[Buildroot] [PATCH] board/qemu/sh4-r2d: fix sh4 kernel bug with Binutils 2.33

Romain Naour romain.naour at gmail.com
Mon Dec 30 14:16:54 UTC 2019


Hello Thomas,

Le 30/12/2019 à 14:53, Thomas Petazzoni a écrit :
> Hello Romain,
> 
> On Mon, 30 Dec 2019 14:36:40 +0100
> Romain Naour <romain.naour at gmail.com> wrote:
> 
>> Remove the Binutils patch reverting [1] that trigger a sh4 kernel bug
>> with Binutils >= 2.33.
>> Add two kernel patch provided by Alan Modra [2] that fix alignment of rodata.
>>
>> [1] https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=ebd2263ba9a9124d93bbc0ece63d7e0fae89b40e
>> [2] https://www.sourceware.org/ml/binutils/2019-12/msg00112.html
>>
>> Signed-off-by: Romain Naour <romain.naour at gmail.com>
> 
> Thanks a lot for those fixes. Have the kernel patches been submitted
> upstream, and accepted?

Not yet, I replied to Alan Modra to do so [1].

> 
> One issue with your change is of course that now the fix only works
> when building the specific qemu_sh4 defconfigs. When not using one of
> those defconfigs, but simply building the regular Linux kernel for SH4,
> will lead to a kernel that no longer boots.

Indeed but I don't think is great to keep the non-upstream patch I've made
reverting [2] for Binutils >= 2.33 for all architectures just to workaround a
sh4 kernel bug. At the time I didn't know that was a sh4 kernel bug...

The two sh4 kernel patches are probably easy to backport as soon as a sh4
toolchain using Binutils >= 2.33 is used.

[1] https://www.sourceware.org/ml/binutils/2019-12/msg00386.html
[2]
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=commitdiff;h=ebd2263ba9a9124d93bbc0ece63d7e0fae89b40e

Best regards,
Romain

> 
> Thomas
> 



More information about the buildroot mailing list