setup_profiling_timer() is not used by any code at this point. Since a
default weak implementation exists, there is no need to keep this
arch-specific definition around. Remove it along with the now-redundant
profile header includes.

Signed-off-by: Anthony Iliopoulos <ailiop@suse.com>
---
 arch/sh/kernel/smp.c  | 9 ---------
 arch/sh/kernel/time.c | 1 -
 2 files changed, 10 deletions(-)

diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c
index 108d808767fa..7209a72134be 100644
--- a/arch/sh/kernel/smp.c
+++ b/arch/sh/kernel/smp.c
@@ -21,7 +21,6 @@
 #include <linux/sched/hotplug.h>
 #include <linux/atomic.h>
 #include <linux/clockchips.h>
-#include <linux/profile.h>
 
 #include <asm/processor.h>
 #include <asm/mmu_context.h>
@@ -322,14 +321,6 @@ void smp_message_recv(unsigned int msg)
 	}
 }
 
-#ifdef CONFIG_PROFILING
-/* Not really SMP stuff ... */
-int setup_profiling_timer(unsigned int multiplier)
-{
-	return 0;
-}
-#endif
-
 #ifdef CONFIG_MMU
 
 static void flush_tlb_all_ipi(void *info)
diff --git a/arch/sh/kernel/time.c b/arch/sh/kernel/time.c
index 821a09cbd605..f3e70d2d573b 100644
--- a/arch/sh/kernel/time.c
+++ b/arch/sh/kernel/time.c
@@ -9,7 +9,6 @@
  */
 #include <linux/kernel.h>
 #include <linux/init.h>
-#include <linux/profile.h>
 #include <linux/timex.h>
 #include <linux/sched.h>
 #include <linux/clockchips.h>
-- 
2.55.0

Hi Anthony,

Thanks for your patch!

On Fri, 21 Aug 2026 at 09:27, Anthony Iliopoulos <ailiop@suse.com> wrote:
> setup_profiling_timer() is not used by any code at this point. Since a

It is used in write_profile()
https://elixir.bootlin.com/linux/v7.2/source/kernel/profile.c#L202

> default weak implementation exists, there is no need to keep this
> arch-specific definition around. Remove it along with the now-redundant
> profile header includes.
>
> Signed-off-by: Anthony Iliopoulos <ailiop@suse.com>

> --- a/arch/sh/kernel/smp.c
> +++ b/arch/sh/kernel/smp.c
> @@ -21,7 +21,6 @@
>  #include <linux/sched/hotplug.h>
>  #include <linux/atomic.h>
>  #include <linux/clockchips.h>
> -#include <linux/profile.h>
>
>  #include <asm/processor.h>
>  #include <asm/mmu_context.h>
> @@ -322,14 +321,6 @@ void smp_message_recv(unsigned int msg)
>         }
>  }
>
> -#ifdef CONFIG_PROFILING
> -/* Not really SMP stuff ... */
> -int setup_profiling_timer(unsigned int multiplier)
> -{
> -       return 0;

This returns zero (success), while the default weak implementation
returns -EINVAL, so your change breaks write_profile().
https://elixir.bootlin.com/linux/v7.2/source/kernel/profile.c#L181

> -}
> -#endif
> -
>  #ifdef CONFIG_MMU
>
>  static void flush_tlb_all_ipi(void *info)

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

On Fri, Aug 21, 2026 at 09:36:46AM +0200, Geert Uytterhoeven wrote:
> Hi Anthony,
> 
> Thanks for your patch!
> 
> On Fri, 21 Aug 2026 at 09:27, Anthony Iliopoulos <ailiop@suse.com> wrote:
> > setup_profiling_timer() is not used by any code at this point. Since a
> 
> It is used in write_profile()
> https://elixir.bootlin.com/linux/v7.2/source/kernel/profile.c#L202
> 
> > default weak implementation exists, there is no need to keep this
> > arch-specific definition around. Remove it along with the now-redundant
> > profile header includes.
> >
> > Signed-off-by: Anthony Iliopoulos <ailiop@suse.com>
> 
> > --- a/arch/sh/kernel/smp.c
> > +++ b/arch/sh/kernel/smp.c
> > @@ -21,7 +21,6 @@
> >  #include <linux/sched/hotplug.h>
> >  #include <linux/atomic.h>
> >  #include <linux/clockchips.h>
> > -#include <linux/profile.h>
> >
> >  #include <asm/processor.h>
> >  #include <asm/mmu_context.h>
> > @@ -322,14 +321,6 @@ void smp_message_recv(unsigned int msg)
> >         }
> >  }
> >
> > -#ifdef CONFIG_PROFILING
> > -/* Not really SMP stuff ... */
> > -int setup_profiling_timer(unsigned int multiplier)
> > -{
> > -       return 0;
> 
> This returns zero (success), while the default weak implementation
> returns -EINVAL, so your change breaks write_profile().
> https://elixir.bootlin.com/linux/v7.2/source/kernel/profile.c#L181

Thanks for the review, yes the plan was to also remove the call entirely
from profile code itself, so that the resetting functionality can remain
and the arch-specific stubs could be dropped. I should clearly have sent
that patch first. I just did it now:

https://lore.kernel.org/all/20260821091744.91496-1-ailiop@suse.com/

Regards,
Anthony

Hi Anthony,

On Thu, 2026-08-20 at 14:39 +0200, Anthony Iliopoulos wrote:
> setup_profiling_timer() is not used by any code at this point. Since a
> default weak implementation exists, there is no need to keep this
> arch-specific definition around. Remove it along with the now-redundant
> profile header includes.

Do you have any clue why it was still implemented for LoongArch which is
a rather new architecture? I'm also seeing that it's not used anywhere,
I'm just wondering why it was not removed before and why it was even
included when LoongArch was added.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Hi Adrian,

On Fri, 21 Aug 2026 at 14:11, John Paul Adrian Glaubitz
<glaubitz@physik.fu-berlin.de> wrote:
> On Thu, 2026-08-20 at 14:39 +0200, Anthony Iliopoulos wrote:
> > setup_profiling_timer() is not used by any code at this point. Since a
> > default weak implementation exists, there is no need to keep this
> > arch-specific definition around. Remove it along with the now-redundant
> > profile header includes.
>
> Do you have any clue why it was still implemented for LoongArch which is
> a rather new architecture? I'm also seeing that it's not used anywhere,
> I'm just wondering why it was not removed before and why it was even
> included when LoongArch was added.

To make it _succeed_ when writing an integer value to that file?
The default weak implementation returns an error code.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

On Thu, Aug 20, 2026 at 04:00:37PM +0200, John Paul Adrian Glaubitz wrote:
> Hi Anthony,
> 
> On Thu, 2026-08-20 at 14:39 +0200, Anthony Iliopoulos wrote:
> > setup_profiling_timer() is not used by any code at this point. Since a
> > default weak implementation exists, there is no need to keep this
> > arch-specific definition around. Remove it along with the now-redundant
> > profile header includes.
> 
> Do you have any clue why it was still implemented for LoongArch which is
> a rather new architecture? I'm also seeing that it's not used anywhere,
> I'm just wondering why it was not removed before and why it was even
> included when LoongArch was added.

Its addition to LoongArch [1] slightly predates converting it to a weak
function [2], so up until that point all archs had to include the stub.

Regards,
Anthony

[1] 46859ac8af52 ("LoongArch: Add multi-processor (SMP) support")
[2] 787dbea11a5d ("profile: setup_profiling_timer() is moslty not implemented")