On 29/04/13(Mon) 13:25, patrick keshishian wrote:
> On Sun, Apr 28, 2013 at 03:44:09PM +0200, Martin Pieuchot wrote:
> > Diff below is a rework of the suspend/resume logic in ehci(4).
> > 
> > If you're not interested in the full story below, please make sure
> > it doesn't introduce any regression on your machine(s) and, in any
> > case report to me (with your dmesg!) thanks :)
> > 
> > Full story:
> > 
> >   So this diff changes the way we supsend/resume ehci(4) in various
> >   way.  It started as a simple rewrite of the ehci_reset() function
> >   to fix a race while working on the rework of the suspend/resume
> >   framework with deraadt@, then later on I discovered that some of
> >   the magic in there was wrong or no longer needed.
> >   
> >   First of all, it removes the logic introduced in r1.46 that places
> >   all the ports into suspend mode because it no longer makes sense
> >   as we *do* detach/reattach USB devices during a suspend/resume 
> >   cycle now.
> 
> Naive question here, but would this be a cause for issue
> reported here:
> 
>     http://marc.info/?l=openbsd-misc&m=136327530312197&w=2

Certainly.

> 
> or at least I'm curious as to why the Wacom tablet does not
> reattach after a resume.

Really?  Are you sure it doesn't reattach or is it just not usable in
X?  It's easy to check, try to suspend/resume in console with the
tablet connected, does it show up again in your dmesg?

And for the crash you can try to add the following in your xorg.conf:

        Section "ServerFlags"
            Option "AllowMouseOpenFail" "True"
        EndSection

I don't know if it will be enough, I'm not sure what's the status of
devices hotplugging in xenocara.  If you want to be more than "curious"
you can build the xf86-input-usbtablet with debug enable and try to find
what the problem is ;)  Diffs are always welcome!

M.

Reply via email to