All of lore.kernel.org
 help / color / mirror / Atom feed
From: Chris J Arges <chris.j.arges@canonical.com>
To: Michal Marek <mmarek@suse.cz>
Cc: Riku Voipio <riku.voipio@linaro.org>,
	linux-kbuild <linux-kbuild@vger.kernel.org>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] builddeb: parallelize debug module installation
Date: Tue, 28 Apr 2015 07:22:59 -0500	[thread overview]
Message-ID: <20150428122258.GA2848@canonical.com> (raw)
In-Reply-To: <553F5B5F.5000709@suse.cz>

On Tue, Apr 28, 2015 at 12:05:19PM +0200, Michal Marek wrote:
> On 2015-04-28 10:57, Riku Voipio wrote:
> > On 27 April 2015 at 19:43, Chris J Arges <chris.j.arges@canonical.com> wrote:
> >> When building the dbg package, we use a large 'for module in $(find' loop that
> >> can be easily parallelized by using 'find | xargs'. This patch modifies this
> >> loop to use the later paradigm.
> >>
> >> In addition, ensure we add '-n1 -P0' to xargs to run as many processes as
> >> possible.
> >>
> >> Signed-off-by: Chris J Arges <chris.j.arges@canonical.com>
> >> ---
> >>  scripts/package/builddeb | 15 ++++++++-------
> >>  1 file changed, 8 insertions(+), 7 deletions(-)
> >>
> >> diff --git a/scripts/package/builddeb b/scripts/package/builddeb
> >> index 88dbf23..538f829 100755
> >> --- a/scripts/package/builddeb
> >> +++ b/scripts/package/builddeb
> >> @@ -152,16 +152,17 @@ if grep -q '^CONFIG_MODULES=y' $KCONFIG_CONFIG ; then
> >>                 rmdir "$tmpdir/lib/modules/$version"
> >>         fi
> >>         if [ -n "$BUILD_DEBUG" ] ; then
> >> -               for module in $(find $tmpdir/lib/modules/ -name *.ko -printf '%P\n'); do
> >> -                       module=lib/modules/$module
> >> -                       mkdir -p $(dirname $dbg_dir/usr/lib/debug/$module)
> >> +               find $tmpdir/lib/modules/ -name *.ko -printf '%P\n' | xargs -n1 -P0 -I {} sh -c '
> > 
> > I would go with -P`getconf _NPROCESSORS_ONLN`. There can be thousands
> > of modules (allmodconfig will make 4500).
> 
> Yep. I was thinking about retrieving the value of the make -j argument
> somehow, but this is not possible and we certainly do not want to
> implement the jobserver protocol in shell :-). So using the number of
> processors is a sensible choice. What can be done is to detect whether
> the -j option is in MAKEFLAGS and only then use multiple instances.
> 
> Michal
>

Michal,

The MAKEFLAGS variable will have the jobserver fd's in it, if we are building
in parallel with -j. I could use this to detect if we are building in parallel
and adjust the '-P' flag in xargs accordingly. I'll work on v3 of this patch,
thanks for the reviews!

--chris 

  reply	other threads:[~2015-04-28 12:23 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-04-24 18:25 [PATCH 0/2] make deb-pkg build speed improvements Chris J Arges
2015-04-24 18:25 ` [PATCH 1/2] package: Makefile: ensure $MAKE can use jobserver Chris J Arges
2015-04-28  8:38   ` Riku Voipio
2015-04-24 18:25 ` [PATCH 2/2] builddeb: parallelize debug module installation Chris J Arges
2015-04-27 14:22   ` Michal Marek
2015-04-27 16:43     ` [PATCH v2] " Chris J Arges
2015-04-28  8:57       ` Riku Voipio
2015-04-28 10:05         ` Michal Marek
2015-04-28 12:22           ` Chris J Arges [this message]
2015-04-28 13:24           ` [PATCH v3] " Chris J Arges

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=20150428122258.GA2848@canonical.com \
    --to=chris.j.arges@canonical.com \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mmarek@suse.cz \
    --cc=riku.voipio@linaro.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.