[ Short story - I already do exactly this. Long story: some of the
differences between the kernel tree and the DRI CVS tree really need
somebody to get rid of the task-queue usage in DRI CVS, and that I don't
have time for myself at least right now ]
On Thu, 12 Dec 2002, Alan Hourihane wrote:
>
> If Linus has already found the time to merge, how long would it be, to
> do a diff and send an email off ? I doubt it's that much extra work.
Oh, I have my own very selfish reasons to keep the kernel and dri cvs
trees close together - it makes it easier for me to continue to merge
stuff (and I don't mind the merging at all, since DRI is something I'm
personally actually _using_ and interested in - I think good 3D support is
very important - even if I myself have no clue about GL etc ;-).
The way I do the merges is to do a context diff of the kernel modules and
the dri tree, and then edit that diff by hand before I apply it to the
kernel tree and do the limited amount of testing I do (I only have
radeons, so..)
Which means that the less differences there are, the easier it is for me.
As fo the last merge, the _only_ differences were in build files (ie the
kernel makefiles are totally different, I'm not even trying to merge those
parts) and in a few one-liners that are fundamentally different in 2.4.x
and 2.5.x (task queues). Everything else has already been merged back.
The 2.4.x and 2.6.x differences are somwehat awkward, but since I detext
having #ifdef's in source trees, I actually prefer haviung the differences
there right now. Some of them should go away if the task-queue handling is
changed to use direct interrupt routines and spinlocks, so the proper fix
is probably _not_ to have the #ifdef's there, but to rewrite the code
somewhat. I wrote a description of that, but quite honestly I don't have
the time to really do the code rewrite and testing myself, so..
Linus
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
