[PATCH 3/4] perf archive: Fix filtering of empty build-ids

[PATCH 3/4] perf archive: Fix filtering of empty build-ids

[Nicholas Fraser]

A non-existent build-id used to be treated as all-zero SHA-1 hash.
Build-ids are now variable width. A non-existent build-id is an empty
string and "perf buildid-list" pads this with spaces.

This fixes "perf-archive" to skip missing build-ids by skipping lines
that start with a padding space rather than with zeroes.

Signed-off-by: Nicholas Fraser <nfraser@codeweavers.com>
---
 tools/perf/perf-archive.sh | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/tools/perf/perf-archive.sh b/tools/perf/perf-archive.sh
index 0cfb3e2cefef..133f0eddbcc4 100644
--- a/tools/perf/perf-archive.sh
+++ b/tools/perf/perf-archive.sh
@@ -20,9 +20,8 @@ else
 fi
 
 BUILDIDS=$(mktemp /tmp/perf-archive-buildids.XXXXXX)
-NOBUILDID=0000000000000000000000000000000000000000
 
-perf buildid-list -i $PERF_DATA --with-hits | grep -v "^$NOBUILDID " > $BUILDIDS
+perf buildid-list -i $PERF_DATA --with-hits | grep -v "^ " > $BUILDIDS
 if [ ! -s $BUILDIDS ] ; then
 	echo "perf archive: no build-ids found"
 	rm $BUILDIDS || true
-- 
2.30.0

Re: [PATCH 3/4] perf archive: Fix filtering of empty build-ids

[Arnaldo Carvalho de Melo]

Em Wed, Feb 10, 2021 at 02:17:49PM -0500, Nicholas Fraser escreveu:
> A non-existent build-id used to be treated as all-zero SHA-1 hash.
> Build-ids are now variable width. A non-existent build-id is an empty
> string and "perf buildid-list" pads this with spaces.
> 
> This fixes "perf-archive" to skip missing build-ids by skipping lines
> that start with a padding space rather than with zeroes.
> 
> Signed-off-by: Nicholas Fraser <nfraser@codeweavers.com>
> ---
>  tools/perf/perf-archive.sh | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/tools/perf/perf-archive.sh b/tools/perf/perf-archive.sh
> index 0cfb3e2cefef..133f0eddbcc4 100644
> --- a/tools/perf/perf-archive.sh
> +++ b/tools/perf/perf-archive.sh
> @@ -20,9 +20,8 @@ else
>  fi
>  
>  BUILDIDS=$(mktemp /tmp/perf-archive-buildids.XXXXXX)
> -NOBUILDID=0000000000000000000000000000000000000000
>  
> -perf buildid-list -i $PERF_DATA --with-hits | grep -v "^$NOBUILDID " > $BUILDIDS
> +perf buildid-list -i $PERF_DATA --with-hits | grep -v "^ " > $BUILDIDS
>  if [ ! -s $BUILDIDS ] ; then
>  	echo "perf archive: no build-ids found"
>  	rm $BUILDIDS || true

What should we do when using older files? We have to support both cases.

- Arnaldo

Re: [PATCH 3/4] perf archive: Fix filtering of empty build-ids

[Nicholas Fraser]

The current version of "perf buildid-list" prints missing build-ids as
blank even when loading perf.data files generated from older versions of
perf. I've tested with a few older versions.

Since perf-archive is an external script that calls back into perf, it
could in theory find a different older version that still returns zeroes.
We could make it look for a matching perf, e.g. "$(dirname "$0")/perf".
Unfortunately that won't work in general because at least on my system
perf-archive is installed to a different path than perf.

Nick


On 2021-02-12 7:29 a.m., Arnaldo Carvalho de Melo wrote:
> Em Wed, Feb 10, 2021 at 02:17:49PM -0500, Nicholas Fraser escreveu:
>> A non-existent build-id used to be treated as all-zero SHA-1 hash.
>> Build-ids are now variable width. A non-existent build-id is an empty
>> string and "perf buildid-list" pads this with spaces.
>>
>> This fixes "perf-archive" to skip missing build-ids by skipping lines
>> that start with a padding space rather than with zeroes.
>>
>> Signed-off-by: Nicholas Fraser <nfraser@codeweavers.com>
>> ---
>>  tools/perf/perf-archive.sh | 3 +--
>>  1 file changed, 1 insertion(+), 2 deletions(-)
>>
>> diff --git a/tools/perf/perf-archive.sh b/tools/perf/perf-archive.sh
>> index 0cfb3e2cefef..133f0eddbcc4 100644
>> --- a/tools/perf/perf-archive.sh
>> +++ b/tools/perf/perf-archive.sh
>> @@ -20,9 +20,8 @@ else
>>  fi
>>  
>>  BUILDIDS=$(mktemp /tmp/perf-archive-buildids.XXXXXX)
>> -NOBUILDID=0000000000000000000000000000000000000000
>>  
>> -perf buildid-list -i $PERF_DATA --with-hits | grep -v "^$NOBUILDID " > $BUILDIDS
>> +perf buildid-list -i $PERF_DATA --with-hits | grep -v "^ " > $BUILDIDS
>>  if [ ! -s $BUILDIDS ] ; then
>>  	echo "perf archive: no build-ids found"
>>  	rm $BUILDIDS || true
> 
> What should we do when using older files? We have to support both cases.
> 
> - Arnaldo
> 

Re: [PATCH 3/4] perf archive: Fix filtering of empty build-ids

[Arnaldo Carvalho de Melo]

Em Fri, Feb 12, 2021 at 01:49:36PM -0500, Nicholas Fraser escreveu:
> The current version of "perf buildid-list" prints missing build-ids as
> blank even when loading perf.data files generated from older versions of
> perf. I've tested with a few older versions.
> 
> Since perf-archive is an external script that calls back into perf, it
> could in theory find a different older version that still returns zeroes.
> We could make it look for a matching perf, e.g. "$(dirname "$0")/perf".
> Unfortunately that won't work in general because at least on my system
> perf-archive is installed to a different path than perf.

It shouldn't try to find a "matching perf" as the userspace perf is not
in lockstep with the kernel perf subsystem :-)

If a current 'perf archive' is called on some perf.data file, it should
know that a zero-filled or a blank filled build-id means the same and
act accordingly.

- Arnaldo
  
> Nick
> 
> 
> On 2021-02-12 7:29 a.m., Arnaldo Carvalho de Melo wrote:
> > Em Wed, Feb 10, 2021 at 02:17:49PM -0500, Nicholas Fraser escreveu:
> >> A non-existent build-id used to be treated as all-zero SHA-1 hash.
> >> Build-ids are now variable width. A non-existent build-id is an empty
> >> string and "perf buildid-list" pads this with spaces.
> >>
> >> This fixes "perf-archive" to skip missing build-ids by skipping lines
> >> that start with a padding space rather than with zeroes.
> >>
> >> Signed-off-by: Nicholas Fraser <nfraser@codeweavers.com>
> >> ---
> >>  tools/perf/perf-archive.sh | 3 +--
> >>  1 file changed, 1 insertion(+), 2 deletions(-)
> >>
> >> diff --git a/tools/perf/perf-archive.sh b/tools/perf/perf-archive.sh
> >> index 0cfb3e2cefef..133f0eddbcc4 100644
> >> --- a/tools/perf/perf-archive.sh
> >> +++ b/tools/perf/perf-archive.sh
> >> @@ -20,9 +20,8 @@ else
> >>  fi
> >>  
> >>  BUILDIDS=$(mktemp /tmp/perf-archive-buildids.XXXXXX)
> >> -NOBUILDID=0000000000000000000000000000000000000000
> >>  
> >> -perf buildid-list -i $PERF_DATA --with-hits | grep -v "^$NOBUILDID " > $BUILDIDS
> >> +perf buildid-list -i $PERF_DATA --with-hits | grep -v "^ " > $BUILDIDS
> >>  if [ ! -s $BUILDIDS ] ; then
> >>  	echo "perf archive: no build-ids found"
> >>  	rm $BUILDIDS || true
> > 
> > What should we do when using older files? We have to support both cases.
> > 
> > - Arnaldo
> > 

-- 

- Arnaldo