From: David Marchand <david.marchand@redhat.com>
To: Agalya Babu RadhaKrishnan <agalyax.babu.radhakrishnan@intel.com>
Cc: dev <dev@dpdk.org>,
reshma.pattan@intel.com,
Bruce Richardson <bruce.richardson@intel.com>
Subject: Re: [dpdk-dev] [PATCH] app/test: enhance meson test run for FreeBSD and Linux
Date: Wed, 3 Apr 2019 14:39:43 +0200 [thread overview]
Message-ID: <CAJFAV8yR7rw=7PYfgL65h690Vn_wyE87vA0PYwpWtX81W+ORjw@mail.gmail.com> (raw)
Message-ID: <20190403123943.8Mp36NCwmt2N0QZTcYx0mgKWIIKA-sOiN_Gq30An0vA@z> (raw)
In-Reply-To: <1551863976-14093-1-git-send-email-agalyax.babu.radhakrishnan@intel.com>
On Wed, Mar 6, 2019 at 10:21 AM Agalya Babu RadhaKrishnan <
agalyax.babu.radhakrishnan@intel.com> wrote:
> From: Agalya Babu RadhaKrishnan <agalyax.babu.radhakrishnan@intel.com>
>
> 1)For linux, running all tests on same cores can increase failure rate.
> So instead run each test case on separate lcore using -l EAL option.
> Lcore for each test case will be selected in the incremental order
> of 0,1,2 ... till it reaches the maximum available cores in the
> system which can be known from 'nproc' command.
> On the other hand, by default meson runs all tests in parallel in
> different threads which can be controlled by MESON_TESTTHREAD option.
> eg. MESON_TESTTHREAD=4 meson test --suite DPDK:fast-tests
>
> 2)FreeBSD dont support file-prefix, so changes are made to run the
> tests without 'file-prefix'option.
> eg. meson test --suite DPDK:fast-tests
>
> Signed-off-by: Agalya Babu RadhaKrishnan <
> agalyax.babu.radhakrishnan@intel.com>
> ---
> app/test/meson.build | 28 +++++++++++++++++++++++-----
> 1 file changed, 23 insertions(+), 5 deletions(-)
>
> diff --git a/app/test/meson.build b/app/test/meson.build
> index 05e5ddeb0..8303aaaa6 100644
> --- a/app/test/meson.build
> +++ b/app/test/meson.build
> @@ -337,11 +337,29 @@ if get_option('tests')
> timeout_seconds_fast = 10
>
> foreach arg : fast_parallel_test_names
> - test(arg, dpdk_test,
> - env : ['DPDK_TEST=' + arg],
> - args : ['-c f','-n 4', '--file-prefix=@0@
> '.format(arg)],
> - timeout : timeout_seconds_fast,
> - suite : 'fast-tests')
> + if host_machine.system() == 'linux'
> + lcore_number = 1
> + cores_running = run_command('nproc')
> + cores_used = cores_running.stdout().to_int()
> + test(arg, dpdk_test,
> + env : ['DPDK_TEST=' + arg],
> + args : ['-l @0@'.format(lcore_number),
> '-n 4', '--file-prefix=@0@'.format(arg)],
> + timeout : timeout_seconds_fast,
> + suite : 'fast-tests')
> + if (lcore_number < cores_used)
> + lcore_number += 1
> + else
> + lcore_number = 1
> + endif
> + else
> + test(arg, dpdk_test,
> + env : ['DPDK_TEST=' + arg],
> + args : ['-n 4'],
> + is_parallel : false,
> + timeout : timeout_seconds_fast,
> + suite : 'fast-tests')
> + endif
> +
> endforeach
>
> foreach arg : fast_non_parallel_test_names
>
We don't know if the tests really need those 4 cores, can you investigate
this point before touching those parameters ?
I am not sure I get all the subtilities in the meson part, but from what I
understood, for Linux, you look at all the cpus available and set the
corelist but for FreeBSD you rely on the EAL autodetection to catch all
cpus available since you don't give a corelist/mask.
To be consistent I would use the same mechanism for both Linux and FreeBSD,
but I would prefer we are sure on the tests needs for cpu resources.
Thanks.
--
David Marchand
next prev parent reply other threads:[~2019-04-03 12:39 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-03-06 9:19 Agalya Babu RadhaKrishnan
2019-04-01 12:32 ` Pattan, Reshma
2019-04-01 12:32 ` Pattan, Reshma
2019-04-03 12:23 ` Pattan, Reshma
2019-04-03 12:23 ` Pattan, Reshma
2019-04-03 12:39 ` David Marchand [this message]
2019-04-03 12:39 ` David Marchand
2019-04-26 14:53 ` Pattan, Reshma
2019-04-26 14:53 ` Pattan, Reshma
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='CAJFAV8yR7rw=7PYfgL65h690Vn_wyE87vA0PYwpWtX81W+ORjw@mail.gmail.com' \
--to=david.marchand@redhat.com \
--cc=agalyax.babu.radhakrishnan@intel.com \
--cc=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
--cc=reshma.pattan@intel.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).