From: Christoph Lameter <cl@linux.com>
To: Thomas Garnier <thgarnie@google.com>
Cc: Pekka Enberg <penberg@kernel.org>,
David Rientjes <rientjes@google.com>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Andrew Morton <akpm@linux-foundation.org>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
gthelen@google.com
Subject: Re: [PATCH v1] memcg: Prevent caches to be both OFF_SLAB & OBJFREELIST_SLAB
Date: Wed, 26 Oct 2016 14:08:07 -0500 (CDT) [thread overview]
Message-ID: <alpine.DEB.2.20.1610261400270.31096@east.gentwo.org> (raw)
In-Reply-To: <1477503688-69191-1-git-send-email-thgarnie@google.com>
Hmmm...Doesnt this belong into memcg_create_kmem_cache() or into
kmem_cache_create() in mm/slab_common.h? Definitely not in an allocator
specific function since this is an issue for all allocators.
memcg_create_kmem_cache() simply assumes that it can pass flags from the
kmem_cache structure to kmem_cache_create(). However, those flags may
contain slab specific options.
kmem_cache_create() could filter out flags that cannot be specified.
Maybe create SLAB_FLAGS_PERMITTED in linux/mm/slab.h and mask other bits
out in kmem_cache_create()?
Slub also has internal flags and those also should not be passed to
kmem_cache_create(). If we define the valid ones we can mask them out.
The cleanest approach would be if kmem_cache_create() would reject invalid
flags and fail and if memcg_create_kmem_cache() would mask out the invalid
flags using SLAB_FLAGS_PERMITTED or so.
On Wed, 26 Oct 2016, Thomas Garnier wrote:
> While testing OBJFREELIST_SLAB integration with pagealloc, we found a
> bug where kmem_cache(sys) would be created with both CFLGS_OFF_SLAB &
> CFLGS_OBJFREELIST_SLAB.
>
> The original kmem_cache is created early making OFF_SLAB not possible.
> When kmem_cache(sys) is created, OFF_SLAB is possible and if pagealloc
> is enabled it will try to enable it first under certain conditions.
> Given kmem_cache(sys) reuses the original flag, you can have both flags
> at the same time resulting in allocation failures and odd behaviors.
>
> The proposed fix removes these flags by default at the entrance of
> __kmem_cache_create. This way the function will define which way the
> freelist should be handled at this stage for the new cache.
>
> Fixes: b03a017bebc4 ("mm/slab: introduce new slab management type, OBJFREELIST_SLAB")
> Signed-off-by: Thomas Garnier <thgarnie@google.com>
> Signed-off-by: Greg Thelen <gthelen@google.com>
> ---
> Based on next-20161025
> ---
> mm/slab.c | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/mm/slab.c b/mm/slab.c
> index 3c83c29..efe280a 100644
> --- a/mm/slab.c
> +++ b/mm/slab.c
> @@ -2027,6 +2027,14 @@ __kmem_cache_create (struct kmem_cache *cachep, unsigned long flags)
> int err;
> size_t size = cachep->size;
>
> + /*
> + * memcg re-creates caches with the flags of the originals. Remove
> + * the freelist related flags to ensure they are re-defined at this
> + * stage. Prevent having both flags on edge cases like with pagealloc
> + * if the original cache was created too early to be OFF_SLAB.
> + */
> + flags &= ~(CFLGS_OBJFREELIST_SLAB|CFLGS_OFF_SLAB);
> +
> #if DEBUG
> #if FORCED_DEBUG
> /*
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2016-10-26 19:09 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-26 17:41 Thomas Garnier
2016-10-26 19:08 ` Christoph Lameter [this message]
2016-10-26 19:22 ` Thomas Garnier
2016-10-26 20:47 ` Christoph Lameter
2016-10-27 7:25 ` Michal Hocko
2016-10-27 14:34 ` Thomas Garnier
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=alpine.DEB.2.20.1610261400270.31096@east.gentwo.org \
--to=cl@linux.com \
--cc=akpm@linux-foundation.org \
--cc=gthelen@google.com \
--cc=iamjoonsoo.kim@lge.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=penberg@kernel.org \
--cc=rientjes@google.com \
--cc=thgarnie@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox