All of lore.kernel.org
 help / color / mirror / Atom feed
From: Richard Purdie <richard.purdie@linuxfoundation.org>
To: niko.mauno@vaisala.com, openembedded-core@lists.openembedded.org
Subject: Re: [OE-core] [PATCH 1/3] ptest.bbclass: Honor PARALLEL_MAKE, PARALLEL_MAKEINST
Date: Sat, 25 Jun 2022 17:55:45 +0100	[thread overview]
Message-ID: <b9d79d0231537cfcb8f75bd0e728d76bbdb64f1a.camel@linuxfoundation.org> (raw)
In-Reply-To: <20220622191615.10168-1-niko.mauno@vaisala.com>

On Wed, 2022-06-22 at 22:16 +0300, niko.mauno@vaisala.com via
lists.openembedded.org wrote:
> Since oe_runmake wrapper doesn't by default involve PARALLEL_MAKE
> outside do_compile() nor PARALLEL_MAKEINST outside do_install(),
> enable parallellized make by default when oe_runmake is invoked from
> do_compile_ptest() or do_install_ptest() by declaring wrapper task
> specific EXTRA_OEMAKE overrides in fashion similar to do_compile and
> do_install overrides in meta/conf/bitbake.conf.
> 
> Parallel make can still be disabled by resetting bbclass specific
> PTEST_PARALLEL_MAKE and PTEST_PARALLEL_MAKEINST variables in recipe
> e.g. if a race issue needs to be avoided without modifying source code.
> 
> Tested by issuing following command sequence on a 32-core build host:
> 
>   $ bitbake -c clean util-linux && bitbake --skip-setscene -c compile util-linux && time bitbake --skip-setscene -c compile_ptest_base util-linux
> 
> and found that before this change the result was
> 
>   real	0m34.684s
>   user	0m0.753s
>   sys	0m0.131s
> 
> and after this change
> 
>   real	0m9.868s
>   user	0m0.749s
>   sys	0m0.150s
> 
> Signed-off-by: Niko Mauno <niko.mauno@vaisala.com>
> ---
>  meta/classes/ptest.bbclass | 4 ++++
>  1 file changed, 4 insertions(+)

This has merged but is causing a few problems. So far, elfutils and
flex both ran into make races with a ton of failed builds as a result.
I've a patch to disable paralell compile for those two recipes in the
ptest compile. Ideally we'd look into and get this issues fixed
upstream.

Some random build failure links (there are many more in that build):

https://autobuilder.yoctoproject.org/typhoon/#/builders/100/builds/3045
https://autobuilder.yoctoproject.org/typhoon/#/builders/65/builds/5446
https://autobuilder.yoctoproject.org/typhoon/#/builders/62/builds/5417
https://autobuilder.yoctoproject.org/typhoon/#/builders/95/builds/3428

Cheers,

Richard




      parent reply	other threads:[~2022-06-25 16:55 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-22 19:16 [PATCH 1/3] ptest.bbclass: Honor PARALLEL_MAKE, PARALLEL_MAKEINST Niko Mauno
2022-06-22 19:16 ` [PATCH 2/3] strace: Drop redundant oe_runmake parameter Niko Mauno
2022-06-23  9:30   ` [OE-core] " Luca Ceresoli
2022-06-23 16:53     ` Niko Mauno
2022-06-24 10:30       ` Luca Ceresoli
2022-06-22 19:16 ` [PATCH 3/3] valgrind: " Niko Mauno
2022-06-25 16:55 ` Richard Purdie [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b9d79d0231537cfcb8f75bd0e728d76bbdb64f1a.camel@linuxfoundation.org \
    --to=richard.purdie@linuxfoundation.org \
    --cc=niko.mauno@vaisala.com \
    --cc=openembedded-core@lists.openembedded.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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.