On Mon, 18 Dec 2006 23:38:23 +0100
"Rafael J. Wysocki" <rjw at sisk.pl> wrote:
> > > Looks like we have a problem with slab shrinking here.
> > >
> > > Could you please use gdb to check what exactly is at shrink_slab+0x9e?
> >
> > Sure, but not till Friday, sorry (I am away).
>
> I reproduced this on one box, but then it turned out that EIP was at line 195
> of mm/vmscan.c where there was
>
> do_div(delta, lru_pages + 1);
That implies that we passed it lru_pages=-1.
Presumably the logic in
vmscanc-account-for-memory-already-freed-in-seeking-to.patch caused that.
> Well, I have no idea how this can lead to a divide error (lru_pages is
> unsigned).
>
> I'm unable to reproduce this on another i386 box, so it seems to be somewhat
> configuration specific.
>
There is one wart in shrink_all_memory() and I think we should fix that in
2.6.20.
Please check the below. I'll drop
vmscanc-account-for-memory-already-freed-in-seeking-to.patch. It has other
stuff in it which we might still need. But altering sc->swap_cluster_max
in that manner looks odd.
From: Andrew Morton <akpm at osdl.org>
At the end of shrink_all_memory() we forget to recalculate lru_pages: it can
be zero.
Fix that up, and add a helper function for this operation too.
Also, recalculate lru_pages each time around the inner loop to get the
balancing correct.
Cc: "Rafael J. Wysocki" <rjw at sisk.pl>
Signed-off-by: Andrew Morton <akpm at osdl.org>
---
mm/vmscan.c | 33 ++++++++++++++++-----------------
1 files changed, 16 insertions(+), 17 deletions(-)
diff -puN mm/vmscan.c~shrink_all_memory-fix-lru_pages-handling mm/vmscan.c
--- a/mm/vmscan.c~shrink_all_memory-fix-lru_pages-handling
+++ a/mm/vmscan.c
@@ -1484,6 +1484,16 @@ static unsigned long shrink_all_zones(un
return ret;
}
+static unsigned long count_lru_pages(void)
+{
+ struct zone *zone;
+ unsigned long ret = 0;
+
+ for_each_zone(zone);
+ ret += zone->nr_active + zone->nr_inactive;
+ return ret;
+}
+
/*
* Try to free `nr_pages' of memory, system-wide, and return the number of
* freed pages.
@@ -1498,7 +1508,6 @@ unsigned long shrink_all_memory(unsigned
unsigned long ret = 0;
int pass;
struct reclaim_state reclaim_state;
- struct zone *zone;
struct scan_control sc = {
.gfp_mask = GFP_KERNEL,
.may_swap = 0,
@@ -1509,10 +1518,7 @@ unsigned long shrink_all_memory(unsigned
current->reclaim_state = &reclaim_state;
- lru_pages = 0;
- for_each_zone(zone)
- lru_pages += zone->nr_active + zone->nr_inactive;
-
+ lru_pages = count_lru_pages();
nr_slab = global_page_state(NR_SLAB_RECLAIMABLE);
/* If slab caches are huge, it's better to hit them first */
while (nr_slab >= lru_pages) {
@@ -1539,13 +1545,6 @@ unsigned long shrink_all_memory(unsigned
for (pass = 0; pass < 5; pass++) {
int prio;
- /* Needed for shrinking slab caches later on */
- if (!lru_pages)
- for_each_zone(zone) {
- lru_pages += zone->nr_active;
- lru_pages += zone->nr_inactive;
- }
-
/* Force reclaiming mapped pages in the passes #3 and #4 */
if (pass > 2) {
sc.may_swap = 1;
@@ -1561,7 +1560,8 @@ unsigned long shrink_all_memory(unsigned
goto out;
reclaim_state.reclaimed_slab = 0;
- shrink_slab(sc.nr_scanned, sc.gfp_mask, lru_pages);
+ shrink_slab(sc.nr_scanned, sc.gfp_mask,
+ count_lru_pages());
ret += reclaim_state.reclaimed_slab;
if (ret >= nr_pages)
goto out;
@@ -1569,20 +1569,19 @@ unsigned long shrink_all_memory(unsigned
if (sc.nr_scanned && prio < DEF_PRIORITY - 2)
congestion_wait(WRITE, HZ / 10);
}
-
- lru_pages = 0;
}
/*
* If ret = 0, we could not shrink LRUs, but there may be something
* in slab caches
*/
- if (!ret)
+ if (!ret) {
do {
reclaim_state.reclaimed_slab = 0;
- shrink_slab(nr_pages, sc.gfp_mask, lru_pages);
+ shrink_slab(nr_pages, sc.gfp_mask, count_lru_pages());
ret += reclaim_state.reclaimed_slab;
} while (ret < nr_pages && reclaim_state.reclaimed_slab > 0);
+ }
out:
current->reclaim_state = NULL;
_
