Skip to contentSkip to conversation

x86: oprofile/op_model_amd.c set return values for op_amd_handle_ibs()

PATCH In review Jaswinder Singh Rajput 3 participants 4 messages Jun 18, 2009 – Jul 31, 2009 x86

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86: oprofile/op_model_amd.c set return values for op_amd_handle_ibs()

op_amd_handle_ibs() should return 0 when IBS is not present or not defined.

Fix compilation warning:
CC [M] arch/x86/oprofile/op_model_amd.o
arch/x86/oprofile/op_model_amd.c: In function ‘op_amd_handle_ibs’:
arch/x86/oprofile/op_model_amd.c:217: warning: no return statement in function returning non-void

---
 arch/x86/oprofile/op_model_amd.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)
arch/x86/oprofile/op_model_amd.c +5 -2
index cc93046..e95268e 100644
@@ -132,7 +132,7 @@ op_amd_handle_ibs(struct pt_regs * const regs,
 	struct op_entry entry;
 
 	if (!has_ibs)
-		return 1;
+		return 0;
 
 	if (ibs_config.fetch_enabled) {
 		rdmsrl(MSR_AMD64_IBSFETCHCTL, ctl);
@@ -214,7 +214,10 @@ static void op_amd_stop_ibs(void)
 #else
 
 static inline int op_amd_handle_ibs(struct pt_regs * const regs,
-				    struct op_msrs const * const msrs) { }
+				    struct op_msrs const * const msrs)
+{
+	return 0;
+}
 static inline void op_amd_start_ibs(void) { }
 static inline void op_amd_stop_ibs(void) { }
 
Signature · 1 line
1.6.0.6
#1 lore raw
Quoted 8 lines · Jaswinder Singh Rajput

op_amd_handle_ibs() should return 0 when IBS is not present or not defined.

Fix compilation warning:
CC [M] arch/x86/oprofile/op_model_amd.o
arch/x86/oprofile/op_model_amd.c: In function ‘op_amd_handle_ibs’:
arch/x86/oprofile/op_model_amd.c:217: warning: no return statement in function returning non-void

Quoted 4 lines
---
 arch/x86/oprofile/op_model_amd.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

Applied to:

git://git.kernel.org/pub/scm/linux/kernel/git/rric/oprofile.git master

Thanks Jaswinder,

-Robert

Signature · 3 lines
Advanced Micro Devices, Inc.
Operating System Research Center
email: [email protected]
#2 lore raw
Quoted 9 lines · On Thu, 18 Jun 2009 16:47:31 +0200 Robert Richter <[email protected]>…
Quoted 8 lines · Jaswinder Singh Rajput

op_amd_handle_ibs() should return 0 when IBS is not present or not defined.

Fix compilation warning:
CC [M] arch/x86/oprofile/op_model_amd.o
arch/x86/oprofile/op_model_amd.c: In function ___op_amd_handle_ibs___:
arch/x86/oprofile/op_model_amd.c:217: warning: no return statement in function returning non-void

Quoted 9 lines
Quoted 4 lines
---
 arch/x86/oprofile/op_model_amd.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

Applied to:

git://git.kernel.org/pub/scm/linux/kernel/git/rric/oprofile.git master

Something has gone badly wrong. This patch took six weeks to turn up in linux-next.

Apart from anything else this led me to have to fix something which was already fixed.

All you guys *saw* that fix and still this didn't prompt anyone to wonder what had gone wrong.

My fix is better, too. The op_amd_handle_ibs() return value is ignored, so it should return void.

#3 lore raw
Quoted 17 lines · Andrew Morton
Quoted 14 lines · On Thu, 18 Jun 2009 16:47:31 +0200 Robert Richter <[email protected]>…
Quoted 12 lines · Jaswinder Singh Rajput

op_amd_handle_ibs() should return 0 when IBS is not present or not defined.

Fix compilation warning:
CC [M] arch/x86/oprofile/op_model_amd.o
arch/x86/oprofile/op_model_amd.c: In function ___op_amd_handle_ibs___:
arch/x86/oprofile/op_model_amd.c:217: warning: no return statement in function returning non-void

---
 arch/x86/oprofile/op_model_amd.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)
Quoted 7 lines
Quoted 4 lines

Applied to:

git://git.kernel.org/pub/scm/linux/kernel/git/rric/oprofile.git master

Something has gone badly wrong. This patch took six weeks to turn up in linux-next.

Yes, at that time it was not yet quite clear of how to send oprofile patches upstream and to linux-next. I was discussing this with Ingo and we finally agreed to send it over tip to linux-next. It took also some weeks to find the right workflow and types of branches for this.

All this is sorted out now. See also my current branches in the oprofile tree:

http://git.kernel.org/?p=linux/kernel/git/rric/oprofile.git;a=summary

In the last days I added oprofile to linux-next as a backup until Ingo will be back. The oprofile for-next branch will work also as backup in the future to make sure, all oprofile patches will be in time in linux-next. I see the things working now.

Again, I also was not happy with the workflow, but I have fixed it together with Ingo in the last weeks. That oprofile pops up in linux-next now is the result of this. So please, see this more as 'things are working now' than 'it took 6 weeks into linux-next'. Also, it is not too late for this, there are some more weeks to got for the next merge window.

Quoted 6 lines

Apart from anything else this led me to have to fix something which was already fixed.

All you guys *saw* that fix and still this didn't prompt anyone to wonder what had gone wrong.

This went wrong, since the fix didn't make it to tip as fast as the patch set itself. But things are fixed now and this shouln't happen again.

Quoted 3 lines

My fix is better, too. The op_amd_handle_ibs() return value is ignored, so it should return void.

There are pros and cons, since the handler actually should tell it did handle the nmi, it should return something. But for some reason this has been activly deactivated (maybe a workaround for a bug). And I decided to not to change the code until I now the reason for it. So the return value is ignored for now. But anyway, there aren't much differences in the end. So I will apply your patch instead.

-Robert

Signature · 3 lines
Advanced Micro Devices, Inc.
Operating System Research Center
email: [email protected]