Messages in this thread | | | Date | Wed, 5 Jan 2011 21:18:01 +0100 | From | Ingo Molnar <> | Subject | Re: [PATCH 2/2] jump label: introduce static_branch() |
| |
* David Daney <ddaney@caviumnetworks.com> wrote:
> On 01/05/2011 11:50 AM, Ingo Molnar wrote: > > > >* David Daney<ddaney@caviumnetworks.com> wrote: > > > >>On 01/05/2011 11:14 AM, Ingo Molnar wrote: > >>> > >>>* H. Peter Anvin<hpa@zytor.com> wrote: > >>> > >>>>On 01/05/2011 09:43 AM, Steven Rostedt wrote: > >>>>>On Wed, 2011-01-05 at 09:32 -0800, David Daney wrote: > >>>>> > >>>>>>This patch will conflict with the MIPS jump label support that Ralf has > >>>>>>queued up for 2.6.38. > >>>>> > >>>>>Can you disable that support for now? As Linus said at Kernel Summit, > >>>>>other archs jumped too quickly onto the jump label band wagon. This > >>>>>change really needs to get in, and IMO, it is more critical to clean up > >>>>>the jump label code than to have other archs implementing it. > >>>>> > >>>> > >>>>Ralf is really good... perhaps we can get the conflicts resolved? > >>> > >>>Yep, the best Git-ish way to handle that is to resolve the conflicts whenever they > >>>happen - i.e. whoever merges his tree upstream later. No need for anyone to 'wait' > >>>or undo anything. > >>> > >> > >>There will be no git conflicts, as the affected files are disjoint. > > > >I regularly resolve semantic conflicts in merge commits - or in the first followup > >commit. > > > > But I am guessing that neither you, nor Linus, regularly build MIPS > kernels with GCC-4.5.x *and* jump label support enabled. [...]
I build MIPS defconfig kernels at least once per day - so at least serious, wide-ranging issues should not slip through. Rarer combos possibly - but that's true of pretty much anything.
> [...] So how would such semantic conflict ever be detected? I would expect the > conflict to first occur when Linus pulls Ralf's tree.
If that slips through then a fix is queued up?
Thanks,
Ingo
| |