lwn.git/tools/vm/slabinfo.c, branch v5.12-rc3

tools/vm/slabinfo: fix sanity checks enabling

2020-01-31T18:30:38+00:00

The sysfs file name for enabling sanity checking is called 'sanity_checks' and not 'sanity'. The name of the file has never changed since the introduction of the slub allocator. Obviously, most people turn the checks on via the command line option and not during runtime using slabinfo. Link: http://lkml.kernel.org/r/20200116131642.642-1-dwagner@suse.de Signed-off-by: Daniel Wagner Reviewed-by: Andrew Morton Cc: "Tobin C. Harding" Cc: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: add sorting info to help menu

2019-07-12T18:05:46+00:00

Passing more than one sorting option has undefined behaviour. Add an explicit statement as such to the help menu, this also has the advantage of highlighting all the sorting options. Link: http://lkml.kernel.org/r/20190426022622.4089-5-tobin@kernel.org Signed-off-by: Tobin C. Harding Cc: Alexander Duyck Cc: Brendan Gregg , Cc: Christoph Lameter Cc: David Rientjes Cc: Jesper Dangaard Brouer Cc: Joonsoo Kim Cc: Mel Gorman Cc: Michal Hocko Cc: Pekka Enberg Cc: Qian Cai Cc: Tejun Heo Cc: Vlastimil Babka Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: add option to sort by partial slabs

2019-07-12T18:05:46+00:00

We would like to get a better view of the level of fragmentation within the SLUB allocator. Total number of partial slabs is an indicator of fragmentation. Add a command line option (-P | --partial) to sort the slab list by total number of partial slabs. Link: http://lkml.kernel.org/r/20190426022622.4089-4-tobin@kernel.org Signed-off-by: Tobin C. Harding Cc: Alexander Duyck Cc: Brendan Gregg , Cc: Christoph Lameter Cc: David Rientjes Cc: Jesper Dangaard Brouer Cc: Joonsoo Kim Cc: Mel Gorman Cc: Michal Hocko Cc: Pekka Enberg Cc: Qian Cai Cc: Tejun Heo Cc: Vlastimil Babka Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: add partial slab listing to -X

2019-07-12T18:05:46+00:00

We would like to see how fragmented the SLUB allocator is, one window into fragmentation is the total number of partial slabs. Currently `slabinfo -X` shows slabs sorted by loss and by size. We can use this option to also show slabs sorted by number of partial slabs. Option '-X' can be used in conjunction with '-N' to control the number of slabs shown e.g. list of top 5 slabs: slabinfo -X -N5 Add list of slabs ordered by number of partial slabs to output of `slabinfo -X`. Link: http://lkml.kernel.org/r/20190426022622.4089-3-tobin@kernel.org Signed-off-by: Tobin C. Harding Cc: Alexander Duyck Cc: Brendan Gregg , Cc: Christoph Lameter Cc: David Rientjes Cc: Jesper Dangaard Brouer Cc: Joonsoo Kim Cc: Mel Gorman Cc: Michal Hocko Cc: Pekka Enberg Cc: Qian Cai Cc: Tejun Heo Cc: Vlastimil Babka Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: order command line options

2019-07-12T18:05:46+00:00

During recent discussion on LKML over SLAB vs SLUB it was suggested by Jesper that it would be nice to have a tool to view the current fragmentation of the slab allocators. CC list for this set is taken from that thread. For SLUB we have all the information for this already exposed by the kernel and also we have a userspace tool for displaying this info: tools/vm/slabinfo.c Extend slabinfo to improve the fragmentation information by enabling sorting of caches by number of partial slabs. Also add cache list sorted in this manner to the output of `slabinfo -X`. This patch (of 4): get_opt() has a spurious character within the option string. Remove it and reorder the options in alphabetic order so that it is easier to keep the options correct. Use the same ordering for command help output and long option handling code. Link: http://lkml.kernel.org/r/20190426022622.4089-2-tobin@kernel.org Signed-off-by: Tobin C. Harding Cc: Jesper Dangaard Brouer Cc: Pekka Enberg Cc: Vlastimil Babka Cc: Christoph Lameter Cc: David Rientjes Cc: Joonsoo Kim Cc: Tejun Heo Cc: Qian Cai Cc: Mel Gorman Cc: Alexander Duyck Cc: Michal Hocko Cc: Brendan Gregg , Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: clean up usage menu debug items

2019-03-06T05:07:20+00:00

Attempt to make the usage comment for debug options a little cleaner. Link: http://lkml.kernel.org/r/20190212001219.27769-5-tobin@kernel.org Signed-off-by: Tobin C. Harding Acked-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: align usage output columns

2019-03-06T05:07:20+00:00

Usage message uses spaces not tabspaces, a few tabspaces have snuck in making the columns not align correctly when output. Align usage output columns using spaces instead of tabspaces. Link: http://lkml.kernel.org/r/20190212001219.27769-4-tobin@kernel.org Signed-off-by: Tobin C. Harding Acked-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: put options in alphabetic order

2019-03-06T05:07:20+00:00

Primarily the usage message lists options in alphabetic order however there are a bunch of the options that are not in alphabetic order. Put options in alphabetic order. Link: http://lkml.kernel.org/r/20190212001219.27769-3-tobin@kernel.org Signed-off-by: Tobin C. Harding Acked-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo: update options in usage message

2019-03-06T05:07:20+00:00

Currently usage message list only a subset of the available options. should list them all. Update options in usage massage to include all available options. Link: http://lkml.kernel.org/r/20190212001219.27769-2-tobin@kernel.org Signed-off-by: Tobin C. Harding Acked-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

tools/vm/slabinfo.c: fix sign-compare warning

2018-09-04T23:45:02+00:00

Currently we get the following compiler warning: slabinfo.c:854:22: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] if (s->object_size < min_objsize) ^ due to the mismatch of signed/unsigned comparison. ->object_size and ->slab_size are never expected to be negative, so let's define them as unsigned int. [n-horiguchi@ah.jp.nec.com: convert everything - none of these can be negative] Link: http://lkml.kernel.org/r/20180826234947.GA9787@hori1.linux.bs1.fc.nec.co.jp Link: http://lkml.kernel.org/r/1535103134-20239-1-git-send-email-n-horiguchi@ah.jp.nec.com Signed-off-by: Naoya Horiguchi Reviewed-by: Andrew Morton Cc: Matthew Wilcox Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds