4.3.3 and 4.3.4 have this regression: https://bugzilla.redhat.com/show_bug.cgi?id=1303270

The system can stay up only 8 hours before crashing with an OOM panic (not due to any user space processes). The BSOD further more leaves the ether nic active in a way that DoSes all devices connected to the same switch.

Stuff seems to work for those 8 hours.

4.3.3 and 4.3.4 have this regression: https://bugzilla.redhat.com/show_bug.cgi?id=1303270

The system can stay up only 8 hours before crashing with an OOM panic (not due to any user space processes). The BSOD further more leaves the ether nic active in a way that DoSes all devices connected to the same switch.

Stuff seems to work for those 8 hours.

4.3.3 and 4.3.4 have this regression: https://bugzilla.redhat.com/show_bug.cgi?id=1303270

The system can stay up only 8 hours before crashing with an OOM panic (not due to any user space processes). The BSOD further more leaves the ether nic active in a way that DoSes all devices connected to the same switch.

Stuff seems to work for those 8 hours.

Works for Me.

Resolving dependencies seems to take much long (100% CPU). Not sure if that is a regression, or just the peculiar state of my packages.

sqlite3_drv passes my test suite. Not using sqlite3_drv in production yet, but stuff still seems to work with hash_drv.

Running all day without having to reboot - immediately ran into bug#1192550, but bug#1165369 seems to be gone. Xorg can now be killed from a text console instead of SSH when it crashes. Yay!

adjtimex fixed. Still getting "wrong connector dpms state" from intel driver, but no crashes or tainting.

i915 modeset errors are gone at boot. This bodes well for fixing bz1165369