Kernel prepatch 3.18-rc7

[Posted December 1, 2014 by corbet]

The 3.18-rc7 prepatch is out. Linus seems happy enough, despite the persistent lockup problem that has defied all debugging attempts so far. "At the same time, with the holidays coming up, and the problem _not_ being a regression, I suspect that what will happen is that I'll release 3.18 on time in a week, because delaying it will either mess up the merge window and the holiday season, or I'd have to delay it a *lot*."

Kernel prepatch 3.18-rc7

Posted Dec 1, 2014 14:33 UTC (Mon) by torquay (guest, #92428) [Link] (2 responses)

the problem _not_ being a regression

Umm, what is Linus smoking? It is still a regression, no matter how you paint it: 3.16 doesn't seem to have this problem. Choosing 3.17 at the reference point and then declaring "no regression" seems rather reckless. It essentially declares any regressions in 3.17 that are detected past the 3.18 merge window as "not really regressions".

Kernel prepatch 3.18-rc7

Posted Dec 1, 2014 14:56 UTC (Mon) by andyc (subscriber, #1130) [Link]

I think what Linus means is that it's not a _new_ regression introduced in 3.18

Kernel prepatch 3.18-rc7

Posted Dec 1, 2014 15:50 UTC (Mon) by welinder (guest, #4699) [Link]

It's just an unfortunate excerpt.

The linked email is clear that the statement is relative to 3.17:
"As it is, that isn't a regression from 3.17, [...]".