1config PAGE_EXTENSION 2 bool "Extend memmap on extra space for more information on page" 3 ---help--- 4 Extend memmap on extra space for more information on page. This 5 could be used for debugging features that need to insert extra 6 field for every page. This extension enables us to save memory 7 by not allocating this extra memory according to boottime 8 configuration. 9 10config DEBUG_PAGEALLOC 11 bool "Debug page memory allocations" 12 depends on DEBUG_KERNEL 13 depends on !HIBERNATION || ARCH_SUPPORTS_DEBUG_PAGEALLOC && !PPC && !SPARC 14 select PAGE_EXTENSION 15 select PAGE_POISONING if !ARCH_SUPPORTS_DEBUG_PAGEALLOC 16 ---help--- 17 Unmap pages from the kernel linear mapping after free_pages(). 18 Depending on runtime enablement, this results in a small or large 19 slowdown, but helps to find certain types of memory corruption. 20 21 For architectures which don't enable ARCH_SUPPORTS_DEBUG_PAGEALLOC, 22 fill the pages with poison patterns after free_pages() and verify 23 the patterns before alloc_pages(). Additionally, 24 this option cannot be enabled in combination with hibernation as 25 that would result in incorrect warnings of memory corruption after 26 a resume because free pages are not saved to the suspend image. 27 28 By default this option will have a small overhead, e.g. by not 29 allowing the kernel mapping to be backed by large pages on some 30 architectures. Even bigger overhead comes when the debugging is 31 enabled by DEBUG_PAGEALLOC_ENABLE_DEFAULT or the debug_pagealloc 32 command line parameter. 33 34config DEBUG_PAGEALLOC_ENABLE_DEFAULT 35 bool "Enable debug page memory allocations by default?" 36 default n 37 depends on DEBUG_PAGEALLOC 38 ---help--- 39 Enable debug page memory allocations by default? This value 40 can be overridden by debug_pagealloc=off|on. 41 42config PAGE_OWNER 43 bool "Track page owner" 44 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT 45 select DEBUG_FS 46 select STACKTRACE 47 select STACKDEPOT 48 select PAGE_EXTENSION 49 help 50 This keeps track of what call chain is the owner of a page, may 51 help to find bare alloc_page(s) leaks. Even if you include this 52 feature on your build, it is disabled in default. You should pass 53 "page_owner=on" to boot parameter in order to enable it. Eats 54 a fair amount of memory if enabled. See tools/vm/page_owner_sort.c 55 for user-space helper. 56 57 If unsure, say N. 58 59config PAGE_POISONING 60 bool "Poison pages after freeing" 61 select PAGE_POISONING_NO_SANITY if HIBERNATION 62 ---help--- 63 Fill the pages with poison patterns after free_pages() and verify 64 the patterns before alloc_pages. The filling of the memory helps 65 reduce the risk of information leaks from freed data. This does 66 have a potential performance impact if enabled with the 67 "page_poison=1" kernel boot option. 68 69 Note that "poison" here is not the same thing as the "HWPoison" 70 for CONFIG_MEMORY_FAILURE. This is software poisoning only. 71 72 If unsure, say N 73 74config PAGE_POISONING_NO_SANITY 75 depends on PAGE_POISONING 76 bool "Only poison, don't sanity check" 77 ---help--- 78 Skip the sanity checking on alloc, only fill the pages with 79 poison on free. This reduces some of the overhead of the 80 poisoning feature. 81 82 If you are only interested in sanitization, say Y. Otherwise 83 say N. 84 85config PAGE_POISONING_ZERO 86 bool "Use zero for poisoning instead of debugging value" 87 depends on PAGE_POISONING 88 ---help--- 89 Instead of using the existing poison value, fill the pages with 90 zeros. This makes it harder to detect when errors are occurring 91 due to sanitization but the zeroing at free means that it is 92 no longer necessary to write zeros when GFP_ZERO is used on 93 allocation. 94 95 If unsure, say N 96 97config DEBUG_PAGE_REF 98 bool "Enable tracepoint to track down page reference manipulation" 99 depends on DEBUG_KERNEL 100 depends on TRACEPOINTS 101 ---help--- 102 This is a feature to add tracepoint for tracking down page reference 103 manipulation. This tracking is useful to diagnose functional failure 104 due to migration failures caused by page reference mismatches. Be 105 careful when enabling this feature because it adds about 30 KB to the 106 kernel code. However the runtime performance overhead is virtually 107 nil until the tracepoints are actually enabled. 108 109config DEBUG_RODATA_TEST 110 bool "Testcase for the marking rodata read-only" 111 depends on STRICT_KERNEL_RWX 112 ---help--- 113 This option enables a testcase for the setting rodata read-only. 114