From: tglozar@redhat.com
To: linux-rt-users@vger.kernel.org
Cc: jkacur@redhat.com, Tomas Glozar <tglozar@redhat.com>
Subject: [PATCH] rteval: Avoid passing None cpulist to measurement
Date: Thu, 6 Jun 2024 12:40:54 +0200 [thread overview]
Message-ID: <20240606104054.111997-1-tglozar@redhat.com> (raw)
From: Tomas Glozar <tglozar@redhat.com>
When rteval is called via the command line, cpulists for both
measurements and loads default to an empty string and are further
processed by parse_cpulist_from_config. However, this is not true when
rteval is used as a module: in that case, neither the default
command-line value is used nor parse_cpulist_from_config is run, leading
to None being set to the config property of measurements which is
explicitely passed down to the corresponding cyclictest config property.
The issue was introduced in 64ce7848dfab ("rteval: Add relative cpulists
for measurements"), where the check for None was removed from
the cyclictest module, assuming the value was processed by
parse_cpulist_from_config.
Avoid this by calling parse_cpulist_from_config with an empty string to
get the default cpulist to pass to cyclictest module if cpulist is empty
in the measurement config.
Note: this is a workaround, in the future, parse_cpulist_from_config
will be called properly in one place only instead of being scattered
throughout the code of rteval. That will also enable relative cpulists
to be used when calling rteval as a module.
Fixes: 64ce7848dfab ("rteval: Add relative cpulists for measurements")
Signed-off-by: Tomas Glozar <tglozar@redhat.com>
---
rteval/modules/measurement/__init__.py | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/rteval/modules/measurement/__init__.py b/rteval/modules/measurement/__init__.py
index 11bd7b0..fda618d 100644
--- a/rteval/modules/measurement/__init__.py
+++ b/rteval/modules/measurement/__init__.py
@@ -148,6 +148,13 @@ measurement profiles, based on their characteristics"""
modcfg = self.__cfg.GetSection("measurement")
cpulist = modcfg.cpulist
run_on_isolcpus = modcfg.run_on_isolcpus
+ if cpulist is None:
+ # Avoid passing None cpulist down to measurement module which might
+ # not expect it
+ # Note: None can appear only when using rteval as module, not when
+ # being called from command line; in the latter, cpulists are all
+ # processed by parse_cpulist_from_config already by this point
+ cpulist = cpulist_utils.collapse_cpulist(parse_cpulist_from_config("", run_on_isolcpus))
for (modname, modtype) in modcfg:
if isinstance(modtype, str) and modtype.lower() == 'module': # Only 'module' will be supported (ds)
--
2.45.0
next reply other threads:[~2024-06-06 10:41 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-06-06 10:40 tglozar [this message]
2024-06-11 4:10 ` [PATCH] rteval: Avoid passing None cpulist to measurement John Kacur
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=20240606104054.111997-1-tglozar@redhat.com \
--to=tglozar@redhat.com \
--cc=jkacur@redhat.com \
--cc=linux-rt-users@vger.kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox