* [PATCH] perf: disable parallelism for 'make clean'
@ 2018-07-05 13:15 Rasmus Villemoes
2018-07-05 13:49 ` Jiri Olsa
2018-08-23 8:37 ` [tip:perf/urgent] perf tools: Disable " tip-bot for Rasmus Villemoes
0 siblings, 2 replies; 8+ messages in thread
From: Rasmus Villemoes @ 2018-07-05 13:15 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Alexander Shishkin, Jiri Olsa, Namhyung Kim
Cc: Rasmus Villemoes, linux-kernel
The Yocto build system does a 'make clean' when rebuilding due to
changed dependencies, and that consistently fails for me (causing the
whole BSP build to fail) with errors such as
| find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
| find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
| find: find: '[...]/perf/1.0-r9/perf-1.0/libtraceevent.a''[...]/perf/1.0-r9/perf-1.0/libtraceevent.a': No such file or directory: No such file or directory
|
[...]
| find: cannot delete '/mnt/xfs/devel/pil/yocto/tmp-glibc/work/wandboard-oe-linux-gnueabi/perf/1.0-r9/perf-1.0/util/.pstack.o.cmd': No such file or directory
Apparently (despite the comment), 'make clean' ends up launching
multiple sub-makes that all want to remove the same things - perhaps
this only happens in combination with a O=... parameter. In any case, we
don't lose much by explicitly disabling the parallelism for the clean
target, and it makes automated builds much more reliable.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
---
Another option is to add -ignore_readdir_race to all find commands,
but that's a lot more invasive and only works properly together with
-delete with a very recent GNU find
(https://savannah.gnu.org/bugs/?52981).
tools/perf/Makefile | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 225454416ed5..7902a5681fc8 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -84,10 +84,10 @@ endif # has_clean
endif # MAKECMDGOALS
#
-# The clean target is not really parallel, don't print the jobs info:
+# Explicitly disable parallelism for the clean target.
#
clean:
- $(make)
+ $(make) -j1
#
# The build-test target is not really parallel, don't print the jobs info,
--
2.16.4
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH] perf: disable parallelism for 'make clean'
2018-07-05 13:15 [PATCH] perf: disable parallelism for 'make clean' Rasmus Villemoes
@ 2018-07-05 13:49 ` Jiri Olsa
2018-08-15 7:30 ` Rasmus Villemoes
2018-08-23 8:37 ` [tip:perf/urgent] perf tools: Disable " tip-bot for Rasmus Villemoes
1 sibling, 1 reply; 8+ messages in thread
From: Jiri Olsa @ 2018-07-05 13:49 UTC (permalink / raw)
To: Rasmus Villemoes
Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Alexander Shishkin, Namhyung Kim, linux-kernel
On Thu, Jul 05, 2018 at 03:15:27PM +0200, Rasmus Villemoes wrote:
> The Yocto build system does a 'make clean' when rebuilding due to
> changed dependencies, and that consistently fails for me (causing the
> whole BSP build to fail) with errors such as
>
> | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
> | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
> | find: find: '[...]/perf/1.0-r9/perf-1.0/libtraceevent.a''[...]/perf/1.0-r9/perf-1.0/libtraceevent.a': No such file or directory: No such file or directory
> |
> [...]
> | find: cannot delete '/mnt/xfs/devel/pil/yocto/tmp-glibc/work/wandboard-oe-linux-gnueabi/perf/1.0-r9/perf-1.0/util/.pstack.o.cmd': No such file or directory
>
> Apparently (despite the comment), 'make clean' ends up launching
> multiple sub-makes that all want to remove the same things - perhaps
> this only happens in combination with a O=... parameter. In any case, we
> don't lose much by explicitly disabling the parallelism for the clean
> target, and it makes automated builds much more reliable.
>
> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Acked-by: Jiri Olsa <jolsa@kernel.org>
thanks,
jirka
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] perf: disable parallelism for 'make clean'
2018-07-05 13:49 ` Jiri Olsa
@ 2018-08-15 7:30 ` Rasmus Villemoes
2018-08-15 19:33 ` Arnaldo Carvalho de Melo
0 siblings, 1 reply; 8+ messages in thread
From: Rasmus Villemoes @ 2018-08-15 7:30 UTC (permalink / raw)
To: Jiri Olsa
Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Alexander Shishkin, Namhyung Kim, linux-kernel
On 2018-07-05 15:49, Jiri Olsa wrote:
> On Thu, Jul 05, 2018 at 03:15:27PM +0200, Rasmus Villemoes wrote:
>> this only happens in combination with a O=... parameter. In any case, we
>> don't lose much by explicitly disabling the parallelism for the clean
>> target, and it makes automated builds much more reliable.
>>
>> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
>
> Acked-by: Jiri Olsa <jolsa@kernel.org>
Was this picked up by anyone? I'm not seeing it in -next?
Rasmus
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] perf: disable parallelism for 'make clean'
2018-08-15 7:30 ` Rasmus Villemoes
@ 2018-08-15 19:33 ` Arnaldo Carvalho de Melo
0 siblings, 0 replies; 8+ messages in thread
From: Arnaldo Carvalho de Melo @ 2018-08-15 19:33 UTC (permalink / raw)
To: Rasmus Villemoes
Cc: Jiri Olsa, Peter Zijlstra, Ingo Molnar, Alexander Shishkin,
Namhyung Kim, linux-kernel
Em Wed, Aug 15, 2018 at 09:30:39AM +0200, Rasmus Villemoes escreveu:
> On 2018-07-05 15:49, Jiri Olsa wrote:
> > On Thu, Jul 05, 2018 at 03:15:27PM +0200, Rasmus Villemoes wrote:
>
> >> this only happens in combination with a O=... parameter. In any case, we
> >> don't lose much by explicitly disabling the parallelism for the clean
> >> target, and it makes automated builds much more reliable.
> >>
> >> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
> >
> > Acked-by: Jiri Olsa <jolsa@kernel.org>
>
> Was this picked up by anyone? I'm not seeing it in -next?
Thanks, applied.
- Arnaldo
^ permalink raw reply [flat|nested] 8+ messages in thread
* [tip:perf/urgent] perf tools: Disable parallelism for 'make clean'
2018-07-05 13:15 [PATCH] perf: disable parallelism for 'make clean' Rasmus Villemoes
2018-07-05 13:49 ` Jiri Olsa
@ 2018-08-23 8:37 ` tip-bot for Rasmus Villemoes
1 sibling, 0 replies; 8+ messages in thread
From: tip-bot for Rasmus Villemoes @ 2018-08-23 8:37 UTC (permalink / raw)
To: linux-tip-commits
Cc: acme, hpa, namhyung, linux, jolsa, alexander.shishkin, tglx,
linux-kernel, peterz, mingo
Commit-ID: da15fc2fa9c07b23db8f5e479bd8a9f0d741ca07
Gitweb: https://git.kernel.org/tip/da15fc2fa9c07b23db8f5e479bd8a9f0d741ca07
Author: Rasmus Villemoes <linux@rasmusvillemoes.dk>
AuthorDate: Thu, 5 Jul 2018 15:15:27 +0200
Committer: Arnaldo Carvalho de Melo <acme@redhat.com>
CommitDate: Mon, 20 Aug 2018 08:54:58 -0300
perf tools: Disable parallelism for 'make clean'
The Yocto build system does a 'make clean' when rebuilding due to
changed dependencies, and that consistently fails for me (causing the
whole BSP build to fail) with errors such as
| find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
| find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
| find: find: '[...]/perf/1.0-r9/perf-1.0/libtraceevent.a''[...]/perf/1.0-r9/perf-1.0/libtraceevent.a': No such file or directory: No such file or directory
|
[...]
| find: cannot delete '/mnt/xfs/devel/pil/yocto/tmp-glibc/work/wandboard-oe-linux-gnueabi/perf/1.0-r9/perf-1.0/util/.pstack.o.cmd': No such file or directory
Apparently (despite the comment), 'make clean' ends up launching
multiple sub-makes that all want to remove the same things - perhaps
this only happens in combination with a O=... parameter. In any case, we
don't lose much by explicitly disabling the parallelism for the clean
target, and it makes automated builds much more reliable.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20180705131527.19749-1-linux@rasmusvillemoes.dk
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
tools/perf/Makefile | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 225454416ed5..7902a5681fc8 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -84,10 +84,10 @@ endif # has_clean
endif # MAKECMDGOALS
#
-# The clean target is not really parallel, don't print the jobs info:
+# Explicitly disable parallelism for the clean target.
#
clean:
- $(make)
+ $(make) -j1
#
# The build-test target is not really parallel, don't print the jobs info,
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH] perf: disable parallelism for 'make clean'
2018-07-05 22:18 ` Andre McCurdy
@ 2018-07-06 7:26 ` Rasmus Villemoes
0 siblings, 0 replies; 8+ messages in thread
From: Rasmus Villemoes @ 2018-07-06 7:26 UTC (permalink / raw)
To: Andre McCurdy; +Cc: OE Core mailing list
On 2018-07-06 00:18, Andre McCurdy wrote:
> On Thu, Jul 5, 2018 at 1:55 PM, Rasmus Villemoes
> <rasmus.villemoes@prevas.dk> wrote:
>> Whenever perf got rebuilt, I was consistently getting errors such as
>>
>> | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
>> | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
>> | find: find: '[...]/perf/1.0-r9/perf-1.0/libtraceevent.a''[...]/perf/1.0-r9/perf-1.0/libtraceevent.a': No such file or directory: No such file or directory
>> |
>> [...]
>> | find: cannot delete '/mnt/xfs/devel/pil/yocto/tmp-glibc/work/wandboard-oe-linux-gnueabi/perf/1.0-r9/perf-1.0/util/.pstack.o.cmd': No such file or directory
>>
>> breaking the whole build. The root cause seems to be that the implicit
>> 'make clean' done during do_configure ends up running in parallel, and
>> thus multiple find commands attempt to stat and/or delete the same
>> file.
>>
>> This patch of tools/perf/Makefile has been ack'ed
>> upstream (lkml.kernel.org/r/20180705134955.GB3686@krava). Since the sed
>> script matches exactly the lines
>>
>> clean:
>> $(make)
>>
>> it should be harmless to have this even with a fixed kernel, and it can
>> be removed if and when all relevant -stable kernels have it.
>
> If it works, then something like this might be a cleaner solution:
>
> EXTRA_OEMAKE_append_task-configure = " -j1"
tools/perf/Makefile explicitly ignores a -j argument, and computes its
own value. But one can pass JOBS=1 instead. I'll try that and resend.
Rasmus
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] perf: disable parallelism for 'make clean'
2018-07-05 20:55 [PATCH] perf: disable " Rasmus Villemoes
@ 2018-07-05 22:18 ` Andre McCurdy
2018-07-06 7:26 ` Rasmus Villemoes
0 siblings, 1 reply; 8+ messages in thread
From: Andre McCurdy @ 2018-07-05 22:18 UTC (permalink / raw)
To: Rasmus Villemoes; +Cc: OE Core mailing list
On Thu, Jul 5, 2018 at 1:55 PM, Rasmus Villemoes
<rasmus.villemoes@prevas.dk> wrote:
> Whenever perf got rebuilt, I was consistently getting errors such as
>
> | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
> | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
> | find: find: '[...]/perf/1.0-r9/perf-1.0/libtraceevent.a''[...]/perf/1.0-r9/perf-1.0/libtraceevent.a': No such file or directory: No such file or directory
> |
> [...]
> | find: cannot delete '/mnt/xfs/devel/pil/yocto/tmp-glibc/work/wandboard-oe-linux-gnueabi/perf/1.0-r9/perf-1.0/util/.pstack.o.cmd': No such file or directory
>
> breaking the whole build. The root cause seems to be that the implicit
> 'make clean' done during do_configure ends up running in parallel, and
> thus multiple find commands attempt to stat and/or delete the same
> file.
>
> This patch of tools/perf/Makefile has been ack'ed
> upstream (lkml.kernel.org/r/20180705134955.GB3686@krava). Since the sed
> script matches exactly the lines
>
> clean:
> $(make)
>
> it should be harmless to have this even with a fixed kernel, and it can
> be removed if and when all relevant -stable kernels have it.
If it works, then something like this might be a cleaner solution:
EXTRA_OEMAKE_append_task-configure = " -j1"
(and if it doesn't work, it would be good to understand why).
> Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
> ---
> meta/recipes-kernel/perf/perf.bb | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/meta/recipes-kernel/perf/perf.bb b/meta/recipes-kernel/perf/perf.bb
> index 51f5597eb9..1cf2666ee5 100644
> --- a/meta/recipes-kernel/perf/perf.bb
> +++ b/meta/recipes-kernel/perf/perf.bb
> @@ -145,6 +145,11 @@ python copy_perf_source_from_kernel() {
> bb.utils.copyfile(src, dest)
> }
>
> +do_configure[prefuncs] += "disable_clean_parallelism"
> +disable_clean_parallelism () {
> + sed -i '/^clean:$/{n;s/^\t\$(make)$/\t\$(make) -j1/}' ${S}/tools/perf/Makefile
> +}
> +
> do_configure_prepend () {
> # If building a multlib based perf, the incorrect library path will be
> # detected by perf, since it triggers via: ifeq ($(ARCH),x86_64). In a 32 bit
> --
> 2.16.4
>
> --
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH] perf: disable parallelism for 'make clean'
@ 2018-07-05 20:55 Rasmus Villemoes
2018-07-05 22:18 ` Andre McCurdy
0 siblings, 1 reply; 8+ messages in thread
From: Rasmus Villemoes @ 2018-07-05 20:55 UTC (permalink / raw)
To: openembedded-core
Whenever perf got rebuilt, I was consistently getting errors such as
| find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
| find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory
| find: find: '[...]/perf/1.0-r9/perf-1.0/libtraceevent.a''[...]/perf/1.0-r9/perf-1.0/libtraceevent.a': No such file or directory: No such file or directory
|
[...]
| find: cannot delete '/mnt/xfs/devel/pil/yocto/tmp-glibc/work/wandboard-oe-linux-gnueabi/perf/1.0-r9/perf-1.0/util/.pstack.o.cmd': No such file or directory
breaking the whole build. The root cause seems to be that the implicit
'make clean' done during do_configure ends up running in parallel, and
thus multiple find commands attempt to stat and/or delete the same
file.
This patch of tools/perf/Makefile has been ack'ed
upstream (lkml.kernel.org/r/20180705134955.GB3686@krava). Since the sed
script matches exactly the lines
clean:
$(make)
it should be harmless to have this even with a fixed kernel, and it can
be removed if and when all relevant -stable kernels have it.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
---
meta/recipes-kernel/perf/perf.bb | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/meta/recipes-kernel/perf/perf.bb b/meta/recipes-kernel/perf/perf.bb
index 51f5597eb9..1cf2666ee5 100644
--- a/meta/recipes-kernel/perf/perf.bb
+++ b/meta/recipes-kernel/perf/perf.bb
@@ -145,6 +145,11 @@ python copy_perf_source_from_kernel() {
bb.utils.copyfile(src, dest)
}
+do_configure[prefuncs] += "disable_clean_parallelism"
+disable_clean_parallelism () {
+ sed -i '/^clean:$/{n;s/^\t\$(make)$/\t\$(make) -j1/}' ${S}/tools/perf/Makefile
+}
+
do_configure_prepend () {
# If building a multlib based perf, the incorrect library path will be
# detected by perf, since it triggers via: ifeq ($(ARCH),x86_64). In a 32 bit
--
2.16.4
^ permalink raw reply related [flat|nested] 8+ messages in thread
end of thread, other threads:[~2018-08-23 8:37 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-07-05 13:15 [PATCH] perf: disable parallelism for 'make clean' Rasmus Villemoes
2018-07-05 13:49 ` Jiri Olsa
2018-08-15 7:30 ` Rasmus Villemoes
2018-08-15 19:33 ` Arnaldo Carvalho de Melo
2018-08-23 8:37 ` [tip:perf/urgent] perf tools: Disable " tip-bot for Rasmus Villemoes
2018-07-05 20:55 [PATCH] perf: disable " Rasmus Villemoes
2018-07-05 22:18 ` Andre McCurdy
2018-07-06 7:26 ` Rasmus Villemoes
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.