lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 24 Jan 2024 09:38:58 +0100
From: Jiri Olsa <olsajiri@...il.com>
To: Ian Rogers <irogers@...gle.com>
Cc: Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>,
	Arnaldo Carvalho de Melo <acme@...nel.org>,
	Namhyung Kim <namhyung@...nel.org>,
	Mark Rutland <mark.rutland@....com>,
	Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
	Adrian Hunter <adrian.hunter@...el.com>,
	Ravi Bangoria <ravi.bangoria@....com>,
	Ross Zwisler <zwisler@...omium.org>,
	Athira Rajeev <atrajeev@...ux.vnet.ibm.com>,
	Shirisha G <shirisha@...ux.ibm.com>,
	Kajol Jain <kjain@...ux.ibm.com>,
	Kan Liang <kan.liang@...ux.intel.com>,
	linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 5/5] perf test: Make daemon signal test less racy

On Tue, Jan 23, 2024 at 08:30:15PM -0800, Ian Rogers wrote:
> The daemon signal test sends signals and then expects files to be
> written. It was observed on an Intel Alderlake that the signals were
> sent too quickly leading to the 3 expected files not appearing. To
> avoid this send the next signal only after the expected previous file
> has appeared. To avoid an infinite loop the number of retries is
> limited.
> 
> Signed-off-by: Ian Rogers <irogers@...gle.com>
> Acked-by: Namhyung Kim <namhyung@...nel.org>

Acked-by: Jiri Olsa <jolsa@...nel.org>

jirka

> ---
>  tools/perf/tests/shell/daemon.sh | 34 ++++++++++++++++++++++----------
>  1 file changed, 24 insertions(+), 10 deletions(-)
> 
> diff --git a/tools/perf/tests/shell/daemon.sh b/tools/perf/tests/shell/daemon.sh
> index 4c598cfc5afa..e5fa8d6f9eb1 100755
> --- a/tools/perf/tests/shell/daemon.sh
> +++ b/tools/perf/tests/shell/daemon.sh
> @@ -414,16 +414,30 @@ EOF
>  	# start daemon
>  	daemon_start ${config} test
>  
> -	# send 2 signals
> -	perf daemon signal --config ${config} --session test
> -	perf daemon signal --config ${config}
> -
> -	# stop daemon
> -	daemon_exit ${config}
> -
> -	# count is 2 perf.data for signals and 1 for perf record finished
> -	count=`ls ${base}/session-test/*perf.data* | wc -l`
> -	if [ ${count} -ne 3 ]; then
> +        # send 2 signals then exit. Do this in a loop watching the number of
> +        # files to avoid races. If the loop retries more than 600 times then
> +        # give up.
> +	local retries=0
> +	local signals=0
> +	local success=0
> +	while [ ${retries} -lt 600 ] && [ ${success} -eq 0 ]; do
> +		local files
> +		files=`ls ${base}/session-test/*perf.data* 2> /dev/null | wc -l`
> +		if [ ${signals} -eq 0 ]; then
> +			perf daemon signal --config ${config} --session test
> +			signals=1
> +		elif [ ${signals} -eq 1 ] && [ $files -ge 1 ]; then
> +			perf daemon signal --config ${config}
> +			signals=2
> +		elif [ ${signals} -eq 2 ] && [ $files -ge 2 ]; then
> +			daemon_exit ${config}
> +			signals=3
> +		elif [ ${signals} -eq 3 ] && [ $files -ge 3 ]; then
> +			success=1
> +		fi
> +		retries=$((${retries} +1))
> +	done
> +	if [ ${success} -eq 0 ]; then
>  		error=1
>  		echo "FAILED: perf data no generated"
>  	fi
> -- 
> 2.43.0.429.g432eaa2c6b-goog
> 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ