Setting modes while the VT is inactive (was: Xserver crash on resume)

[Keith Packard]

On Mon, 2007-11-26 at 14:03 +0100, Xavier Bestel wrote:

> This sounds like a recipe for deadlock (imagine a script or something
> waiting for a setmode to do a chvt).

Yeah, that sounds like a poor plan.

> Sounds better (maybe using some kernel-mode modesetting ?).
> Or maybe just fail, but do not hang.

Kernel mode setting would need to deal with validating modes when
inactive as well, so I don't think that changes anything.

Changing DIX to always fail would make it really hard to ever fix this
(drivers wouldn't be tested to see mode setting while inactive). So, how
about fixing the drivers to fail at this point, and let them succeed
when they're ready? Most of the rest of the RandR driver API has
explicit comments about being used even when vtSema is FALSE; the mode
setting and screen size setting ones appear unique in missing this
information.

-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg/attachments/20071126/e28b2427/attachment.pgp>