Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
[PATCH] mm: slub: fix ALLOC_SLOWPATH stat
2 views
Skip to first unread message
Dave Hansen
Jan 6, 2014, 3:50:01 PM1/6/14
to
From: Dave Hansen <dave....@linux.intel.com>
There used to be only one path out of __slab_alloc(), and
ALLOC_SLOWPATH got bumped in that exit path. Now there are two,
and a bunch of gotos. ALLOC_SLOWPATH can now get set more than once
during a single call to __slab_alloc() which is pretty bogus.
Here's the sequence:
1. Enter __slab_alloc(), fall through all the way to the
stat(s, ALLOC_SLOWPATH);
2. hit 'if (!freelist)', and bump DEACTIVATE_BYPASS, jump to
new_slab (goto #1)
3. Hit 'if (c->partial)', bump CPU_PARTIAL_ALLOC, goto redo
(goto #2)
4. Fall through in the same path we did before all the way to
stat(s, ALLOC_SLOWPATH)
5. bump ALLOC_REFILL stat, then return
Doing this is obviously bogus. It keeps us from being able to
accurately compare ALLOC_SLOWPATH vs. ALLOC_FASTPATH. It also
means that the total number of allocs always exceeds the total
number of frees.
This patch moves stat(s, ALLOC_SLOWPATH) to be called from the
same place that __slab_alloc() is. This makes it much less
likely that ALLOC_SLOWPATH will get botched again in the
spaghetti-code inside __slab_alloc().
Signed-off-by: Dave Hansen <dave....@linux.intel.com>
---
linux.git-davehans/mm/slub.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff -puN mm/slub.c~slub-ALLOC_SLOWPATH-stat mm/slub.c
--- linux.git/mm/slub.c~slub-ALLOC_SLOWPATH-stat 2014-01-06 12:39:28.148072544 -0800
+++ linux.git-davehans/mm/slub.c 2014-01-06 12:39:28.155072860 -0800
@@ -2301,8 +2301,6 @@ redo:
if (freelist)
goto load_freelist;
- stat(s, ALLOC_SLOWPATH);
-
freelist = get_freelist(s, page);
if (!freelist) {
@@ -2409,10 +2407,10 @@ redo:
object = c->freelist;
page = c->page;
- if (unlikely(!object || !node_match(page, node)))
+ if (unlikely(!object || !node_match(page, node))) {
object = __slab_alloc(s, gfpflags, node, addr, c);
-
- else {
+ stat(s, ALLOC_SLOWPATH);
+ } else {
void *next_object = get_freepointer_safe(s, object);
/*
_
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majo...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Christoph Lameter
Jan 6, 2014, 9:50:02 PM1/6/14
to
On Mon, 6 Jan 2014, Dave Hansen wrote:
> This patch moves stat(s, ALLOC_SLOWPATH) to be called from the
> same place that __slab_alloc() is. This makes it much less
> likely that ALLOC_SLOWPATH will get botched again in the
> spaghetti-code inside __slab_alloc().
Acked-by: Christoph Lameter <c...@linux.com>
> This patch moves stat(s, ALLOC_SLOWPATH) to be called from the
> same place that __slab_alloc() is. This makes it much less
> likely that ALLOC_SLOWPATH will get botched again in the
> spaghetti-code inside __slab_alloc().
David Rientjes
Jan 8, 2014, 9:30:01 PM1/8/14
to
On Mon, 6 Jan 2014, Dave Hansen wrote:
> There used to be only one path out of __slab_alloc(), and
> ALLOC_SLOWPATH got bumped in that exit path. Now there are two,
> and a bunch of gotos. ALLOC_SLOWPATH can now get set more than once
> during a single call to __slab_alloc() which is pretty bogus.
> Here's the sequence:
>
> 1. Enter __slab_alloc(), fall through all the way to the
> stat(s, ALLOC_SLOWPATH);
> 2. hit 'if (!freelist)', and bump DEACTIVATE_BYPASS, jump to
> new_slab (goto #1)
> 3. Hit 'if (c->partial)', bump CPU_PARTIAL_ALLOC, goto redo
> (goto #2)
> 4. Fall through in the same path we did before all the way to
> stat(s, ALLOC_SLOWPATH)
> 5. bump ALLOC_REFILL stat, then return
>
> Doing this is obviously bogus. It keeps us from being able to
> accurately compare ALLOC_SLOWPATH vs. ALLOC_FASTPATH. It also
> means that the total number of allocs always exceeds the total
> number of frees.
>
> This patch moves stat(s, ALLOC_SLOWPATH) to be called from the
> same place that __slab_alloc() is. This makes it much less
> likely that ALLOC_SLOWPATH will get botched again in the
> spaghetti-code inside __slab_alloc().
>
> Signed-off-by: Dave Hansen <dave....@linux.intel.com>
Acked-by: David Rientjes <rien...@google.com>
> ALLOC_SLOWPATH got bumped in that exit path. Now there are two,
> and a bunch of gotos. ALLOC_SLOWPATH can now get set more than once
> during a single call to __slab_alloc() which is pretty bogus.
> Here's the sequence:
>
> 1. Enter __slab_alloc(), fall through all the way to the
> stat(s, ALLOC_SLOWPATH);
> 2. hit 'if (!freelist)', and bump DEACTIVATE_BYPASS, jump to
> new_slab (goto #1)
> 3. Hit 'if (c->partial)', bump CPU_PARTIAL_ALLOC, goto redo
> (goto #2)
> 4. Fall through in the same path we did before all the way to
> stat(s, ALLOC_SLOWPATH)
> 5. bump ALLOC_REFILL stat, then return
>
> Doing this is obviously bogus. It keeps us from being able to
> accurately compare ALLOC_SLOWPATH vs. ALLOC_FASTPATH. It also
> means that the total number of allocs always exceeds the total
> number of frees.
>
> This patch moves stat(s, ALLOC_SLOWPATH) to be called from the
> same place that __slab_alloc() is. This makes it much less
> likely that ALLOC_SLOWPATH will get botched again in the
> spaghetti-code inside __slab_alloc().
>
> Signed-off-by: Dave Hansen <dave....@linux.intel.com>
David Rientjes
Jan 21, 2014, 5:20:05 PM1/21/14
to
nice to have for 3.14 for those of us who use the stats.
David Rientjes
May 7, 2014, 11:10:01 PM5/7/14
to
0 new messages