linux-kernel - Re: [PATCH v2] x86/nmi: Add an emergency handler in nmi_desc & use it in nmi_shootdown

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: <9765a61a-e832-4491-af02-97b8736411ef@redhat.com>
Date: Wed, 4 Dec 2024 23:01:38 -0500
From: Waiman Long <llong@...hat.com>
To: Thomas Gleixner <tglx@...utronix.de>, Waiman Long <llong@...hat.com>,
 Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
 Dave Hansen <dave.hansen@...ux.intel.com>,
 Peter Zijlstra <peterz@...radead.org>
Cc: x86@...nel.org, linux-kernel@...r.kernel.org,
 "H. Peter Anvin" <hpa@...or.com>
Subject: Re: [PATCH v2] x86/nmi: Add an emergency handler in nmi_desc & use it
 in nmi_shootdown_cpus()

On 12/4/24 2:28 PM, Waiman Long wrote:
>
> On 12/4/24 1:03 PM, Thomas Gleixner wrote:
>> On Wed, Dec 04 2024 at 12:23, Waiman Long wrote:
>>> On 12/4/24 8:10 AM, Thomas Gleixner wrote:
>>>> On Tue, Dec 03 2024 at 10:07, Waiman Long wrote:
>>>>> +    /*
>>>>> +     * Call the emergency handler first, if set
>>>>> +     * Emergency handler is not traced or checked by 
>>>>> nmi_check_duration().
>>>>> +     */
>>>>> +    ehandler = READ_ONCE(desc->emerg_handler);
>>>>> +    if (ehandler)
>>>>> +        handled = ehandler(type, regs);
>>>> Shouldn't this just stop processing right here?
>>> Yes in the case of crash_nmi_callback(). I suppose it is a no-return
>>> call. As the emergency handler is supposed to be a general mechanism in
>>> design, I don't want to make too many assumptions of what will happen
>>> when the handler is invoked.
>> I'm not convinced that this should be used as a general mechanism. It's
>> for emergency situations and that's where it stops. If the thing
>> returns, it's a bug IMO.
>
> OK, I am fine with that. I will put a BUG_ON() after that in the next 
> version.

Actually, crash_nmi_callback() can return in the case of the crashing 
CPUs, though all the other CPUs will not return once called. So I 
believe the current form is correct. I will update the comment to 
reflect that.

Cheers,
Longman