linux-kernel - Re: [RFC PATCH v2 1/8] arm64: Implement stack trace termination record

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: <20210318150905.GL5469@sirena.org.uk>
Date:   Thu, 18 Mar 2021 15:09:05 +0000
From:   Mark Brown <broonie@...nel.org>
To:     madvenka@...ux.microsoft.com
Cc:     mark.rutland@....com, jpoimboe@...hat.com, jthierry@...hat.com,
        catalin.marinas@....com, will@...nel.org,
        linux-arm-kernel@...ts.infradead.org,
        live-patching@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH v2 1/8] arm64: Implement stack trace termination
 record

On Mon, Mar 15, 2021 at 11:57:53AM -0500, madvenka@...ux.microsoft.com wrote:

> In summary, task pt_regs->stackframe is where a successful stack trace ends.

>         .if \el == 0
> -       mov     x29, xzr
> +       stp     xzr, xzr, [sp, #S_STACKFRAME]
>         .else
>         stp     x29, x22, [sp, #S_STACKFRAME]
> -       add     x29, sp, #S_STACKFRAME
>         .endif
> +       add     x29, sp, #S_STACKFRAME

For both user and kernel threads this patch (at least by itself) results
in an additional record being reported in stack traces with a NULL
function pointer since it keeps the existing record where it is and adds
this new fixed record below it.  This is addressed for the kernel later
in the series, by "arm64: Terminate the stack trace at TASK_FRAME and
EL0_FRAME", but will still be visible to other unwinders such as
debuggers.  I'm not sure that this *matters* but it might and should at
least be called out more explicitly.

If we are going to add the extra record there would probably be less
potential for confusion if we pointed it at some sensibly named dummy
function so anything or anyone that does see it on the stack doesn't get
confused by a NULL.

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)