Hi all,
On 14-06-17 11:34, Michael Thayer wrote:
Hello Sean,
13.06.2017 20:03, Sean Paul wrote:
[Discussion of vboxvideo driver clean-up.]
First, thank you for your submission.
Thank you for your feedback.
[Discussion of the OS-independent code in the driver submission.]
I took a quick skim through your driver, and there doesn't seem to be much
secret sauce there that will be tough to keep up-to-date across platforms.
I have two particular concerns there: first if we add new functionality
(which we would do out of tree first) it will need porting over.
Acceleration support is the most likely candidate. And if someone does
make fixes to that part of the code in the kernel tree they will also
need porting over. I agree, that concern is probably overblown, and
best addressed by keeping that part of the code as close to our tree as
possible while still meeting kernel standards (hence my question as to
what that would be).
The second concern is not relevant to DRI, but it concerns our other
guest drivers (not the host one with the C++ in it Greg!) which Hans
also expressed interest in putting upstream after seeing how vboxvideo
fares. The OS-independent part is quite a bit larger and more volatile,
though it has thankfully stablised a lot. That concern is probably also
overblown, though I do wonder whether upstreaming those driver is the
best solution (that would be Hans's call though).
As I already mentioned in previous mails on this, for the vboxguest driver
my plan is to simply do a fork and remove anything related to the
portability. It currently weighs in at 100000+ lines of code which is a bit
much for what it does I believe I will be able to get a Linux only version
of it down to a small fraction of that and the result will be a much cleaner
and better driver.
FWIW I've already stared looking into cleaning up the vboxguest driver and
my first target is to remove all dependencies on the r0drv code.
Regards,
Hans
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
