Re: [ 13/40] vmscan: shrinker->nr updates race and go wrong

From: Mel Gorman
Date: Mon Jul 30 2012 - 05:07:20 EST


On Sun, Jul 29, 2012 at 09:29:30PM +0100, Ben Hutchings wrote:
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -251,17 +251,29 @@ unsigned long shrink_slab(struct shrink_
> > unsigned long total_scan;
> [...]
> > + total_scan += delta;
> > + if (total_scan < 0) {
> [...]
>
> This condition is never true since total_scan is unsigned.
> I think 3.0.y needs this as well:
>
> commit 635697c663f38106063d5659f0cf2e45afcd4bb5
> Author: Konstantin Khlebnikov <khlebnikov@xxxxxxxxxx>
> Date: Thu Dec 8 14:33:51 2011 -0800
>
> vmscan: fix initial shrinker size handling
>

You're right, thanks for pointing it out as I missed it. Greg, this commit
is already in 3.2 but can you pick it up for 3.0-stable please?

---8<---
vmscan: fix initial shrinker size handling

commit 635697c663f38106063d5659f0cf2e45afcd4bb5 upstream.

Stable note: The commit [acf92b48: vmscan: shrinker->nr updates race and
go wrong] aimed to reduce excessive reclaim of slab objects but
had bug in how it treated shrinker functions that returned -1.

A shrinker function can return -1, means that it cannot do anything
without a risk of deadlock. For example prune_super() does this if it
cannot grab a superblock refrence, even if nr_to_scan=0. Currently we
interpret this -1 as a ULONG_MAX size shrinker and evaluate `total_scan'
according to this. So the next time around this shrinker can cause
really big pressure. Let's skip such shrinkers instead.

Also make total_scan signed, otherwise the check (total_scan < 0) below
never works.

Signed-off-by: Konstantin Khlebnikov <khlebnikov@xxxxxxxxxx>
Cc: Dave Chinner <david@xxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Mel Gorman <mgorman@xxxxxxx>
---
mm/vmscan.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index e0afff3..bd8a540 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -248,12 +248,16 @@ unsigned long shrink_slab(struct shrink_control *shrink,

list_for_each_entry(shrinker, &shrinker_list, list) {
unsigned long long delta;
- unsigned long total_scan;
- unsigned long max_pass;
+ long total_scan;
+ long max_pass;
int shrink_ret = 0;
long nr;
long new_nr;

+ max_pass = do_shrinker_shrink(shrinker, shrink, 0);
+ if (max_pass <= 0)
+ continue;
+
/*
* copy the current shrinker scan count into a local variable
* and zero it so that other concurrent shrinker invocations
@@ -264,7 +268,6 @@ unsigned long shrink_slab(struct shrink_control *shrink,
} while (cmpxchg(&shrinker->nr, nr, 0) != nr);

total_scan = nr;
- max_pass = do_shrinker_shrink(shrinker, shrink, 0);
delta = (4 * nr_pages_scanned) / shrinker->seeks;
delta *= max_pass;
do_div(delta, lru_pages + 1);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/