Messages in this thread | | | Date | Fri, 4 Apr 2014 12:12:26 -0700 | Subject | Re: [tip:x86/reboot] [PATCH] x86: Try the BIOS reboot method before the PCI reboot method | From | Linus Torvalds <> |
| |
On Fri, Apr 4, 2014 at 10:48 AM, Ingo Molnar <mingo@kernel.org> wrote: > > 'Breaks a box' is more than enough justification to revert a patch...
Indeed. That's especially true when "breaks a box" was found within a day of the patch being merged.
The default thinking should be: "If we found _one_ box that broke during the merge window, that probably means that there are at least ten thousand boxes that would break if the change actually hit a major distribution kernel".
Yes, developer boxes are sometimes "special". But the fact is, I doubt that's the case. Not when this kind of breakage (oops, reboot doesn't work) is actually not all that unusual. The quirk entries we have for it are *not* developer boxes.
Linus
| |