[PATCH 3/3] xfree86: Add a conservative SIGIO handler path

[Adam Jackson]

On Tue, 2010-10-19 at 18:08 -0200, Vignatti Tiago (Nokia-MS/Helsinki)
wrote:
> On Tue, Oct 19, 2010 at 09:41:20PM +0200, ext Adam Jackson wrote:
> > Which should be a material improvement for everybody, even if it doesn't
> > reduce the number of code paths it's at least no more complex than what
> > we have.
> 
> I'm not that convinced yet that conservative's SIGIO approach performs better
> than the method using main thread handler. For instance, I'm afraid that it
> could penalize too much other running X requests.

Easy enough to test - make a uevent mouse device and jiggle it 500 times
a second and see if your x11perf numbers change - but I would be really
surprised.  Also seems like a weird thing to care about, I'm willing to
pay 3% in worst case rendering speed to get the cursor to move smoothly.

> Anyway, I'd like to ask if you can hang these patches for more one week until
> I come with numbers benchmarking all these different approaches.

Well, the libc5 bit is merged, but otherwise sure.  I was mostly looking
for feedback from non-Linux people, since I doubt the conservative path
will see much use on Linux once input threads lands (in the same way
that -nosilk is basically unused).

- ajax
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg-devel/attachments/20101020/2f426b25/attachment.pgp>