1. Patchew
  2. linux
  3. [PATCH 0/3] unwind_user: Cleanups
  4. View patch

[PATCH 2/3] unwind_user/fp: Use dummies instead of ifdef

Jens Remus posted 3 patches 6 days, 7 hours ago
There is a newer version of this series
[PATCH 2/3] unwind_user/fp: Use dummies instead of ifdef
Posted by Jens Remus 6 days, 7 hours ago 
This simplifies the code.   unwind_user_next_fp() does not need to
return -EINVAL if config option HAVE_UNWIND_USER_FP is disabled, as
unwind_user_start() will then not select this unwind method and
unwind_user_next() will therefore not call it.

Note that enabling the config option HAVE_UNWIND_USER_FP without
defining ARCH_INIT_USER_FP_FRAME, ARCH_INIT_USER_FP_ENTRY_FRAME, and
unwind_user_at_function_start() will result in a compile error, which
is helpful when implementing support for unwind user fp in an
architecture.

Signed-off-by: Jens Remus <jremus@linux.ibm.com>
---
 include/linux/unwind_user.h | 14 +++++++++++---
 kernel/unwind/user.c        |  4 ----
 2 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/include/linux/unwind_user.h b/include/linux/unwind_user.h
index 7f7282516bf5..c3ff690a43e2 100644
--- a/include/linux/unwind_user.h
+++ b/include/linux/unwind_user.h
@@ -5,9 +5,17 @@
 #include <linux/unwind_user_types.h>
 #include <asm/unwind_user.h>
 
-#ifndef ARCH_INIT_USER_FP_FRAME
- #define ARCH_INIT_USER_FP_FRAME
-#endif
+#ifndef CONFIG_HAVE_UNWIND_USER_FP
+
+#define ARCH_INIT_USER_FP_FRAME
+#define ARCH_INIT_USER_FP_ENTRY_FRAME
+
+static inline bool unwind_user_at_function_start(struct pt_regs *regs)
+{
+	return false;
+}
+
+#endif /* !CONFIG_HAVE_UNWIND_USER_FP */
 
 int unwind_user(struct unwind_stacktrace *trace, unsigned int max_entries);
 
diff --git a/kernel/unwind/user.c b/kernel/unwind/user.c
index 0ca434f86e73..90ab3c1a205e 100644
--- a/kernel/unwind/user.c
+++ b/kernel/unwind/user.c
@@ -67,7 +67,6 @@ static int unwind_user_next_common(struct unwind_user_state *state,
 
 static int unwind_user_next_fp(struct unwind_user_state *state)
 {
-#ifdef CONFIG_HAVE_UNWIND_USER_FP
 	struct pt_regs *regs = task_pt_regs(current);
 
 	if (state->topmost && unwind_user_at_function_start(regs)) {
@@ -81,9 +80,6 @@ static int unwind_user_next_fp(struct unwind_user_state *state)
 		ARCH_INIT_USER_FP_FRAME(state->ws)
 	};
 	return unwind_user_next_common(state, &fp_frame);
-#else
-	return -EINVAL;
-#endif
 }
 
 static int unwind_user_next(struct unwind_user_state *state)
-- 
2.51.0
Re: [PATCH 2/3] unwind_user/fp: Use dummies instead of ifdef
Posted by Jens Remus 4 days, 7 hours ago 
On 11/25/2025 5:43 PM, Jens Remus wrote:
> This simplifies the code.   unwind_user_next_fp() does not need to
> return -EINVAL if config option HAVE_UNWIND_USER_FP is disabled, as
> unwind_user_start() will then not select this unwind method and
> unwind_user_next() will therefore not call it.
> 
> Note that enabling the config option HAVE_UNWIND_USER_FP without
> defining ARCH_INIT_USER_FP_FRAME, ARCH_INIT_USER_FP_ENTRY_FRAME, and
> unwind_user_at_function_start() will result in a compile error, which
> is helpful when implementing support for unwind user fp in an
> architecture.
> 
> Signed-off-by: Jens Remus <jremus@linux.ibm.com>

> diff --git a/include/linux/unwind_user.h b/include/linux/unwind_user.h

> @@ -5,9 +5,17 @@
>  #include <linux/unwind_user_types.h>
>  #include <asm/unwind_user.h>
>  
> -#ifndef ARCH_INIT_USER_FP_FRAME
> - #define ARCH_INIT_USER_FP_FRAME
> -#endif
> +#ifndef CONFIG_HAVE_UNWIND_USER_FP
> +
> +#define ARCH_INIT_USER_FP_FRAME
> +#define ARCH_INIT_USER_FP_ENTRY_FRAME

Will fix this as follows in the next version:

#define ARCH_INIT_USER_FP_FRAME(ws)
#define ARCH_INIT_USER_FP_ENTRY_FRAME(ws)

> +
> +static inline bool unwind_user_at_function_start(struct pt_regs *regs)
> +{
> +	return false;
> +}

Would it be better to provide a generic dummy implementation (see below)
or should each arch implement that if it cannot tell whether the topmost
frame is at function start? If so, would it move from linux/unwind_user.h
to asm-generic/unwind_user.h?  Either way it would need to be outside of
the !CONFIG_HAVE_UNWIND_USER_FP guard.

#ifndef unwind_user_at_function_start
static inline bool unwind_user_at_function_start(struct pt_regs *regs)
{
	return false;
}
#define unwind_user_at_function_start unwind_user_at_function_start
#endif

If doing so ARCH_INIT_USER_FP_ENTRY_FRAME should be handled similar, so
that archs do not need to provide their own dummy either:

#ifndef ARCH_INIT_USER_FP_ENTRY_FRAME
#define ARCH_INIT_USER_FP_ENTRY_FRAME(ws)
#endif

In that case only ARCH_INIT_USER_FP_FRAME would remain guarded by
!CONFIG_HAVE_UNWIND_USER_FP, so that compile would fail, if enabling
CONFIG_HAVE_UNWIND_USER_FP without providing ARCH_INIT_USER_FP_FRAME:

#ifndef CONFIG_HAVE_UNWIND_USER_FP

#define ARCH_INIT_USER_FP_FRAME(ws)

#endif /* !CONFIG_HAVE_UNWIND_USER_FP */

> +
> +#endif /* !CONFIG_HAVE_UNWIND_USER_FP */
>  
>  int unwind_user(struct unwind_stacktrace *trace, unsigned int max_entries);
>  
Thanks and regards,
Jens
-- 
Jens Remus
Linux on Z Development (D3303)
+49-7031-16-1128 Office
jremus@de.ibm.com

IBM

IBM Deutschland Research & Development GmbH; Vorsitzender des Aufsichtsrats: Wolfgang Wendt; Geschäftsführung: David Faller; Sitz der Gesellschaft: Böblingen; Registergericht: Amtsgericht Stuttgart, HRB 243294
IBM Data Privacy Statement: https://www.ibm.com/privacy/
Re: [PATCH 2/3] unwind_user/fp: Use dummies instead of ifdef
Posted by Peter Zijlstra 3 days, 13 hours ago 
On Thu, Nov 27, 2025 at 05:51:47PM +0100, Jens Remus wrote:
> On 11/25/2025 5:43 PM, Jens Remus wrote:
> > This simplifies the code.   unwind_user_next_fp() does not need to
> > return -EINVAL if config option HAVE_UNWIND_USER_FP is disabled, as
> > unwind_user_start() will then not select this unwind method and
> > unwind_user_next() will therefore not call it.
> > 
> > Note that enabling the config option HAVE_UNWIND_USER_FP without
> > defining ARCH_INIT_USER_FP_FRAME, ARCH_INIT_USER_FP_ENTRY_FRAME, and
> > unwind_user_at_function_start() will result in a compile error, which
> > is helpful when implementing support for unwind user fp in an
> > architecture.
> > 
> > Signed-off-by: Jens Remus <jremus@linux.ibm.com>
> 
> > diff --git a/include/linux/unwind_user.h b/include/linux/unwind_user.h
> 
> > @@ -5,9 +5,17 @@
> >  #include <linux/unwind_user_types.h>
> >  #include <asm/unwind_user.h>
> >  
> > -#ifndef ARCH_INIT_USER_FP_FRAME
> > - #define ARCH_INIT_USER_FP_FRAME
> > -#endif
> > +#ifndef CONFIG_HAVE_UNWIND_USER_FP
> > +
> > +#define ARCH_INIT_USER_FP_FRAME
> > +#define ARCH_INIT_USER_FP_ENTRY_FRAME
> 
> Will fix this as follows in the next version:
> 
> #define ARCH_INIT_USER_FP_FRAME(ws)
> #define ARCH_INIT_USER_FP_ENTRY_FRAME(ws)
> 
> > +
> > +static inline bool unwind_user_at_function_start(struct pt_regs *regs)
> > +{
> > +	return false;
> > +}
> 
> Would it be better to provide a generic dummy implementation (see below)
> or should each arch implement that if it cannot tell whether the topmost
> frame is at function start? If so, would it move from linux/unwind_user.h
> to asm-generic/unwind_user.h?  Either way it would need to be outside of
> the !CONFIG_HAVE_UNWIND_USER_FP guard.

I suppose a common fallback would be fine.

Patchew 2.1-devel-b67e4c2

© 2016 - 2025 Red Hat, Inc.