[GIT,PULL] Meson clock fixes for 4.13-rc4

Message ID ce41382f-fd93-1dc9-0dba-1484ac555945@baylibre.com
State New
Headers show

Pull-request

git://github.com/baylibre/clk-meson.git tags/meson-clk-fixes-for-4.13-rc4

Message

Neil Armstrong July 31, 2017, 9:25 a.m.
Hi Mike, Stephen,

Here is some fixes for Meson GXBB and Meson8 drivers, that could go for 4.13-rc4.

Thanks,
Neil

The following changes since commit 5771a8c08880cdca3bfb4a3fc6d309d6bba20877:

  Linux v4.13-rc1 (2017-07-15 15:22:10 -0700)

are available in the git repository at:

  git://github.com/baylibre/clk-meson.git tags/meson-clk-fixes-for-4.13-rc4

for you to fetch changes up to c1922eed58f38f3d7876f53e1000e3c3e3024e54:

  clk: meson: mpll: fix mpll0 fractional part ignored (2017-07-31 10:08:01 +0200)

----------------------------------------------------------------
Meson Clock fixes:
- gxbb: fix meson cts_amclk divider flags
- gxbb: fix clk_mclk_i958 divider flags
- meson8b: fix protection against undefined clks
- gxbb: fix protection against undefined clks
- mpll: fix mpll0 fractional part ignored

----------------------------------------------------------------
Jerome Brunet (5):
      clk: meson: gxbb: fix meson cts_amclk divider flags
      clk: meson: gxbb: fix clk_mclk_i958 divider flags
      clk: meson: meson8b: fix protection against undefined clks
      clk: meson: gxbb: fix protection against undefined clks
      clk: meson: mpll: fix mpll0 fractional part ignored

 drivers/clk/meson/clk-mpll.c |  7 +++++++
 drivers/clk/meson/clkc.h     |  1 +
 drivers/clk/meson/gxbb.c     | 17 +++++++++++++----
 drivers/clk/meson/meson8b.c  |  6 ++++++
 4 files changed, 27 insertions(+), 4 deletions(-)

Comments

Stephen Boyd July 31, 2017, 8:20 p.m. | #1
On 07/31, Neil Armstrong wrote:
> Hi Mike, Stephen,
> 
> Here is some fixes for Meson GXBB and Meson8 drivers, that could go for 4.13-rc4.
> 

Are any of these critical fixes for regressions in the v4.13-rc
series? All the Fixes tags look like they go back to before
v4.13-rc1, so they don't look critical at first glance. If
anything, they're mostly nice to have non-critical fixes? If so,
can we defer these until the next merge window?
Jerome Brunet July 31, 2017, 9:43 p.m. | #2
On Mon, 2017-07-31 at 13:20 -0700, Stephen Boyd wrote:
> On 07/31, Neil Armstrong wrote:
> > Hi Mike, Stephen,
> > 
> > Here is some fixes for Meson GXBB and Meson8 drivers, that could go for
> > 4.13-rc4.
> > 
> 
> Are any of these critical fixes for regressions in the v4.13-rc
> series? All the Fixes tags look like they go back to before
> v4.13-rc1, so they don't look critical at first glance. If
> anything, they're mostly nice to have non-critical fixes? If so,
> can we defer these until the next merge window?

I agree that these fixes hardly qualify as critical, except maybe for the mpll
fix. The clock rate actually applied can be really off compared to what CCF will
 except to have set.

These fixes are quite simple and safe, it would be nice if we could avoid having
 the next version tagged with known issues.

>
Stephen Boyd Aug. 1, 2017, 1:38 a.m. | #3
On 07/31, Jerome Brunet wrote:
> On Mon, 2017-07-31 at 13:20 -0700, Stephen Boyd wrote:
> > On 07/31, Neil Armstrong wrote:
> > > Hi Mike, Stephen,
> > > 
> > > Here is some fixes for Meson GXBB and Meson8 drivers, that could go for
> > > 4.13-rc4.
> > > 
> > 
> > Are any of these critical fixes for regressions in the v4.13-rc
> > series? All the Fixes tags look like they go back to before
> > v4.13-rc1, so they don't look critical at first glance. If
> > anything, they're mostly nice to have non-critical fixes? If so,
> > can we defer these until the next merge window?
> 
> I agree that these fixes hardly qualify as critical, except maybe for the mpll
> fix. The clock rate actually applied can be really off compared to what CCF will
>  except to have set.

Ok. So perhaps just the mpll fix would be appropriate then?

> 
> These fixes are quite simple and safe, it would be nice if we could avoid having
>  the next version tagged with known issues.
> 

I agree they're simple and safe, but they're also not necessary
to the release we're working on unless they're causing some
problem with code that was introduced this merge window. We try
to avoid merging non-critical fixes unless they're really causing
someone some sort of problems. If they aren't, we can stick them
into linux-next and wait until next release. Of course, they'll
be backported to stable trees as well with the Fixes tag, so
everything works out in the end.
Neil Armstrong Aug. 1, 2017, 12:11 p.m. | #4
On 08/01/2017 03:38 AM, Stephen Boyd wrote:
> On 07/31, Jerome Brunet wrote:
>> On Mon, 2017-07-31 at 13:20 -0700, Stephen Boyd wrote:
>>> On 07/31, Neil Armstrong wrote:
>>>> Hi Mike, Stephen,
>>>>
>>>> Here is some fixes for Meson GXBB and Meson8 drivers, that could go for
>>>> 4.13-rc4.
>>>>
>>>
>>> Are any of these critical fixes for regressions in the v4.13-rc
>>> series? All the Fixes tags look like they go back to before
>>> v4.13-rc1, so they don't look critical at first glance. If
>>> anything, they're mostly nice to have non-critical fixes? If so,
>>> can we defer these until the next merge window?
>>
>> I agree that these fixes hardly qualify as critical, except maybe for the mpll
>> fix. The clock rate actually applied can be really off compared to what CCF will
>>  except to have set.
> 
> Ok. So perhaps just the mpll fix would be appropriate then?
> 
>>
>> These fixes are quite simple and safe, it would be nice if we could avoid having
>>  the next version tagged with known issues.
>>
> 
> I agree they're simple and safe, but they're also not necessary
> to the release we're working on unless they're causing some
> problem with code that was introduced this merge window. We try
> to avoid merging non-critical fixes unless they're really causing
> someone some sort of problems. If they aren't, we can stick them
> into linux-next and wait until next release. Of course, they'll
> be backported to stable trees as well with the Fixes tag, so
> everything works out in the end.
> 

OK, I'll resend a v2 PR with only the mpll fix.

The other will be pushed on our next/drivers branch.

Neil