radeon_cursor

[Michel Dänzer]

On Thu, 2005-01-13 at 11:35 +0100, Thomas Winischhofer wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Michel Dänzer wrote:
> | On Sun, 2005-01-02 at 15:19 -0500, Vladimir Dergachev wrote:
> |
> |>  My understanding is that the cursor update runs in a separate thread
> (as
> |>the mouse still works when the server is otherwise locked up).
> |
> |
> | I've been fooled by this before as well, but no. Only mouse movement is
> | handled asynchronously.
> 
> 
> So, is it possible that other driver functions are being interrupted by
> that mouse thread?

Yeah, although as Adam pointed out, it's not a thread, just a signal
handler.

> I have a slightly different problem: It currently write an X driver for
> USB2VGA dongles. The kernel driver for this device provides a device
> file to/from which all hardware access goes to/comes from.
> 
> Using the HW cursor, I get lockups quite frequently (because the kernel
> device blocks concurrent read/writes). Using the SW cursor works just fine.
> 
> My suspiction is that the HW cursor's position setting gets hung in the
> wait queue, producing a "classic" race.

Quite possibly. If you can't make the cursor position updates signal
handler safe, you'll have to disable Silken Mouse (i.e. not enable it in
the first place :) for that driver.


-- 
Earthling Michel Dänzer      |     Debian (powerpc), X and DRI developer
Libre software enthusiast    |   http://svcs.affero.net/rm.php?r=daenzer