Commit fbe96f29 (perf tools: Make perf.data more self-descriptive)
read '/proc/cpuinfo' to form cpu descriptor. For ARM, it finds
'Processor' field. It is correct when the patch merged, but due to
commit b4b8f770 (ARM: kernel: update cpuinfo to print all online CPUs
features), the corresponding information becomes 'model name' field.
This patch simply corrects it.
Signed-off-by: Wang Nan <[email protected]>
---
tools/perf/perf-sys.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/perf/perf-sys.h b/tools/perf/perf-sys.h
index 937e432..4293970 100644
--- a/tools/perf/perf-sys.h
+++ b/tools/perf/perf-sys.h
@@ -113,7 +113,7 @@
#define mb() ((void(*)(void))0xffff0fa0)()
#define wmb() ((void(*)(void))0xffff0fa0)()
#define rmb() ((void(*)(void))0xffff0fa0)()
-#define CPUINFO_PROC "Processor"
+#define CPUINFO_PROC "model name"
#endif
#ifdef __aarch64__
--
1.8.4
Em Wed, Oct 15, 2014 at 11:28:53AM +0800, Wang Nan escreveu:
> Commit fbe96f29 (perf tools: Make perf.data more self-descriptive)
> read '/proc/cpuinfo' to form cpu descriptor. For ARM, it finds
> 'Processor' field. It is correct when the patch merged, but due to
> commit b4b8f770 (ARM: kernel: update cpuinfo to print all online CPUs
> features), the corresponding information becomes 'model name' field.
>
> This patch simply corrects it.
It doesn't :-)
It makes it work with kernels >= b4b8f770, and fail with older kernels.
We need to somehow figure out where the right information is regardless
of the kernel.
- Arnaldo
> Signed-off-by: Wang Nan <[email protected]>
> ---
> tools/perf/perf-sys.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tools/perf/perf-sys.h b/tools/perf/perf-sys.h
> index 937e432..4293970 100644
> --- a/tools/perf/perf-sys.h
> +++ b/tools/perf/perf-sys.h
> @@ -113,7 +113,7 @@
> #define mb() ((void(*)(void))0xffff0fa0)()
> #define wmb() ((void(*)(void))0xffff0fa0)()
> #define rmb() ((void(*)(void))0xffff0fa0)()
> -#define CPUINFO_PROC "Processor"
> +#define CPUINFO_PROC "model name"
> #endif
>
> #ifdef __aarch64__
> --
> 1.8.4
On 2014/10/15 23:13, Arnaldo Carvalho de Melo wrote:
> Em Wed, Oct 15, 2014 at 11:28:53AM +0800, Wang Nan escreveu:
>> Commit fbe96f29 (perf tools: Make perf.data more self-descriptive)
>> read '/proc/cpuinfo' to form cpu descriptor. For ARM, it finds
>> 'Processor' field. It is correct when the patch merged, but due to
>> commit b4b8f770 (ARM: kernel: update cpuinfo to print all online CPUs
>> features), the corresponding information becomes 'model name' field.
>>
>> This patch simply corrects it.
>
> It doesn't :-)
>
> It makes it work with kernels >= b4b8f770, and fail with older kernels.
>
> We need to somehow figure out where the right information is regardless
> of the kernel.
>
> - Arnaldo
>
However it is a real problem. Look at output on arm platform at the
bottom of this mail, especially "cpudesc", "total memory" and "cmdline" field.
By further debugging I found that the real problem resides in return value checking
when write header error. Please see my other patches:
https://lkml.org/lkml/2014/10/15/612
https://lkml.org/lkml/2014/10/15/611
I think at least one of them is required.
Thanks.
>> Signed-off-by: Wang Nan <[email protected]>
>> ---
>> tools/perf/perf-sys.h | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/perf-sys.h b/tools/perf/perf-sys.h
>> index 937e432..4293970 100644
>> --- a/tools/perf/perf-sys.h
>> +++ b/tools/perf/perf-sys.h
>> @@ -113,7 +113,7 @@
>> #define mb() ((void(*)(void))0xffff0fa0)()
>> #define wmb() ((void(*)(void))0xffff0fa0)()
>> #define rmb() ((void(*)(void))0xffff0fa0)()
>> -#define CPUINFO_PROC "Processor"
>> +#define CPUINFO_PROC "model name"
>> #endif
>>
>> #ifdef __aarch64__
>> --
>> 1.8.4
bash-4.2# perf record ls
...
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
bash-4.2# perf report --stdio --header
Error:
The perf.data file has no samples!
# ========
# captured on: Fri Sep 12 10:09:10 2014
# hostname : arma15el
# os release : 3.17.0+
# perf version : 3.10.53
# arch : armv7l
# nrcpus online : 4
# nrcpus avail : 1
# cpudesc : (null)
# total memory : 0 kB
# cmdline :
# event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
# pmu mappings: not available
# ========
#
Em Thu, Oct 16, 2014 at 11:21:13AM +0800, Wang Nan escreveu:
> On 2014/10/15 23:13, Arnaldo Carvalho de Melo wrote:
> > Em Wed, Oct 15, 2014 at 11:28:53AM +0800, Wang Nan escreveu:
> >> Commit fbe96f29 (perf tools: Make perf.data more self-descriptive)
> >> read '/proc/cpuinfo' to form cpu descriptor. For ARM, it finds
> >> 'Processor' field. It is correct when the patch merged, but due to
> >> commit b4b8f770 (ARM: kernel: update cpuinfo to print all online CPUs
> >> features), the corresponding information becomes 'model name' field.
> >>
> >> This patch simply corrects it.
> >
> > It doesn't :-)
> >
> > It makes it work with kernels >= b4b8f770, and fail with older kernels.
> >
> > We need to somehow figure out where the right information is regardless
> > of the kernel.
> >
> > - Arnaldo
> >
>
> However it is a real problem. Look at output on arm platform at the
> bottom of this mail, especially "cpudesc", "total memory" and "cmdline" field.
Yeah, I haven't said it wasn't a problem, just that the proposed fix
wasn't enough.
- Arnaldo
> By further debugging I found that the real problem resides in return value checking
> when write header error. Please see my other patches:
>
> https://lkml.org/lkml/2014/10/15/612
> https://lkml.org/lkml/2014/10/15/611
>
> I think at least one of them is required.
>
> Thanks.
>
> >> Signed-off-by: Wang Nan <[email protected]>
> >> ---
> >> tools/perf/perf-sys.h | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/tools/perf/perf-sys.h b/tools/perf/perf-sys.h
> >> index 937e432..4293970 100644
> >> --- a/tools/perf/perf-sys.h
> >> +++ b/tools/perf/perf-sys.h
> >> @@ -113,7 +113,7 @@
> >> #define mb() ((void(*)(void))0xffff0fa0)()
> >> #define wmb() ((void(*)(void))0xffff0fa0)()
> >> #define rmb() ((void(*)(void))0xffff0fa0)()
> >> -#define CPUINFO_PROC "Processor"
> >> +#define CPUINFO_PROC "model name"
> >> #endif
> >>
> >> #ifdef __aarch64__
> >> --
> >> 1.8.4
>
>
>
> bash-4.2# perf record ls
> ...
> [ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
> bash-4.2# perf report --stdio --header
> Error:
> The perf.data file has no samples!
> # ========
> # captured on: Fri Sep 12 10:09:10 2014
> # hostname : arma15el
> # os release : 3.17.0+
> # perf version : 3.10.53
> # arch : armv7l
> # nrcpus online : 4
> # nrcpus avail : 1
> # cpudesc : (null)
> # total memory : 0 kB
> # cmdline :
> # event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
> # pmu mappings: not available
> # ========
> #
>
>
On 2014/10/16 22:55, Arnaldo Carvalho de Melo wrote:
> Em Thu, Oct 16, 2014 at 11:21:13AM +0800, Wang Nan escreveu:
>> On 2014/10/15 23:13, Arnaldo Carvalho de Melo wrote:
>>> Em Wed, Oct 15, 2014 at 11:28:53AM +0800, Wang Nan escreveu:
>>>> Commit fbe96f29 (perf tools: Make perf.data more self-descriptive)
>>>> read '/proc/cpuinfo' to form cpu descriptor. For ARM, it finds
>>>> 'Processor' field. It is correct when the patch merged, but due to
>>>> commit b4b8f770 (ARM: kernel: update cpuinfo to print all online CPUs
>>>> features), the corresponding information becomes 'model name' field.
>>>>
>>>> This patch simply corrects it.
>>>
>>> It doesn't :-)
>>>
>>> It makes it work with kernels >= b4b8f770, and fail with older kernels.
>>>
>>> We need to somehow figure out where the right information is regardless
>>> of the kernel.
>>>
>>> - Arnaldo
>>>
>>
>> However it is a real problem. Look at output on arm platform at the
>> bottom of this mail, especially "cpudesc", "total memory" and "cmdline" field.
>
> Yeah, I haven't said it wasn't a problem, just that the proposed fix
> wasn't enough.
>
> - Arnaldo
>
>> By further debugging I found that the real problem resides in return value checking
>> when write header error. Please see my other patches:
>>
>> https://lkml.org/lkml/2014/10/15/612
>> https://lkml.org/lkml/2014/10/15/611
So what's your opinion on these two patches?
>>
>> I think at least one of them is required.
>>
>> Thanks.
>>
>>>> Signed-off-by: Wang Nan <[email protected]>
>>>> ---
>>>> tools/perf/perf-sys.h | 2 +-
>>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/tools/perf/perf-sys.h b/tools/perf/perf-sys.h
>>>> index 937e432..4293970 100644
>>>> --- a/tools/perf/perf-sys.h
>>>> +++ b/tools/perf/perf-sys.h
>>>> @@ -113,7 +113,7 @@
>>>> #define mb() ((void(*)(void))0xffff0fa0)()
>>>> #define wmb() ((void(*)(void))0xffff0fa0)()
>>>> #define rmb() ((void(*)(void))0xffff0fa0)()
>>>> -#define CPUINFO_PROC "Processor"
>>>> +#define CPUINFO_PROC "model name"
>>>> #endif
>>>>
>>>> #ifdef __aarch64__
>>>> --
>>>> 1.8.4
>>
>>
>>
>> bash-4.2# perf record ls
>> ...
>> [ perf record: Woken up 1 times to write data ]
>> [ perf record: Captured and wrote 0.001 MB perf.data (~36 samples) ]
>> bash-4.2# perf report --stdio --header
>> Error:
>> The perf.data file has no samples!
>> # ========
>> # captured on: Fri Sep 12 10:09:10 2014
>> # hostname : arma15el
>> # os release : 3.17.0+
>> # perf version : 3.10.53
>> # arch : armv7l
>> # nrcpus online : 4
>> # nrcpus avail : 1
>> # cpudesc : (null)
>> # total memory : 0 kB
>> # cmdline :
>> # event : name = cycles, type = 0, config = 0x0, config1 = 0x0, config2 = 0x0, excl_usr = 0, excl_kern = 0, excl_host = 0, excl_guest = 1, precise_ip = 0
>> # pmu mappings: not available
>> # ========
>> #
>>
>>