Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 29 Dec 2004 21:33:03 -0800
From:      Eric Anholt <eta@lclark.edu>
To:        Dave Horsfall <dave@horsfall.org>
Cc:        freebsd-x11@freebsd.org
Subject:   Re: Upgrade to 5.3-STABLE broke X?
Message-ID:  <1104384783.854.62.camel@leguin>
In-Reply-To: <Pine.BSI.4.61.0412301543530.22225@dave.horsfall.org>
References:  <Pine.BSI.4.61.0412301543530.22225@dave.horsfall.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 2004-12-30 at 16:11 +1100, Dave Horsfall wrote:
> (Originally posted to -questions, and someone said this was a known issue 
> and to post here instead.  Edited version, as I've learned a bit more 
> since the original post.)
> 
> Was running 5.3-RELEASE (from CD) on R31 Thinkpad, where Xorg worked just 
> fine.
> 
> I decided to give -STABLE a spin, followed by a "portupgrade".  Now, it's 
> unclear at what point the breakage happened, because one followed the 
> other straight away without trying X (I know, a big mistake).
> 
> Anyway, after what seemed like an eternity (and probably maxing out my 
> ADSL link in the process) everything finished, and I fired up "X" (X.org, 
> which I had originally, so no issues with XFree86 etc as mentioned in 
> UPDATING).  This is when I discovered that a) 6.7.0 got itself upgraded to 
> 6.8.1 when I was asleep, and b) it didn't work.
> 
> Obscure error messages:
> 
> (EE) I810(0): Failed to allocate HW (ARGB) cursor space.
> (EE) I810(0): [dri] DRIScreenInit failed. Disabling DRI.
> Fatal server error:
> Caught signal 11.  Server aborting
> 
> The first message seemed to be because hardware cursors are not supported 
> on the i830M, and the second because DRI wanted the first. I disabled 
> both, but I still get a SIGSEGV.  Running GDB on the coredump (with and 
> without intercepting signals) shows a corrupted stack.
> 
> This used to work in 6.7.0, so what broke?  Note that I can't even run 
> xorgcfg either, for the same reason.

I'm not sure yet.  I was assuming it was some regression in 6.8.0/1
which would have been fixed in 6.8.2, so I made a patch to pull in the
updates, http://people.freebsd.org/~anholt/X/patch-i810-6_8 which would
be dropped in x11-server/xorg-server/files/ before you rebuild/install
it.  The associated PR is
http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/75425 and has
information on other workarounds.

If you'd like to contribute to the debugging process, please note what
hardware you're seeing the issue on in a followup to ports/75425, and
try out patch-i810-6_8 to see if it helps.  One person has said it
didn't help them, but I'd like to see two before I give up on this
angle.

I'm also concerned about the V_BIOS warnings I've seen, which have meant
AGP issues in the past, though it seems odd that it would crop up now.
Getting people to say what hardware they're seeing the issue on
(something that's been rather lacking until now) could help a lot here.

> I poked around wiki.x.org hoping to get release notes etc, to no avail - 
> the links are broken.  It also offered the advice of running "xorgcfg" 
> (see above), or sending mail to xorg@freedesktop.org (which turned out to 
> be a mailing list to which one has to subscribe; it could at least say 
> so).

http://bugs.freedesktop.org/ would be the place to complain about
that :)

-- 
Eric Anholt                                eta@lclark.edu          
http://people.freebsd.org/~anholt/         anholt@FreeBSD.org



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1104384783.854.62.camel>