<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<p><font size="4">Hi!</font><br>
</p>
<div class="moz-cite-prefix">On 1/21/22 11:16, Petr Vorel wrote:<br>
</div>
<blockquote type="cite" cite="mid:YeqIB+Y4rO2XK57i@pevik">
<pre class="moz-quote-pre" wrap="">Hi all,
[Cc Li ]
</pre>
<blockquote type="cite">
<pre class="moz-quote-pre" wrap="">tmpfs does not support it.
</pre>
</blockquote>
<pre class="moz-quote-pre" wrap="">Verifying with .all_filesystems, that only tmpfs is problematic
- tested on exfat, ntfs over FUSE, vfat (common linux filesystems obviously work)
That's why I avoided opening file with O_DIRECT in setup().
BTW I wonder what is the reason of duplicate entries in ltp-aiodio.part4?
i.e. dio_sparse, aiodio_append, dio_append, dio_truncate...
To create bigger load?</pre>
</blockquote>
<p>To duplicate entries in ltp-aiodio.part4 doesn't make much sense
and I perfectly agree with you. Once the aiodio testing suite is
finished, we do need to change runtest file in order to have
multiple scenarios with different files/buffers size with a
different number of sub-processes.<br>
</p>
<blockquote type="cite" cite="mid:YeqIB+Y4rO2XK57i@pevik">
<pre class="moz-quote-pre" wrap="">
IMHO it'd make sense to run at least one test on .all_filesystems to check more
filesystems (dio_append seems to be quick), but I wouldn't done it when there
are duplicate entries.
Kind regards,
Petr
</pre>
</blockquote>
</body>
</html>