linux-kernel - Re: [PATCH] mm: shrinkers: fix race condition on debugfs cleanup

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 3 Jul 2023 12:14:22 +0100
From:   Steven Price <steven.price@....com>
To:     Joan Bruguera Micó <joanbrugueram@...il.com>,
        zhengqi.arch@...edance.com,
        "Matthew Wilcox (Oracle)" <willy@...radead.org>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Roman Gushchin <roman.gushchin@...ux.dev>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm: shrinkers: fix race condition on debugfs cleanup

On 21/05/2023 14:57, Joan Bruguera Micó wrote:
> On 2023/5/3 13:37, Qi Zheng wrote:
>>> +void shrinker_debugfs_remove(struct dentry *debugfs_entry, int debugfs_id)
>>> +{
>>
>> It would be better to add a check:
>>
>> 	if (!debugfs_entry)
>> 		return;
>>
>>> +	debugfs_remove_recursive(debugfs_entry);
>>> +	ida_free(&shrinker_debugfs_ida, debugfs_id);
>>> +}
> 
> As a practical matter, both `debugfs_remove_recursive(NULL)` and
> `ida_free(_, -1);` are documented as no-ops, see:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/fs/debugfs/inode.c?id=0dd2a6fb1e34d6dcb96806bc6b111388ad324722#n748
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=fc82bbf4dede758007763867d0282353c06d1121
> 
> Sorry for the late reply (the patch already reached the mainline tree).

Note that commit 69cb69ea5542 ("ida: Remove assertions that an ID was
allocated") currently in linux-next makes ida_free(..., -1) illegal.

I see a crash on boot on my test platform (Firefly-RK3288) with
linux-next because of this. Qi's suggested change fixes this.

I'm not sure whether Matthew's change removing the ((int)id < 0) check
was intentional or not. Reinstating that check would also fix the crash.

Steve