1sphinx.addnodesdocument)}( rawsourcechildren]( translations LanguagesNode)}(hhh](h pending_xref)}(hhh]docutils.nodesTextChinese (Simplified)}parenthsba attributes}(ids]classes]names]dupnames]backrefs] refdomainstdreftypedoc reftarget'/translations/zh_CN/admin-guide/mm/slabmodnameN classnameN refexplicitutagnamehhh ubh)}(hhh]hChinese (Traditional)}hh2sbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/zh_TW/admin-guide/mm/slabmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hItalian}hhFsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/it_IT/admin-guide/mm/slabmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hJapanese}hhZsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/ja_JP/admin-guide/mm/slabmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hKorean}hhnsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/ko_KR/admin-guide/mm/slabmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hPortuguese (Brazilian)}hhsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/pt_BR/admin-guide/mm/slabmodnameN classnameN refexplicituh1hhh ubh)}(hhh]hSpanish}hhsbah}(h]h ]h"]h$]h&] refdomainh)reftypeh+ reftarget'/translations/sp_SP/admin-guide/mm/slabmodnameN classnameN refexplicituh1hhh ubeh}(h]h ]h"]h$]h&]current_languageEnglishuh1h hh _documenthsourceNlineNubhsection)}(hhh](htitle)}(h(Short users guide for the slab allocatorh]h(Short users guide for the slab allocator}(hhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhhhA/var/lib/git/docbuild/linux/Documentation/admin-guide/mm/slab.rsthKubh paragraph)}(hX6The slab allocator includes full debugging support (when built with CONFIG_SLUB_DEBUG=y) but it is off by default (unless built with CONFIG_SLUB_DEBUG_ON=y). You can enable debugging only for selected slabs in order to avoid an impact on overall system performance which may make a bug more difficult to find.h]hX6The slab allocator includes full debugging support (when built with CONFIG_SLUB_DEBUG=y) but it is off by default (unless built with CONFIG_SLUB_DEBUG_ON=y). You can enable debugging only for selected slabs in order to avoid an impact on overall system performance which may make a bug more difficult to find.}(hhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhhhhubh)}(hIn order to switch debugging on one can add an option ``slab_debug`` to the kernel command line. That will enable full debugging for all slabs.h](h6In order to switch debugging on one can add an option }(hhhhhNhNubhliteral)}(h``slab_debug``h]h slab_debug}(hhhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhubhK to the kernel command line. That will enable full debugging for all slabs.}(hhhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhK hhhhubh)}(hXTypically one would then use the ``slabinfo`` command to get statistical data and perform operation on the slabs. By default ``slabinfo`` only lists slabs that have data in them. See "slabinfo -h" for more options when running the command. ``slabinfo`` can be compiled with ::h](h!Typically one would then use the }(hhhhhNhNubh)}(h ``slabinfo``h]hslabinfo}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhubhP command to get statistical data and perform operation on the slabs. By default }(hhhhhNhNubh)}(h ``slabinfo``h]hslabinfo}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhubhk only lists slabs that have data in them. See “slabinfo -h” for more options when running the command. }(hhhhhNhNubh)}(h ``slabinfo``h]hslabinfo}(hj)hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhubh can be compiled with}(hhhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhhhhubh literal_block)}(h#gcc -o slabinfo tools/mm/slabinfo.ch]h#gcc -o slabinfo tools/mm/slabinfo.c}hjCsbah}(h]h ]h"]h$]h&] xml:spacepreserveuh1jAhhhKhhhhubh)}(hSome of the modes of operation of ``slabinfo`` require that slub debugging be enabled on the command line. F.e. no tracking information will be available without debugging on and validation can only partially be performed if debugging was not switched on.h](h"Some of the modes of operation of }(hjShhhNhNubh)}(h ``slabinfo``h]hslabinfo}(hj[hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjSubh require that slub debugging be enabled on the command line. F.e. no tracking information will be available without debugging on and validation can only partially be performed if debugging was not switched on.}(hjShhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhhhhubh)}(hhh](h)}(h+Some more sophisticated uses of slab_debug:h]h+Some more sophisticated uses of slab_debug:}(hjvhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjshhhhhKubh)}(hgParameters may be given to ``slab_debug``. If none is specified then full debugging is enabled. Format:h](hParameters may be given to }(hjhhhNhNubh)}(h``slab_debug``h]h slab_debug}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh>. If none is specified then full debugging is enabled. Format:}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjshhubhdefinition_list)}(hhh](hdefinition_list_item)}(h8slab_debug= Enable options for all slabs h](hterm)}(hslab_debug=h]hslab_debug=}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhK#hjubh definition)}(hhh]h)}(hEnable options for all slabsh]hEnable options for all slabs}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK#hjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhK#hjubj)}(hxslab_debug=,,,... Enable options only for select slabs (no spaces after a comma) h](j)}(h8slab_debug=,,,...h]h8slab_debug=,,,...}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhK'hjubj)}(hhh]h)}(h>Enable options only for select slabs (no spaces after a comma)h]h>Enable options only for select slabs (no spaces after a comma)}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK&hjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhK'hjhhubeh}(h]h ]h"]h$]h&]uh1jhjshhhhhNubh)}(hX6Multiple blocks of options for all slabs or selected slabs can be given, with blocks of options delimited by ';'. The last of "all slabs" blocks is applied to all slabs except those that match one of the "select slabs" block. Options of the first "select slabs" blocks that matches the slab's name are applied.h]hXHMultiple blocks of options for all slabs or selected slabs can be given, with blocks of options delimited by ‘;’. The last of “all slabs” blocks is applied to all slabs except those that match one of the “select slabs” block. Options of the first “select slabs” blocks that matches the slab’s name are applied.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK)hjshhubh)}(hPossible debug options are::h]hPossible debug options are:}(hj!hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK.hjshhubjB)}(hXFF Sanity checks on (enables SLAB_DEBUG_CONSISTENCY_CHECKS Sorry SLAB legacy issues) Z Red zoning P Poisoning (object and padding) U User tracking (free and alloc) T Trace (please only use on single slabs) A Enable failslab filter mark for the cache O Switch debugging off for caches that would have caused higher minimum slab orders - Switch all debugging off (useful if the kernel is configured with CONFIG_SLUB_DEBUG_ON)h]hXFF Sanity checks on (enables SLAB_DEBUG_CONSISTENCY_CHECKS Sorry SLAB legacy issues) Z Red zoning P Poisoning (object and padding) U User tracking (free and alloc) T Trace (please only use on single slabs) A Enable failslab filter mark for the cache O Switch debugging off for caches that would have caused higher minimum slab orders - Switch all debugging off (useful if the kernel is configured with CONFIG_SLUB_DEBUG_ON)}hj/sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhK0hjshhubh)}(hPF.e. in order to boot just with sanity checks and red zoning one would specify::h]hOF.e. in order to boot just with sanity checks and red zoning one would specify:}(hj=hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjshhubh)}(h2Trying to find an issue in the dentry cache? Try::h]h1Trying to find an issue in the dentry cache? Try:}(hjYhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK@hjshhubjB)}(hslab_debug=,dentryh]hslab_debug=,dentry}hjgsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKBhjshhubh)}(hto only enable debugging on the dentry cache. You may use an asterisk at the end of the slab name, in order to cover all slabs with the same prefix. For example, here's how you can poison the dentry cache as well as all kmalloc slabs::h]hto only enable debugging on the dentry cache. You may use an asterisk at the end of the slab name, in order to cover all slabs with the same prefix. For example, here’s how you can poison the dentry cache as well as all kmalloc slabs:}(hjuhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKDhjshhubjB)}(hslab_debug=P,kmalloc-*,dentryh]hslab_debug=P,kmalloc-*,dentry}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKIhjshhubh)}(hiRed zoning and tracking may realign the slab. We can just apply sanity checks to the dentry cache with::h]hhRed zoning and tracking may realign the slab. We can just apply sanity checks to the dentry cache with:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKKhjshhubjB)}(hslab_debug=F,dentryh]hslab_debug=F,dentry}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKNhjshhubh)}(hXeDebugging options may require the minimum possible slab order to increase as a result of storing the metadata (for example, caches with PAGE_SIZE object sizes). This has a higher likelihood of resulting in slab allocation errors in low memory situations or if there's high fragmentation of memory. To switch off debugging for such caches by default, use::h]hXfDebugging options may require the minimum possible slab order to increase as a result of storing the metadata (for example, caches with PAGE_SIZE object sizes). This has a higher likelihood of resulting in slab allocation errors in low memory situations or if there’s high fragmentation of memory. To switch off debugging for such caches by default, use:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKPhjshhubjB)}(h slab_debug=Oh]h slab_debug=O}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKVhjshhubh)}(hYou can apply different options to different list of slab names, using blocks of options. This will enable red zoning for dentry and user tracking for kmalloc. All other slabs will not get any debugging enabled::h]hYou can apply different options to different list of slab names, using blocks of options. This will enable red zoning for dentry and user tracking for kmalloc. All other slabs will not get any debugging enabled:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKXhjshhubjB)}(hslab_debug=Z,dentry;U,kmalloc-*h]hslab_debug=Z,dentry;U,kmalloc-*}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhK\hjshhubh)}(hYou can also enable options (e.g. sanity checks and poisoning) for all caches except some that are deemed too performance critical and don't need to be debugged by specifying global debug options followed by a list of slab names with "-" as options::h]hYou can also enable options (e.g. sanity checks and poisoning) for all caches except some that are deemed too performance critical and don’t need to be debugged by specifying global debug options followed by a list of slab names with “-” as options:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhK^hjshhubjB)}(h slab_debug=FZ;-,zs_handle,zspageh]h slab_debug=FZ;-,zs_handle,zspage}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKchjshhubh)}(hVThe state of each debug option for a slab can be found in the respective files under::h]hUThe state of each debug option for a slab can be found in the respective files under:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKehjshhubjB)}(h/sys/kernel/slab//h]h/sys/kernel/slab//}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKhhjshhubh)}(hIf the file contains 1, the option is enabled, 0 means disabled. The debug options from the ``slab_debug`` parameter translate to the following files::h](h\If the file contains 1, the option is enabled, 0 means disabled. The debug options from the }(hjhhhNhNubh)}(h``slab_debug``h]h slab_debug}(hj%hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh, parameter translate to the following files:}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKjhjshhubjB)}(hgF sanity_checks Z red_zone P poison U store_user T trace A failslabh]hgF sanity_checks Z red_zone P poison U store_user T trace A failslab}hj=sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKmhjshhubh)}(hfailslab file is writable, so writing 1 or 0 will enable or disable the option at runtime. Write returns -EINVAL if cache is an alias. Careful with tracing: It may spew out lots of information and never stop if used on the wrong slab.h]hfailslab file is writable, so writing 1 or 0 will enable or disable the option at runtime. Write returns -EINVAL if cache is an alias. Careful with tracing: It may spew out lots of information and never stop if used on the wrong slab.}(hjKhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKthjshhubh)}(hhh](h)}(h Slab mergingh]h Slab merging}(hj\hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjYhhhhhKzubh)}(hIf no debug options are specified then SLUB may merge similar slabs together in order to reduce overhead and increase cache hotness of objects. ``slabinfo -a`` displays which slabs were merged together.h](hIf no debug options are specified then SLUB may merge similar slabs together in order to reduce overhead and increase cache hotness of objects. }(hjjhhhNhNubh)}(h``slabinfo -a``h]h slabinfo -a}(hjrhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjjubh+ displays which slabs were merged together.}(hjjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhK|hjYhhubeh}(h] slab-mergingah ]h"] slab mergingah$]h&]uh1hhjshhhhhKzubh)}(hhh](h)}(hSlab validationh]hSlab validation}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKubh)}(hSLUB can validate all object if the kernel was booted with slab_debug. In order to do so you must have the ``slabinfo`` tool. Then you can do ::h](hkSLUB can validate all object if the kernel was booted with slab_debug. In order to do so you must have the }(hjhhhNhNubh)}(h ``slabinfo``h]hslabinfo}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh tool. Then you can do}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjhhubjB)}(h slabinfo -vh]h slabinfo -v}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKhjhhubh)}(hDwhich will test all objects. Output will be generated to the syslog.h]hDwhich will test all objects. Output will be generated to the syslog.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubh)}(hThis also works in a more limited way if boot was without slab debug. In that case ``slabinfo -v`` simply tests all reachable objects. Usually these are in the cpu slabs and the partial slabs. Full slabs are not tracked by SLUB in a non debug situation.h](hSThis also works in a more limited way if boot was without slab debug. In that case }(hjhhhNhNubh)}(h``slabinfo -v``h]h slabinfo -v}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh simply tests all reachable objects. Usually these are in the cpu slabs and the partial slabs. Full slabs are not tracked by SLUB in a non debug situation.}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjhhubeh}(h]slab-validationah ]h"]slab validationah$]h&]uh1hhjshhhhhKubh)}(hhh](h)}(hGetting more performanceh]hGetting more performance}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKubh)}(hTo some degree SLUB's performance is limited by the need to take the list_lock once in a while to deal with partial slabs. That overhead is governed by the order of the allocation for each slab. The allocations can be influenced by kernel parameters:h]hTo some degree SLUB’s performance is limited by the need to take the list_lock once in a while to deal with partial slabs. That overhead is governed by the order of the allocation for each slab. The allocations can be influenced by kernel parameters:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubhcomment)}(hNslab_min_objects=x (default: automatically scaled by number of cpus)h]hNslab_min_objects=x (default: automatically scaled by number of cpus)}hj(sbah}(h]h ]h"]h$]h&]jQjRuh1j&hjhhhhhKubj')}(h(slab_min_order=x (default 0)h]h(slab_min_order=x (default 0)}hj6sbah}(h]h ]h"]h$]h&]jQjRuh1j&hjhhhhhKubj')}(hBslab_max_order=x (default 3 (PAGE_ALLOC_COSTLY_ORDER))h]hBslab_max_order=x (default 3 (PAGE_ALLOC_COSTLY_ORDER))}hjDsbah}(h]h ]h"]h$]h&]jQjRuh1j&hjhhhhhKubj)}(hhh](j)}(hX+``slab_min_objects`` allows to specify how many objects must at least fit into one slab in order for the allocation order to be acceptable. In general slub will be able to perform this number of allocations on a slab without consulting centralized resources (list_lock) where contention may occur. h](j)}(h``slab_min_objects``h]h)}(hj[h]hslab_min_objects}(hj]hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjYubah}(h]h ]h"]h$]h&]uh1jhhhKhjUubj)}(hhh]h)}(hXallows to specify how many objects must at least fit into one slab in order for the allocation order to be acceptable. In general slub will be able to perform this number of allocations on a slab without consulting centralized resources (list_lock) where contention may occur.h]hXallows to specify how many objects must at least fit into one slab in order for the allocation order to be acceptable. In general slub will be able to perform this number of allocations on a slab without consulting centralized resources (list_lock) where contention may occur.}(hjshhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjpubah}(h]h ]h"]h$]h&]uh1jhjUubeh}(h]h ]h"]h$]h&]uh1jhhhKhjRubj)}(hc``slab_min_order`` specifies a minimum order of slabs. A similar effect like ``slab_min_objects``. h](j)}(h``slab_min_order``h]h)}(hjh]hslab_min_order}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubah}(h]h ]h"]h$]h&]uh1jhhhKhjubj)}(hhh]h)}(hOspecifies a minimum order of slabs. A similar effect like ``slab_min_objects``.h](h:specifies a minimum order of slabs. A similar effect like }(hjhhhNhNubh)}(h``slab_min_objects``h]hslab_min_objects}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh.}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhKhjRhhubj)}(hX``slab_max_order`` specified the order at which ``slab_min_objects`` should no longer be checked. This is useful to avoid SLUB trying to generate super large order pages to fit ``slab_min_objects`` of a slab cache with large object sizes into one high order page. Setting command line parameter ``debug_guardpage_minorder=N`` (N > 0), forces setting ``slab_max_order`` to 0, what cause minimum possible order of slabs allocation. h](j)}(h``slab_max_order``h]h)}(hjh]hslab_max_order}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubah}(h]h ]h"]h$]h&]uh1jhhhKhjubj)}(hhh]h)}(hXspecified the order at which ``slab_min_objects`` should no longer be checked. This is useful to avoid SLUB trying to generate super large order pages to fit ``slab_min_objects`` of a slab cache with large object sizes into one high order page. Setting command line parameter ``debug_guardpage_minorder=N`` (N > 0), forces setting ``slab_max_order`` to 0, what cause minimum possible order of slabs allocation.h](hspecified the order at which }(hjhhhNhNubh)}(h``slab_min_objects``h]hslab_min_objects}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubhm should no longer be checked. This is useful to avoid SLUB trying to generate super large order pages to fit }(hjhhhNhNubh)}(h``slab_min_objects``h]hslab_min_objects}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubhb of a slab cache with large object sizes into one high order page. Setting command line parameter }(hjhhhNhNubh)}(h``debug_guardpage_minorder=N``h]hdebug_guardpage_minorder=N}(hj!hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh (N > 0), forces setting }(hjhhhNhNubh)}(h``slab_max_order``h]hslab_max_order}(hj3hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh= to 0, what cause minimum possible order of slabs allocation.}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhKhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhKhjRhhubj)}(hX``slab_strict_numa`` Enables the application of memory policies on each allocation. This results in more accurate placement of objects which may result in the reduction of accesses to remote nodes. The default is to only apply memory policies at the folio level when a new folio is acquired or a folio is retrieved from the lists. Enabling this option reduces the fastpath performance of the slab allocator. h](j)}(h``slab_strict_numa``h]h)}(hj]h]hslab_strict_numa}(hj_hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj[ubah}(h]h ]h"]h$]h&]uh1jhhhKhjWubj)}(hhh]h)}(hXEnables the application of memory policies on each allocation. This results in more accurate placement of objects which may result in the reduction of accesses to remote nodes. The default is to only apply memory policies at the folio level when a new folio is acquired or a folio is retrieved from the lists. Enabling this option reduces the fastpath performance of the slab allocator.h]hXEnables the application of memory policies on each allocation. This results in more accurate placement of objects which may result in the reduction of accesses to remote nodes. The default is to only apply memory policies at the folio level when a new folio is acquired or a folio is retrieved from the lists. Enabling this option reduces the fastpath performance of the slab allocator.}(hjuhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjrubah}(h]h ]h"]h$]h&]uh1jhjWubeh}(h]h ]h"]h$]h&]uh1jhhhKhjRhhubeh}(h]h ]h"]h$]h&]uh1jhjhhhhhNubeh}(h]getting-more-performanceah ]h"]getting more performanceah$]h&]uh1hhjshhhhhKubh)}(hhh](h)}(hSLUB Debug outputh]hSLUB Debug output}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhKubh)}(h'Here is a sample of slub debug output::h]h&Here is a sample of slub debug output:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubjB)}(hX4==================================================================== BUG kmalloc-8: Right Redzone overwritten -------------------------------------------------------------------- INFO: 0xc90f6d28-0xc90f6d2b. First byte 0x00 instead of 0xcc INFO: Slab 0xc528c530 flags=0x400000c3 inuse=61 fp=0xc90f6d58 INFO: Object 0xc90f6d20 @offset=3360 fp=0xc90f6d58 INFO: Allocated in get_modalias+0x61/0xf5 age=53 cpu=1 pid=554 Bytes b4 (0xc90f6d10): 00 00 00 00 00 00 00 00 5a 5a 5a 5a 5a 5a 5a 5a ........ZZZZZZZZ Object (0xc90f6d20): 31 30 31 39 2e 30 30 35 1019.005 Redzone (0xc90f6d28): 00 cc cc cc . Padding (0xc90f6d50): 5a 5a 5a 5a 5a 5a 5a 5a ZZZZZZZZ [] dump_trace+0x63/0x1eb [] show_trace_log_lvl+0x1a/0x2f [] show_trace+0x12/0x14 [] dump_stack+0x16/0x18 [] object_err+0x143/0x14b [] check_object+0x66/0x234 [] __slab_free+0x239/0x384 [] kfree+0xa6/0xc6 [] get_modalias+0xb9/0xf5 [] dmi_dev_uevent+0x27/0x3c [] dev_uevent+0x1ad/0x1da [] kobject_uevent_env+0x20a/0x45b [] kobject_uevent+0xa/0xf [] store_uevent+0x4f/0x58 [] dev_attr_store+0x29/0x2f [] sysfs_write_file+0x16e/0x19c [] vfs_write+0xd1/0x15a [] sys_write+0x3d/0x72 [] sysenter_past_esp+0x5f/0x99 [] 0xb7f7b410 ======================= FIX kmalloc-8: Restoring Redzone 0xc90f6d28-0xc90f6d2b=0xcch]hX4==================================================================== BUG kmalloc-8: Right Redzone overwritten -------------------------------------------------------------------- INFO: 0xc90f6d28-0xc90f6d2b. First byte 0x00 instead of 0xcc INFO: Slab 0xc528c530 flags=0x400000c3 inuse=61 fp=0xc90f6d58 INFO: Object 0xc90f6d20 @offset=3360 fp=0xc90f6d58 INFO: Allocated in get_modalias+0x61/0xf5 age=53 cpu=1 pid=554 Bytes b4 (0xc90f6d10): 00 00 00 00 00 00 00 00 5a 5a 5a 5a 5a 5a 5a 5a ........ZZZZZZZZ Object (0xc90f6d20): 31 30 31 39 2e 30 30 35 1019.005 Redzone (0xc90f6d28): 00 cc cc cc . Padding (0xc90f6d50): 5a 5a 5a 5a 5a 5a 5a 5a ZZZZZZZZ [] dump_trace+0x63/0x1eb [] show_trace_log_lvl+0x1a/0x2f [] show_trace+0x12/0x14 [] dump_stack+0x16/0x18 [] object_err+0x143/0x14b [] check_object+0x66/0x234 [] __slab_free+0x239/0x384 [] kfree+0xa6/0xc6 [] get_modalias+0xb9/0xf5 [] dmi_dev_uevent+0x27/0x3c [] dev_uevent+0x1ad/0x1da [] kobject_uevent_env+0x20a/0x45b [] kobject_uevent+0xa/0xf [] store_uevent+0x4f/0x58 [] dev_attr_store+0x29/0x2f [] sysfs_write_file+0x16e/0x19c [] vfs_write+0xd1/0x15a [] sys_write+0x3d/0x72 [] sysenter_past_esp+0x5f/0x99 [] 0xb7f7b410 ======================= FIX kmalloc-8: Restoring Redzone 0xc90f6d28-0xc90f6d2b=0xcc}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKhjhhubh)}(hIf SLUB encounters a corrupted object (full detection requires the kernel to be booted with slab_debug) then the following output will be dumped into the syslog:h]hIf SLUB encounters a corrupted object (full detection requires the kernel to be booted with slab_debug) then the following output will be dumped into the syslog:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjhhubhenumerated_list)}(hhh](h list_item)}(hXDescription of the problem encountered This will be a message in the system log starting with:: =============================================== BUG : ----------------------------------------------- INFO: - INFO: Slab
INFO: Object
INFO: Allocated in age= cpu= pid= INFO: Freed in age= cpu= pid= (Object allocation / free information is only available if SLAB_STORE_USER is set for the slab. slab_debug sets that option) h](h)}(h&Description of the problem encounteredh]h&Description of the problem encountered}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubh)}(h8This will be a message in the system log starting with::h]h7This will be a message in the system log starting with:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubjB)}(hX=============================================== BUG : ----------------------------------------------- INFO: - INFO: Slab
INFO: Object
INFO: Allocated in age= cpu= pid= INFO: Freed in age= cpu= pid=h]hX=============================================== BUG : ----------------------------------------------- INFO: - INFO: Slab
INFO: Object
INFO: Allocated in age= cpu= pid= INFO: Freed in age= cpu= pid=}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhKhjubh)}(h|(Object allocation / free information is only available if SLAB_STORE_USER is set for the slab. slab_debug sets that option)h]h|(Object allocation / free information is only available if SLAB_STORE_USER is set for the slab. slab_debug sets that option)}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhjubeh}(h]h ]h"]h$]h&]uh1jhjhhhhhNubj)}(hXcThe object contents if an object was involved. Various types of lines can follow the BUG SLUB line: Bytes b4
: Shows a few bytes before the object where the problem was detected. Can be useful if the corruption does not stop with the start of the object. Object
: The bytes of the object. If the object is inactive then the bytes typically contain poison values. Any non-poison value shows a corruption by a write after free. Redzone
: The Redzone following the object. The Redzone is used to detect writes after the object. All bytes should always have the same value. If there is any deviation then it is due to a write after the object boundary. (Redzone information is only available if SLAB_RED_ZONE is set. slab_debug sets that option) Padding
: Unused data to fill up the space in order to get the next object properly aligned. In the debug case we make sure that there are at least 4 bytes of padding. This allows the detection of writes before the object. h](h)}(h.The object contents if an object was involved.h]h.The object contents if an object was involved.}(hj%hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhj!ubh)}(h4Various types of lines can follow the BUG SLUB line:h]h4Various types of lines can follow the BUG SLUB line:}(hj3hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhKhj!ubj)}(hhh](j)}(hBytes b4
: Shows a few bytes before the object where the problem was detected. Can be useful if the corruption does not stop with the start of the object. h](j)}(hBytes b4
h]hBytes b4
}(hjHhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhMhjDubh classifier)}(hh]h}(hjXhhhNhNubah}(h]h ]h"]h$]h&]uh1jVhjDhhubj)}(hhh]h)}(hShows a few bytes before the object where the problem was detected. Can be useful if the corruption does not stop with the start of the object.h]hShows a few bytes before the object where the problem was detected. Can be useful if the corruption does not stop with the start of the object.}(hjihhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjfubah}(h]h ]h"]h$]h&]uh1jhjDubeh}(h]h ]h"]h$]h&]uh1jhhhMhjAubj)}(hObject
: The bytes of the object. If the object is inactive then the bytes typically contain poison values. Any non-poison value shows a corruption by a write after free. h](j)}(hObject
h]hObject
}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhMhjubjW)}(hh]h}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jVhjhhubj)}(hhh]h)}(hThe bytes of the object. If the object is inactive then the bytes typically contain poison values. Any non-poison value shows a corruption by a write after free.h]hThe bytes of the object. If the object is inactive then the bytes typically contain poison values. Any non-poison value shows a corruption by a write after free.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhMhjAubj)}(hXORedzone
: The Redzone following the object. The Redzone is used to detect writes after the object. All bytes should always have the same value. If there is any deviation then it is due to a write after the object boundary. (Redzone information is only available if SLAB_RED_ZONE is set. slab_debug sets that option) h](j)}(hRedzone
h]hRedzone
}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhMhjubjW)}(hh]h}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jVhjhhubj)}(hhh](h)}(hThe Redzone following the object. The Redzone is used to detect writes after the object. All bytes should always have the same value. If there is any deviation then it is due to a write after the object boundary.h]hThe Redzone following the object. The Redzone is used to detect writes after the object. All bytes should always have the same value. If there is any deviation then it is due to a write after the object boundary.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM hjubh)}(h\(Redzone information is only available if SLAB_RED_ZONE is set. slab_debug sets that option)h]h\(Redzone information is only available if SLAB_RED_ZONE is set. slab_debug sets that option)}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjubeh}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]uh1jhhhMhjAubj)}(hPadding
: Unused data to fill up the space in order to get the next object properly aligned. In the debug case we make sure that there are at least 4 bytes of padding. This allows the detection of writes before the object. h](j)}(hPadding
h]hPadding
}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jhhhMhj ubjW)}(hh]h}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1jVhj hhubj)}(hhh]h)}(hUnused data to fill up the space in order to get the next object properly aligned. In the debug case we make sure that there are at least 4 bytes of padding. This allows the detection of writes before the object.h]hUnused data to fill up the space in order to get the next object properly aligned. In the debug case we make sure that there are at least 4 bytes of padding. This allows the detection of writes before the object.}(hj.hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj+ubah}(h]h ]h"]h$]h&]uh1jhj ubeh}(h]h ]h"]h$]h&]uh1jhhhMhjAubeh}(h]h ]h"]h$]h&]uh1jhj!ubeh}(h]h ]h"]h$]h&]uh1jhjhhhNhNubj)}(hA stackdump The stackdump describes the location where the error was detected. The cause of the corruption is may be more likely found by looking at the function that allocated or freed the object. h](h)}(h A stackdumph]h A stackdump}(hjXhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjTubh)}(hThe stackdump describes the location where the error was detected. The cause of the corruption is may be more likely found by looking at the function that allocated or freed the object.h]hThe stackdump describes the location where the error was detected. The cause of the corruption is may be more likely found by looking at the function that allocated or freed the object.}(hjfhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjTubeh}(h]h ]h"]h$]h&]uh1jhjhhhhhNubj)}(hXReport on how the problem was dealt with in order to ensure the continued operation of the system. These are messages in the system log beginning with:: FIX : In the above sample SLUB found that the Redzone of an active object has been overwritten. Here a string of 8 characters was written into a slab that has the length of 8 characters. However, a 8 character string needs a terminating 0. That zero has overwritten the first byte of the Redzone field. After reporting the details of the issue encountered the FIX SLUB message tells us that SLUB has restored the Redzone to its proper value and then system operations continue. h](h)}(hbReport on how the problem was dealt with in order to ensure the continued operation of the system.h]hbReport on how the problem was dealt with in order to ensure the continued operation of the system.}(hj~hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjzubh)}(h5These are messages in the system log beginning with::h]h4These are messages in the system log beginning with:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM"hjzubjB)}(h4FIX : h]h4FIX : }hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhM$hjzubh)}(hXIn the above sample SLUB found that the Redzone of an active object has been overwritten. Here a string of 8 characters was written into a slab that has the length of 8 characters. However, a 8 character string needs a terminating 0. That zero has overwritten the first byte of the Redzone field. After reporting the details of the issue encountered the FIX SLUB message tells us that SLUB has restored the Redzone to its proper value and then system operations continue.h]hXIn the above sample SLUB found that the Redzone of an active object has been overwritten. Here a string of 8 characters was written into a slab that has the length of 8 characters. However, a 8 character string needs a terminating 0. That zero has overwritten the first byte of the Redzone field. After reporting the details of the issue encountered the FIX SLUB message tells us that SLUB has restored the Redzone to its proper value and then system operations continue.}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM&hjzubeh}(h]h ]h"]h$]h&]uh1jhjhhhhhNubeh}(h]h ]h"]h$]h&]enumtypearabicprefixhsuffix.uh1jhjhhhhhKubeh}(h]slub-debug-outputah ]h"]slub debug outputah$]h&]uh1hhjshhhhhKubh)}(hhh](h)}(hEmergency operationsh]hEmergency operations}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjhhhhhM/ubh)}(hHMinimal debugging (sanity checks alone) can be enabled by booting with::h]hGMinimal debugging (sanity checks alone) can be enabled by booting with:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM1hjhhubjB)}(h slab_debug=Fh]h slab_debug=F}hjsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhM3hjhhubh)}(hXThis will be generally be enough to enable the resiliency features of slub which will keep the system running even if a bad kernel component will keep corrupting objects. This may be important for production systems. Performance will be impacted by the sanity checks and there will be a continual stream of error messages to the syslog but no additional memory will be used (unlike full debugging).h]hXThis will be generally be enough to enable the resiliency features of slub which will keep the system running even if a bad kernel component will keep corrupting objects. This may be important for production systems. Performance will be impacted by the sanity checks and there will be a continual stream of error messages to the syslog but no additional memory will be used (unlike full debugging).}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM5hjhhubh)}(hNo guarantees. The kernel component still needs to be fixed. Performance may be optimized further by locating the slab that experiences corruption and enabling debugging only for that cacheh]hNo guarantees. The kernel component still needs to be fixed. Performance may be optimized further by locating the slab that experiences corruption and enabling debugging only for that cache}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM<hjhhubh)}(hI.e.::h]hI.e.:}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM@hjhhubjB)}(hslab_debug=F,dentryh]hslab_debug=F,dentry}hj&sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhMBhjhhubh)}(hIf the corruption occurs by writing after the end of the object then it may be advisable to enable a Redzone to avoid corrupting the beginning of other objects::h]hIf the corruption occurs by writing after the end of the object then it may be advisable to enable a Redzone to avoid corrupting the beginning of other objects:}(hj4hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMDhjhhubjB)}(hslab_debug=FZ,dentryh]hslab_debug=FZ,dentry}hjBsbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhMHhjhhubeh}(h]emergency-operationsah ]h"]emergency operationsah$]h&]uh1hhjshhhhhM/ubh)}(hhh](h)}(h#Extended slabinfo mode and plottingh]h#Extended slabinfo mode and plotting}(hj[hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjXhhhhhMKubj)}(hhh]j)}(hThe ``slabinfo`` tool has a special 'extended' ('-X') mode that includes: - Slabcache Totals - Slabs sorted by size (up to -N slabs, default 1) - Slabs sorted by loss (up to -N slabs, default 1) h](j)}(hIThe ``slabinfo`` tool has a special 'extended' ('-X') mode that includes:h](hThe }(hjphhhNhNubh)}(h ``slabinfo``h]hslabinfo}(hjxhhhNhNubah}(h]h ]h"]h$]h&]uh1hhjpubhA tool has a special ‘extended’ (‘-X’) mode that includes:}(hjphhhNhNubeh}(h]h ]h"]h$]h&]uh1jhhhMPhjlubj)}(hhh]h bullet_list)}(hhh](j)}(hSlabcache Totalsh]h)}(hjh]hSlabcache Totals}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMNhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(h6Slabs sorted by size (up to -N slabs, default 1)h]h)}(hjh]h6Slabs sorted by size (up to -N slabs, default 1)}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMOhjubah}(h]h ]h"]h$]h&]uh1jhjubj)}(h7Slabs sorted by loss (up to -N slabs, default 1) h]h)}(h6Slabs sorted by loss (up to -N slabs, default 1)h]h6Slabs sorted by loss (up to -N slabs, default 1)}(hjhhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMPhjubah}(h]h ]h"]h$]h&]uh1jhjubeh}(h]h ]h"]h$]h&]bullet-uh1jhhhMNhjubah}(h]h ]h"]h$]h&]uh1jhjlubeh}(h]h ]h"]h$]h&]uh1jhhhMPhjiubah}(h]h ]h"]h$]h&]uh1jhjXhhhNhNubh)}(hX Additionally, in this mode ``slabinfo`` does not dynamically scale sizes (G/M/K) and reports everything in bytes (this functionality is also available to other slabinfo modes via '-B' option) which makes reporting more precise and accurate. Moreover, in some sense the `-X' mode also simplifies the analysis of slabs' behaviour, because its output can be plotted using the ``slabinfo-gnuplot.sh`` script. So it pushes the analysis from looking through the numbers (tons of numbers) to something easier -- visual analysis.h](hAdditionally, in this mode }(hjhhhNhNubh)}(h ``slabinfo``h]hslabinfo}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjubh does not dynamically scale sizes (G/M/K) and reports everything in bytes (this functionality is also available to other slabinfo modes via ‘-B’ option) which makes reporting more precise and accurate. Moreover, in some sense the }(hjhhhNhNubhtitle_reference)}(h`-X' mode also simplifies the analysis of slabs' behaviour, because its output can be plotted using the ``slabinfo-gnuplot.sh``h]h-X’ mode also simplifies the analysis of slabs’ behaviour, because its output can be plotted using the ``slabinfo-gnuplot.sh`}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1j hjubh} script. So it pushes the analysis from looking through the numbers (tons of numbers) to something easier -- visual analysis.}(hjhhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMRhjXhhubh)}(hTo generate plots:h]hTo generate plots:}(hj, hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM[hjXhhubj)}(hhh](j)}(hocollect slabinfo extended records, for example:: while [ 1 ]; do slabinfo -X >> FOO_STATS; sleep 1; done h](h)}(h0collect slabinfo extended records, for example::h]h/collect slabinfo extended records, for example:}(hjA hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhM]hj= ubjB)}(h7while [ 1 ]; do slabinfo -X >> FOO_STATS; sleep 1; doneh]h7while [ 1 ]; do slabinfo -X >> FOO_STATS; sleep 1; done}hjO sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhM_hj= ubeh}(h]h ]h"]h$]h&]uh1jhj: hhhhhNubj)}(hXpass stats file(-s) to ``slabinfo-gnuplot.sh`` script:: slabinfo-gnuplot.sh FOO_STATS [FOO_STATS2 .. FOO_STATSN] The ``slabinfo-gnuplot.sh`` script will pre-processes the collected records and generates 3 png files (and 3 pre-processing cache files) per STATS file: - Slabcache Totals: FOO_STATS-totals.png - Slabs sorted by size: FOO_STATS-slabs-by-size.png - Slabs sorted by loss: FOO_STATS-slabs-by-loss.png h](h)}(h7pass stats file(-s) to ``slabinfo-gnuplot.sh`` script::h](hpass stats file(-s) to }(hjg hhhNhNubh)}(h``slabinfo-gnuplot.sh``h]hslabinfo-gnuplot.sh}(hjo hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjg ubh script:}(hjg hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMahjc ubjB)}(h8slabinfo-gnuplot.sh FOO_STATS [FOO_STATS2 .. FOO_STATSN]h]h8slabinfo-gnuplot.sh FOO_STATS [FOO_STATS2 .. FOO_STATSN]}hj sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhMchjc ubh)}(hX)The ``slabinfo-gnuplot.sh`` script will pre-processes the collected records and generates 3 png files (and 3 pre-processing cache files) per STATS file: - Slabcache Totals: FOO_STATS-totals.png - Slabs sorted by size: FOO_STATS-slabs-by-size.png - Slabs sorted by loss: FOO_STATS-slabs-by-loss.pngh](hThe }(hj hhhNhNubh)}(h``slabinfo-gnuplot.sh``h]hslabinfo-gnuplot.sh}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubhX script will pre-processes the collected records and generates 3 png files (and 3 pre-processing cache files) per STATS file: - Slabcache Totals: FOO_STATS-totals.png - Slabs sorted by size: FOO_STATS-slabs-by-size.png - Slabs sorted by loss: FOO_STATS-slabs-by-loss.png}(hj hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMehjc ubeh}(h]h ]h"]h$]h&]uh1jhj: hhhhhNubeh}(h]h ]h"]h$]h&]j loweralphajhj)uh1jhjXhhhhhM]ubh)}(hX6Another use case, when ``slabinfo-gnuplot.sh`` can be useful, is when you need to compare slabs' behaviour "prior to" and "after" some code modification. To help you out there, ``slabinfo-gnuplot.sh`` script can 'merge' the `Slabcache Totals` sections from different measurements. To visually compare N plots:h](hAnother use case, when }(hj hhhNhNubh)}(h``slabinfo-gnuplot.sh``h]hslabinfo-gnuplot.sh}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubh can be useful, is when you need to compare slabs’ behaviour “prior to” and “after” some code modification. To help you out there, }(hj hhhNhNubh)}(h``slabinfo-gnuplot.sh``h]hslabinfo-gnuplot.sh}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubh script can ‘merge’ the }(hj hhhNhNubj )}(h`Slabcache Totals`h]hSlabcache Totals}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1j hj ubhC sections from different measurements. To visually compare N plots:}(hj hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMlhjXhhubj)}(hhh](j)}(h{Collect as many STATS1, STATS2, .. STATSN files as you need:: while [ 1 ]; do slabinfo -X >> STATS; sleep 1; done h](h)}(h=Collect as many STATS1, STATS2, .. STATSN files as you need::h]h> STATS; sleep 1; doneh]h6while [ 1 ]; do slabinfo -X >> STATS; sleep 1; done}hj sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhMthj ubeh}(h]h ]h"]h$]h&]uh1jhj hhhhhNubj)}(hRPre-process those STATS files:: slabinfo-gnuplot.sh STATS1 STATS2 .. STATSN h](h)}(hPre-process those STATS files::h]hPre-process those STATS files:}(hj4 hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMvhj0 ubjB)}(h+slabinfo-gnuplot.sh STATS1 STATS2 .. STATSNh]h+slabinfo-gnuplot.sh STATS1 STATS2 .. STATSN}hjB sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhMxhj0 ubeh}(h]h ]h"]h$]h&]uh1jhj hhhhhNubj)}(hXExecute ``slabinfo-gnuplot.sh`` in '-t' mode, passing all of the generated pre-processed \*-totals:: slabinfo-gnuplot.sh -t STATS1-totals STATS2-totals .. STATSN-totals This will produce a single plot (png file). Plots, expectedly, can be large so some fluctuations or small spikes can go unnoticed. To deal with that, ``slabinfo-gnuplot.sh`` has two options to 'zoom-in'/'zoom-out': a) ``-s %d,%d`` -- overwrites the default image width and height b) ``-r %d,%d`` -- specifies a range of samples to use (for example, in ``slabinfo -X >> FOO_STATS; sleep 1;`` case, using a ``-r 40,60`` range will plot only samples collected between 40th and 60th seconds). h](h)}(hdExecute ``slabinfo-gnuplot.sh`` in '-t' mode, passing all of the generated pre-processed \*-totals::h](hExecute }(hjZ hhhNhNubh)}(h``slabinfo-gnuplot.sh``h]hslabinfo-gnuplot.sh}(hjb hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjZ ubhH in ‘-t’ mode, passing all of the generated pre-processed *-totals:}(hjZ hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMzhjV ubjB)}(hCslabinfo-gnuplot.sh -t STATS1-totals STATS2-totals .. STATSN-totalsh]hCslabinfo-gnuplot.sh -t STATS1-totals STATS2-totals .. STATSN-totals}hjz sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhM}hjV ubh)}(h+This will produce a single plot (png file).h]h+This will produce a single plot (png file).}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjV ubh)}(hPlots, expectedly, can be large so some fluctuations or small spikes can go unnoticed. To deal with that, ``slabinfo-gnuplot.sh`` has two options to 'zoom-in'/'zoom-out':h](hjPlots, expectedly, can be large so some fluctuations or small spikes can go unnoticed. To deal with that, }(hj hhhNhNubh)}(h``slabinfo-gnuplot.sh``h]hslabinfo-gnuplot.sh}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubh1 has two options to ‘zoom-in’/’zoom-out’:}(hj hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMhjV ubj)}(hhh](j)}(h=``-s %d,%d`` -- overwrites the default image width and heighth]h)}(hj h](h)}(h ``-s %d,%d``h]h-s %d,%d}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubh1 -- overwrites the default image width and height}(hj hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubj)}(h``-r %d,%d`` -- specifies a range of samples to use (for example, in ``slabinfo -X >> FOO_STATS; sleep 1;`` case, using a ``-r 40,60`` range will plot only samples collected between 40th and 60th seconds). h]h)}(h``-r %d,%d`` -- specifies a range of samples to use (for example, in ``slabinfo -X >> FOO_STATS; sleep 1;`` case, using a ``-r 40,60`` range will plot only samples collected between 40th and 60th seconds).h](h)}(h ``-r %d,%d``h]h-r %d,%d}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubh9 -- specifies a range of samples to use (for example, in }(hj hhhNhNubh)}(h&``slabinfo -X >> FOO_STATS; sleep 1;``h]h"slabinfo -X >> FOO_STATS; sleep 1;}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubh case, using a }(hj hhhNhNubh)}(h ``-r 40,60``h]h-r 40,60}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhj ubhG range will plot only samples collected between 40th and 60th seconds).}(hj hhhNhNubeh}(h]h ]h"]h$]h&]uh1hhhhMhj ubah}(h]h ]h"]h$]h&]uh1jhj ubeh}(h]h ]h"]h$]h&]jj jhjj uh1jhjV ubeh}(h]h ]h"]h$]h&]uh1jhj hhhNhNubeh}(h]h ]h"]h$]h&]jj jhjj uh1jhjXhhhhhMrubeh}(h]#extended-slabinfo-mode-and-plottingah ]h"]#extended slabinfo mode and plottingah$]h&]uh1hhjshhhhhMKubh)}(hhh](h)}(hDebugFS files for SLUBh]hDebugFS files for SLUB}(hjE hhhNhNubah}(h]h ]h"]h$]h&]uh1hhjB hhhhhMubh)}(hX.For more information about current state of SLUB caches with the user tracking debug option enabled, debugfs files are available, typically under /sys/kernel/debug/slab// (created only for caches with enabled user tracking). There are 2 types of these files with the following debug information:h]hX.For more information about current state of SLUB caches with the user tracking debug option enabled, debugfs files are available, typically under /sys/kernel/debug/slab// (created only for caches with enabled user tracking). There are 2 types of these files with the following debug information:}(hjS hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjB hhubj)}(hhh](j)}(hXzalloc_traces:: Prints information about unique allocation traces of the currently allocated objects. The output is sorted by frequency of each trace. Information in the output: Number of objects, allocating function, possible memory wastage of kmalloc objects(total/per-object), minimal/average/maximal jiffies since alloc, pid range of the allocating processes, cpu mask of allocating cpus, numa node mask of origins of memory, and stack trace. Example::: 338 pci_alloc_dev+0x2c/0xa0 waste=521872/1544 age=290837/291891/293509 pid=1 cpus=106 nodes=0-1 __kmem_cache_alloc_node+0x11f/0x4e0 kmalloc_trace+0x26/0xa0 pci_alloc_dev+0x2c/0xa0 pci_scan_single_device+0xd2/0x150 pci_scan_slot+0xf7/0x2d0 pci_scan_child_bus_extend+0x4e/0x360 acpi_pci_root_create+0x32e/0x3b0 pci_acpi_scan_root+0x2b9/0x2d0 acpi_pci_root_add.cold.11+0x110/0xb0a acpi_bus_attach+0x262/0x3f0 device_for_each_child+0xb7/0x110 acpi_dev_for_each_child+0x77/0xa0 acpi_bus_attach+0x108/0x3f0 device_for_each_child+0xb7/0x110 acpi_dev_for_each_child+0x77/0xa0 acpi_bus_attach+0x108/0x3f0 h](h)}(halloc_traces::h]h alloc_traces:}(hjh hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjd ubjB)}(hXPPrints information about unique allocation traces of the currently allocated objects. The output is sorted by frequency of each trace. Information in the output: Number of objects, allocating function, possible memory wastage of kmalloc objects(total/per-object), minimal/average/maximal jiffies since alloc, pid range of the allocating processes, cpu mask of allocating cpus, numa node mask of origins of memory, and stack trace. Example::: 338 pci_alloc_dev+0x2c/0xa0 waste=521872/1544 age=290837/291891/293509 pid=1 cpus=106 nodes=0-1 __kmem_cache_alloc_node+0x11f/0x4e0 kmalloc_trace+0x26/0xa0 pci_alloc_dev+0x2c/0xa0 pci_scan_single_device+0xd2/0x150 pci_scan_slot+0xf7/0x2d0 pci_scan_child_bus_extend+0x4e/0x360 acpi_pci_root_create+0x32e/0x3b0 pci_acpi_scan_root+0x2b9/0x2d0 acpi_pci_root_add.cold.11+0x110/0xb0a acpi_bus_attach+0x262/0x3f0 device_for_each_child+0xb7/0x110 acpi_dev_for_each_child+0x77/0xa0 acpi_bus_attach+0x108/0x3f0 device_for_each_child+0xb7/0x110 acpi_dev_for_each_child+0x77/0xa0 acpi_bus_attach+0x108/0x3f0h]hXPPrints information about unique allocation traces of the currently allocated objects. The output is sorted by frequency of each trace. Information in the output: Number of objects, allocating function, possible memory wastage of kmalloc objects(total/per-object), minimal/average/maximal jiffies since alloc, pid range of the allocating processes, cpu mask of allocating cpus, numa node mask of origins of memory, and stack trace. Example::: 338 pci_alloc_dev+0x2c/0xa0 waste=521872/1544 age=290837/291891/293509 pid=1 cpus=106 nodes=0-1 __kmem_cache_alloc_node+0x11f/0x4e0 kmalloc_trace+0x26/0xa0 pci_alloc_dev+0x2c/0xa0 pci_scan_single_device+0xd2/0x150 pci_scan_slot+0xf7/0x2d0 pci_scan_child_bus_extend+0x4e/0x360 acpi_pci_root_create+0x32e/0x3b0 pci_acpi_scan_root+0x2b9/0x2d0 acpi_pci_root_add.cold.11+0x110/0xb0a acpi_bus_attach+0x262/0x3f0 device_for_each_child+0xb7/0x110 acpi_dev_for_each_child+0x77/0xa0 acpi_bus_attach+0x108/0x3f0 device_for_each_child+0xb7/0x110 acpi_dev_for_each_child+0x77/0xa0 acpi_bus_attach+0x108/0x3f0}hjv sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhMhjd ubeh}(h]h ]h"]h$]h&]uh1jhja hhhhhNubj)}(hXfree_traces:: Prints information about unique freeing traces of the currently allocated objects. The freeing traces thus come from the previous life-cycle of the objects and are reported as not available for objects allocated for the first time. The output is sorted by frequency of each trace. Information in the output: Number of objects, freeing function, minimal/average/maximal jiffies since free, pid range of the freeing processes, cpu mask of freeing cpus, and stack trace. Example::: 1980 age=4294912290 pid=0 cpus=0 51 acpi_ut_update_ref_count+0x6a6/0x782 age=236886/237027/237772 pid=1 cpus=1 kfree+0x2db/0x420 acpi_ut_update_ref_count+0x6a6/0x782 acpi_ut_update_object_reference+0x1ad/0x234 acpi_ut_remove_reference+0x7d/0x84 acpi_rs_get_prt_method_data+0x97/0xd6 acpi_get_irq_routing_table+0x82/0xc4 acpi_pci_irq_find_prt_entry+0x8e/0x2e0 acpi_pci_irq_lookup+0x3a/0x1e0 acpi_pci_irq_enable+0x77/0x240 pcibios_enable_device+0x39/0x40 do_pci_enable_device.part.0+0x5d/0xe0 pci_enable_device_flags+0xfc/0x120 pci_enable_device+0x13/0x20 virtio_pci_probe+0x9e/0x170 local_pci_probe+0x48/0x80 pci_device_probe+0x105/0x1c0 h](h)}(h free_traces::h]h free_traces:}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhj ubjB)}(hXPrints information about unique freeing traces of the currently allocated objects. The freeing traces thus come from the previous life-cycle of the objects and are reported as not available for objects allocated for the first time. The output is sorted by frequency of each trace. Information in the output: Number of objects, freeing function, minimal/average/maximal jiffies since free, pid range of the freeing processes, cpu mask of freeing cpus, and stack trace. Example::: 1980 age=4294912290 pid=0 cpus=0 51 acpi_ut_update_ref_count+0x6a6/0x782 age=236886/237027/237772 pid=1 cpus=1 kfree+0x2db/0x420 acpi_ut_update_ref_count+0x6a6/0x782 acpi_ut_update_object_reference+0x1ad/0x234 acpi_ut_remove_reference+0x7d/0x84 acpi_rs_get_prt_method_data+0x97/0xd6 acpi_get_irq_routing_table+0x82/0xc4 acpi_pci_irq_find_prt_entry+0x8e/0x2e0 acpi_pci_irq_lookup+0x3a/0x1e0 acpi_pci_irq_enable+0x77/0x240 pcibios_enable_device+0x39/0x40 do_pci_enable_device.part.0+0x5d/0xe0 pci_enable_device_flags+0xfc/0x120 pci_enable_device+0x13/0x20 virtio_pci_probe+0x9e/0x170 local_pci_probe+0x48/0x80 pci_device_probe+0x105/0x1c0h]hXPrints information about unique freeing traces of the currently allocated objects. The freeing traces thus come from the previous life-cycle of the objects and are reported as not available for objects allocated for the first time. The output is sorted by frequency of each trace. Information in the output: Number of objects, freeing function, minimal/average/maximal jiffies since free, pid range of the freeing processes, cpu mask of freeing cpus, and stack trace. Example::: 1980 age=4294912290 pid=0 cpus=0 51 acpi_ut_update_ref_count+0x6a6/0x782 age=236886/237027/237772 pid=1 cpus=1 kfree+0x2db/0x420 acpi_ut_update_ref_count+0x6a6/0x782 acpi_ut_update_object_reference+0x1ad/0x234 acpi_ut_remove_reference+0x7d/0x84 acpi_rs_get_prt_method_data+0x97/0xd6 acpi_get_irq_routing_table+0x82/0xc4 acpi_pci_irq_find_prt_entry+0x8e/0x2e0 acpi_pci_irq_lookup+0x3a/0x1e0 acpi_pci_irq_enable+0x77/0x240 pcibios_enable_device+0x39/0x40 do_pci_enable_device.part.0+0x5d/0xe0 pci_enable_device_flags+0xfc/0x120 pci_enable_device+0x13/0x20 virtio_pci_probe+0x9e/0x170 local_pci_probe+0x48/0x80 pci_device_probe+0x105/0x1c0 }hj sbah}(h]h ]h"]h$]h&]jQjRuh1jAhhhMhj ubeh}(h]h ]h"]h$]h&]uh1jhja hhhhhNubeh}(h]h ]h"]h$]h&]jjjhjjuh1jhjB hhhhhMubh)}(hDChristoph Lameter, May 30, 2007 Sergey Senozhatsky, October 23, 2015h]hDChristoph Lameter, May 30, 2007 Sergey Senozhatsky, October 23, 2015}(hj hhhNhNubah}(h]h ]h"]h$]h&]uh1hhhhMhjB hhubeh}(h]debugfs-files-for-slubah ]h"]debugfs files for slubah$]h&]uh1hhjshhhhhMubeh}(h]*some-more-sophisticated-uses-of-slab-debugah ]h"]+some more sophisticated uses of slab_debug:ah$]h&]uh1hhhhhhhhKubeh}(h](short-users-guide-for-the-slab-allocatorah ]h"](short users guide for the slab allocatorah$]h&]uh1hhhhhhhhKubeh}(h]h ]h"]h$]h&]sourcehuh1hcurrent_sourceN current_lineNsettingsdocutils.frontendValues)}(hN generatorN datestampN source_linkN source_urlN toc_backlinksentryfootnote_backlinksK sectnum_xformKstrip_commentsNstrip_elements_with_classesN strip_classesN report_levelK halt_levelKexit_status_levelKdebugNwarning_streamN tracebackinput_encoding utf-8-siginput_encoding_error_handlerstrictoutput_encodingutf-8output_encoding_error_handlerj error_encodingutf-8error_encoding_error_handlerbackslashreplace language_codeenrecord_dependenciesNconfigN id_prefixhauto_id_prefixid dump_settingsNdump_internalsNdump_transformsNdump_pseudo_xmlNexpose_internalsNstrict_visitorN_disable_configN_sourcehʌ _destinationN _config_files]7/var/lib/git/docbuild/linux/Documentation/docutils.confafile_insertion_enabled raw_enabledKline_length_limitM'pep_referencesN pep_base_urlhttps://peps.python.org/pep_file_url_templatepep-%04drfc_referencesN rfc_base_url&https://datatracker.ietf.org/doc/html/ tab_widthKtrim_footnote_reference_spacesyntax_highlightlong smart_quotessmartquotes_locales]character_level_inline_markupdoctitle_xform docinfo_xformKsectsubtitle_xform image_loadinglinkembed_stylesheetcloak_email_addressessection_self_linkenvNubreporterNindirect_targets]substitution_defs}substitution_names}refnames}refids}nameids}(j j j j jjjjjjjjjUjRj? j< j j u nametypes}(j j jjjjjUj? j uh}(j hj jsjjYjjjjjjjRjj< jXj jB u footnote_refs} citation_refs} autofootnotes]autofootnote_refs]symbol_footnotes]symbol_footnote_refs] footnotes] citations]autofootnote_startKsymbol_footnote_startK id_counter collectionsCounter}Rparse_messages]transform_messages] transformerN include_log] decorationNhhub.