On Tue, Feb 09, 2021 at 01:24:30PM +0000, Mark Rutland wrote:
> On Tue, Feb 09, 2021 at 09:34:10AM +0100, Peter Zijlstra wrote:
> >
> > Subject: lockdep: Noinstr annotate warn_bogus_irq_restore()
> > From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> > Date: Tue Feb 9 09:30:03 CET 2021
> >
> > vmlinux.o: warning: objtool: lock_is_held_type()+0x107: call to warn_bogus_irq_restore() leaves .noinstr.text section
> >
> > As per the general rule that WARNs are allowed to violate noinstr to
> > get out, annotate it away.
> >
> > Fixes: 997acaf6b4b5 ("lockdep: report broken irq restoration")
> > Reported-by: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
> > Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
>
> Whoops; sorry for missing that!
>
> Acked-by: Mark Rutland <mark.rutland@xxxxxxx>
This runs into sodding header hell on mips (and possibly others)
How's this then?
---
Subject: lockdep: Noinstr annotate warn_bogus_irq_restore()
From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Date: Tue Feb 9 09:30:03 CET 2021
vmlinux.o: warning: objtool: lock_is_held_type()+0x107: call to warn_bogus_irq_restore() leaves .noinstr.text section
As per the general rule that WARNs are allowed to violate noinstr to
get out, annotate it away.
Fixes: 997acaf6b4b5 ("lockdep: report broken irq restoration")
Reported-by: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Acked-by: Mark Rutland <mark.rutland@xxxxxxx>
---
include/linux/irqflags.h | 5 ++++-
kernel/locking/irqflag-debug.c | 4 +++-
2 files changed, 7 insertions(+), 2 deletions(-)
--- a/kernel/locking/irqflag-debug.c
+++ b/kernel/locking/irqflag-debug.c
@@ -4,8 +4,10 @@
#include <linux/export.h>
#include <linux/irqflags.h>
-void warn_bogus_irq_restore(void)
+noinstr void warn_bogus_irq_restore(void)
{
+ instrumentation_begin();
WARN_ONCE(1, "raw_local_irq_restore() called with IRQs enabled\n");
+ instrumentation_end();
}
EXPORT_SYMBOL(warn_bogus_irq_restore);
