From: Vlastimil Babka <vbabka@suse.cz>
To: Sachin Sant <sachinp@linux.vnet.ibm.com>
Cc: bharata@linux.ibm.com, Nathan Lynch <nathanl@linux.ibm.com>,
Srikar Dronamraju <srikar@linux.vnet.ibm.com>,
linuxppc-dev@lists.ozlabs.org, Michal Hocko <mhocko@kernel.org>,
Pekka Enberg <penberg@kernel.org>,
linux-mm@kvack.org, Kirill Tkhai <ktkhai@virtuozzo.com>,
David Rientjes <rientjes@google.com>,
Christopher Lameter <cl@linux.com>,
Mel Gorman <mgorman@techsingularity.net>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Michael Ellerman <mpe@ellerman.id.au>
Subject: Re: [RFC 1/2] mm, slub: prevent kmalloc_node crashes and memory leaks
Date: Thu, 19 Mar 2020 14:47:58 +0100 [thread overview]
Message-ID: <339cf655-393e-c48e-4797-86f61df56c35@suse.cz> (raw)
In-Reply-To: <658E6AB8-581F-4722-BCBB-4BDD2245D265@linux.vnet.ibm.com>
On 3/19/20 2:26 PM, Sachin Sant wrote:
>
>
>> On 19-Mar-2020, at 6:53 PM, Vlastimil Babka <vbabka@suse.cz> wrote:
>>
>> On 3/19/20 9:52 AM, Sachin Sant wrote:
>>>
>>>> OK how about this version? It's somewhat ugly, but important is that the fast
>>>> path case (c->page exists) is unaffected and another common case (c->page is
>>>> NULL, but node is NUMA_NO_NODE) is just one extra check - impossible to avoid at
>>>> some point anyway.
>>>>
>>>
>>> I attempted the suggested tests.
>>>
>>> Test 1: March 18 linux-next + Patch 1 [1] + Patch 2 [2]
>>>
>>> Machine boots fine. numactl o/p after boot:
>>
>> Great, thanks! Can I add your Tested-by: then?
>
> Sure.
> Tested-by: Sachin Sant <sachinp@linux.vnet.ibm.com>
>
> Thank you for the fix.
Thanks! Sorry to bother, but in the end I decided to do further change so I
would appreciate verification if it still works as intended.
The point is to use node_state(N_NORMAL_MEMORY) instead of node_present_pages(),
as that is really what SLUB uses to decide whether to allocate the
kmem_cache_node. So we should match this properly given the opportunity.
I have also again removed the node_online() check in alloc_slab_page() as it
really shouldn't be reachable with an offline node - everything is taken care of
in ___slab_alloc, or callers use NUMA_NO_NODE.
----8<----
diff --git a/mm/slub.c b/mm/slub.c
index 17dc00e33115..7113b1f9cd77 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -1973,8 +1973,6 @@ static void *get_partial(struct kmem_cache *s, gfp_t flags, int node,
if (node == NUMA_NO_NODE)
searchnode = numa_mem_id();
- else if (!node_present_pages(node))
- searchnode = node_to_mem_node(node);
object = get_partial_node(s, get_node(s, searchnode), c, flags);
if (object || node != NUMA_NO_NODE)
@@ -2563,17 +2561,27 @@ static void *___slab_alloc(struct kmem_cache *s, gfp_t gfpflags, int node,
struct page *page;
page = c->page;
- if (!page)
+ if (!page) {
+ /*
+ * if the node is not online or has no normal memory, just
+ * ignore the node constraint
+ */
+ if (unlikely(node != NUMA_NO_NODE &&
+ !node_state(node, N_NORMAL_MEMORY)))
+ node = NUMA_NO_NODE;
goto new_slab;
+ }
redo:
if (unlikely(!node_match(page, node))) {
- int searchnode = node;
-
- if (node != NUMA_NO_NODE && !node_present_pages(node))
- searchnode = node_to_mem_node(node);
-
- if (unlikely(!node_match(page, searchnode))) {
+ /*
+ * same as above but node_match() being false already
+ * implies node != NUMA_NO_NODE
+ */
+ if (!node_state(node, N_NORMAL_MEMORY)) {
+ node = NUMA_NO_NODE;
+ goto redo;
+ } else {
stat(s, ALLOC_NODE_MISMATCH);
deactivate_slab(s, page, c->freelist, c);
goto new_slab;
--
2.25.1
next prev parent reply other threads:[~2020-03-19 13:48 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-18 14:42 Vlastimil Babka
2020-03-18 14:42 ` [RFC 2/2] Revert "topology: add support for node_to_mem_node() to determine the fallback node" Vlastimil Babka
2020-03-18 16:06 ` [RFC 1/2] mm, slub: prevent kmalloc_node crashes and memory leaks Bharata B Rao
2020-03-18 16:10 ` Vlastimil Babka
2020-03-18 16:51 ` Vlastimil Babka
2020-03-19 8:52 ` Sachin Sant
2020-03-19 13:23 ` Vlastimil Babka
2020-03-19 13:26 ` Sachin Sant
2020-03-19 13:47 ` Vlastimil Babka [this message]
2020-03-19 14:05 ` Srikar Dronamraju
2020-03-19 14:10 ` Vlastimil Babka
2020-03-20 7:46 ` Srikar Dronamraju
2020-03-20 8:43 ` Vlastimil Babka
2020-03-20 10:10 ` Srikar Dronamraju
2020-03-19 14:59 ` Sachin Sant
2020-03-20 3:42 ` Bharata B Rao
2020-03-20 8:37 ` Vlastimil Babka
2020-03-20 8:44 ` Bharata B Rao
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=339cf655-393e-c48e-4797-86f61df56c35@suse.cz \
--to=vbabka@suse.cz \
--cc=bharata@linux.ibm.com \
--cc=cl@linux.com \
--cc=iamjoonsoo.kim@lge.com \
--cc=ktkhai@virtuozzo.com \
--cc=linux-mm@kvack.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=mgorman@techsingularity.net \
--cc=mhocko@kernel.org \
--cc=mpe@ellerman.id.au \
--cc=nathanl@linux.ibm.com \
--cc=penberg@kernel.org \
--cc=rientjes@google.com \
--cc=sachinp@linux.vnet.ibm.com \
--cc=srikar@linux.vnet.ibm.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