linux-kernel - Re: [PATCH v2 07/10] objtool: check: Allow save/restore hint in non standard function symbols

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]

Date:   Wed, 1 Apr 2020 15:06:46 +0100
From:   Julien Thierry <jthierry@...hat.com>
To:     Miroslav Benes <mbenes@...e.cz>
Cc:     linux-kernel@...r.kernel.org, jpoimboe@...hat.com,
        peterz@...radead.org, raphael.gault@....com
Subject: Re: [PATCH v2 07/10] objtool: check: Allow save/restore hint in non
 standard function symbols



On 4/1/20 2:54 PM, Miroslav Benes wrote:
> On Fri, 27 Mar 2020, Julien Thierry wrote:
> 
>> The kernel code base provides CODE_SYM_START/END to declare assembly
>> code sequences that don't follow function standard calling conventions.
>>
>> As non-C/non-standard code, these sequences can very much benefit from
>> unwind hints. However, when a restore unwind_hint is used in a
>> non-function code sequence, objtool will crash when looking for the
>> corresponding save hint.
>>
>> Record the code symbol an instruction belongs to and look for save hints
>> belonging to the same code symbol as the restore hint.
>>
>> Signed-off-by: Julien Thierry <jthierry@...hat.com>
> 
> Looks ok, but save/restore hints are about to go away. See
> https://lore.kernel.org/lkml/20200331222703.GH2452@worktop.programming.kicks-ass.net/
> 

Ah, just as I started bringing unwind hints to the arm64 side...

I'll have to scratch my head a bit more over this then. Thanks for 
bringing it to my attention.

Thanks,

-- 
Julien Thierry