linux-kernel - Re: [PATCH UPDATED 3/7] workqueue: Remove cpus_read_lock() from apply_wqattrs

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <xesrnvzs4xahy4473mp5vzxemjdtrci457x76si6gqbte3xx7k@ubfomqay7tck>
Date: Mon, 15 Jul 2024 11:13:56 -0400
From: Daniel Jordan <daniel.m.jordan@...cle.com>
To: Tejun Heo <tj@...nel.org>
Cc: Lai Jiangshan <jiangshanlai@...il.com>, linux-kernel@...r.kernel.org,
        Lai Jiangshan <jiangshan.ljs@...group.com>
Subject: Re: [PATCH UPDATED 3/7] workqueue: Remove cpus_read_lock() from
 apply_wqattrs_lock()

On Thu, Jul 11, 2024 at 07:11:16AM GMT, Tejun Heo wrote:
> From 49d94fbe1f8c275c793e298e8a38278e90029cf6 Mon Sep 17 00:00:00 2001
> From: Lai Jiangshan <jiangshan.ljs@...group.com>
> Date: Thu, 11 Jul 2024 16:35:43 +0800
> Subject: [PATCH] workqueue: Remove cpus_read_lock() from apply_wqattrs_lock()
> 
> 1726a1713590 ("workqueue: Put PWQ allocation and WQ enlistment in the same
> lock C.S.") led to the following possible deadlock:
> 
>   WARNING: possible recursive locking detected
>   6.10.0-rc5-00004-g1d4c6111406c #1 Not tainted
>    --------------------------------------------
>    swapper/0/1 is trying to acquire lock:
>    c27760f4 (cpu_hotplug_lock){++++}-{0:0}, at: alloc_workqueue (kernel/workqueue.c:5152 kernel/workqueue.c:5730)
> 
>    but task is already holding lock:
>    c27760f4 (cpu_hotplug_lock){++++}-{0:0}, at: padata_alloc (kernel/padata.c:1007)
>    ...
>    stack backtrace:
>    ...
>    cpus_read_lock (include/linux/percpu-rwsem.h:53 kernel/cpu.c:488)
>    alloc_workqueue (kernel/workqueue.c:5152 kernel/workqueue.c:5730)
>    padata_alloc (kernel/padata.c:1007 (discriminator 1))
>    pcrypt_init_padata (crypto/pcrypt.c:327 (discriminator 1))
>    pcrypt_init (crypto/pcrypt.c:353)
>    do_one_initcall (init/main.c:1267)
>    do_initcalls (init/main.c:1328 (discriminator 1) init/main.c:1345 (discriminator 1))
>    kernel_init_freeable (init/main.c:1364)
>    kernel_init (init/main.c:1469)
>    ret_from_fork (arch/x86/kernel/process.c:153)
>    ret_from_fork_asm (arch/x86/entry/entry_32.S:737)
>    entry_INT80_32 (arch/x86/entry/entry_32.S:944)
> 
> This is caused by pcrypt allocating a workqueue while holding
> cpus_read_lock(), so workqueue code can't do it again as that can lead to
> deadlocks if down_write starts after the first down_read.
> 
> The pwq creations and installations have been reworked based on
> wq_online_cpumask rather than cpu_online_mask making cpus_read_lock() is
> unneeded during wqattrs changes. Fix the deadlock by removing
> cpus_read_lock() from apply_wqattrs_lock().

pcrypt/padata doesn't need to hold cpus_read_lock during
alloc_workqueue.

I didn't look closely at the workqueue changes that avoid this issue,
but I can remove the restriction in padata if it helps to reduce
complexity in workqueue.

I saw the recent wq pull though, so if it's fine as is, that's ok too.