[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ZheioNCrbeT+0UHR@gmail.com>
Date: Thu, 11 Apr 2024 10:43:12 +0200
From: Ingo Molnar <mingo@...nel.org>
To: Nikita Kiryushin <kiryushin@...ud.ru>
Cc: Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>,
Borislav Petkov <bp@...en8.de>,
Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
"H. Peter Anvin" <hpa@...or.com>,
Peter Zijlstra <peterz@...radead.org>,
Ashok Raj <ashok.raj@...el.com>,
David Woodhouse <dwmw@...zon.co.uk>, linux-kernel@...r.kernel.org,
lvc-project@...uxtesting.org
Subject: Re: [PATCH] x86/smpboot: Add map vars allocation check in
smp_prepare_cpus_common
* Nikita Kiryushin <kiryushin@...ud.ru> wrote:
> But I have some questions considering __GFP_NOFAIL. It clearly shows,
> that allocation will not fail/is not expected to fail, does it not try
> making allocation until it succeeds? Would not it make the system hang in
> a problematic case?
Yes, which is better than undefined behavior. This case should only happen
on hopelessly buggy kernels or hopelessly buggy hardware - which will very
likely crash way sooner before they get into this part of the init
sequence.
This whole problem is a non-issue.
Thanks,
Ingo
Powered by blists - more mailing lists