Skip to contentSkip to conversation

ipmi: avoid atomic_inc in exit function

PATCH In review Arnd Bergmann 3 participants 5 messages Apr 15, 2019 – Apr 16, 2019 IPMI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ipmi: avoid atomic_inc in exit function

This causes a link failure on ARM in certain configurations, when we reference each atomic operation from .alt.smp.init in order to patch out atomics on non-SMP systems:

.exit.text' referenced in section .alt.smp.init' of drivers/char/ipmi/ipmi_msghandler.o: defined in discarded section `.exit.text' of drivers/char/ipmi/ipmi_msghandler.o

In this case, we can trivially replace the atomic_inc() with an atomic_set() that has the same effect and does not require a fixup.

---
 drivers/char/ipmi/ipmi_msghandler.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/ipmi/ipmi_msghandler.c +1 -1
index e8ba67834746..c48198eef510 100644
@@ -5164,7 +5164,7 @@ static void __exit cleanup_ipmi(void)
 		 * avoids problems with race conditions removing the timer
 		 * here.
 		 */
-		atomic_inc(&stop_operation);
+		atomic_set(&stop_operation, 1);
 		del_timer_sync(&ipmi_timer);
 
 		initialized = false;
Signature · 1 line
2.20.0
#1 lore raw
Quoted 9 lines · Arnd Bergmann

This causes a link failure on ARM in certain configurations, when we reference each atomic operation from .alt.smp.init in order to patch out atomics on non-SMP systems:

.exit.text' referenced in section .alt.smp.init' of drivers/char/ipmi/ipmi_msghandler.o: defined in discarded section `.exit.text' of drivers/char/ipmi/ipmi_msghandler.o

In this case, we can trivially replace the atomic_inc() with an atomic_set() that has the same effect and does not require a fixup.

I'd rather fіx the arm section management. Using atomic in exit routines is perfectly valid, and it would seem odd to forbid it.

#2 lore raw
Quoted 5 lines · Christoph Hellwig
Quoted 4 lines · Arnd Bergmann

This causes a link failure on ARM in certain configurations, when we reference each atomic operation from .alt.smp.init in order to patch out atomics on non-SMP systems:

Quoted 8 lines
Quoted 5 lines

.exit.text' referenced in section .alt.smp.init' of drivers/char/ipmi/ipmi_msghandler.o: defined in discarded section `.exit.text' of drivers/char/ipmi/ipmi_msghandler.o

In this case, we can trivially replace the atomic_inc() with an atomic_set() that has the same effect and does not require a fixup.

I'd rather fіx the arm section management. Using atomic in exit routines is perfectly valid, and it would seem odd to forbid it.

That was my first thought, too. It's kind of hard to believe that the IPMI driver is the only thing that does an atomic_inc() in the exit code.

-corey

#3 lore raw
Quoted 13 lines · Corey Minyard
Quoted 11 lines · Christoph Hellwig
Quoted 9 lines · Arnd Bergmann

This causes a link failure on ARM in certain configurations, when we reference each atomic operation from .alt.smp.init in order to patch out atomics on non-SMP systems:

.exit.text' referenced in section .alt.smp.init' of drivers/char/ipmi/ipmi_msghandler.o: defined in discarded section `.exit.text' of drivers/char/ipmi/ipmi_msghandler.o

In this case, we can trivially replace the atomic_inc() with an atomic_set() that has the same effect and does not require a fixup.

Quoted 6 lines
Quoted 2 lines

I'd rather fіx the arm section management. Using atomic in exit routines is perfectly valid, and it would seem odd to forbid it.

That was my first thought, too. It's kind of hard to believe that the IPMI driver is the only thing that does an atomic_inc() in the exit code.

That's what I had thought as well at first, and I carried a patch to work around this by not dropping the .text.exit section on ARM when SMP patching is enabled for a few years. I never sent this because that can waste a significant amount of kernel memory, and I knew the warning is harmless.

When revisiting it now, I found that this one was the only instance I ever hit. It seems to be that using atomics in module_exit() is indeed odd, because the function is rarely concurrent with anything else.

Arnd

#4 lore raw
Quoted 27 lines · Arnd Bergmann
Quoted 13 lines · Corey Minyard
Quoted 11 lines · Christoph Hellwig
Quoted 9 lines · Arnd Bergmann

This causes a link failure on ARM in certain configurations, when we reference each atomic operation from .alt.smp.init in order to patch out atomics on non-SMP systems:

.exit.text' referenced in section .alt.smp.init' of drivers/char/ipmi/ipmi_msghandler.o: defined in discarded section `.exit.text' of drivers/char/ipmi/ipmi_msghandler.o

In this case, we can trivially replace the atomic_inc() with an atomic_set() that has the same effect and does not require a fixup.

Quoted 6 lines
Quoted 2 lines

I'd rather fіx the arm section management. Using atomic in exit routines is perfectly valid, and it would seem odd to forbid it.

That was my first thought, too. It's kind of hard to believe that the IPMI driver is the only thing that does an atomic_inc() in the exit code.

That's what I had thought as well at first, and I carried a patch to work around this by not dropping the .text.exit section on ARM when SMP patching is enabled for a few years. I never sent this because that can waste a significant amount of kernel memory, and I knew the warning is harmless.

Quoted 4 lines

When revisiting it now, I found that this one was the only instance I ever hit. It seems to be that using atomics in module_exit() is indeed odd, because the function is rarely concurrent with anything else.

I've added the change to my tree; it actually makes a little more sense, so I'm ok with it.

I guess it's up to you to deal with any new ones that happen in the future ;-).

-corey