From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Stancek Date: Wed, 15 Feb 2017 11:50:47 -0500 (EST) Subject: [LTP] [PATCH] syscalls/mq_open: fix limits for queues_max In-Reply-To: <20170215152406.GB29673@rei.lan> References: <1486384930-28588-1-git-send-email-yangx.jy@cn.fujitsu.com> <20170215152406.GB29673@rei.lan> Message-ID: <956936356.5228861.1487177447794.JavaMail.zimbra@redhat.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ltp@lists.linux.it ----- Original Message ----- > From: "Cyril Hrubis" > To: "Xiao Yang" > Cc: ltp@lists.linux.it, "Jan Stancek" > Sent: Wednesday, 15 February, 2017 4:24:06 PM > Subject: Re: [LTP] [PATCH] syscalls/mq_open: fix limits for queues_max > > Hi! > > This case fails on RHEL6.8GA and RHEL6.9Beta, because setting > > queues_max to 0 is invalid. the minimum value of queues_max > > has been limitted to 1 on some distributions, Please see the > > following kernel commit: > > > > commit 5b5c4d1a1440e94994c73dddbad7be0676cd8b9a since RHEL6.2 > > Author: Doug Ledford > > Date: Thu May 31 16:26:30 2012 -0700 > > > > ipc/mqueue: update maximums for the mqueue subsystem > > > > We set queues_max to 1 instead of 0, so this case can work on RHEL6. > > First I think that you have wrong commit, the one that actually added > the limits was: > > commit 93e6f119c0ce8a1bba6e81dc8dd97d67be360844 since RHEL6.2 > Author: Doug Ledford > Date: Thu May 31 16:26:28 2012 -0700 > > ipc/mqueue: cleanup definition names and locations > > Since this commit changes the proc handler in sysctl table and adds the > minimum > and maximum values there. > > Secondly these limits were removed in: > > commit a5c5928b759d8c7382cccc7be36769a68046cc58 doesn't seem to present in any RHEL6 kernel > Author: Joe Perches > Date: Fri Jun 6 14:38:07 2014 -0700 > > ipc: convert use of typedef ctl_table to struct ctl_table > > > I'm not sure if this is worth the effort. > > Jan what do you think, should we apply this to fix the test on RHEL6? I would like this to work with RHEL6, but it be good to understand why failures started for Xiao only now. I'm looking at results of kernel-2.6.32-694.el6 (6.9), and it's passing: <<>> tag=mq_open01 stime=1487065446 cmdline="mq_open01" contacts="" analysis=exit <<>> EXPECT: return value(ret)=(N >= 0) errno=0 (Success) RESULT: return value(ret)= 7 errno=0 (Success) mq_open01 0 TINFO : (case00) START mq_open01 0 TINFO : mq_maxmsg E:20, R:20 mq_open01 0 TINFO : mq_msgsize E:16384, R:16384 mq_open01 0 TINFO : (case00) END => OK mq_open01 1 TPASS : mq_open call succeeded <<>> initiation_status="ok" duration=0 termination_type=exited termination_id=0 corefile=no cutime=0 cstime=0 <<>> I'll try 6.8 too. Regards, Jan