[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZoR1GSV4_IWrNMK3@agluck-desk3.sc.intel.com>
Date: Tue, 2 Jul 2024 14:46:01 -0700
From: Tony Luck <tony.luck@...el.com>
To: Reinette Chatre <reinette.chatre@...el.com>
Cc: "Wieczor-Retman, Maciej" <maciej.wieczor-retman@...el.com>,
"shuah@...nel.org" <shuah@...nel.org>,
"Yu, Fenghua" <fenghua.yu@...el.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-kselftest@...r.kernel.org" <linux-kselftest@...r.kernel.org>,
"ilpo.jarvinen@...ux.intel.com" <ilpo.jarvinen@...ux.intel.com>
Subject: Re: [PATCH v3 2/2] selftests/resctrl: Adjust SNC support messages
On Tue, Jul 02, 2024 at 02:26:25PM -0700, Reinette Chatre wrote:
> Hi Tony,
>
> On 7/1/24 9:04 AM, Luck, Tony wrote:
> > +static bool cpus_offline_empty(void)
> > +{
> > + char offline_cpus_str[64];
> > + FILE *fp;
> > +
> > + fp = fopen("/sys/devices/system/cpu/offline", "r");
> >
> > Check for fp == NULL before using it.
> >
> > + if (fscanf(fp, "%s", offline_cpus_str) < 0) {
> >
> > fscanf() seems like a heavy hammer.
>
> Do you perhaps have any recommendations that should be used instead of
> fscanf()? I checked with stat() but could not see a difference between
> file with a CPU and a file without. Other alternative is
> open()/read()/close()? Looks like when there are no offline CPUs then
> the file will only contain '\n' so it may be possible to read one byte
> from the file and confirm it is '\n' as a check for "cpus_offline_empty()".
Sorry. I replied with Outlook and didn't quote things properly
so my alternate suggestion didn't stand out. Here it is again:
if (fgets(offline_cpus_str, sizeof(offline_cpus_str), fp) == NULL) {
fclose(fp);
return true;
}
But that was with the assumption that /sys/devices/system/cpu/offline
would be empty. Not that it would conatain a single "\n" as you say
above.
So fgets( ...) followed with "if (offline_cpus_str[0] == '\n') "?
-Tony
Powered by blists - more mailing lists