perf: omap3/omap4: enable OMAP3_EMU to support perf

05-20-2011, 02:59 AM

See the below dependency:

config CPU_HAS_PMU
depends on (CPU_V6 || CPU_V6K || CPU_V7 || XSCALE_PMU) &&
(!ARCH_OMAP3 || OMAP3_EMU)
default y
bool

so we have to enable OMAP3_EMU that CPU_HAS_PMU can be selected to
make pmu.o built into kernel, then perf is supported on omap3 and omap4.

No doubt, perf may work well on omap3 based board. For omap4, perf still
does work with the patches:

http://lists.infradead.org/pipermail/linux-arm-kernel/2011-March/045283.html

LP#702999 have some kind of track about this.

Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
debian.ti-omap4/config/config.common.ubuntu | 4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/debian.ti-omap4/config/config.common.ubuntu b/debian.ti-omap4/config/config.common.ubuntu
index 3e3725f..262324b 100644
--- a/debian.ti-omap4/config/config.common.ubuntu
+++ b/debian.ti-omap4/config/config.common.ubuntu
@@ -480,6 +480,7 @@ CONFIG_CPU_CP15_MMU=y
# CONFIG_CPU_DCACHE_DISABLE is not set
# CONFIG_CPU_FREQ is not set
CONFIG_CPU_HAS_ASID=y
+CONFIG_CPU_HAS_PMU=y
# CONFIG_CPU_ICACHE_DISABLE is not set
CONFIG_CPU_IDLE=y
CONFIG_CPU_IDLE_GOV_LADDER=y
@@ -1068,6 +1069,7 @@ CONFIG_HWMON_VID=m
CONFIG_HWSPINLOCK=y
CONFIG_HWSPINLOCK_OMAP=y
CONFIG_HW_CONSOLE=y
+CONFIG_HW_PERF_EVENTS=y
CONFIG_HW_RANDOM=y
# CONFIG_HW_RANDOM_TIMERIOMEM is not set
CONFIG_HZ=128
@@ -2098,7 +2100,7 @@ CONFIG_OMAP2_DSS_MIN_FCK_PER_PCK=0
CONFIG_OMAP2_VRAM=y
CONFIG_OMAP2_VRAM_SIZE=32
CONFIG_OMAP2_VRFB=y
-# CONFIG_OMAP3_EMU is not set
+CONFIG_OMAP3_EMU=y
# CONFIG_OMAP3_L2_AUX_SECURE_SAVE_RESTORE is not set
# CONFIG_OMAP3_SDRC_AC_TIMING is not set
CONFIG_OMAP4_DSS_HDMI=y
--
1.7.4.1

thanks,
--
Ming Lei

--
kernel-team mailing list
kernel-team@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/kernel-team

05-20-2011, 03:10 AM

05-20-2011, 03:29 AM

On Fri, May 20, 2011 at 11:10 AM, Ming Lei <ming.lei@canonical.com> wrote:
> From 38737ec07a2aa013015d9f0088e2912a3d090f6f Mon Sep 17 00:00:00 2001
> From: Ming Lei <ming.lei@canonical.com>
> Date: Fri, 20 May 2011 10:59:59 +0800
> Subject: [PATCH] perf: omap3/omap4: enable OMAP3_EMU to support perf

For Natty, we are currently in SRU. Please read the wiki here:
https://wiki.ubuntu.com/KernelTeam/KernelUpdates

So the title should be "[Natty] [SRU] perf: ......."
and it's better to provide a similar patch for ti-omap4 branch like
"[Natty/ti-omap4] [SRU] perf: ......"

Moreover, you need to provide SRU justification description in this
git log message.

>
> See the below dependency:
>
> * * * *config CPU_HAS_PMU
> * * * * * * * *depends on (CPU_V6 || CPU_V6K || CPU_V7 || XSCALE_PMU) &&
> * * * * * * * * * * * * * (!ARCH_OMAP3 || OMAP3_EMU)
> * * * * * * * *default y
> * * * * * * * *bool
>
> so we have to enable OMAP3_EMU that CPU_HAS_PMU can be selected to
> make pmu.o built into kernel, then perf is supported on omap3 and omap4.
>
> No doubt, perf may work well on omap3 based board. For omap4, perf still
> does work with the patches:
>
> * * * *http://lists.infradead.org/pipermail/linux-arm-kernel/2011-March/045283.html
>

Yeah, so you need 2 versions of this patch.
1 for Natty master which supports omap3
2 for Natty ti-omap4 for omap4 supporting and you might provide more
perf patches to ti-omap4

> LP#702999 have some kind of track about this.

Should use BugLink for the bug tracking

>
> Signed-off-by: Ming Lei <ming.lei@canonical.com>

Actually, it is a very good start for you. I followed the email you
posted to ARM mail list and some discussion with Will Deacon.
I think the patch content is OK, just need some modification for our SRU policy.

Thanks,
-Bryan
> ---
> *debian.ti-omap4/config/config.common.ubuntu | * *4 +++-
> *1 files changed, 3 insertions(+), 1 deletions(-)
>
> diff --git a/debian.ti-omap4/config/config.common.ubuntu b/debian.ti-omap4/config/config.common.ubuntu
> index 3e3725f..262324b 100644
> --- a/debian.ti-omap4/config/config.common.ubuntu
> +++ b/debian.ti-omap4/config/config.common.ubuntu
> @@ -480,6 +480,7 @@ CONFIG_CPU_CP15_MMU=y
> *# CONFIG_CPU_DCACHE_DISABLE is not set
> *# CONFIG_CPU_FREQ is not set
> *CONFIG_CPU_HAS_ASID=y
> +CONFIG_CPU_HAS_PMU=y
> *# CONFIG_CPU_ICACHE_DISABLE is not set
> *CONFIG_CPU_IDLE=y
> *CONFIG_CPU_IDLE_GOV_LADDER=y
> @@ -1068,6 +1069,7 @@ CONFIG_HWMON_VID=m
> *CONFIG_HWSPINLOCK=y
> *CONFIG_HWSPINLOCK_OMAP=y
> *CONFIG_HW_CONSOLE=y
> +CONFIG_HW_PERF_EVENTS=y
> *CONFIG_HW_RANDOM=y
> *# CONFIG_HW_RANDOM_TIMERIOMEM is not set
> *CONFIG_HZ=128
> @@ -2098,7 +2100,7 @@ CONFIG_OMAP2_DSS_MIN_FCK_PER_PCK=0
> *CONFIG_OMAP2_VRAM=y
> *CONFIG_OMAP2_VRAM_SIZE=32
> *CONFIG_OMAP2_VRFB=y
> -# CONFIG_OMAP3_EMU is not set
> +CONFIG_OMAP3_EMU=y
> *# CONFIG_OMAP3_L2_AUX_SECURE_SAVE_RESTORE is not set
> *# CONFIG_OMAP3_SDRC_AC_TIMING is not set
> *CONFIG_OMAP4_DSS_HDMI=y
> --
> 1.7.4.1
>
> thanks,
> --
> Ming Lei
>
>
>
>
>
> --
> kernel-team mailing list
> kernel-team@lists.ubuntu.com
> https://lists.ubuntu.com/mailman/listinfo/kernel-team
>

--
kernel-team mailing list
kernel-team@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/kernel-team

05-20-2011, 04:09 AM

Hi Bryan,

Thanks for your detailed comments.

On Fri, 20 May 2011 11:29:29 +0800
Bryan Wu <bryan.wu@canonical.com> wrote:

> On Fri, May 20, 2011 at 11:10 AM, Ming Lei <ming.lei@canonical.com>
> wrote:
> > From 38737ec07a2aa013015d9f0088e2912a3d090f6f Mon Sep 17 00:00:00
> > 2001 From: Ming Lei <ming.lei@canonical.com>
> > Date: Fri, 20 May 2011 10:59:59 +0800
> > Subject: [PATCH] perf: omap3/omap4: enable OMAP3_EMU to support perf
>
> For Natty, we are currently in SRU. Please read the wiki here:
> https://wiki.ubuntu.com/KernelTeam/KernelUpdates

After reading through above, I am not sure if the patch is a SRU:

Without this patch, perf can't work on both OMAP3 and OMAP4. With
this patch, perf can work well on omap3. For OMPAP4, this patch is
required too, but another 3 patches are needed to make perf doable.

If you think this patch is a SRU, I will follow Bryan's comment to
prepare for new post.

> So the title should be "[Natty] [SRU] perf: ......."
> and it's better to provide a similar patch for ti-omap4 branch like
> "[Natty/ti-omap4] [SRU] perf: ......"
>
> Moreover, you need to provide SRU justification description in this
> git log message.
>
> >
> > See the below dependency:
> >
> > * * * *config CPU_HAS_PMU
> > * * * * * * * *depends on (CPU_V6 || CPU_V6K || CPU_V7 ||
> > XSCALE_PMU) && (!ARCH_OMAP3 || OMAP3_EMU)
> > * * * * * * * *default y
> > * * * * * * * *bool
> >
> > so we have to enable OMAP3_EMU that CPU_HAS_PMU can be selected to
> > make pmu.o built into kernel, then perf is supported on omap3 and
> > omap4.
> >
> > No doubt, perf may work well on omap3 based board. For omap4, perf
> > still does work with the patches:
> >
> > * * * *http://lists.infradead.org/pipermail/linux-arm-kernel/2011-March/045283.html
> >
>
> Yeah, so you need 2 versions of this patch.
> 1 for Natty master which supports omap3
> 2 for Natty ti-omap4 for omap4 supporting and you might provide more
> perf patches to ti-omap4
>
> > LP#702999 have some kind of track about this.
>
> Should use BugLink for the bug tracking
>
> >
> > Signed-off-by: Ming Lei <ming.lei@canonical.com>
>
> Actually, it is a very good start for you. I followed the email you
> posted to ARM mail list and some discussion with Will Deacon.
> I think the patch content is OK, just need some modification for our
> SRU policy.
>
> Thanks,
> -Bryan
> > ---
> > *debian.ti-omap4/config/config.common.ubuntu | * *4 +++-
> > *1 files changed, 3 insertions(+), 1 deletions(-)
> >
> > diff --git a/debian.ti-omap4/config/config.common.ubuntu
> > b/debian.ti-omap4/config/config.common.ubuntu index
> > 3e3725f..262324b 100644 ---
> > a/debian.ti-omap4/config/config.common.ubuntu +++
> > b/debian.ti-omap4/config/config.common.ubuntu @@ -480,6 +480,7 @@
> > CONFIG_CPU_CP15_MMU=y # CONFIG_CPU_DCACHE_DISABLE is not set
> > *# CONFIG_CPU_FREQ is not set
> > *CONFIG_CPU_HAS_ASID=y
> > +CONFIG_CPU_HAS_PMU=y
> > *# CONFIG_CPU_ICACHE_DISABLE is not set
> > *CONFIG_CPU_IDLE=y
> > *CONFIG_CPU_IDLE_GOV_LADDER=y
> > @@ -1068,6 +1069,7 @@ CONFIG_HWMON_VID=m
> > *CONFIG_HWSPINLOCK=y
> > *CONFIG_HWSPINLOCK_OMAP=y
> > *CONFIG_HW_CONSOLE=y
> > +CONFIG_HW_PERF_EVENTS=y
> > *CONFIG_HW_RANDOM=y
> > *# CONFIG_HW_RANDOM_TIMERIOMEM is not set
> > *CONFIG_HZ=128
> > @@ -2098,7 +2100,7 @@ CONFIG_OMAP2_DSS_MIN_FCK_PER_PCK=0
> > *CONFIG_OMAP2_VRAM=y
> > *CONFIG_OMAP2_VRAM_SIZE=32
> > *CONFIG_OMAP2_VRFB=y
> > -# CONFIG_OMAP3_EMU is not set
> > +CONFIG_OMAP3_EMU=y
> > *# CONFIG_OMAP3_L2_AUX_SECURE_SAVE_RESTORE is not set
> > *# CONFIG_OMAP3_SDRC_AC_TIMING is not set
> > *CONFIG_OMAP4_DSS_HDMI=y
> > --
> > 1.7.4.1
> >
> > thanks,
> > --
> > Ming Lei
> >
> >
> >
> >
> >
> > --
> > kernel-team mailing list
> > kernel-team@lists.ubuntu.com
> > https://lists.ubuntu.com/mailman/listinfo/kernel-team
> >

--
Ming Lei

--
kernel-team mailing list
kernel-team@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/kernel-team

05-20-2011, 04:38 AM

On Fri, May 20, 2011 at 12:09 PM, Ming Lei <ming.lei@canonical.com> wrote:
>
> Hi Bryan,
>
> Thanks for your detailed comments.
>
> On Fri, 20 May 2011 11:29:29 +0800
> Bryan Wu <bryan.wu@canonical.com> wrote:
>
>> On Fri, May 20, 2011 at 11:10 AM, Ming Lei <ming.lei@canonical.com>
>> wrote:
>> > From 38737ec07a2aa013015d9f0088e2912a3d090f6f Mon Sep 17 00:00:00
>> > 2001 From: Ming Lei <ming.lei@canonical.com>
>> > Date: Fri, 20 May 2011 10:59:59 +0800
>> > Subject: [PATCH] perf: omap3/omap4: enable OMAP3_EMU to support perf
>>
>> For Natty, we are currently in SRU. Please read the wiki here:
>> https://wiki.ubuntu.com/KernelTeam/KernelUpdates
>
> After reading through above, I am not sure if the patch is a SRU:
>
> Without this patch, perf can't work on both OMAP3 and OMAP4. *With
> this patch, perf can work well on omap3. For OMPAP4, this patch is
> required too, but another 3 patches are needed to make perf doable.
>
> If you think this patch is a SRU, I will follow Bryan's comment to
> prepare for new post.

I think this is OK as an SRU for OMAP, although it's actually a feature-add,
yet it seems OK not to bring any regression.

Just go ahead and post it again, so others can Ack/Nack.

>
>> So the title should be "[Natty] [SRU] perf: ......."
>> and it's better to provide a similar patch for ti-omap4 branch like
>> "[Natty/ti-omap4] [SRU] perf: ......"
>>
>> Moreover, you need to provide SRU justification description in this
>> git log message.
>>
>> >
>> > See the below dependency:
>> >
>> > * * * *config CPU_HAS_PMU
>> > * * * * * * * *depends on (CPU_V6 || CPU_V6K || CPU_V7 ||
>> > XSCALE_PMU) && (!ARCH_OMAP3 || OMAP3_EMU)
>> > * * * * * * * *default y
>> > * * * * * * * *bool
>> >
>> > so we have to enable OMAP3_EMU that CPU_HAS_PMU can be selected to
>> > make pmu.o built into kernel, then perf is supported on omap3 and
>> > omap4.
>> >
>> > No doubt, perf may work well on omap3 based board. For omap4, perf
>> > still does work with the patches:
>> >
>> > * * * *http://lists.infradead.org/pipermail/linux-arm-kernel/2011-March/045283.html
>> >
>>
>> Yeah, so you need 2 versions of this patch.
>> 1 for Natty master which supports omap3
>> 2 for Natty ti-omap4 for omap4 supporting and you might provide more
>> perf patches to ti-omap4
>>
>> > LP#702999 have some kind of track about this.
>>
>> Should use BugLink for the bug tracking
>>
>> >
>> > Signed-off-by: Ming Lei <ming.lei@canonical.com>
>>
>> Actually, it is a very good start for you. I followed the email you
>> posted to ARM mail list and some discussion with Will Deacon.
>> I think the patch content is OK, just need some modification for our
>> SRU policy.
>>
>> Thanks,
>> -Bryan
>> > ---
>> > *debian.ti-omap4/config/config.common.ubuntu | * *4 +++-
>> > *1 files changed, 3 insertions(+), 1 deletions(-)
>> >
>> > diff --git a/debian.ti-omap4/config/config.common.ubuntu
>> > b/debian.ti-omap4/config/config.common.ubuntu index
>> > 3e3725f..262324b 100644 ---
>> > a/debian.ti-omap4/config/config.common.ubuntu +++
>> > b/debian.ti-omap4/config/config.common.ubuntu @@ -480,6 +480,7 @@
>> > CONFIG_CPU_CP15_MMU=y # CONFIG_CPU_DCACHE_DISABLE is not set
>> > *# CONFIG_CPU_FREQ is not set
>> > *CONFIG_CPU_HAS_ASID=y
>> > +CONFIG_CPU_HAS_PMU=y
>> > *# CONFIG_CPU_ICACHE_DISABLE is not set
>> > *CONFIG_CPU_IDLE=y
>> > *CONFIG_CPU_IDLE_GOV_LADDER=y
>> > @@ -1068,6 +1069,7 @@ CONFIG_HWMON_VID=m
>> > *CONFIG_HWSPINLOCK=y
>> > *CONFIG_HWSPINLOCK_OMAP=y
>> > *CONFIG_HW_CONSOLE=y
>> > +CONFIG_HW_PERF_EVENTS=y
>> > *CONFIG_HW_RANDOM=y
>> > *# CONFIG_HW_RANDOM_TIMERIOMEM is not set
>> > *CONFIG_HZ=128
>> > @@ -2098,7 +2100,7 @@ CONFIG_OMAP2_DSS_MIN_FCK_PER_PCK=0
>> > *CONFIG_OMAP2_VRAM=y
>> > *CONFIG_OMAP2_VRAM_SIZE=32
>> > *CONFIG_OMAP2_VRFB=y
>> > -# CONFIG_OMAP3_EMU is not set
>> > +CONFIG_OMAP3_EMU=y
>> > *# CONFIG_OMAP3_L2_AUX_SECURE_SAVE_RESTORE is not set
>> > *# CONFIG_OMAP3_SDRC_AC_TIMING is not set
>> > *CONFIG_OMAP4_DSS_HDMI=y
>> > --
>> > 1.7.4.1
>> >
>> > thanks,
>> > --
>> > Ming Lei
>> >
>> >
>> >
>> >
>> >
>> > --
>> > kernel-team mailing list
>> > kernel-team@lists.ubuntu.com
>> > https://lists.ubuntu.com/mailman/listinfo/kernel-team
>> >
>
>
>
> --
> Ming Lei
>
> --
> kernel-team mailing list
> kernel-team@lists.ubuntu.com
> https://lists.ubuntu.com/mailman/listinfo/kernel-team
>

--
kernel-team mailing list
kernel-team@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/kernel-team

05-20-2011, 06:40 AM

Hi Eric,

On Fri, 20 May 2011 12:38:49 +0800
Eric Miao <eric.miao@canonical.com> wrote:

>
> I think this is OK as an SRU for OMAP, although it's actually a
> feature-add, yet it seems OK not to bring any regression.
>
> Just go ahead and post it again, so others can Ack/Nack.

OK, I will prepare a patch for ti-omap4, but not required for master
since the option is enabled already in master branch.

thanks,
--
Ming Lei

--
kernel-team mailing list
kernel-team@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/kernel-team