* Heiko Carstens <heiko.carst...@de.ibm.com> wrote:
> On Mon, Sep 21, 2009 at 09:30:43AM +0200, Ingo Molnar wrote:
> >
> > * Metzger, Markus T <markus.t.metz...@intel.com> wrote:
> >
> > > >-----Original Message-----
> > > >From: Paul Mackerras [mailto:pau...@samba.org]
> > > >Sent: Monday, September 21, 2009 8:45 AM
> > >
> > >
> > > >Markus, please take care in future to mention it in the changelog if
> > > >your patches touch definitions used by other architectures. If you
> > > >could go so far as to use grep a bit more and fix up other
> > > >architectures' callsites for the things you're changing, that would be
> > > >very much appreciated. Thanks.
> > >
> > > I'm sorry I missed that.
> > >
> > > There's one more place in arch/sparc/.
> > > The below patch should fix it, but I have no means to test it.
> >
> > You also missed a third thing:
> >
> > +static inline int
> > +perf_output_begin(struct perf_output_handle *handle, struct perf_counter
> > *c,
> > + unsigned int size, int nmi, int sample) { }
> >
> > an 'int' function returning void ...
> >
> > Plus all the !PERF_COUNTERS branch of empty inlines is pointless - these
> > facilities are used by perfcounters code only. I fixed that too.
>
> Hi Ingo,
>
> did you fix all of these warnings for !PERF_COUNTERS?
>
> include/linux/perf_counter.h: In function 'perf_output_begin':
> include/linux/perf_counter.h:854: warning: no return statement in function
> returning non-void
> include/linux/perf_counter.h: At top level:
> include/linux/perf_counter.h:863: warning: 'struct perf_sample_data' declared
> inside parameter list
> include/linux/perf_counter.h:863: warning: its scope is only this definition
> or declaration, which is probably not what you want
> include/linux/perf_counter.h:868: warning: 'struct perf_sample_data' declared
> inside parameter list
Yes. The full commit is below.
Ingo
---------------->
>From cd74c86bdf705f824d494a2bbda393d1d562b40a Mon Sep 17 00:00:00 2001
From: Paul Mackerras <pau...@samba.org>
Date: Mon, 21 Sep 2009 16:44:32 +1000
Subject: [PATCH] perf_counter, powerpc, sparc: Fix compilation after
perf_counter_overflow() change
Commit 5622f295 ("x86, perf_counter, bts: Optimize BTS overflow
handling") removed the regs field from struct perf_sample_data and
added a regs parameter to perf_counter_overflow(). This breaks the
build on powerpc (and Sparc) as reported by Sachin Sant:
arch/powerpc/kernel/perf_counter.c: In function 'record_and_restart':
arch/powerpc/kernel/perf_counter.c:1165: error: unknown field 'regs'
specified in initializer
This adjusts arch/powerpc/kernel/perf_counter.c to correspond with the
new struct perf_sample_data and perf_counter_overflow().
[ v2: also fix Sparc, Markus Metzger <markus.t.metz...@intel.com> ]
Reported-by: Sachin Sant <sach...@in.ibm.com>
Signed-off-by: Paul Mackerras <pau...@samba.org>
Cc: Markus Metzger <markus.t.metz...@intel.com>
Cc: David S. Miller <da...@davemloft.net>
Cc: b...@kernel.crashing.org
Cc: linuxppc-...@ozlabs.org
Cc: Peter Zijlstra <a.p.zijls...@chello.nl>
LKML-Reference: <19127.8400.376239.586...@drongo.ozlabs.ibm.com>
Signed-off-by: Ingo Molnar <mi...@elte.hu>
---
arch/powerpc/kernel/perf_counter.c | 3 +--
arch/sparc/kernel/perf_counter.c | 3 +--
include/linux/perf_counter.h | 17 -----------------
3 files changed, 2 insertions(+), 21 deletions(-)
diff --git a/arch/powerpc/kernel/perf_counter.c
b/arch/powerpc/kernel/perf_counter.c
index 7ceefaf..5ccf9bc 100644
--- a/arch/powerpc/kernel/perf_counter.c
+++ b/arch/powerpc/kernel/perf_counter.c
@@ -1162,7 +1162,6 @@ static void record_and_restart(struct perf_counter
*counter, unsigned long val,
*/
if (record) {
struct perf_sample_data data = {
- .regs = regs,
.addr = 0,
.period = counter->hw.last_period,
};
@@ -1170,7 +1169,7 @@ static void record_and_restart(struct perf_counter
*counter, unsigned long val,
if (counter->attr.sample_type & PERF_SAMPLE_ADDR)
perf_get_data_addr(regs, &data.addr);
- if (perf_counter_overflow(counter, nmi, &data)) {
+ if (perf_counter_overflow(counter, nmi, &data, regs)) {
/*
* Interrupts are coming too fast - throttle them
* by setting the counter to 0, so it will be
diff --git a/arch/sparc/kernel/perf_counter.c b/arch/sparc/kernel/perf_counter.c
index 09de403..b1265ce 100644
--- a/arch/sparc/kernel/perf_counter.c
+++ b/arch/sparc/kernel/perf_counter.c
@@ -493,7 +493,6 @@ static int __kprobes perf_counter_nmi_handler(struct
notifier_block *self,
regs = args->regs;
- data.regs = regs;
data.addr = 0;
cpuc = &__get_cpu_var(cpu_hw_counters);
@@ -513,7 +512,7 @@ static int __kprobes perf_counter_nmi_handler(struct
notifier_block *self,
if (!sparc_perf_counter_set_period(counter, hwc, idx))
continue;
- if (perf_counter_overflow(counter, 1, &data))
+ if (perf_counter_overflow(counter, 1, &data, regs))
sparc_pmu_disable_counter(hwc, idx);
}
diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h
index bd34100..740caad 100644
--- a/include/linux/perf_counter.h
+++ b/include/linux/perf_counter.h
@@ -849,23 +849,6 @@ static inline void perf_counter_comm(struct task_struct
*tsk) { }
static inline void perf_counter_fork(struct task_struct *tsk) { }
static inline void perf_counter_init(void) { }
-static inline int
-perf_output_begin(struct perf_output_handle *handle, struct perf_counter *c,
- unsigned int size, int nmi, int sample) { }
-static inline void perf_output_end(struct perf_output_handle *handle) { }
-static inline void
-perf_output_copy(struct perf_output_handle *handle,
- const void *buf, unsigned int len) { }
-static inline void
-perf_output_sample(struct perf_output_handle *handle,
- struct perf_event_header *header,
- struct perf_sample_data *data,
- struct perf_counter *counter) { }
-static inline void
-perf_prepare_sample(struct perf_event_header *header,
- struct perf_sample_data *data,
- struct perf_counter *counter,
- struct pt_regs *regs) { }
#endif
#define perf_output_put(handle, x) \
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev
