Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Arvind Sankar <nivedita@alum.mit.edu>
Date:   Sun, 5 Apr 2020 19:18:45 -0400
To:     Sergey Shatunov <me@prok.pw>
Cc:     nivedita@alum.mit.edu, bp@alien8.de, hpa@zytor.com,
        linux-kernel@vger.kernel.org, mingo@redhat.com, tglx@linutronix.de,
        x86@kernel.org, linux-efi@vger.kernel.org
Subject: Re: [PATCH 1/2] x86/boot/compressed/64: Remove .bss/.pgtable from
 bzImage
Message-ID: <20200405231845.GA3095309@rani.riverdale.lan>
References: <20200109150218.16544-1-nivedita@alum.mit.edu>
 <20200405154245.11972-1-me@prok.pw>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20200405154245.11972-1-me@prok.pw>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Sun, Apr 05, 2020 at 10:42:46PM +0700, Sergey Shatunov wrote:
> This patch causes some strange things happens with my laptop.
> 
> Cold boot crashed in some early initilization logic with message 'Failed to execute /esp/.../linux.efi: Buffer Too Small'.
> After couple reboots into firmware setup (bios) or hot reboot from other working kernel (without that commit) helps it to boot.
> During bisecting couple times I saw different message: 'exit_efi() failed; efi_main() failed', but above tricks helps it too.
> So bisect points to that commit and I tried to compile kernel with that commit reverted and it works flawlessly.
> 
> Some notes about my setup:
> Kernel tree I used: Torvalds git (which recently pulls that commit).
> Kernel itself with initrd and cmdline packed into systemd-boot stub (probably here can be issues too, not sure).
> Secure boot enabled with custom keyring.
> 
> I can provide more info or change my setup (for example get rid of systemd-boot stub) if needed for sure.

I'm not familiar with systemd-boot: when you say systemd-boot stub, is
that something different from the kernel's EFI_STUB option? Or is it
just a kernel with EFI_STUB enabled and with builtin initramfs + builtin
cmdline? If it is different, can you point me to the tool that creates
it?

The first error message you mention should be from systemd-boot after
the kernel exits with an EFI_BUFFER_TOO_SMALL error status. The second
one -- is that 'exit_boot() failed' rather than 'exit_efi() failed'? I
can't find the latter string in the tree.

Can you also specify a commit tag that is broken + works with this patch
reverted, just to make sure we're looking at the same code?

Thanks

Cc linux-efi.