Messages in this thread | | | Subject | Re: Linux 3.0.8 | From | Mike Galbraith <> | Date | Wed, 26 Oct 2011 10:14:10 +0200 |
| |
On Tue, 2011-10-25 at 04:41 -0400, Valdis.Kletnieks@vt.edu wrote: > On Tue, 25 Oct 2011 09:02:40 +0200, Greg KH said: > > On Mon, Oct 24, 2011 at 10:40:44PM -0700, Randy Dunlap wrote: > > > On 10/24/11 22:19, Greg KH wrote: > > > > All users of the 3.0 kernel series must upgrade. > > > I noticed that "must" word in 3.0.7 also. Please explain. > > Nothing new, I've been saying that for years. > > I think Randy means that out of the 29 commits for .7->.8, there's only one > (Hugh Dicken's mremap fix) and *maybe* Peter Zijlstra's cputimer fix, which > even *possibly* rise to the "all users" level, as everything else is against > XFS or CIFS or ARM or something else that's not an across-the-board issue. > > And the changelog for the mremap commit says that "but this lucky report > hints", which sort of implies it's a very hard-to-hit bug that nobody even knew > was there. > > So it's not obvious from just the shortlog that it's a "MUST upgrade" category.
Seems like perfectly fine boilerplate to me. If I want the warmest fuzzy I can get, I must upgrade. Peter's patch fixed a box killer that I don't care to hit (dead bugs being kinda boring), so I must upgrade.
-Mike
| |