Re: linux-next: build failure after merge of the pm tree

Re: linux-next: build failure after merge of the pm tree

[Zhang Rui]

Resend with linux-pm mailing list CCed.

On Thu, 2022-04-28 at 11:00 +1000, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the pm tree, today's linux-next build (x86_64
> allmodconfig)
> failed like this:
> 
> drivers/idle/intel_idle.c: In function 'adl_idle_state_table_update':
> drivers/idle/intel_idle.c:1701:17: error: 'disable_promotion_to_c1e'
> undeclared (first use in this function)
>  1701 |                 disable_promotion_to_c1e = true;
>       |                 ^~~~~~~~~~~~~~~~~~~~~~~~
> drivers/idle/intel_idle.c:1701:17: note: each undeclared identifier
> is reported only once for each function it appears in
> drivers/idle/intel_idle.c:1706:9: error: implicit declaration of
> function 'c1e_promotion_enable' [-Werror=implicit-function-
> declaration]
>  1706 |         c1e_promotion_enable();
>       |         ^~~~~~~~~~~~~~~~~~~~
> drivers/idle/intel_idle.c: At top level:
> drivers/idle/intel_idle.c:1854:13: error: conflicting types for
> 'c1e_promotion_enable'; have 'void(void)' [-Werror]
>  1854 | static void c1e_promotion_enable(void)
>       |             ^~~~~~~~~~~~~~~~~~~~
> drivers/idle/intel_idle.c:1854:13: error: static declaration of
> 'c1e_promotion_enable' follows non-static declaration
> drivers/idle/intel_idle.c:1706:9: note: previous implicit declaration
> of 'c1e_promotion_enable' with type 'void(void)'
>  1706 |         c1e_promotion_enable();
>       |         ^~~~~~~~~~~~~~~~~~~~
> 
> Caused by commit
> 
>   39c184a6a9a7 ("intel_idle: Fix the 'preferred_cstates' module
> parameter")
> 
> interacting with commit
> 
>   cc6e234b8264 ("intel_idle: Add AlderLake support")
> 
> Presumably this should have been fixed up in commit
> 
>   55ecda6f25ef ("Merge branch 'intel-idle' into linux-next")
> 
> I have used the pm tree from next-20220427 for today.
> 
TBH, I'm not quite sure about the Fixes tag below.

Although commit 39c184a6a9a7 is merged later and breaks commit
cc6e234b8264, but given that commit 39c184a6a9a7 is for -rc material
and commit cc6e234b8264 and this patch are for next merge window, so I
still use the orginal ADL intel_idle commit for Fixes tag.

thanks,
rui

From 9b784d7f9ea5593a92eb6d616523b5f47464e981 Mon Sep 17 00:00:00 2001
From: Zhang Rui <rui.zhang@intel.com>
Date: Thu, 28 Apr 2022 09:37:10 +0800
Subject: [PATCH] intel_idle: fix C1E handling for AlderLake

commit cc6e234b8264 ("intel_idle: Add AlderLake support") disables the C1E
promotion using the 'disable_promotion_to_c1e' variable, but enables the
the C1E promotion by invoking c1e_promotion_enable() directly.

Then, commit 39c184a6a9a7 ("intel_idle: Fix the 'preferred_cstates' module
parameter") removes the 'disable_promotion_to_c1e' variable and introduces
a new tri-state 'c1e_promotion' variable that can be used for both
enabling and disabling C1E promotion, on a per CPU basis.

Switch to use the new 'c1e_promotion' variable to fix the build failure,
and also to do C1E promotion bit update on all CPUs.

Fixes: commit cc6e234b8264 ("intel_idle: Add AlderLake support")
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
---
 drivers/idle/intel_idle.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
index 7c081ed26b64..2de6e0a2d9a1 100644
--- a/drivers/idle/intel_idle.c
+++ b/drivers/idle/intel_idle.c
@@ -1698,12 +1698,12 @@ static void __init adl_idle_state_table_update(void)
 		cpuidle_state_table[1].flags |= CPUIDLE_FLAG_UNUSABLE;
 
 		/* Disable C1E by clearing the "C1E promotion" bit. */
-		disable_promotion_to_c1e = true;
+		c1e_promotion = C1E_PROMOTION_DISABLE;
 		return;
 	}
 end:
 	/* Make sure C1E is enabled by default */
-	c1e_promotion_enable();
+	c1e_promotion = C1E_PROMOTION_ENABLE;
 }
 
 /**
-- 
2.17.1

Re: linux-next: build failure after merge of the pm tree

[Rafael J. Wysocki]

On Thu, Apr 28, 2022 at 10:56 AM Zhang Rui <rui.zhang@intel.com> wrote:
>
> Resend with linux-pm mailing list CCed.
>
> On Thu, 2022-04-28 at 11:00 +1000, Stephen Rothwell wrote:
> > Hi all,
> >
> > After merging the pm tree, today's linux-next build (x86_64
> > allmodconfig)
> > failed like this:
> >
> > drivers/idle/intel_idle.c: In function 'adl_idle_state_table_update':
> > drivers/idle/intel_idle.c:1701:17: error: 'disable_promotion_to_c1e'
> > undeclared (first use in this function)
> >  1701 |                 disable_promotion_to_c1e = true;
> >       |                 ^~~~~~~~~~~~~~~~~~~~~~~~
> > drivers/idle/intel_idle.c:1701:17: note: each undeclared identifier
> > is reported only once for each function it appears in
> > drivers/idle/intel_idle.c:1706:9: error: implicit declaration of
> > function 'c1e_promotion_enable' [-Werror=implicit-function-
> > declaration]
> >  1706 |         c1e_promotion_enable();
> >       |         ^~~~~~~~~~~~~~~~~~~~
> > drivers/idle/intel_idle.c: At top level:
> > drivers/idle/intel_idle.c:1854:13: error: conflicting types for
> > 'c1e_promotion_enable'; have 'void(void)' [-Werror]
> >  1854 | static void c1e_promotion_enable(void)
> >       |             ^~~~~~~~~~~~~~~~~~~~
> > drivers/idle/intel_idle.c:1854:13: error: static declaration of
> > 'c1e_promotion_enable' follows non-static declaration
> > drivers/idle/intel_idle.c:1706:9: note: previous implicit declaration
> > of 'c1e_promotion_enable' with type 'void(void)'
> >  1706 |         c1e_promotion_enable();
> >       |         ^~~~~~~~~~~~~~~~~~~~
> >
> > Caused by commit
> >
> >   39c184a6a9a7 ("intel_idle: Fix the 'preferred_cstates' module
> > parameter")
> >
> > interacting with commit
> >
> >   cc6e234b8264 ("intel_idle: Add AlderLake support")
> >
> > Presumably this should have been fixed up in commit
> >
> >   55ecda6f25ef ("Merge branch 'intel-idle' into linux-next")
> >
> > I have used the pm tree from next-20220427 for today.
> >
> TBH, I'm not quite sure about the Fixes tag below.
>
> Although commit 39c184a6a9a7 is merged later and breaks commit
> cc6e234b8264, but given that commit 39c184a6a9a7 is for -rc material
> and commit cc6e234b8264 and this patch are for next merge window, so I
> still use the orginal ADL intel_idle commit for Fixes tag.

Thanks for the fix.

I think I'll redo the ADL support commit with the below folded in.

> From 9b784d7f9ea5593a92eb6d616523b5f47464e981 Mon Sep 17 00:00:00 2001
> From: Zhang Rui <rui.zhang@intel.com>
> Date: Thu, 28 Apr 2022 09:37:10 +0800
> Subject: [PATCH] intel_idle: fix C1E handling for AlderLake
>
> commit cc6e234b8264 ("intel_idle: Add AlderLake support") disables the C1E
> promotion using the 'disable_promotion_to_c1e' variable, but enables the
> the C1E promotion by invoking c1e_promotion_enable() directly.
>
> Then, commit 39c184a6a9a7 ("intel_idle: Fix the 'preferred_cstates' module
> parameter") removes the 'disable_promotion_to_c1e' variable and introduces
> a new tri-state 'c1e_promotion' variable that can be used for both
> enabling and disabling C1E promotion, on a per CPU basis.
>
> Switch to use the new 'c1e_promotion' variable to fix the build failure,
> and also to do C1E promotion bit update on all CPUs.
>
> Fixes: commit cc6e234b8264 ("intel_idle: Add AlderLake support")
> Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> ---
>  drivers/idle/intel_idle.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
> index 7c081ed26b64..2de6e0a2d9a1 100644
> --- a/drivers/idle/intel_idle.c
> +++ b/drivers/idle/intel_idle.c
> @@ -1698,12 +1698,12 @@ static void __init adl_idle_state_table_update(void)
>                 cpuidle_state_table[1].flags |= CPUIDLE_FLAG_UNUSABLE;
>
>                 /* Disable C1E by clearing the "C1E promotion" bit. */
> -               disable_promotion_to_c1e = true;
> +               c1e_promotion = C1E_PROMOTION_DISABLE;
>                 return;
>         }
>  end:
>         /* Make sure C1E is enabled by default */
> -       c1e_promotion_enable();
> +       c1e_promotion = C1E_PROMOTION_ENABLE;
>  }
>
>  /**
> --
> 2.17.1
>
>

Re: linux-next: build failure after merge of the pm tree

[Wysocki, Rafael J]

Daniel,

On 7/14/2023 3:00 AM, Stephen Rothwell wrote:
> Hi all,
>
> After merging the pm tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
>
> In file included from drivers/thermal/tegra/soctherm.c:36:
> drivers/thermal/tegra/../thermal_core.h:18:2: error: #warning This header can only be included by the thermal core code [-Werror-cpp]
>     18 | #warning This header can only be included by the thermal core code
>        |  ^~~~~~~
>
> Introduced by commit
>
>    bc1bb350b534 ("thermal/core: Hardening the self-encapsulation")

I'll drop this patch.  I think that the rest of the series doesn't 
really depend on it.


> I have used the pm tree from next-20230713 for today.

Thanks!

Re: linux-next: build failure after merge of the pm tree

[Rafael J. Wysocki]

On Fri, Jul 14, 2023 at 8:20 PM Wysocki, Rafael J
<rafael.j.wysocki@intel.com> wrote:
>
> Daniel,
>
> On 7/14/2023 3:00 AM, Stephen Rothwell wrote:
> > Hi all,
> >
> > After merging the pm tree, today's linux-next build (x86_64
> > allmodconfig) failed like this:
> >
> > In file included from drivers/thermal/tegra/soctherm.c:36:
> > drivers/thermal/tegra/../thermal_core.h:18:2: error: #warning This header can only be included by the thermal core code [-Werror-cpp]
> >     18 | #warning This header can only be included by the thermal core code
> >        |  ^~~~~~~
> >
> > Introduced by commit
> >
> >    bc1bb350b534 ("thermal/core: Hardening the self-encapsulation")
>
> I'll drop this patch.  I think that the rest of the series doesn't
> really depend on it.

The second patch in the series does depend on it, so I had to drop
them both.  Please fix and resend.

Thanks!

Re: linux-next: build failure after merge of the pm tree

[Heiner Kallweit]

On 04.01.2022 09:02, Stephen Rothwell wrote:
> Hi Heiner,
> 
> On Tue, 4 Jan 2022 08:29:28 +0100 Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>
>> The patch in the pm tree annotating pm_runtime_resume_and_get() as __must_check
>> follows some fixes of pm_runtime_resume_and_get() usage that went through other
>> trees. These fixes are in linux-next but don't seem to be in the pm tree.
>> We talk about:
>> f04b4fb47d83 ("ASoC: sh: rz-ssi: Check return value of pm_runtime_resume_and_get()")
> 
> In the sound-asoc tree.
> 
>> 3d6b661330a7 ("crypto: stm32 - Revert broken pm_runtime_resume_and_get changes")
> 
> In the crypto tree.
> 
> Both those are merged into linux-next after the pm tree.  If Linus did
> the same, the pm tree commit would break his build.  The only way you
> can have that pm tree commit in linux-next is to ask Andrew Morton to
> put it in the post linux-next part of his patch series.  Otherwise, it
> needs to be removed form the pm tree and wait until after the next
> merge window closes (or at least both the above trees have been merged
> by Linus).
> 
Rafael,
can you take care of this?
To avoid such issues I think next time I'd route all dependent patches
through one tree and just get the ACK from the other involved maintainers.

Thank you, Heiner

Re: linux-next: build failure after merge of the pm tree

[Rafael J. Wysocki]

On Tue, Jan 4, 2022 at 3:43 PM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>
> On 04.01.2022 09:02, Stephen Rothwell wrote:
> > Hi Heiner,
> >
> > On Tue, 4 Jan 2022 08:29:28 +0100 Heiner Kallweit <hkallweit1@gmail.com> wrote:
> >>
> >> The patch in the pm tree annotating pm_runtime_resume_and_get() as __must_check
> >> follows some fixes of pm_runtime_resume_and_get() usage that went through other
> >> trees. These fixes are in linux-next but don't seem to be in the pm tree.
> >> We talk about:
> >> f04b4fb47d83 ("ASoC: sh: rz-ssi: Check return value of pm_runtime_resume_and_get()")
> >
> > In the sound-asoc tree.
> >
> >> 3d6b661330a7 ("crypto: stm32 - Revert broken pm_runtime_resume_and_get changes")
> >
> > In the crypto tree.
> >
> > Both those are merged into linux-next after the pm tree.  If Linus did
> > the same, the pm tree commit would break his build.  The only way you
> > can have that pm tree commit in linux-next is to ask Andrew Morton to
> > put it in the post linux-next part of his patch series.  Otherwise, it
> > needs to be removed form the pm tree and wait until after the next
> > merge window closes (or at least both the above trees have been merged
> > by Linus).
> >
> Rafael,
> can you take care of this?

Done.

I've dropped the commit in question from pm-core and I'm going to add
it back and submit by the end of the merge window.