Messages in this thread Patch in this message | | | Date | Tue, 1 Sep 2020 12:06:30 -0300 | From | Arnaldo Carvalho de Melo <> | Subject | Re: [PATCH] perf: correct SNOOPX field offset |
| |
Em Tue, Sep 01, 2020 at 12:02:25PM -0300, Arnaldo Carvalho de Melo escreveu: > Em Wed, Aug 26, 2020 at 03:33:35PM +0100, Al Grant escreveu: > > On 26/08/2020 15:26, Arnaldo Carvalho de Melo wrote: > > > Em Tue, Aug 25, 2020 at 10:40:43AM -0700, Andi Kleen escreveu: > > > > On Mon, Aug 24, 2020 at 10:28:34AM +0100, Al Grant wrote: > > > > > perf_event.h has macros that define the field offsets in the > > > > > data_src bitmask in perf records. The SNOOPX and REMOTE offsets > > > > > were both 37. These are distinct fields, and the bitfield layout > > > > > in perf_mem_data_src confirms that SNOOPX should be at offset 38. > > > > > > > > Looks good. > > > > > > > > Reviewed-by: Andi Kleen <ak@linux.intel.com> > > > > > > > > Probably should have a Fixes: header > > > > > > Please do so, find the patch that introduced the error, add the Fixes > > > tag, will help me not having to do it myself :-) > > > > Fixes: 52839e653b562 ("perf tools: Add support for printing new mem_info > > encodings") > > Ok, I'll add that, thanks. > > But you forgot to add your Signed-off-by:, can you please provide it? > > There was also a minor problem in the patch, there was no separation of > --- from the patch comment section to the patch itself, I'll fix that as > well.
Also you mixed up tools/ with include/ things, the perf part of the kernel is maintained by Ingo, PeterZ.
Peter, the patch is the one below, I'll collect the tools/include/uapi/linux/perf_event.h bit as it fixes the tooling, please consider taking the kernel part.
Thanks,
- Arnaldo
---
From: Al Grant <al.grant@foss.arm.com> Subject: [PATCH] perf: correct SNOOPX field offset Message-ID: <9974f2d0-bf7f-518e-d9f7-4520e5ff1bb0@foss.arm.com> Date: Mon, 24 Aug 2020 10:28:34 +0100
perf_event.h has macros that define the field offsets in the data_src bitmask in perf records. The SNOOPX and REMOTE offsets were both 37. These are distinct fields, and the bitfield layout in perf_mem_data_src confirms that SNOOPX should be at offset 38.
Signed-off-by: Al Grant <al.grant@arm.com>
---
include/uapi/linux/perf_event.h | 2 +- tools/include/uapi/linux/perf_event.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
---
diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h index 077e7ee69e3d..3e5dcdd48a49 100644 --- a/include/uapi/linux/perf_event.h +++ b/include/uapi/linux/perf_event.h @@ -1196,7 +1196,7 @@ union perf_mem_data_src {
#define PERF_MEM_SNOOPX_FWD 0x01 /* forward */ /* 1 free */ -#define PERF_MEM_SNOOPX_SHIFT 37 +#define PERF_MEM_SNOOPX_SHIFT 38
/* locked instruction */ #define PERF_MEM_LOCK_NA 0x01 /* not available */ diff --git a/tools/include/uapi/linux/perf_event.h b/tools/include/uapi/linux/perf_event.h index 077e7ee69e3d..3e5dcdd48a49 100644 --- a/tools/include/uapi/linux/perf_event.h +++ b/tools/include/uapi/linux/perf_event.h @@ -1196,7 +1196,7 @@ union perf_mem_data_src {
#define PERF_MEM_SNOOPX_FWD 0x01 /* forward */ /* 1 free */ -#define PERF_MEM_SNOOPX_SHIFT 37 +#define PERF_MEM_SNOOPX_SHIFT 38
/* locked instruction */ #define PERF_MEM_LOCK_NA 0x01 /* not available */
| |