[GIT PULL 2/4] ARM: tegra: core SoC enhancements for 3.12

[Stephen Warren]

On 08/20/2013 07:28 PM, Mike Turquette wrote:
> Quoting Stephen Warren (2013-08-20 16:28:25)
>> On 08/20/2013 05:24 PM, Kevin Hilman wrote:
>>> + Mike Turquette
>>>
>>> Stephen Warren <swarren at wwwdotorg.org> writes:
>>>
>>>> This branch includes a number of enhancements to core SoC support for
>>>> Tegra devices. The major new features are:
>>>>
>>>> * Adds a new CPU-power-gated cpuidle state for Tegra114.
>>>> * Adds initial system suspend support for Tegra114, initially supporting
>>>>   just CPU-power-gating during suspend.
>>>> * Adds "LP1" suspend mode support for all of Tegra20/30/114. This mode
>>>>   both gates CPU power, and places the DRAM into self-refresh mode.
>>>> * A new DT-driven PCIe driver to Tegra20/30. The driver is also moved
>>>>   from arch/arm/mach-tegra/ to drivers/pci/host/.
>>>>
>>>> The PCIe driver work depends on the following tag from Thomas Petazzoni:
>>>> git://git.infradead.org/linux-mvebu.git mis-3.12.2
>>>> ... which is merged into the middle of this pull request.
>>>>
>>>> ----------------------------------------------------------------
>>>>
>>>> (If the location of the merge into this branch is problematic, just let
>>>> me know and I'll rebase everything on top of the merge)
>>>
>>> The merge looks fine, but...
>>>
>>>> The following changes since commit ad81f0545ef01ea651886dddac4bef6cec930092:
>>>>
>>>>   Linux 3.11-rc1
>>>>
>>>> are available in the git repository at:
>>>>
>>>>   git://git.kernel.org/pub/scm/linux/kernel/git/swarren/linux-tegra.git tegra-for-3.12-soc
>>>>
>>>> for you to fetch changes up to b4f173752a56187bd55752b0474429202f2ab1d3:
>>>>
>>>>   ARM: tegra: disable LP2 cpuidle state if PCIe is enabled
>>>>
>>>> ----------------------------------------------------------------
>>>>
>>>> Jay Agarwal (1):
>>>>       PCI: tegra: Add Tegra 30 PCIe support
>>>>
>>>> Joseph Lo (21):
>>>>       ARM: tegra: enable Cortex-A15 erratum 798181
>>>>       Revert "ARM: tegra: add cpu_disable for hotplug"
>>>>       ARM: tegra114: Reprogram GIC CPU interface to bypass IRQ on CPU PM entry
>>>>       ARM: tegra114: add low level support for CPU idle powered-down mode
>>>>       ARM: tegra114: cpuidle: add powered-down state
>>>>       ARM: tegra: do v7_invalidate_l1 only when CPU is Cortex-A9
>>>>       ARM: tegra: add a flag for tegra_disable_clean_inv_dcache to do LoUIS or ALL
>>>>       ARM: tegra: set up the correct L2 data RAM latency for Cortex-A15
>>>>       ARM: tegra: add low level code for Tegra114 cluster power down
>>>>       ARM: tegra: shut off the CPU rail when the last CPU in suspend
>>>>       ARM: tegra: hook tegra_tear_down_cpu function
>>>>       ARM: tegra: flowctrl: add support for cpu_suspend_enter/exit
>>>>       clk: tegra: add suspend/resume function for tegra_cpu_car_ops
>>>
>>> This one...
>>>
>>>>       ARM: tegra: remove the limitation that Tegra114 can't support suspend
>>>>       ARM: tegra: add common resume handling code for LP1 resuming
>>>>       ARM: tegra: config the polarity of the request of sys clock
>>>>       clk: tegra114: add LP1 suspend/resume support
>>>
>>> ...and this one are drivers/clk and I don't see an ack from the
>>> clock framework maintainer (Mike Turquette) on either.
>>>
>>> If Mike is OK for them to go via arm-soc, that should be specified in
>>> the description above (and he should ack them), otherwise they should be
>>> split out and sent via Mike.
>>>
>>> A quick glance suggests there are no direct dependencies, so it's
>>> probably best if those changes go through Mike.
>>
>> There are at least run-time dependencies here, although it's possible
>> they're limited to "suspend won't work on Tegra114 unless all those
>> patches are present" rather than "if all these patches aren't present,
>> something other than a new feature will regress due to this series".
>>
>> Mike's certainly been CC'd on the patches many times as quite a few
>> versions of each were posted, and I asked for acks, but got no response.
>> In the past, that's meant he's fine with the patches...
> 
> Hi Stephen,
> 
> Let's not set the precedent of silence conferring an Ack. For the patch
> "clk: tegra: add suspend/resume function for tegra_cpu_car_ops" I gave
> the V1 patch a "looks good to me" response but I seemed to have missed
> the V2 patch:
> http://article.gmane.org/gmane.linux.ports.tegra/12213

Yes, it looks like Joseph simply didn't add the ack to the commit. I'll
follow up internally to make sure people get the process right.

> For the patch "clk: tegra114: add LP1 suspend/resume support" I hadn't
> bothered to review since there were plenty of comments on V1 & V2. The
> same day V3 was posted it was merged into your tree. I should have given
> it a post-merge Ack but didn't.

Sorry to be picky, but I don't think there were any comments on this
specific patch in V1. Admittedly there were some on V2, and by the time
V3 had come around I had forgotten that there were actually changes to
the clk patch within the series. All these series that require so many
revisions wear me down:-(

It'd be really useful if you could reply to the patches and let us know
you've seen them and are deliberately holding off on reviewing them.
That way I'll know you aren't just ignoring them due to e.g. too much
mail. Or, review v1 rather than waiting for any other issues to be
resolved in the series. In the past, I got the impression that you cared
mostly about the core common clock implementation, and didn't really
care what happened inside the individual clock drivers, and were happy
to just have the relevant SoC maintainers review them. Hence, I have
treated your ack on drivers/clk/tegra/* as a bonus more than a necessity.

> Anyways the changes are fine by me. Kevin if you don't mind applying it:
> Acked-by: Mike Turquette <mturquette at linaro.org>

Thanks.