[PATCH] arm64: errata: Check for --fix-cortex-a53-843419 and --fix-cortex-a53

On 4 January 2017 at 14:39, Florian Fainelli [off-list ref] wrote:

On 01/04/2017 03:49 AM, Will Deacon wrote:

quoted

Hi Florian,

On Wed, Dec 28, 2016 at 12:17:23PM -0800, Florian Fainelli wrote:

quoted

On 11/03/2016 10:20 AM, Florian Fainelli wrote:

quoted

On 11/03/2016 07:16 AM, Will Deacon wrote:

quoted

If you can't change toolchain and you want this worked around, why can't you
either build gold with it enabled by default, or pass the extra flag on the
command line to the kernel build system?

Because that creates a distribution problem and now we have to document
this for people who want to build a kernel on their own, without
necessarily understanding if this is something they might need, or why
this is needed, and why the kernel is not taking care of that on its
own? So yes, this comes down to who is responsible for what, in that
case the kernel's Makefile is the best place where to put such knowledge
as to which workaround needs to be enabled by the linker and it
simplifies things a lot for people.

Was this convincing enough for Catalin to pick Markus' patch or does
that mean this patch needs to remain out of tree for us because of using
a slightly older toolchain?

I thought more about this last night, and there are two questions that
might sway me:

  1. How prevalent is the binary toolchain with this issue? Is it, for
     example, shipping as part of a publicly available LTS distribution?
     I know you quoted some Linaro build, but I can't actually find those
     binaries on their website.

The toolchain is available for download from Broadcom's website at
https://www.broadcom.com/support/download-search/?pg=Broadband:+CPE-Gateway,+Infrastructure,+and+Set-top+Box&pf=Set-top+Box+Solutions
(working on the download agreement as we speak) and is used by the large
majority of our customers today, it's hard to give you numbers, but
several hundreds if not more people definitively use it AFAICT.

quoted

  2. Could we extend the Makefile magic to detect that, not only is
     --fix-cortex-a53-843419 unsupported, but also that the linker is
     in fact gold?

I suppose we could do that, Markus do you mind update your original
patch? Thanks!

Hm. Unfortunately, the linker doesn't identify as "gold." I could try
to use "Linaro 2014" or similar from the version string as an
additional test, but I don't know if that would catch all cases. It
would certainly catch ours, so it may be good enough.

$ aarch64-linux-gnu-ld -v
GNU ld (GNU Binutils) Linaro 2014.11-2 2.24.0.20141017

$ aarch64-linux-gnu-ld --fix-cortex-a53-843419
aarch64-linux-gnu-ld: unrecognized option '--fix-cortex-a53-843419'
aarch64-linux-gnu-ld: use the --help option for usage information

$ aarch64-linux-gnu-ld --fix-cortex-a53
aarch64-linux-gnu-ld: no input files

Regards,
-Markus

--
Florian