linux-kernel - Re: [patch V2 26/45] genirq/chip: Rework irq_set

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: <87bjrxlllc.ffs@tglx>
Date: Mon, 12 May 2025 20:39:43 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Jiri Slaby <jirislaby@...nel.org>, Nathan Chancellor
 <nathan@...nel.org>, Miguel Ojeda <ojeda@...nel.org>
Cc: linux-kernel@...r.kernel.org, peterz@...radead.org
Subject: Re: [patch V2 26/45] genirq/chip: Rework irq_set_handler() variants

On Mon, May 12 2025 at 09:25, Jiri Slaby wrote:
> On 11. 05. 25, 19:49, Nathan Chancellor wrote:
>>   {
>> -	class_irqdesc_lock_t _t = {
>> -		.bus	= bus,
>> -		.lock	= __irq_get_desc_lock(irq, &_t.flags, bus, check),
>
> I assume the value stored by __irq_get_desc_lock() to &_t.flags is 
> overwritten by 0 by the initializer. class_irqdesc_lock_t::flags is 
> later than ::lock in the structure, so __irq_get_desc_lock() should be 
> called, setting ::flags, then the initializer should set flags to 0.

That's what you'd assume. But indeed the compiler might decide that
flags is uninitialized and then overwrite it after the initialized
fields :(

Nathan, can you please send a patch with a proper change log?

Thanks,

        tglx