Another useful dichotomy would be to separate the changes that are required for Xen into two buckets: Changes/files that impact existing Linux files, and files that (eventually) end up in a subdirectory that is distinct from any existing Linux subdirectory. Presumably the latter class will not require (nearly as much) review. On a somewhat self-serving note, the first class of changes will almost certainly be guaranteed to be arch-neutral (else there will be howls from all of the linux arch maintainers). We must also work to ensure that the second class are also arch-neutral... they currently are not (even files that are currently targeted for "linux/drivers/xen/core"). Dan > -----Original Message----- > From: virtualization-bounces@xxxxxxxxxxxxxx > [mailto:virtualization-bounces@xxxxxxxxxxxxxx] On Behalf Of > Martin J. Bligh > Sent: Wednesday, September 28, 2005 12:41 PM > To: Walker, Bruce J (HP-Labs); Pratap Subrahmanyam; Rik van Riel > Cc: cgriffin@xxxxxxxxxx; Chris Wright; steven@xxxxxxxxxxxxx; > virtualization@xxxxxxxxxxxxxx > Subject: RE: Phone meeting about kernel virtualization hooks > > > I'm very encouraged that "we all know what needs doing" and > I'm quite > > happy not to have the call. I am concerned, however, that progress > > appears to have been very slow (Chris is currently on leave and very > > unavailable). At the Xen mini-summit at OLS, Ian said it > was urgent for > > Xen's success to get the hooks into the kernel and pleaded for help. > > Since then little has visibly happened (Zach has made a little > > progress). How is this going to get done? > > > > So. I'm cancelling the call but plead for someone to lead > the effort > > (or at least explain what needs doing). We are willing to > help but need > > some direction. > > If "we all know what needs doing", might be useful to list that out in > email just to make sure. Very very roughly, what I was thinking was to > make (and keep maintained a stack of (in order)): > > mainline kernel + things going upstream + rest of xen > > And keep that stack working. To put it another way, we need > to keep slicing > stuff off the bottom of the Xen patch and merging it. > > I don't think we have a working stack in that form at the > moment (at least > not on top of any recent mainline kernel), which is my main > concern right > now (can't really do much without it). > > Would be splendid if we can keep that up-to-date with every > major release, > plus -rc releases if possible. I know it's a lot of work ... > > M. > >
