linux-next: build failure after merge of the percpu tree

linux-next: build failure after merge of the percpu tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1578 bytes --]

Hi all,

After merging the percpu tree, today's linux-next build (sparc64
defconfig) failed like this:

mm/percpu.c: In function 'pcpu_page_first_chunk':
mm/percpu.c:3336:17: error: implicit declaration of function 'flush_cache_vmap_early'; did you mean 'flush_cache_vmap'? [-Werror=implicit-function-declaration]
 3336 |                 flush_cache_vmap_early(unit_addr, unit_addr + ai->unit_size); 
      |                 ^~~~~~~~~~~~~~~~~~~~~~
      |                 flush_cache_vmap
cc1: some warnings being treated as errors

Caused by commit

  a95c15a43f4a ("mm: Introduce flush_cache_vmap_early() and its riscv implementation")

I have applied the following fix patch for today.  Are there other
archs that don't use asm-generic/cacheflush.h?

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 11 Dec 2023 16:57:00 +1100
Subject: [PATCH] fix up for "mm: Introduce flush_cache_vmap_early() and its riscv implementation"

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 arch/sparc/include/asm/cacheflush.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/sparc/include/asm/cacheflush.h b/arch/sparc/include/asm/cacheflush.h
index 881ac76eab93..9d87b2bcb217 100644
--- a/arch/sparc/include/asm/cacheflush.h
+++ b/arch/sparc/include/asm/cacheflush.h
@@ -10,4 +10,11 @@
 #else
 #include <asm/cacheflush_32.h>
 #endif
+
+#ifndef __ASSEMBLY__
+static inline void flush_cache_vmap_early(unsigned long start, unsigned long end)
+{
+}
+#endif
+
 #endif
-- 
2.40.1

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

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

[Alexandre Ghiti]

Hi Stephen,

On Mon, Dec 11, 2023 at 7:14 AM Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Hi all,
>
> After merging the percpu tree, today's linux-next build (sparc64
> defconfig) failed like this:
>
> mm/percpu.c: In function 'pcpu_page_first_chunk':
> mm/percpu.c:3336:17: error: implicit declaration of function 'flush_cache_vmap_early'; did you mean 'flush_cache_vmap'? [-Werror=implicit-function-declaration]
>  3336 |                 flush_cache_vmap_early(unit_addr, unit_addr + ai->unit_size);
>       |                 ^~~~~~~~~~~~~~~~~~~~~~
>       |                 flush_cache_vmap
> cc1: some warnings being treated as errors
>
> Caused by commit
>
>   a95c15a43f4a ("mm: Introduce flush_cache_vmap_early() and its riscv implementation")
>
> I have applied the following fix patch for today.  Are there other
> archs that don't use asm-generic/cacheflush.h?

It seems like most archs do not include this file, I should have
checked. As I'm a bit scared of the possible side-effects of including
asm-generic/cacheflush.h, I'll define flush_cache_vmap_early() on all
archs that do define flush_cache_vmap().

Stephen, do you want a patch fix? Or do you want me to send a new
version of the current patches so that you can drop them for now?

Sorry for the oversight,

Thanks,

Alex

>
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Mon, 11 Dec 2023 16:57:00 +1100
> Subject: [PATCH] fix up for "mm: Introduce flush_cache_vmap_early() and its riscv implementation"
>
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> ---
>  arch/sparc/include/asm/cacheflush.h | 7 +++++++
>  1 file changed, 7 insertions(+)
>
> diff --git a/arch/sparc/include/asm/cacheflush.h b/arch/sparc/include/asm/cacheflush.h
> index 881ac76eab93..9d87b2bcb217 100644
> --- a/arch/sparc/include/asm/cacheflush.h
> +++ b/arch/sparc/include/asm/cacheflush.h
> @@ -10,4 +10,11 @@
>  #else
>  #include <asm/cacheflush_32.h>
>  #endif
> +
> +#ifndef __ASSEMBLY__
> +static inline void flush_cache_vmap_early(unsigned long start, unsigned long end)
> +{
> +}
> +#endif
> +
>  #endif
> --
> 2.40.1
>
> --
> Cheers,
> Stephen Rothwell

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

[Dennis Zhou]

Hello,

On Mon, Dec 11, 2023 at 09:31:25AM +0100, Alexandre Ghiti wrote:
> Hi Stephen,
> 
> On Mon, Dec 11, 2023 at 7:14 AM Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> >
> > Hi all,
> >
> > After merging the percpu tree, today's linux-next build (sparc64
> > defconfig) failed like this:
> >
> > mm/percpu.c: In function 'pcpu_page_first_chunk':
> > mm/percpu.c:3336:17: error: implicit declaration of function 'flush_cache_vmap_early'; did you mean 'flush_cache_vmap'? [-Werror=implicit-function-declaration]
> >  3336 |                 flush_cache_vmap_early(unit_addr, unit_addr + ai->unit_size);
> >       |                 ^~~~~~~~~~~~~~~~~~~~~~
> >       |                 flush_cache_vmap
> > cc1: some warnings being treated as errors
> >
> > Caused by commit
> >
> >   a95c15a43f4a ("mm: Introduce flush_cache_vmap_early() and its riscv implementation")
> >
> > I have applied the following fix patch for today.  Are there other
> > archs that don't use asm-generic/cacheflush.h?
> 

I'm surprised automation didn't catch this as this should have failed
for any sparc build? It passed `sparc allmodconfig gcc` on my branches.

> It seems like most archs do not include this file, I should have
> checked. As I'm a bit scared of the possible side-effects of including
> asm-generic/cacheflush.h, I'll define flush_cache_vmap_early() on all
> archs that do define flush_cache_vmap().
> 

Hmmm. That makes sense, but we'd still need to check so we have the
generic #ifndef definition included everywhere too.

> Stephen, do you want a patch fix? Or do you want me to send a new
> version of the current patches so that you can drop them for now?
> 

The for-next tree gets recreated from pulls of the maintainers' trees.
I'm going to drop the series from percpu and then we can go again with a
v2.

> Sorry for the oversight,
> 

All good it happens. It's why the automation is there.

Thanks,
Dennis

> Thanks,
> 
> Alex
> 
> >
> > From: Stephen Rothwell <sfr@canb.auug.org.au>
> > Date: Mon, 11 Dec 2023 16:57:00 +1100
> > Subject: [PATCH] fix up for "mm: Introduce flush_cache_vmap_early() and its riscv implementation"
> >
> > Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> > ---
> >  arch/sparc/include/asm/cacheflush.h | 7 +++++++
> >  1 file changed, 7 insertions(+)
> >
> > diff --git a/arch/sparc/include/asm/cacheflush.h b/arch/sparc/include/asm/cacheflush.h
> > index 881ac76eab93..9d87b2bcb217 100644
> > --- a/arch/sparc/include/asm/cacheflush.h
> > +++ b/arch/sparc/include/asm/cacheflush.h
> > @@ -10,4 +10,11 @@
> >  #else
> >  #include <asm/cacheflush_32.h>
> >  #endif
> > +
> > +#ifndef __ASSEMBLY__
> > +static inline void flush_cache_vmap_early(unsigned long start, unsigned long end)
> > +{
> > +}
> > +#endif
> > +
> >  #endif
> > --
> > 2.40.1
> >
> > --
> > Cheers,
> > Stephen Rothwell

linux-next: build failure after merge of the percpu tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1203 bytes --]

Hi all,

After merging the percpu tree, today's linux-next build (sparc defconfig)
failed like this:

mm/percpu.c: In function 'pcpu_balance_workfn':
mm/percpu.c:1613:3: error: implicit declaration of function 'cond_resched'; did you mean 'should_resched'? [-Werror=implicit-function-declaration]
   cond_resched();
   ^~~~~~~~~~~~
   should_resched

Caused by commit

  accd4f36a7d1 ("percpu: add a schedule point in pcpu_balance_workfn()")

I have added this patch for today:

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 26 Feb 2018 14:47:39 +1100
Subject: [PATCH] percpu: include sched.h for cond_resched()

Fixes: accd4f36a7d1 ("percpu: add a schedule point in pcpu_balance_workfn()")
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 mm/percpu.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/mm/percpu.c b/mm/percpu.c
index 36e7b65ba6cf..15a398c00791 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -80,6 +80,7 @@
 #include <linux/vmalloc.h>
 #include <linux/workqueue.h>
 #include <linux/kmemleak.h>
+#include <linux/sched.h>
 
 #include <asm/cacheflush.h>
 #include <asm/sections.h>
-- 
2.16.1

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

linux-next: build failure after merge of the percpu tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1151 bytes --]

Hi all,

After merging the percpu tree, today's linux-next build (powerpc
ppc64_defconfig) failed like this:

In file included from arch/powerpc/include/asm/xics.h:9:0,
                 from arch/powerpc/kernel/asm-offsets.c:47:
include/linux/interrupt.h:372:0: warning: "set_softirq_pending" redefined
 #define set_softirq_pending(x) (local_softirq_pending() = (x))
 ^
In file included from include/linux/hardirq.h:8:0,
                 from include/linux/memcontrol.h:24,
                 from include/linux/swap.h:8,
                 from include/linux/suspend.h:4,
                 from arch/powerpc/kernel/asm-offsets.c:24:
arch/powerpc/include/asm/hardirq.h:25:0: note: this is the location of the previous definition
 #define set_softirq_pending(x) __this_cpu_write(irq_stat._softirq_pending, (x))
 ^

I got lots (and lots :-() of these and some were considered errors
(powerpc is built with -Werr in arch/powerpc).

Caused by commit 5828f666c069 ("powerpc: Replace __get_cpu_var uses").

I have used the percpu tree from next-20140826 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

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

[Tejun Heo]

Hello, again.

On Thu, Jan 06, 2011 at 03:18:15PM +1100, Stephen Rothwell wrote:
> > My apologies.  I forgot an earlier patch to introduce this_cpu_has()
> > macro.  I've reverted the offending commit.
> 
> But that revert is not in your published for-next branch yet.  I have
> manually reverted that commit for today.

Ah, embarrassing.  I updated for-2.6.38 but forgot to update for-next
accordingly.  Fixed.

Thank you.

-- 
tejun

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

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1077 bytes --]

Hi Tejun,

On Tue, 4 Jan 2011 06:13:14 +0100 Tejun Heo <tj@kernel.org> wrote:
>
> On Tue, Jan 04, 2011 at 03:21:05PM +1100, Stephen Rothwell wrote:
> > After merging the percpu tree, today's linux-next build (x86_64
> > allmodconfig) failed like this:
> > 
> > arch/x86/kernel/cpu/mcheck/therm_throt.c: In function 'intel_thermal_interrupt':
> > arch/x86/kernel/cpu/mcheck/therm_throt.c:368: error: implicit declaration of function 'this_cpu_has'
> > 
> > Caused by commit 6ac0bb7148b93fb40bccba5dff06d51a3e3ea283 ("x86: use
> > this_cpu_has for thermal_interrupt").
> > 
> > this_cpu_has() does not exist anywhere except in this introduced usage.
> > 
> > I have used the percpu tree from next-20101231 for today.
> 
> My apologies.  I forgot an earlier patch to introduce this_cpu_has()
> macro.  I've reverted the offending commit.

But that revert is not in your published for-next branch yet.  I have
manually reverted that commit for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

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

[Tejun Heo]

Hello, Stephen.

On Tue, Jan 04, 2011 at 03:21:05PM +1100, Stephen Rothwell wrote:
> After merging the percpu tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
> 
> arch/x86/kernel/cpu/mcheck/therm_throt.c: In function 'intel_thermal_interrupt':
> arch/x86/kernel/cpu/mcheck/therm_throt.c:368: error: implicit declaration of function 'this_cpu_has'
> 
> Caused by commit 6ac0bb7148b93fb40bccba5dff06d51a3e3ea283 ("x86: use
> this_cpu_has for thermal_interrupt").
> 
> this_cpu_has() does not exist anywhere except in this introduced usage.
> 
> I have used the percpu tree from next-20101231 for today.

My apologies.  I forgot an earlier patch to introduce this_cpu_has()
macro.  I've reverted the offending commit.

Thank you.

--
tejun

linux-next: build failure after merge of the percpu tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 649 bytes --]

Hi all,

After merging the percpu tree, today's linux-next build (x86_64
allmodconfig) failed like this:

arch/x86/kernel/cpu/mcheck/therm_throt.c: In function 'intel_thermal_interrupt':
arch/x86/kernel/cpu/mcheck/therm_throt.c:368: error: implicit declaration of function 'this_cpu_has'

Caused by commit 6ac0bb7148b93fb40bccba5dff06d51a3e3ea283 ("x86: use
this_cpu_has for thermal_interrupt").

this_cpu_has() does not exist anywhere except in this introduced usage.

I have used the percpu tree from next-20101231 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

linux-next: build failure after merge of the percpu tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1108 bytes --]

Hi all,

After merging the percpu tree, today's linux-next build (x86_64
allmodconfig) failed like this:

In file included from drivers/vhost/net.c:26:
include/linux/if_macvlan.h: In function 'macvlan_count_rx':
include/linux/if_macvlan.h:69: error: read-only variable 'tcp_ptr__' used as 'asm' output
In file included from drivers/net/macvlan.c:30:
include/linux/if_macvlan.h: In function 'macvlan_count_rx':
include/linux/if_macvlan.h:69: error: read-only variable 'tcp_ptr__' used as 'asm' output
In file included from drivers/net/macvtap.c:2:
include/linux/if_macvlan.h: In function 'macvlan_count_rx':
include/linux/if_macvlan.h:69: error: read-only variable 'tcp_ptr__' used as 'asm' output

Caused by commit 7f56c13698abfb1bdf6cf2b5cfcd80626d44c12e ("x86, percpu:
Optimize this_cpu_ptr").

That line in if_macvlan.h is:

	rx_stats = this_cpu_ptr(vlan->rx_stats);

Where vlan is a "const struct macvlan_dev *".

I have used the percpu tree from next-20100909 for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]