Eric, I updated my 5.1-RELEASE system to CURRENT dated today at
approx. 9:10 CDT to give your changes a try. I had a bit of a fright at
first with kernel panics right at the end of the boot sequence but it turned
out I had forgotten to disable the ltmdm code -- the kernel module
compiled under -RELEASE wasn't friendly to -CURRENT.
I've got just a basic install with my custom kernel. I'm using the
packages
for X from the 5.1-RELEASE cd running twm. Hangs on restart are gone!
I restarted about 10 times in a row and ran glxinfo and glxgears each time
to verify DRI was still activated and working -- no issues. VT switches are
fine (even while running glxgears). The one thing that does not work is
resume from acpiconf -s 4 (disk) -- there is a failure to refresh in X and no
ability *apparently* to switch to a VT; the keystrokes just generate beeps.
Interestingly, the cursor still changed between the different modes when
mousing over the xterm and onto the background. Also, Alt-Cntl-Del did
work just fine.
The only other thing I noticed is that there seems to be a syslog entry for
every instance of running glxgears that reads:
[MP SAFE] drm0
Is this expected behavior? I noticed that same message (in brackets) in
front of each of my disks as they were probed during boot.
Any further info you'd like or more testing I could do to help?
Sean
-----Original Message-----
From: Eric Anholt <eta@lclark.edu>
Sent: Oct 23, 2003 9:09 PM
To: Glenn Johnson <glennpj@charter.net>
Cc: Vulpes Velox <kitbsdlists@HotPOP.com>, freebsd-stable@freebsd.org,
freebsd-current@freebsd.org, Sean_Welch@alum.wofford.org,
welchsm@earthlink.net
Subject: Re: Radeon 7500 w/ DRI locking on restart of X
On Tue, 2003-08-26 at 20:37, Glenn Johnson wrote:> On Tue, Aug 26, 2003 at 01:27:11PM -0700, Eric Anholt wrote:
>
> > On Tue, 2003-08-26 at 18:05, Vulpes Velox wrote:
> >
> > > I had similar problem with a 7200 and OGL. I solved the problem
by
> > > turning off some of the options in the X config.
> > >
> > > On Tue, 26 Aug 2003 12:21:56 -0500 (GMT) Sean Welch
> > > <welchsm@earthlink.net> wrote:
> > >
> > > > Is anyone else seeing this issue? I'm running into it
on desktop
> > > > boxes and a laptop running 4.8-RELEASE with up to date ports
> > > > collections and various versions of DRI installed over a
ports
> > > > version of X. I'm also seeing this under 5.1-RELEASE on
the
> > > > laptop.
> > > >
> > > > Everything works perfectly unless/until I restart the X
server.
> > > > This appears to be initiated automatically when running GDM
-- ie,
> > > > GDM starts, you log in using that X session, you log out and
the
> > > > session stops, GDM starts X again and displays the login
screen.
> >
> > Everyone that's experiencing this and is using the DRI, what
version
> > of the radeon DRM is loaded? (dmesg | grep drm) Is anyone experiencing
> > this without the DRI loaded? The ForcePCIMode workaround is
> > interesting, I'll take a look at what could be going on there.
>
> I did some googling tonight and found out this problem is supposedly
> fixed in XFree86-4.3.99 although I do not see any specific mention of
> this problem in the Changelog. See:
>
> http://www.knoppix.net/forum/viewtopic.php?t=2504&highlight
That patch has been in our XFree86 for quite a while. For those of you
using -current, could you try with the latest DRM which I committed to
FreeBSD CVS a few minutes ago?
--
Eric Anholt eta@lclark.edu
http://people.freebsd.org/~anholt/ anholt@FreeBSD.org
On Fri, 2003-10-24 at 11:06, Sean Welch wrote:> Eric, I updated my 5.1-RELEASE system to CURRENT dated today at > approx. 9:10 CDT to give your changes a try. I had a bit of a fright at > first with kernel panics right at the end of the boot sequence but it turned > out I had forgotten to disable the ltmdm code -- the kernel module > compiled under -RELEASE wasn't friendly to -CURRENT. > > I've got just a basic install with my custom kernel. I'm using the packages > for X from the 5.1-RELEASE cd running twm. Hangs on restart are gone! > I restarted about 10 times in a row and ran glxinfo and glxgears each time > to verify DRI was still activated and working -- no issues. VT switches are > fine (even while running glxgears). The one thing that does not work is > resume from acpiconf -s 4 (disk) -- there is a failure to refresh in X and no > ability *apparently* to switch to a VT; the keystrokes just generate beeps. > Interestingly, the cursor still changed between the different modes when > mousing over the xterm and onto the background. Also, Alt-Cntl-Del did > work just fine.Suspend/resume is doesn't work well. You might get lucky with the -current DRM with XFree86-4-Server-snap.> The only other thing I noticed is that there seems to be a syslog entry for > every instance of running glxgears that reads: > > [MP SAFE] drm0 > > Is this expected behavior? I noticed that same message (in brackets) in > front of each of my disks as they were probed during boot.That printout happens whenever an mpsafe interrupt handler is installed. It gets installed upon initializing the DRI in the server. I don't think any drivers do it more often. It may happen frequently if glxgears is your only app running, in which case you'll get a server regenerate on exit of glxgears. -- Eric Anholt eta@lclark.edu http://people.freebsd.org/~anholt/ anholt@FreeBSD.org
Okay. I won't worry about the messages then.
I mentioned the suspend/resume problem only because it
actually worked correctly when using option ForcePCIMode;
I was shocked when it did and was hoping that your fix for
AGP accelerated DRI would pull off the same trick.
I should note here that I didn't see any of the panics other
people did before your last iteration of the fix. I'm guessing
it is because I didn't actually compile the code into my kernel
but instead let it build new modules. Any idea why that would
make the difference?
Sean
-----Original Message-----
From: Eric Anholt <eta@lclark.edu>
Sent: Oct 27, 2003 11:48 PM
To: Sean Welch <Sean_Welch@alum.wofford.org>
Cc: freebsd-stable@freebsd.org, freebsd-current@freebsd.org
Subject: Re: Radeon 7500 w/ DRI locking on restart of X
On Fri, 2003-10-24 at 11:06, Sean Welch wrote:> Eric, I updated my 5.1-RELEASE system to CURRENT dated today at
> approx. 9:10 CDT to give your changes a try. I had a bit of a fright at
> first with kernel panics right at the end of the boot sequence but it
turned
> out I had forgotten to disable the ltmdm code -- the kernel module
> compiled under -RELEASE wasn't friendly to -CURRENT.
>
> I've got just a basic install with my custom kernel. I'm using the
packages
> for X from the 5.1-RELEASE cd running twm. Hangs on restart are gone!
> I restarted about 10 times in a row and ran glxinfo and glxgears each time
> to verify DRI was still activated and working -- no issues. VT switches
are
> fine (even while running glxgears). The one thing that does not work is
> resume from acpiconf -s 4 (disk) -- there is a failure to refresh in X and
no
> ability *apparently* to switch to a VT; the keystrokes just generate beeps.
> Interestingly, the cursor still changed between the different modes when
> mousing over the xterm and onto the background. Also, Alt-Cntl-Del did
> work just fine.
Suspend/resume is doesn't work well. You might get lucky with the
-current DRM with XFree86-4-Server-snap.
> The only other thing I noticed is that there seems to be a syslog entry for
> every instance of running glxgears that reads:
>
> [MP SAFE] drm0
>
> Is this expected behavior? I noticed that same message (in brackets) in
> front of each of my disks as they were probed during boot.
That printout happens whenever an mpsafe interrupt handler is
installed. It gets installed upon initializing the DRI in the server.
I don't think any drivers do it more often. It may happen frequently if
glxgears is your only app running, in which case you'll get a server
regenerate on exit of glxgears.
--
Eric Anholt eta@lclark.edu
http://people.freebsd.org/~anholt/ anholt@FreeBSD.org