bus: arm-ccn: Fix cpu notifier priority
PATCH 1/1 bus: arm-ccn: Fix cpu notifier priority
Fix cpu priority of notifier block that was erroneously set to return value of next statement.
--- drivers/bus/arm-ccn.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/bus/arm-ccn.c +1 -1
@@ -1271,7 +1271,7 @@ static int arm_ccn_pmu_init(struct arm_ccn *ccn)* picked to have a chance to migrate events before perf is notified. */ ccn->dt.cpu_nb.notifier_call = arm_ccn_pmu_cpu_notifier; - ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 , + ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 ; err =register_cpu_notifier (&ccn->dt.cpu_nb);if (err) goto error_cpu_notifier;
Signature · 1 line
1.9.1
PATCH 1/1 bus: arm-ccn: Fix cpu notifier priority
Quoted 16 lines · Jan Glauber
Fix cpu priority of notifier block that was erroneously set to return value of next statement.
Signed-off-by Jan Glauber <[email protected]>--- drivers/bus/arm-ccn.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/bus/arm-ccn.c +0
-0@@ -1271,7 +1271,7 @@ static int arm_ccn_pmu_init(struct arm_ccn *ccn) * picked to have a chance to migrate events before perf isnotified.
drivers/bus/arm-ccn.c static int arm_ccn_pmu_init(struct arm_ccn *ccn) · line 1277
Quoted 7 lines
drivers/bus/arm-ccn.c +1
-1*/ ccn->dt.cpu_nb.notifier_call = arm_ccn_pmu_cpu_notifier; - ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 , + ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 ; err =register_cpu_notifier (&ccn->dt.cpu_nb);if (err) goto error_cpu_notifier;
Damn, of course! I have never made (or heard about) such a typo before
- I wish the compiler warned about it :-(
By all means:
Arnd, would you be still able to pick it up in this cycle?
Thanks Both!
Pawe?
PATCH 1/1 bus: arm-ccn: Fix cpu notifier priority
Quoted 19 lines · Pawel Moll
Quoted 18 lines · Jan Glauber
Fix cpu priority of notifier block that was erroneously set to return value of next statement.
Signed-off-by Jan Glauber <[email protected]>--- drivers/bus/arm-ccn.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/bus/arm-ccn.c +0
-0@@ -1271,7 +1271,7 @@ static int arm_ccn_pmu_init(struct arm_ccn *ccn) * picked to have a chance to migrate events before perf isnotified.
drivers/bus/arm-ccn.c +0
-0*/ ccn->dt.cpu_nb.notifier_call = arm_ccn_pmu_cpu_notifier;
Quoted 8 lines
drivers/bus/arm-ccn.c static int arm_ccn_pmu_init(struct arm_ccn *ccn) · line 1277
Quoted 5 lines
drivers/bus/arm-ccn.c +1
-1- ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 , + ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 ; err =register_cpu_notifier (&ccn->dt.cpu_nb);if (err) goto error_cpu_notifier;Damn, of course! I have never made (or heard about) such a typo before
- I wish the compiler warned about it :-(
Yeah, I was also surprised that we have plenty of cosmetical checks but nothing that catches this bug.
Maybe it's not too hard to catch the simple case, just check for a line with assignment and no brackets / parenthesis ended by a comma?
Jan
Quoted 10 lines
By all means:
Acked-by Pawel Moll <[email protected]>Cc <[email protected]> # 4.2+Arnd, would you be still able to pick it up in this cycle?
Thanks Both!
Pawe?
PATCH 1/1 bus: arm-ccn: Fix cpu notifier priority
In reply to Pawel MollQuoted 19 lines · Pawel Moll
Quoted 18 lines · Jan Glauber
Fix cpu priority of notifier block that was erroneously set to return value of next statement.
Signed-off-by Jan Glauber <[email protected]>--- drivers/bus/arm-ccn.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/bus/arm-ccn.c +0
-0@@ -1271,7 +1271,7 @@ static int arm_ccn_pmu_init(struct arm_ccn *ccn) * picked to have a chance to migrate events before perf isnotified.
drivers/bus/arm-ccn.c +0
-0*/ ccn->dt.cpu_nb.notifier_call = arm_ccn_pmu_cpu_notifier;
Quoted 8 lines
drivers/bus/arm-ccn.c static int arm_ccn_pmu_init(struct arm_ccn *ccn) · line 1277
Quoted 5 lines
drivers/bus/arm-ccn.c +1
-1- ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 , + ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 ; err =register_cpu_notifier (&ccn->dt.cpu_nb);if (err) goto error_cpu_notifier;Damn, of course! I have never made (or heard about) such a typo before
- I wish the compiler warned about it :-(
Hi Pawel,
actually my patch description is wrong. The compiler evaluates both statements but does _not_ overwrite the value of the first statement (would require parenthesis to do so). So it is just a cosmetical change, feel free to drop it.
Sorry for the noise,
Jan
Quoted 10 lines
By all means:
Acked-by Pawel Moll <[email protected]>Cc <[email protected]> # 4.2+Arnd, would you be still able to pick it up in this cycle?
Thanks Both!
Pawe?
PATCH 1/1 bus: arm-ccn: Fix cpu notifier priority
Quoted 31 lines · Jan Glauber
Quoted 21 lines · Pawel Moll
Quoted 20 lines · Jan Glauber
Fix cpu priority of notifier block that was erroneously set to return value of next statement.
Signed-off-by Jan Glauber <[email protected]>--- drivers/bus/arm-ccn.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/bus/arm-ccn.c +0
-0@@ -1271,7 +1271,7 @@ static int arm_ccn_pmu_init(struct arm_ccn *ccn) * picked to have a chance to migrate events before perfis
notified.drivers/bus/arm-ccn.c +0
-1*/ ccn->dt.cpu_nb.notifier_call = arm_ccn_pmu_cpu_notifier; - ccn->dt.cpu_nb.priority = CPU_PRI_PERF + 1,Quoted 8 lines
drivers/bus/arm-ccn.c static int arm_ccn_pmu_init(struct arm_ccn *ccn) · line 1277
Quoted 4 lines
drivers/bus/arm-ccn.c +1
-0+ ccn->dt.cpu_nb.priority =CPU_PRI_PERF +1 ; err =register_cpu_notifier (&ccn->dt.cpu_nb);if (err) goto error_cpu_notifier;Damn, of course! I have never made (or heard about) such a typo before
- I wish the compiler warned about it :-(
Quoted 3 lines
actually my patch description is wrong. The compiler evaluates both statements but does _not_ overwrite the value of the first statement (would require parenthesis to do so).
I must admit I looked at the code change rather than at the description, but thinking about it now, comma has the lowest precedence, so the assignments will win. Still, with gcc 6 now warning about suspicious indentations, my wish still stays a wish ;-)
Quoted 1 line
So it is just a cosmetical change, feel free to drop it.
Oh no, it's still a bug (only maybe not a "stable-class" one). And one I'll remember :-)
Thanks!
Pawel