[U-Boot] [PATCH v3 11/11] sunxi: Add PSCI implementation in C

[Hans de Goede]

Hi,

On 13-06-16 12:58, Chen-Yu Tsai wrote:
> On Mon, Jun 13, 2016 at 5:49 PM, Hans de Goede <hdegoede at redhat.com> wrote:
>> Hi,
>>
>> On 13-06-16 04:50, Chen-Yu Tsai wrote:
>>>
>>> On Fri, Jun 10, 2016 at 9:40 PM, Hans de Goede <hdegoede at redhat.com>
>>> wrote:
>>>>
>>>> Hi,
>>>>
>>>> On 07-06-16 04:54, Chen-Yu Tsai wrote:
>>>>>
>>>>>
>>>>> To make the PSCI backend more maintainable and easier to port to newer
>>>>> SoCs, rewrite the current PSCI implementation in C.
>>>>>
>>>>> Some inline assembly bits are required to access coprocessor registers.
>>>>> PSCI stack setup is the only part left completely in assembly. In theory
>>>>> this part could be split out of psci_arch_init into a separate common
>>>>> function, and psci_arch_init could be completely in C.
>>>>>
>>>>> Signed-off-by: Chen-Yu Tsai <wens at csie.org>
>>>>
>>>>
>>>>
>>>> I tried merging this in my tree to add it to u-boot-sunxi/next, but
>>>> unfortunately it triggers a bug in gcc-6.1, I've filed a bug with
>>>> gcc to get this fixed:
>>>>
>>>> https://bugzilla.redhat.com/show_bug.cgi?id=1344717
>>>
>>>
>>> Interesting. Seems like the compiler should be emitting LDMFD instead
>>> of POP, or a POP followed by a SUBS/MOVS.
>>>
>>>>
>>>> Also cp15_read_scr / cp15_write_scr are missing __secure notations.
>>>
>>>
>>> Do you want me to send a new version for this?
>>
>>
>> No, I already have a local branch with this fixed.
>>
>> As for getting this merged and Tom's "go for it"
>> comment earlier, I consider this a blocker for merging, sorry.
>>
>> I hope that the gcc folks can resolve this soon and that I can
>> get the Fedora devs to deploy a fix soon-ish. I'm not sure what
>> other distros are doing wrt jumping to gcc-6.1, but having code
>> in u-boot master which does not build with the latest gcc seems
>> like a bad idea.
>
> This is marked as resolved upstream:
>
>     https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70830
>
> And Debian/Ubuntu seems to have pulled the fix into their latest
> unstable package:
>
>     http://metadata.ftp-master.debian.org/changelogs//main/g/gcc-6/gcc-6_6.1.1-6_changelog
>
> I'm not familiar with Fedora, though you could reference the GCC
> PR I listed above and have them pull in the fix.

Yes it indeed is that bug, thanks for finding that. I've done a local
Fedora cross-gcc build with the patches from that bug applied and that
fixes things.

I've added these patches to my sunxi-wip branch now and I'll send
them out in my next pull-req, after the dust surrounding the
BOOTDELAY changes settles.

> Interesting that Fedora only provides one version of GCC though.

Fedora always contains the latest gcc, glibc, etc. Given that it is
Red Hat's community distro and that we employ a lot of gcc / glibc
developers we try to follow upstream as close as possible.


Regards,

Hans