[PATCH] more generosity for defer_open ??

Michael Adam obnox at samba.org
Thu Sep 11 12:15:29 MDT 2014

While I trying to get the infamous defer_open test
more reliable, I raised the absolute upper generousity
from + 1 to + 1.5. Due to the scaling of 1s <=> 0.1s
in torture, this allows for up to 1.65s in real time
of delay for SHARING_VIOLATION. (where we actually mean 1.5
scaled seconds, i.e. 1.65 = 1.5 * 0.1 + 1.5).

Now in my defer_open test loop, after some 1600 successful
runs, I had a failure on a busy system with 1.82 s (real time).

Should we add the above generosity to an absolute 2 s?
(The lower bound is more important anyways.)

Patch attached for consideration.
And sorry that this is so much back and forth. 
This is not exact science. .. :)

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-s4-torture-base-be-even-more-generous-in-the-defer_o.patch
Type: text/x-diff
Size: 1207 bytes
Desc: not available
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20140911/9bb1a223/attachment.patch>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20140911/9bb1a223/attachment.pgp>

More information about the samba-technical mailing list