mbox series

[GIT,PULL,6/9] ARM: tegra: Core changes for v5.1-rc1

Message ID 20190208150935.32108-7-thierry.reding@gmail.com
State New
Headers show
Series Tegra changes for v5.1-rc1 | expand

Pull-request

git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git tags/tegra-for-5.1-arm-core

Message

Thierry Reding Feb. 8, 2019, 3:09 p.m. UTC
Hi ARM SoC maintainers,

The following changes since commit bfeffd155283772bbe78c6a05dec7c0128ee500c:

  Linux 5.0-rc1 (2019-01-06 17:08:20 -0800)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git tags/tegra-for-5.1-arm-core

for you to fetch changes up to 1c6279b49d3ffff38357e93d112512fecf3711a7:

  ARM: tegra: Restore memory arbitration on resume from LP1 on Tegra30+ (2019-01-16 13:22:04 +0100)

Thanks,
Thierry

----------------------------------------------------------------
ARM: tegra: Core changes for v5.1-rc1

This contains three fixes for resume from LP1 on Tegra30.

----------------------------------------------------------------
Dmitry Osipenko (3):
      ARM: tegra: Fix missed EMC registers latching on resume from LP1 on Tegra30+
      ARM: tegra: Fix DRAM refresh-interval clobbering on resume from LP1 on Tegra30
      ARM: tegra: Restore memory arbitration on resume from LP1 on Tegra30+

 arch/arm/mach-tegra/iomap.h         |  9 +++++++++
 arch/arm/mach-tegra/sleep-tegra20.S |  2 --
 arch/arm/mach-tegra/sleep-tegra30.S | 25 +++++++++++++++++++++++--
 3 files changed, 32 insertions(+), 4 deletions(-)

Comments

Arnd Bergmann Feb. 15, 2019, 4:48 p.m. UTC | #1
On Fri, Feb 8, 2019 at 4:09 PM Thierry Reding <thierry.reding@gmail.com> wrote:
>
> Hi ARM SoC maintainers,
>
> The following changes since commit bfeffd155283772bbe78c6a05dec7c0128ee500c:
>
>   Linux 5.0-rc1 (2019-01-06 17:08:20 -0800)
>
> are available in the Git repository at:
>
>   git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git tags/tegra-for-5.1-arm-core
>
> for you to fetch changes up to 1c6279b49d3ffff38357e93d112512fecf3711a7:
>
>   ARM: tegra: Restore memory arbitration on resume from LP1 on Tegra30+ (2019-01-16 13:22:04 +0100)
>
> Thanks,
> Thierry
>
> ----------------------------------------------------------------
> ARM: tegra: Core changes for v5.1-rc1
>
> This contains three fixes for resume from LP1 on Tegra30.
>

Pulled into arm/soc?

Should these bugfixes have been marked Cc:stable?

      Arnd
Thierry Reding Feb. 15, 2019, 9:48 p.m. UTC | #2
On Fri, Feb 15, 2019 at 05:48:25PM +0100, Arnd Bergmann wrote:
> On Fri, Feb 8, 2019 at 4:09 PM Thierry Reding <thierry.reding@gmail.com> wrote:
> >
> > Hi ARM SoC maintainers,
> >
> > The following changes since commit bfeffd155283772bbe78c6a05dec7c0128ee500c:
> >
> >   Linux 5.0-rc1 (2019-01-06 17:08:20 -0800)
> >
> > are available in the Git repository at:
> >
> >   git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git tags/tegra-for-5.1-arm-core
> >
> > for you to fetch changes up to 1c6279b49d3ffff38357e93d112512fecf3711a7:
> >
> >   ARM: tegra: Restore memory arbitration on resume from LP1 on Tegra30+ (2019-01-16 13:22:04 +0100)
> >
> > Thanks,
> > Thierry
> >
> > ----------------------------------------------------------------
> > ARM: tegra: Core changes for v5.1-rc1
> >
> > This contains three fixes for resume from LP1 on Tegra30.
> >
> 
> Pulled into arm/soc?
> 
> Should these bugfixes have been marked Cc:stable?

I don't think that'd be worth it. As far as I can tell these have been
broken since forever. Backporting this would be fairly involved and I'm
not sure if this hurts enough to warrant the additional work, given that
it's not been reported in all this time.

Dmitry, can you clarify whether you want these to be backported?

Thierry
Dmitry Osipenko Feb. 16, 2019, 12:03 a.m. UTC | #3
16.02.2019 0:48, Thierry Reding пишет:
> On Fri, Feb 15, 2019 at 05:48:25PM +0100, Arnd Bergmann wrote:
>> On Fri, Feb 8, 2019 at 4:09 PM Thierry Reding <thierry.reding@gmail.com> wrote:
>>>
>>> Hi ARM SoC maintainers,
>>>
>>> The following changes since commit bfeffd155283772bbe78c6a05dec7c0128ee500c:
>>>
>>>   Linux 5.0-rc1 (2019-01-06 17:08:20 -0800)
>>>
>>> are available in the Git repository at:
>>>
>>>   git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git tags/tegra-for-5.1-arm-core
>>>
>>> for you to fetch changes up to 1c6279b49d3ffff38357e93d112512fecf3711a7:
>>>
>>>   ARM: tegra: Restore memory arbitration on resume from LP1 on Tegra30+ (2019-01-16 13:22:04 +0100)
>>>
>>> Thanks,
>>> Thierry
>>>
>>> ----------------------------------------------------------------
>>> ARM: tegra: Core changes for v5.1-rc1
>>>
>>> This contains three fixes for resume from LP1 on Tegra30.
>>>
>>
>> Pulled into arm/soc?
>>
>> Should these bugfixes have been marked Cc:stable?
> 
> I don't think that'd be worth it. As far as I can tell these have been
> broken since forever. Backporting this would be fairly involved and I'm
> not sure if this hurts enough to warrant the additional work, given that
> it's not been reported in all this time.
> 
> Dmitry, can you clarify whether you want these to be backported?

Technically it fixes a real problem, but realistically there is no real need to backport these patches because of other suspend-resume troubles that also existed all the time and haven't been resolved yet (likely fixes won't be backportable). In general suspend-resume kinda works and fails mostly under some load (like stress testing), at least that's the case for T30.
Dmitry Osipenko Feb. 16, 2019, 1 a.m. UTC | #4
16.02.2019 3:03, Dmitry Osipenko пишет:
> 16.02.2019 0:48, Thierry Reding пишет:
>> On Fri, Feb 15, 2019 at 05:48:25PM +0100, Arnd Bergmann wrote:
>>> On Fri, Feb 8, 2019 at 4:09 PM Thierry Reding <thierry.reding@gmail.com> wrote:
>>>>
>>>> Hi ARM SoC maintainers,
>>>>
>>>> The following changes since commit bfeffd155283772bbe78c6a05dec7c0128ee500c:
>>>>
>>>>   Linux 5.0-rc1 (2019-01-06 17:08:20 -0800)
>>>>
>>>> are available in the Git repository at:
>>>>
>>>>   git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git tags/tegra-for-5.1-arm-core
>>>>
>>>> for you to fetch changes up to 1c6279b49d3ffff38357e93d112512fecf3711a7:
>>>>
>>>>   ARM: tegra: Restore memory arbitration on resume from LP1 on Tegra30+ (2019-01-16 13:22:04 +0100)
>>>>
>>>> Thanks,
>>>> Thierry
>>>>
>>>> ----------------------------------------------------------------
>>>> ARM: tegra: Core changes for v5.1-rc1
>>>>
>>>> This contains three fixes for resume from LP1 on Tegra30.
>>>>
>>>
>>> Pulled into arm/soc?
>>>
>>> Should these bugfixes have been marked Cc:stable?
>>
>> I don't think that'd be worth it. As far as I can tell these have been
>> broken since forever. Backporting this would be fairly involved and I'm
>> not sure if this hurts enough to warrant the additional work, given that
>> it's not been reported in all this time.
>>
>> Dmitry, can you clarify whether you want these to be backported?
> 
> Technically it fixes a real problem, but realistically there is no real need to backport these patches because of other suspend-resume troubles that also existed all the time and haven't been resolved yet (likely fixes won't be backportable). In general suspend-resume kinda works and fails mostly under some load (like stress testing), at least that's the case for T30.

Actually no, I looked up the other T30 fixes and they are pretty trivial. Moreover turned out that I already sent them out before [0] and forgot about them.. probably it's a bit too late for v5.1 now.

Still it's not really worth to backport the patches given that vanilla upstream kernel can boot only on dev boards in a case of Tegra30 because there is no support for the Trusted Foundations T30 firmware yet. Hopefully we will manage to get all the fixes and the firmware support ready for v5.2 and then everything will be in place.

[0] https://lore.kernel.org/patchwork/project/lkml/list/?series=364448