[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Ulf Hansson]

From: Ulf Hansson <ulf.hansson@linaro.org>

The apb clock was before the "common" clock driver for ux500 was merged,
handled internally by the clock driver. Now this clock needs to be managed
from the mtu driver as a separate clock.

This patches is based in 3.7 rc2.

It is important the "ARM nomadik patch" is merged together with the
clock patch to not break boot. Therefore I suggest this series to go
through Mike Turquettes clock tree.

Ulf Hansson (2):
  clk: ux500: Register mtu apb_pclocks
  ARM: plat-nomadik: Use apb_pclock in mtu driver

 arch/arm/plat-nomadik/timer.c |    8 +++++++-
 drivers/clk/ux500/u8500_clk.c |    9 +++------
 2 files changed, 10 insertions(+), 7 deletions(-)

-- 
1.7.10

[PATCH 1/2] clk: ux500: Register mtu apb_pclocks

[Ulf Hansson]

From: Ulf Hansson <ulf.hansson@linaro.org>

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
---
 drivers/clk/ux500/u8500_clk.c |    9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/drivers/clk/ux500/u8500_clk.c b/drivers/clk/ux500/u8500_clk.c
index 4ec6f60..8ca3227 100644
--- a/drivers/clk/ux500/u8500_clk.c
+++ b/drivers/clk/ux500/u8500_clk.c
@@ -160,12 +160,6 @@ void u8500_clk_init(void)
 	clk = clk_reg_prcmu_gate("uiccclk", NULL, PRCMU_UICCCLK, CLK_IS_ROOT);
 	clk_register_clkdev(clk, NULL, "uicc");
 
-	/*
-	 * FIXME: The MTU clocks might need some kind of "parent muxed join"
-	 * and these have no K-clocks. For now, we ignore the missing
-	 * connection to the corresponding P-clocks, p6_mtu0_clk and
-	 * p6_mtu1_clk. Instead timclk is used which is the valid parent.
-	 */
 	clk = clk_reg_prcmu_gate("timclk", NULL, PRCMU_TIMCLK, CLK_IS_ROOT);
 	clk_register_clkdev(clk, NULL, "mtu0");
 	clk_register_clkdev(clk, NULL, "mtu1");
@@ -402,8 +396,11 @@ void u8500_clk_init(void)
 
 	clk = clk_reg_prcc_pclk("p6_pclk6", "per6clk", U8500_CLKRST6_BASE,
 				BIT(6), 0);
+	clk_register_clkdev(clk, "apb_pclk", "mtu0");
+
 	clk = clk_reg_prcc_pclk("p6_pclk7", "per6clk", U8500_CLKRST6_BASE,
 				BIT(7), 0);
+	clk_register_clkdev(clk, "apb_pclk", "mtu1");
 
 	/* PRCC K-clocks
 	 *
-- 
1.7.10

[PATCH 2/2] ARM: plat-nomadik: Use apb_pclock in mtu driver

[Ulf Hansson]

From: Ulf Hansson <ulf.hansson@linaro.org>

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
---
 arch/arm/plat-nomadik/timer.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/arm/plat-nomadik/timer.c b/arch/arm/plat-nomadik/timer.c
index 9222e55..17c690d 100644
--- a/arch/arm/plat-nomadik/timer.c
+++ b/arch/arm/plat-nomadik/timer.c
@@ -177,9 +177,15 @@ void nmdk_clksrc_reset(void)
 void __init nmdk_timer_init(void __iomem *base)
 {
 	unsigned long rate;
-	struct clk *clk0;
+	struct clk *clk0, *pclk0;
 
 	mtu_base = base;
+
+	pclk0 = clk_get_sys("mtu0", "apb_pclk");
+	BUG_ON(IS_ERR(pclk0));
+	BUG_ON(clk_prepare(pclk0) < 0);
+	BUG_ON(clk_enable(pclk0) < 0);
+
 	clk0 = clk_get_sys("mtu0", NULL);
 	BUG_ON(IS_ERR(clk0));
 	BUG_ON(clk_prepare(clk0) < 0);
-- 
1.7.10

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Linus Walleij]

On Wed, Oct 24, 2012 at 2:13 PM, Ulf Hansson <ulf.hansson@stericsson.com> wrote:

> From: Ulf Hansson <ulf.hansson@linaro.org>
>
> The apb clock was before the "common" clock driver for ux500 was merged,
> handled internally by the clock driver. Now this clock needs to be managed
> from the mtu driver as a separate clock.
>
> This patches is based in 3.7 rc2.
>
> It is important the "ARM nomadik patch" is merged together with the
> clock patch to not break boot. Therefore I suggest this series to go
> through Mike Turquettes clock tree.

OK go ahead:
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Mike Turquette]

Quoting Linus Walleij (2012-10-24 10:25:26)
> On Wed, Oct 24, 2012 at 2:13 PM, Ulf Hansson <ulf.hansson@stericsson.com> wrote:
> 
> > From: Ulf Hansson <ulf.hansson@linaro.org>
> >
> > The apb clock was before the "common" clock driver for ux500 was merged,
> > handled internally by the clock driver. Now this clock needs to be managed
> > from the mtu driver as a separate clock.
> >
> > This patches is based in 3.7 rc2.
> >
> > It is important the "ARM nomadik patch" is merged together with the
> > clock patch to not break boot. Therefore I suggest this series to go
> > through Mike Turquettes clock tree.
> 
> OK go ahead:
> Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
> 

I've taken these into clk-next.  While the patches are trivial I'd still
prefer a changelog in the future.  For instance, it would be nice to
explain why removing the block comment is OK in the first patch.

Thanks,
Mike

> Yours,
> Linus Walleij

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Linus Walleij]

On Sat, Nov 10, 2012 at 2:24 AM, Mike Turquette <mturquette@linaro.org> wrote:
> Quoting Linus Walleij (2012-10-24 10:25:26)
>> On Wed, Oct 24, 2012 at 2:13 PM, Ulf Hansson <ulf.hansson@stericsson.com> wrote:
>>
>> > From: Ulf Hansson <ulf.hansson@linaro.org>
>> >
>> > The apb clock was before the "common" clock driver for ux500 was merged,
>> > handled internally by the clock driver. Now this clock needs to be managed
>> > from the mtu driver as a separate clock.
>> >
>> > This patches is based in 3.7 rc2.
>> >
>> > It is important the "ARM nomadik patch" is merged together with the
>> > clock patch to not break boot. Therefore I suggest this series to go
>> > through Mike Turquettes clock tree.
>>
>> OK go ahead:
>> Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
>
> I've taken these into clk-next.  While the patches are trivial I'd still
> prefer a changelog in the future.  For instance, it would be nice to
> explain why removing the block comment is OK in the first patch.

So there may be some collision in linux-next due to this,
because I've moved the nomadik timer to
drivers/clksrc and deleted the entire plat-nomadik
directory.

I don't know quite how ingenious git is in detecting
patches on moved files, but if it's causing trouble
I think it's better if you could ACK them and we
could reubmit them to ARM SoC on the multiplatform
branch. i.e. this one:
http://git.kernel.org/?p=linux/kernel/git/arm/arm-soc.git;a=shortlog;h=refs/heads/next/multiplatform

Yours,
Linus Walleij

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Mike Turquette]

Quoting Linus Walleij (2012-11-11 10:47:46)
> On Sat, Nov 10, 2012 at 2:24 AM, Mike Turquette <mturquette@linaro.org> wrote:
> > Quoting Linus Walleij (2012-10-24 10:25:26)
> >> On Wed, Oct 24, 2012 at 2:13 PM, Ulf Hansson <ulf.hansson@stericsson.com> wrote:
> >>
> >> > From: Ulf Hansson <ulf.hansson@linaro.org>
> >> >
> >> > The apb clock was before the "common" clock driver for ux500 was merged,
> >> > handled internally by the clock driver. Now this clock needs to be managed
> >> > from the mtu driver as a separate clock.
> >> >
> >> > This patches is based in 3.7 rc2.
> >> >
> >> > It is important the "ARM nomadik patch" is merged together with the
> >> > clock patch to not break boot. Therefore I suggest this series to go
> >> > through Mike Turquettes clock tree.
> >>
> >> OK go ahead:
> >> Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
> >
> > I've taken these into clk-next.  While the patches are trivial I'd still
> > prefer a changelog in the future.  For instance, it would be nice to
> > explain why removing the block comment is OK in the first patch.
> 
> So there may be some collision in linux-next due to this,
> because I've moved the nomadik timer to
> drivers/clksrc and deleted the entire plat-nomadik
> directory.
> 
> I don't know quite how ingenious git is in detecting
> patches on moved files, but if it's causing trouble
> I think it's better if you could ACK them and we
> could reubmit them to ARM SoC on the multiplatform
> branch. i.e. this one:
> http://git.kernel.org/?p=linux/kernel/git/arm/arm-soc.git;a=shortlog;h=refs/heads/next/multiplatform
> 

Those patches have my ACK.  Do you want to wait to see if something goes
boom or would you rather I drop them from clk-next preemptively?

Regards,
Mike

> Yours,
> Linus Walleij
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Linus Walleij]

On Mon, Nov 12, 2012 at 7:36 PM, Mike Turquette <mturquette@ti.com> wrote:
> Quoting Linus Walleij (2012-11-11 10:47:46)

>> I don't know quite how ingenious git is in detecting
>> patches on moved files, but if it's causing trouble
>> I think it's better if you could ACK them and we
>> could reubmit them to ARM SoC on the multiplatform
>> branch. i.e. this one:
>> http://git.kernel.org/?p=linux/kernel/git/arm/arm-soc.git;a=shortlog;h=refs/heads/next/multiplatform
>>
>
> Those patches have my ACK.  Do you want to wait to see if something goes
> boom or would you rather I drop them from clk-next preemptively?

Hm that's basically a question of how smart git is when
merging these trees together.

Arnd, Olof: what is your experience? If one tree is moving
files and another tree is patching them, should we try to
rebase these onto your branch in ARM SoC and submit
them there instead?

Yours,
Linus Walleij

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Arnd Bergmann]

On Thursday 15 November 2012, Linus Walleij wrote:
> On Mon, Nov 12, 2012 at 7:36 PM, Mike Turquette <mturquette@ti.com> wrote:
> > Quoting Linus Walleij (2012-11-11 10:47:46)
> 
> >> I don't know quite how ingenious git is in detecting
> >> patches on moved files, but if it's causing trouble
> >> I think it's better if you could ACK them and we
> >> could reubmit them to ARM SoC on the multiplatform
> >> branch. i.e. this one:
> >> http://git.kernel.org/?p=linux/kernel/git/arm/arm-soc.git;a=shortlog;h=refs/heads/next/multiplatform
> >>
> >
> > Those patches have my ACK.  Do you want to wait to see if something goes
> > boom or would you rather I drop them from clk-next preemptively?
> 
> Hm that's basically a question of how smart git is when
> merging these trees together.
> 
> Arnd, Olof: what is your experience? If one tree is moving
> files and another tree is patching them, should we try to
> rebase these onto your branch in ARM SoC and submit
> them there instead?

rename vs change conflicts are usually handled, but not always.
The one thing that never works is when part of one file is moved
into another file (or two files get merged into one) and another
patch changes the same file.

If I understand the problem at hand correctly, the safe solution
should be to base the ux500 clk changes on top of c3b9d1db23c4e,
which is in the nomadik/plat-removal branch in arm-soc that got
merged into next/multiplatform. I would prefer if you use that
as a base instead of the larger next/multiplatform branch that
also contains vt8500 and other changes.

	Arnd

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Linus Walleij]

On Thu, Nov 15, 2012 at 1:21 PM, Arnd Bergmann <arnd@arndb.de> wrote:

> If I understand the problem at hand correctly, the safe solution
> should be to base the ux500 clk changes on top of c3b9d1db23c4e,
> which is in the nomadik/plat-removal branch in arm-soc that got
> merged into next/multiplatform. I would prefer if you use that
> as a base instead of the larger next/multiplatform branch that
> also contains vt8500 and other changes.

OK I'll take a stab at this and see if I can fix it...

Yours,
Linus Walleij

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Mike Turquette]

Quoting Linus Walleij (2012-11-15 05:29:32)
> On Thu, Nov 15, 2012 at 1:21 PM, Arnd Bergmann <arnd@arndb.de> wrote:
> 
> > If I understand the problem at hand correctly, the safe solution
> > should be to base the ux500 clk changes on top of c3b9d1db23c4e,
> > which is in the nomadik/plat-removal branch in arm-soc that got
> > merged into next/multiplatform. I would prefer if you use that
> > as a base instead of the larger next/multiplatform branch that
> > also contains vt8500 and other changes.
> 
> OK I'll take a stab at this and see if I can fix it...
> 

I've dropped both of these patches from clk-next.

Regards,
Mike

> Yours,
> Linus Walleij

[PATCH 0/2] clk: ux500: Make mtu driver use apb_pclock

[Ulf Hansson]

On 15 November 2012 19:54, Mike Turquette <mturquette@ti.com> wrote:
> Quoting Linus Walleij (2012-11-15 05:29:32)
>> On Thu, Nov 15, 2012 at 1:21 PM, Arnd Bergmann <arnd@arndb.de> wrote:
>>
>> > If I understand the problem at hand correctly, the safe solution
>> > should be to base the ux500 clk changes on top of c3b9d1db23c4e,
>> > which is in the nomadik/plat-removal branch in arm-soc that got
>> > merged into next/multiplatform. I would prefer if you use that
>> > as a base instead of the larger next/multiplatform branch that
>> > also contains vt8500 and other changes.
>>
>> OK I'll take a stab at this and see if I can fix it...

Thanks Linus!

>>
>
> I've dropped both of these patches from clk-next.
>

Thanks Mike!

> Regards,
> Mike
>
>> Yours,
>> Linus Walleij


Kind regards
Ulf Hansson