufficient for dosemu and X and all common users of
  15. /dev/mem.
  16. If in doubt, say Y.
  17. config EARLY_PRINTK
  18. bool "Early printk" if EMBEDDED
  19. default y
  20. help
  21. Write kernel log output directly into the VGA buffer or to a serial
  22. port.
  23. This is useful for kernel debugging when your machine crashes very
  24. early before the console code is initialized. For normal operation
  25. it is not recommended because it looks ugly and doesn't cooperate
  26. with klogd/syslogd or the X server. You should normally N here,
  27. unless you want to debug such a crash.
  28. config DEBUG_STACKOVERFLOW
  29. bool "Check for stack overflows"
  30. depends on DEBUG_KERNEL
  31. help
  32. This option will cause messages to be printed if free stack space
  33. drops below a certain limit.
  34. config DEBUG_STACK_USAGE
  35. bool "Stack utilization instrumentation"
  36. depends on DEBUG_KERNEL
  37. help
  38. Enables the display of the minimum amount of free stack which each
  39. task has ever had available in the sysrq-T and sysrq-P debug output.
  40. This option will slow down process creation somewhat.
  41. config DEBUG_PAGEALLOC
  42. bool "Debug page memory allocations"
  43. depends on DEBUG_KERNEL
  44. help
  45. Unmap pages from the kernel linear mapping after free_pages().
  46. This results in a large slowdown, but helps to find certain types
  47. of memory corruptions.
  48. config DEBUG_PER_CPU_MAPS
  49. bool "Debug access to per_cpu maps"
  50. depends on DEBUG_KERNEL
  51. depends on X86_64_SMP
  52. default n
  53. help
  54. Say Y to verify that the per_cpu map being accessed has
  55. been setup. Adds a fair amount of code to kernel memory
  56. and decreases performance.
  57. Say N if unsure.
  58. config X86_PTDUMP
  59. bool "Export kernel pagetable layout to userspace via debugfs"
  60. depends on DEBUG_KERNEL
  61. select DEBUG_FS
  62. help
  63. Say Y here if you want to show the kernel pagetable layout in a
  64. debugfs file. This information is only useful for kernel developers
  65. who are working in architecture specific areas of the kernel.
  66. It is probably not a good idea to enable this feature in a production
  67. kernel.
  68. If in doubt, say "N"
  69. config DEBUG_RODATA
  70. bool "Write protect kernel read-only data structures"
  71. default y
  72. depends on DEBUG_KERNEL
  73. help
  74. Mark the kernel read-only data as write-protected in the pagetables,
  75. in order to catch accidental (and incorrect) writes to such const
  76. data. This is recommended so that we can catch kernel bugs sooner.
  77. If in doubt, say "Y".
  78. config DIRECT_GBPAGES
  79. bool "Enable gbpages-mapped kernel pagetables"
  80. depends on DEBUG_KERNEL && EXPERIMENTAL && X86_64
  81. help
  82. Enable gigabyte pages support (if the CPU supports it). This can
  83. improve the kernel's performance a tiny bit by reducing TLB
  84. pressure.
  85. This is experimental code.
  86. If in doubt, say "N".
  87. config DEBUG_RODATA_TEST
  88. bool "Testcase for the DEBUG_RODATA feature"
  89. depends on DEBUG_RODATA
  90. help
  91. This option enables a testcase for the DEBUG_RODATA
  92. feature as well as for the change_page_attr() infrastructure.
  93. If in doubt, say "N"
  94. config DEBUG_NX_TEST
  95. tristate "Testcase for the NX non-executable stack feature"
  96. depends on DEBUG_KERNEL && m
  97. help
  98. This option enables a testcase for the CPU NX capability
  99. and the software setup of this feature.
  100. If in doubt, say "N"
  101. config 4KSTACKS
  102. bool "Use 4Kb for kernel stacks instead of 8Kb"
  103. depends on X86_32
  104. help
  105. If you say Y here the kernel will use a 4Kb stacksize for the
  106. kernel stack attached to each process/thread. This facilitates
  107. running more threads on a system and also reduces the pressure
  108. on the VM subsystem for higher order allocations. This option
  109. will also use IRQ stacks to compensate for the reduced stackspace.
  110. config X86_FIND_SMP_CONFIG
  111. def_bool y
  112. depends on X86_LOCAL_APIC || X86_VOYAGER
  113. depends on X86_32
  114. config X86_MPPARSE
  115. def_bool y
  116. depends on (X86_32 && (X86_LOCAL_APIC && !X86_VISWS)) || X86_64
  117. config DOUBLEFAULT
  118. default y
  119. bool "Enable doublefault exception handler" if EMBEDDED
  120. depends on X86_32
  121. help
  122. This option allows trapping of rare doublefault exceptions that
  123. would otherwise cause a system to silently reboot. Disabling this
  124. option saves about 4k and might cause you much additional grey
  125. hair.
  126. config IOMMU_DEBUG
  127. bool "Enable IOMMU debugging"
  128. depends on GART_IOMMU && DEBUG_KERNEL
  129. depends on X86_64
  130. help
  131. Force the IOMMU to on even when you have less than 4GB of
  132. memory and add debugging code. On overflow always panic. And
  133. allow to enable IOMMU leak tracing. Can be disabled at boot
  134. time with iommu=noforce. This will also enable scatter gather
  135. list merging. Currently not recommended for production
  136. code. When you use it make sure you have a big enough
  137. IOMMU/AGP aperture. Most of the options enabled by this can
  138. be set more finegrained using the iommu= command line
  139. options. See Documentation/x86_64/boot-options.txt for more
  140. details.
  141. config IOMMU_LEAK
  142. bool "IOMMU leak tracing"
  143. depends on DEBUG_KERNEL
  144. depends on IOMMU_DEBUG
  145. help
  146. Add a simple leak tracer to the IOMMU code. This is useful when you
  147. are debugging a buggy device driver that leaks IOMMU mappings.
  148. config PAGE_FAULT_HANDLERS
  149. bool "Custom page fault handlers"
  150. depends on DEBUG_KERNEL
  151. help
  152. Allow the use of custom page fault handlers. A kernel module may
  153. register a function that is called on every page fault. Custom
  154. handlers are used by some debugging and reverse engineering tools.
  155. #
  156. # IO delay types:
  157. #
  158. config IO_DELAY_TYPE_0X80
  159. int
  160. default "0"
  161. config IO_DELAY_TYPE_0XED
  162. int
  163. default "1"
  164. config IO_DELAY_TYPE_UDELAY
  165. int
  166. default "2"
  167. config IO_DELAY_TYPE_NONE
  168. int
  169. default "3"
  170. choice
  171. prompt "IO delay type"
  172. default IO_DELAY_0X80
  173. config IO_DELAY_0X80
  174. bool "port 0x80 based port-IO delay [recommended]"
  175. help
  176. This is the traditional Linux IO delay used for in/out_p.
  177. It is the most tested hence safest selection here.
  178. config IO_DELAY_0XED
  179. bool "port 0xed based port-IO delay"
  180. help
  181. Use port 0xed as the IO delay. This frees up port 0x80 which is
  182. often used as a hardware-debug port.
  183. config IO_DELAY_UDELAY
  184. bool "udelay based port-IO delay"
  185. help
  186. Use udelay(2) as the IO delay method. This provides the delay
  187. while not having any side-effect on the IO port space.
  188. config IO_DELAY_NONE
  189. bool "no port-IO delay"
  190. help
  191. No port-IO delay. Will break on old boxes that require port-IO
  192. delay for certain operations. Should work on most new machines.
  193. endchoice
  194. if IO_DELAY_0X80
  195. config DEFAULT_IO_DELAY_TYPE
  196. int
  197. default IO_DELAY_TYPE_0X80
  198. endif
  199. if IO_DELAY_0XED
  200. config DEFAULT_IO_DELAY_TYPE
  201. int
  202. default IO_DELAY_TYPE_0XED
  203. endif
  204. if IO_DELAY_UDELAY
  205. config DEFAULT_IO_DELAY_TYPE
  206. int
  207. default IO_DELAY_TYPE_UDELAY
  208. endif
  209. if IO_DELAY_NONE
  210. config DEFAULT_IO_DELAY_TYPE
  211. int
  212. default IO_DELAY_TYPE_NONE
  213. endif
  214. config DEBUG_BOOT_PARAMS
  215. bool "Debug boot parameters"
  216. depends on DEBUG_KERNEL
  217. depends on DEBUG_FS
  218. help
  219. This option will cause struct boot_params to be exported via debugfs.
  220. config CPA_DEBUG
  221. bool "CPA self-test code"
  222. depends on DEBUG_KERNEL
  223. help
  224. Do change_page_attr() self-tests every 30 seconds.
  225. config OPTIMIZE_INLINING
  226. bool "Allow gcc to uninline functions marked 'inline'"
  227. depends on BROKEN
  228. help
  229. This option determines if the kernel forces gcc to inline the functions
  230. developers have marked 'inline'. Doing so takes away freedom from gcc to
  231. do what it thinks is best, which is desirable for the gcc 3.x series of
  232. compilers. The gcc 4.x series have a rewritten inlining algorithm and
  233. disabling this option will generate a smaller kernel there. Hopefully
  234. this algorithm is so good that allowing gcc4 to make the decision can
  235. become the default in the future, until then this option is there to
  236. test gcc for this.
  237. endmenu