On Tue 22-04-14 14:52:57, Vladimir Davydov wrote:
> On 04/22/2014 02:30 PM, Michal Hocko wrote:
> > On Fri 18-04-14 13:08:17, Andrew Morton wrote:
> > [...]
> >> From: Vladimir Davydov <vdavydov@xxxxxxxxxxxxx>
> >> Subject: slub: fix memcg_propagate_slab_attrs
> >>
> >> After creating a cache for a memcg we should initialize its sysfs attrs
> >> with the values from its parent. That's what memcg_propagate_slab_attrs
> >> is for. Currently it's broken - we clearly muddled root-vs-memcg caches
> >> there. Let's fix it up.
> >
> > Andrew didn't so I'll do. What is the effect of the mismatch? I am
> > really drowning in that code...
>
> If we tune a kmem cache's params via sysfs and then create a memcg that
> wants to allocate from the cache, the memcg's copy of the cache will
> have default values of the sysfs params instead of those of the global
> cache.
Ahh, ok, I see. Thanks for the clarification.
> >> Signed-off-by: Vladimir Davydov <vdavydov@xxxxxxxxxxxxx>
> >> Cc: Christoph Lameter <cl@xxxxxxxxx>
> >> Cc: Pekka Enberg <penberg@xxxxxxxxxx>
> >> Cc: Michal Hocko <mhocko@xxxxxxx>
> >> Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
> >> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> >> ---
> >>
> >> mm/slub.c | 11 +++++++----
> >> 1 file changed, 7 insertions(+), 4 deletions(-)
> >>
> >> diff -puN mm/slub.c~slub-fix-memcg_propagate_slab_attrs mm/slub.c
> >> --- a/mm/slub.c~slub-fix-memcg_propagate_slab_attrs
> >> +++ a/mm/slub.c
> >> @@ -5071,15 +5071,18 @@ static void memcg_propagate_slab_attrs(s
> >> #ifdef CONFIG_MEMCG_KMEM
> >> int i;
> >> char *buffer = NULL;
> >> + struct kmem_cache *root_cache;
> >>
> >> - if (!is_root_cache(s))
> >> + if (is_root_cache(s))
> >> return;
> >>
> >> + root_cache = s->memcg_params->root_cache;
> >> +
> >> /*
> >> * This mean this cache had no attribute written. Therefore, no point
> >> * in copying default values around
> >> */
> >> - if (!s->max_attr_size)
> >> + if (!root_cache->max_attr_size)
> >> return;
> >>
> >> for (i = 0; i < ARRAY_SIZE(slab_attrs); i++) {
> >> @@ -5101,7 +5104,7 @@ static void memcg_propagate_slab_attrs(s
> >> */
> >> if (buffer)
> >> buf = buffer;
> >> - else if (s->max_attr_size < ARRAY_SIZE(mbuf))
> >> + else if (root_cache->max_attr_size < ARRAY_SIZE(mbuf))
> >> buf = mbuf;
> >> else {
> >> buffer = (char *) get_zeroed_page(GFP_KERNEL);
> >> @@ -5110,7 +5113,7 @@ static void memcg_propagate_slab_attrs(s
> >> buf = buffer;
> >> }
> >>
> >> - attr->show(s->memcg_params->root_cache, buf);
> >> + attr->show(root_cache, buf);
> >> attr->store(s, buf, strlen(buf));
> >> }
> >>
> >> _
> >>
> >> Patches currently in -mm which might be from vdavydov@xxxxxxxxxxxxx are
> >>
> >> slub-fix-memcg_propagate_slab_attrs.patch
> >> slb-charge-slabs-to-kmemcg-explicitly.patch
> >> mm-get-rid-of-__gfp_kmemcg.patch
> >> mm-get-rid-of-__gfp_kmemcg-fix.patch
> >> slab-document-kmalloc_order.patch
> >>
> >
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@xxxxxxxxx. For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>
--
Michal Hocko
SUSE Labs
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>