linux-kernel - Re: [patch V2 06/45] genirq/proc: Switch to lock guards

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: <878qm1p7l3.ffs@tglx>
Date: Mon, 09 Jun 2025 11:50:16 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Zenghui Yu <yuzenghui@...wei.com>
Cc: LKML <linux-kernel@...r.kernel.org>, Jiri Slaby <jirislaby@...nel.org>,
 Peter Zijlstra <peterz@...radead.org>
Subject: Re: [patch V2 06/45] genirq/proc: Switch to lock guards

On Sun, Jun 08 2025 at 20:45, Zenghui Yu wrote:
> On 2025/4/29 14:54, Thomas Gleixner wrote:
>> -	raw_spin_lock_irqsave(&desc->lock, flags);
>> -	if (desc->affinity_hint)
>> -		cpumask_copy(mask, desc->affinity_hint);
>> -	raw_spin_unlock_irqrestore(&desc->lock, flags);
>> +	scoped_guard(raw_spinlock_irq, &desc->lock) {
>
> Any reason it has been switched to a raw_spinlock_irq?

Yes. This code is always thread context and can never be invoked with
interrupts disabled. So there is zero reason to use irqsave().

> I've hit some random Oops with the backtrace looks like:
>
>  Call trace:
>   string+0x110/0x3b8 (P)
>   vsnprintf+0x2f0/0xac8
>   seq_printf+0x180/0x220
>   show_interrupts+0x4e0/0x7e0
>   seq_read_iter+0x350/0xd80
>   proc_reg_read_iter+0x194/0x248
>   vfs_read+0x5b0/0x940
>   ksys_read+0xf0/0x1e8
>   __arm64_sys_read+0x74/0xb0
>   invoke_syscall+0x74/0x270
>   el0_svc_common.constprop.0+0xb4/0x240
>   do_el0_svc+0x48/0x68
>   el0_svc+0x4c/0xe8
>   el0t_64_sync_handler+0xc8/0xd0
>   el0t_64_sync+0x1ac/0x1b0
>
> I haven't dig further. But it looks to me that this patch had introduced
> functional change and I'm planning to give the following diff a go on
> the same box.

That won't help at all because the actual crash is within show_interrupts()....

Thanks,

        tglx