linux-kernel - Re: [PATCH 2/2] arm64/mm: Update create_kpti_ng_temp_pgd() to handle pgtable

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] [day] [month] [year] [list]

Message-ID: <17e83e3c-1314-4198-82eb-ffc18454e344@gmail.com>
Date: Tue, 19 Aug 2025 12:44:09 +0400
From: Giorgi Tchankvetadze <giorgitchankvetadze1997@...il.com>
To: David Hildenbrand <david@...hat.com>,
 Chaitanya S Prakash <chaitanyas.prakash@....com>,
 linux-arm-kernel@...ts.infradead.org, linux-mm@...ck.org,
 linux-kernel@...r.kernel.org
Cc: Ryan Roberts <ryan.roberts@....com>,
 Yang Shi <yang@...amperecomputing.com>,
 Catalin Marinas <catalin.marinas@....com>, Will Deacon <will@...nel.org>,
 Kevin Brodsky <kevin.brodsky@....com>,
 Anshuman Khandual <anshuman.khandual@....com>,
 Andrew Morton <akpm@...ux-foundation.org>,
 Zhenhua Huang <quic_zhenhuah@...cinc.com>, Joey Gouly <joey.gouly@....com>
Subject: Re: [PATCH 2/2] arm64/mm: Update create_kpti_ng_temp_pgd() to handle
 pgtable_alloc failure

Smart! BUG_ON() ends up in the generic “kernel bug” path, which on many 
distros is configured to continue after printing the back-trace (e.g. 
panic_on_oops=0).
Since a memory-allocation failure in early boot is unrecoverable , we 
must force a halt.

On 8/19/2025 11:41 AM, David Hildenbrand wrote:
> On 13.08.25 16:56, Chaitanya S Prakash wrote:
>> create_kpti_ng_temp_pgd() was created as an alias for void returning
>> __create_pgd_mapping_locked() and relied on pgtable_alloc() to BUG_ON()
>> if an allocation failure occurred. But as __create_pgd_mapping_locked()
>> has been updated as a part of the error propagation patch to return a
>> non-void value, update create_kpti_ng_temp_pgd() to act as a wrapper
>> around __create_pgd_mapping_locked() and BUG_ON() on ret being a non
>> zero value.
> 
> If  my memory serves me right, panic() is preferred in such unexpected 
> early-boot scenarios (BUG_ON is frowned upon), where you can actually 
> print what is going wrong.