| Message ID | 20240906005337.1220091-1-kent.overstreet@linux.dev (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | mm/show_mem.c: report alloc tags in human readable units | expand |
On Thu, Sep 5, 2024 at 5:53 PM Kent Overstreet <kent.overstreet@linux.dev> wrote: > > We already do this when reporting slab info - more consistent and more > readable. Hi Kent, I remember we discussed this before and agreed upon Pasha's suggestion that if needed one could do: # sort -g /proc/allocinfo|tail|numfmt --to=iec 2.8M 22648 fs/kernfs/dir.c:615 func:__kernfs_new_node 3.8M 953 mm/memory.c:4214 func:alloc_anon_folio 4.0M 1010 drivers/staging/ctagmod/ctagmod.c:20 [ctagmod] func:ctagmod_start 4.1M 4 net/netfilter/nf_conntrack_core.c:2567 func:nf_ct_alloc_hashtable 6.0M 1532 mm/filemap.c:1919 func:__filemap_get_folio 8.8M 2785 kernel/fork.c:307 func:alloc_thread_stack_node 13M 234 block/blk-mq.c:3421 func:blk_mq_alloc_rqs 14M 3520 mm/mm_init.c:2530 func:alloc_large_system_hash 15M 3656 mm/readahead.c:247 func:page_cache_ra_unbounded 55M 4887 mm/slub.c:2259 func:alloc_slab_page 122M 31168 mm/page_ext.c:270 func:alloc_page_ext That's even documented here: https://elixir.bootlin.com/linux/v6.10.8/source/Documentation/mm/allocation-profiling.rst#L36 Did something change? > > Cc: Suren Baghdasaryan <surenb@google.com> > Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev> > --- > mm/show_mem.c | 11 +++++++---- > 1 file changed, 7 insertions(+), 4 deletions(-) > > diff --git a/mm/show_mem.c b/mm/show_mem.c > index 691e1b457d04..1b448e1ebd09 100644 > --- a/mm/show_mem.c > +++ b/mm/show_mem.c > @@ -459,15 +459,18 @@ void __show_mem(unsigned int filter, nodemask_t *nodemask, int max_zone_idx) > struct codetag *ct = tags[i].ct; > struct alloc_tag *tag = ct_to_alloc_tag(ct); > struct alloc_tag_counters counter = alloc_tag_read(tag); > + char bytes[10]; > + > + string_get_size(counter.bytes, 1, STRING_UNITS_2, bytes, sizeof(bytes)); > > /* Same as alloc_tag_to_text() but w/o intermediate buffer */ > if (ct->modname) > - pr_notice("%12lli %8llu %s:%u [%s] func:%s\n", > - counter.bytes, counter.calls, ct->filename, > + pr_notice("%12s %8llu %s:%u [%s] func:%s\n", > + bytes, counter.calls, ct->filename, > ct->lineno, ct->modname, ct->function); > else > - pr_notice("%12lli %8llu %s:%u func:%s\n", > - counter.bytes, counter.calls, ct->filename, > + pr_notice("%12s %8llu %s:%u func:%s\n", > + bytes, counter.calls, ct->filename, > ct->lineno, ct->function); > } > } > -- > 2.45.2 >
On Thu, Sep 05, 2024 at 08:32:24PM GMT, Suren Baghdasaryan wrote: > On Thu, Sep 5, 2024 at 5:53 PM Kent Overstreet > <kent.overstreet@linux.dev> wrote: > > > > We already do this when reporting slab info - more consistent and more > > readable. > > Hi Kent, > I remember we discussed this before and agreed upon Pasha's suggestion > that if needed one could do: > > # sort -g /proc/allocinfo|tail|numfmt --to=iec > 2.8M 22648 fs/kernfs/dir.c:615 func:__kernfs_new_node > 3.8M 953 mm/memory.c:4214 func:alloc_anon_folio > 4.0M 1010 drivers/staging/ctagmod/ctagmod.c:20 [ctagmod] > func:ctagmod_start > 4.1M 4 net/netfilter/nf_conntrack_core.c:2567 > func:nf_ct_alloc_hashtable > 6.0M 1532 mm/filemap.c:1919 func:__filemap_get_folio > 8.8M 2785 kernel/fork.c:307 func:alloc_thread_stack_node > 13M 234 block/blk-mq.c:3421 func:blk_mq_alloc_rqs > 14M 3520 mm/mm_init.c:2530 func:alloc_large_system_hash > 15M 3656 mm/readahead.c:247 func:page_cache_ra_unbounded > 55M 4887 mm/slub.c:2259 func:alloc_slab_page > 122M 31168 mm/page_ext.c:270 func:alloc_page_ext > > That's even documented here: > https://elixir.bootlin.com/linux/v6.10.8/source/Documentation/mm/allocation-profiling.rst#L36 > Did something change? That's for /proc/allocinfo, which this patch doesn't change - this is the oom report.
On Thu, Sep 5, 2024 at 8:52 PM Kent Overstreet <kent.overstreet@linux.dev> wrote: > > On Thu, Sep 05, 2024 at 08:32:24PM GMT, Suren Baghdasaryan wrote: > > On Thu, Sep 5, 2024 at 5:53 PM Kent Overstreet > > <kent.overstreet@linux.dev> wrote: > > > > > > We already do this when reporting slab info - more consistent and more > > > readable. > > > > Hi Kent, > > I remember we discussed this before and agreed upon Pasha's suggestion > > that if needed one could do: > > > > # sort -g /proc/allocinfo|tail|numfmt --to=iec > > 2.8M 22648 fs/kernfs/dir.c:615 func:__kernfs_new_node > > 3.8M 953 mm/memory.c:4214 func:alloc_anon_folio > > 4.0M 1010 drivers/staging/ctagmod/ctagmod.c:20 [ctagmod] > > func:ctagmod_start > > 4.1M 4 net/netfilter/nf_conntrack_core.c:2567 > > func:nf_ct_alloc_hashtable > > 6.0M 1532 mm/filemap.c:1919 func:__filemap_get_folio > > 8.8M 2785 kernel/fork.c:307 func:alloc_thread_stack_node > > 13M 234 block/blk-mq.c:3421 func:blk_mq_alloc_rqs > > 14M 3520 mm/mm_init.c:2530 func:alloc_large_system_hash > > 15M 3656 mm/readahead.c:247 func:page_cache_ra_unbounded > > 55M 4887 mm/slub.c:2259 func:alloc_slab_page > > 122M 31168 mm/page_ext.c:270 func:alloc_page_ext > > > > That's even documented here: > > https://elixir.bootlin.com/linux/v6.10.8/source/Documentation/mm/allocation-profiling.rst#L36 > > Did something change? > > That's for /proc/allocinfo, which this patch doesn't change - this is > the oom report. Ah, got it! Looks reasonable to me. Reviewed-by: Suren Baghdasaryan <surenb@google.com>
diff --git a/mm/show_mem.c b/mm/show_mem.c index 691e1b457d04..1b448e1ebd09 100644 --- a/mm/show_mem.c +++ b/mm/show_mem.c @@ -459,15 +459,18 @@ void __show_mem(unsigned int filter, nodemask_t *nodemask, int max_zone_idx) struct codetag *ct = tags[i].ct; struct alloc_tag *tag = ct_to_alloc_tag(ct); struct alloc_tag_counters counter = alloc_tag_read(tag); + char bytes[10]; + + string_get_size(counter.bytes, 1, STRING_UNITS_2, bytes, sizeof(bytes)); /* Same as alloc_tag_to_text() but w/o intermediate buffer */ if (ct->modname) - pr_notice("%12lli %8llu %s:%u [%s] func:%s\n", - counter.bytes, counter.calls, ct->filename, + pr_notice("%12s %8llu %s:%u [%s] func:%s\n", + bytes, counter.calls, ct->filename, ct->lineno, ct->modname, ct->function); else - pr_notice("%12lli %8llu %s:%u func:%s\n", - counter.bytes, counter.calls, ct->filename, + pr_notice("%12s %8llu %s:%u func:%s\n", + bytes, counter.calls, ct->filename, ct->lineno, ct->function); } }
We already do this when reporting slab info - more consistent and more readable. Cc: Suren Baghdasaryan <surenb@google.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev> --- mm/show_mem.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-)