linux-kernel - Re: [PATCH 1/2] x86/boot/compressed/64: Remove .bss/.pgtable from bzImage

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:   Mon, 6 Apr 2020 10:47:38 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     Ard Biesheuvel <ardb@...nel.org>
Cc:     Arvind Sankar <nivedita@...m.mit.edu>,
        Sergey Shatunov <me@...k.pw>, hpa@...or.com,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        mingo@...hat.com, Thomas Gleixner <tglx@...utronix.de>,
        x86@...nel.org, linux-efi <linux-efi@...r.kernel.org>,
        initramfs@...r.kernel.org,
        Donovan Tremura <neurognostic@...tonmail.ch>,
        Harald Hoyer <harald@...er.xyz>
Subject: Re: [PATCH 1/2] x86/boot/compressed/64: Remove .bss/.pgtable from
 bzImage

On Mon, Apr 06, 2020 at 09:32:47AM +0200, Ard Biesheuvel wrote:
> The EFI handover protocol strikes again :-(
> 
> It seems we did not include any guidance in the documentation in
> Documentation/x86/boot.rst regarding zero-initializing BSS, and come
> to think of it, we don't include any other requirements either, i.e.,
> regarding placement wrt section alignment etc. This is a serious bug.
> Even though EFI usually lays out PE/COFF images in files the exact way
> they appear in memory, this is not actually required by the spec. Most
> notably, the virtual size can be smaller than the file size, and the
> loader is expected to zero-initialize the difference as well.

Is that expectation stated explicitly somewhere?

> Since the EFI handover protocol should be considered deprecated at
> this point (and is never going to be supported in upstream GRUB
> either, for instance), I would recommend the systemd-boot developers
> to start looking into deprecating this as well, and switch to the
> ordinary PE/COFF entry point, and use the new initrd callback protocol
> for initrd loading.

Any pointers to that new initrd callback protocol?

In any case, I'd really appreciate a patch to boot.rst formulating those
requirements so that they're written down and people can find them.

> On the Linux/x86 side, we should at least add some code to the EFI
> handover protocol entry point to zero initialize BSS, and ensure that
> it is either not needed in other places, or add the code to deal with
> those as well.

Sounds like a simple fix, if that would fix it.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette