From: Yonghong Song <yhs@fb•com>
To: Daniel Borkmann <daniel@iogearbox•net>, <ast@fb•com>,
<malat@debian•org>, <netdev@vger•kernel.org>
Cc: <kernel-team@fb•com>
Subject: Re: [PATCH bpf] bpf: fix memory leak in lpm_trie map_free callback function
Date: Tue, 13 Feb 2018 17:46:23 -0800 [thread overview]
Message-ID: <99733349-d2fd-99bf-ff18-94f65deb02fe@fb.com> (raw)
In-Reply-To: <b9174898-fe82-0088-3b5d-cf8e5daa829b@iogearbox.net>
On 2/13/18 5:11 PM, Daniel Borkmann wrote:
> Hi Yonghong,
>
> On 02/12/2018 10:58 PM, Yonghong Song wrote:
>> There is a memory leak happening in lpm_trie map_free callback
>> function trie_free. The trie structure itself does not get freed.
>>
>> Also, trie_free function did not do synchronize_rcu before freeing
>> various data structures. This is incorrect as some rcu_read_lock
>> region(s) for lookup, update, delete or get_next_key may not complete yet.
>> The fix is to add synchronize_rcu in the beginning of trie_free.
>> The useless spin_lock is removed from this function as well.
>>
>> Fixes: b95a5c4db09b ("bpf: add a longest prefix match trie map implementation")
>> Reported-by: Mathieu Malaterre <malat@debian•org>
>> Reported-by: Alexei Starovoitov <ast@kernel•org>
>> Tested-by: Mathieu Malaterre <malat@debian•org>
>> Signed-off-by: Yonghong Song <yhs@fb•com>
>> ---
>> kernel/bpf/lpm_trie.c | 9 +++++++--
>> 1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/bpf/lpm_trie.c b/kernel/bpf/lpm_trie.c
>> index 7b469d1..9b41ea4 100644
>> --- a/kernel/bpf/lpm_trie.c
>> +++ b/kernel/bpf/lpm_trie.c
>> @@ -555,7 +555,12 @@ static void trie_free(struct bpf_map *map)
>> struct lpm_trie_node __rcu **slot;
>> struct lpm_trie_node *node;
>>
>> - raw_spin_lock(&trie->lock);
>> + /* at this point bpf_prog->aux->refcnt == 0 and this map->refcnt == 0,
>> + * so the programs (can be more than one that used this map) were
>> + * disconnected from events. Wait for outstanding programs to complete
>> + * update/lookup/delete/get_next_key and free the trie.
>> + */
>
> I think the first part of the comment is slightly misleading, e.g. map refcount
> could drop to zero also for various other reasons, not strictly due to prog
> refcnt dropping to 0, so I would just keep the second part with 'Wait for
Oh, yes. Make sense. Copy-paste without thinking.
We have similar comments in virtually all other places of using
synchronize_rcu under kernel/bpf. This can be cleaned up later though.
> outstanding [...]' which is okay since this is eventually what is relevant in
> this context.
Will do.
>
>> + synchronize_rcu();
>>
>> /* Always start at the root and walk down to a node that has no
>> * children. Then free that node, nullify its reference in the parent
>> @@ -588,7 +593,7 @@ static void trie_free(struct bpf_map *map)
>> }
>>
>> unlock:
>> - raw_spin_unlock(&trie->lock);
>
> Could you do a minor change here: since we get rid of the locking as there's
> no user left anymore after grace period, it would be great if you could also
> change the 'unlock' label name above.
Will do.
>
> Other than that, good to go, thanks!
>
>> + kfree(trie);
>> }
>>
>> static int trie_get_next_key(struct bpf_map *map, void *_key, void *_next_key)
>>
>
> Thanks,
> Daniel
>
prev parent reply other threads:[~2018-02-14 1:47 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-02-12 21:58 [PATCH bpf] bpf: fix memory leak in lpm_trie map_free callback function Yonghong Song
2018-02-12 22:15 ` Eric Dumazet
2018-02-12 22:20 ` Eric Dumazet
2018-02-14 1:11 ` Daniel Borkmann
2018-02-14 1:46 ` Yonghong Song [this message]
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=99733349-d2fd-99bf-ff18-94f65deb02fe@fb.com \
--to=yhs@fb$(echo .)com \
--cc=ast@fb$(echo .)com \
--cc=daniel@iogearbox$(echo .)net \
--cc=kernel-team@fb$(echo .)com \
--cc=malat@debian$(echo .)org \
--cc=netdev@vger$(echo .)kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
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