We keep metrics on our binaries with respect to how much memory they take up at startup time. So for instance if you run ./a.pl --basemem, it grabs the VmRSS value out of /proc/$$/status and exits. When we added Red Hat 7 support to our product, we noticed that this value was VERY unstable. After much poking, I accidentally discovered that this was a side effect of having selinux enabled when doing the check. When I disable selinux in /etc/selinux/config and reboot, I found that ./a.pl --basemem became much more stable. I am guessing something changed between Red Hat 6 and 7 which caused libselinux to map differently into the address space. I'm however not ruling out a libc change. Can anyone shed light on this or point me to something that would indicate why? For reference, this is libselinux 2.2.2 vs 2.0.94. Thanks, Todd Rinaldo
Attachment:
signature.asc
Description: Message signed with OpenPGP using GPGMail
_______________________________________________ Selinux mailing list Selinux@xxxxxxxxxxxxx To unsubscribe, send email to Selinux-leave@xxxxxxxxxxxxx. To get help, send an email containing "help" to Selinux-request@xxxxxxxxxxxxx.
