linux-kernel - Re: [PATCH 2/3] efi: stmm: Use EFI return code of setup_mm

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: <CAC_iWjJ5bogQwhqGRVBB59=J6wcB0q5h5B-LkiTkZOmwN4z-2Q@mail.gmail.com>
Date: Thu, 21 Aug 2025 16:26:33 +0300
From: Ilias Apalodimas <ilias.apalodimas@...aro.org>
To: Jan Kiszka <jan.kiszka@...mens.com>
Cc: Ard Biesheuvel <ardb@...nel.org>, Masahisa Kojima <masahisa.kojima@...aro.org>, 
	linux-efi@...r.kernel.org, linux-kernel@...r.kernel.org, 
	Sumit Garg <sumit.garg@...aro.org>, Jens Wiklander <jens.wiklander@...aro.org>
Subject: Re: [PATCH 2/3] efi: stmm: Use EFI return code of setup_mm_hdr

On Thu, 21 Aug 2025 at 16:00, Jan Kiszka <jan.kiszka@...mens.com> wrote:
>
> On 20.08.25 16:59, Jan Kiszka wrote:
> > On 20.08.25 09:10, Ilias Apalodimas wrote:
> >> Hi Jan
> >>
> >> On Fri, 15 Aug 2025 at 22:12, Jan Kiszka <jan.kiszka@...mens.com> wrote:
> >>>
> >>> From: Jan Kiszka <jan.kiszka@...mens.com>
> >>>
> >>> If a too large payload_size is passed to setup_mm_hdr, callers will
> >>> returned EFI_OUT_OF_RESOURCES rather than EFI_INVALID_PARAMETER that is
> >>> passed down via ret. No need to fold errors here.
> >>
> >> Apart from not folding the error here, the current code kind of
> >> violates the EFI spec.
> >> If you look at GetVariable, GetNextVariable, SetVariable, and
> >> QueryVariableInfo only SetVariable is supposed to return
> >> EFI_OUT_OF_RESOURCES, if there's no storage space left.
> >
> > And with storage space is likely meant the persistent part of it. ENOMEM
> > is something different.
> >
> >>
> >> Should we also change setup_mm_hdr() and return EFI_INVALID_PARAMETER
> >> always? It's still not ideal, but much closer to the spec.
> >
> > EFI_DEVICE_ERROR? The "hardware" is has a problem by not providing us
> > enough RAM. Yeah, not optimal either. But invalid parameter is clearly
> > described, and nothing fits.
> >
>
> If there are no concerns, I will switch to EFI_DEVICE_ERROR and even
> drop the error "ret" argument in v2.

Yea, I don't think there's an ideal scenario and the EFI spec doesn't
cover the case where some allocation failed, but please add this info
on the commit message.

Thanks
/Ilias
>
> Jan
>
> --
> Siemens AG, Foundational Technologies
> Linux Expert Center