From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Stancek Date: Tue, 26 Apr 2016 05:47:34 -0400 (EDT) Subject: [LTP] [PATCH v3 5/5] kernel/syscalls: add new test with 'open() + O_TMPFILE' In-Reply-To: <20160425150735.GE20158@rei.lan> References: <1454069150-27889-1-git-send-email-alexey.kodanev@oracle.com> <1454069150-27889-6-git-send-email-alexey.kodanev@oracle.com> <1689676802.79541.1461594932412.JavaMail.zimbra@redhat.com> <571E33E0.5040807@oracle.com> <20160425150735.GE20158@rei.lan> Message-ID: <1962140078.296306.1461664054799.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: "Alexey Kodanev" > Cc: "Jan Stancek" , "vasily isaenko" , ltp@lists.linux.it > Sent: Monday, 25 April, 2016 5:07:35 PM > Subject: Re: [LTP] [PATCH v3 5/5] kernel/syscalls: add new test with 'open() + O_TMPFILE' > > Hi! > > If the issue with mode... it can be because glibc doesn't process mode > > argument unless O_CREAT is found. In our case we have O_TMPFILE instead. > > > > Here is the link to bugzilla: > > > > "open() and openat() ignore 'mode' with O_TMPFILE" > > https://sourceware.org/bugzilla/show_bug.cgi?id=17523 Thanks. Jeff Bastian independently found this BZ as well. systemtap confirms, that mode remains 0, even when testcase is passing non-zero values. sys_openat ., dfd=0xffffffffffffff9c filename=0x40ac50 flags=0x404002 mode=0x0 > > What I see matches the the behavior in Comment 5. I'm on aarch64, so that may explain why both fail for me. After applying this patch: https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=65f6f938cd562a614a68e15d0581a34b177ec29d both testcases passed. Regards, Jan