From: Adrian Bunk <bunk@stusta.de>
To: Mike Looijmans <mike.looijmans@topic.nl>
Cc: openembedded-core@lists.openembedded.org
Subject: Re: [PATCH] classes/populate_sdk_base: Implement xz compression options
Date: Thu, 19 Mar 2020 13:37:36 +0200 [thread overview]
Message-ID: <20200319113736.GA28081@localhost> (raw)
In-Reply-To: <20200318122156.23599-1-mike.looijmans@topic.nl>
On Wed, Mar 18, 2020 at 01:21:56PM +0100, Mike Looijmans wrote:
> Building an SDK on a machine with 8GB RAM resulted in excessive swapping
> due to the xz compressor using ~20GB of memory. This is because xz is
> being called with "-T 0 -9".
>
> To allow tuning the compression versus memory usage, introduce a variable
> named SDK_XZ_OPTIONS that defaults to a more sane default:
> SDK_XZ_OPTIONS ?= "${XZ_DEFAULTS} ${XZ_COMPRESSION_LEVEL}"
> Thus, inherit any XZ tuning already done, and allow users to specify
> overrides for this task in their config by supplying SDK_XZ_OPTIONS.
> Since XZ_COMPRESSION_LEVEL defaults to -9 this does not change the standard
> behavior.
>...
> +SDK_XZ_OPTIONS ?= "${XZ_DEFAULTS} ${XZ_COMPRESSION_LEVEL}"
A problem is that XZ_COMPRESSION_LEVEL compression would now be used for
unrelated usecases, and lowering the compression for being able to boot
on low-end targets (<= 64 MB RAM) shouldn't impact the SDK.
>...
> - d.setVar('SDK_ARCHIVE_CMD', 'cd ${SDK_OUTPUT}/${SDKPATH}; tar ${SDKTAROPTS} -cf - . | xz -T 0 -9 > ${SDKDEPLOYDIR}/${TOOLCHAIN_OUTPUTNAME}.${SDK_ARCHIVE_TYPE}')
> + d.setVar('SDK_ARCHIVE_CMD', 'cd ${SDK_OUTPUT}/${SDKPATH}; tar ${SDKTAROPTS} -cf - . | xz ${SDK_XZ_OPTIONS} > ${SDKDEPLOYDIR}/${TOOLCHAIN_OUTPUTNAME}.${SDK_ARCHIVE_TYPE}')
>...
Replacing "-T 0" with ${XZ_DEFAULTS} should solve your problem.
This already takes both cpus and memory into account.
cu
Adrian
next prev parent reply other threads:[~2020-03-19 11:37 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-18 12:21 [PATCH] classes/populate_sdk_base: Implement xz compression options Mike Looijmans
2020-03-19 11:37 ` Adrian Bunk [this message]
2020-03-20 14:48 ` Mike Looijmans
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=20200319113736.GA28081@localhost \
--to=bunk@stusta.de \
--cc=mike.looijmans@topic.nl \
--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.