Bug#253088: radeonfb corruption when killing X from outside X.
On Fri, Dec 10, 2004 at 03:52:47AM -0500, Branden Robinson wrote:
> tag 253088 = help upstream moerinfo
> thanks
>
> On Thu, Nov 11, 2004 at 12:41:35PM +0100, Sven Luther wrote:
> > I am seing this also, sortof, when i VT switch away from X, and then skill or
> > stop X in some way, the radeonfb text console is corrupted. When i restart X
> > blindly, then the console is ok again. When i kill X from inside X, then there
> > is no problem.
>
> What exact video card are you using, according to the XFree86 log file?
$ lspci -s 01:08
0001:01:08.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01)
0001:01:08.1 Display controller: ATI Technologies Inc RV280 [Radeon 9200] (Secondary) (rev 01)
$ lspci -n -s 01:08
0001:01:08.0 0300: 1002:5961 (rev 01)
0001:01:08.1 0380: 1002:5941 (rev 01)
(II) Primary Device is: PCI 01:08:0
(--) Assigning device section with no busID to primary device
(WW) RADEON: No matching Device section for instance (BusID PCI:1:8:1) found
(--) Chipset ATI Radeon 9200 5961 (AGP) found
> > I may have misunderstood the dontswitchvt option, but i feel that forbidding
> > VT switching is no real option here, but then the option probably does
> > something more subtle :).
>
> The idea there was to propose a crude workaround as a diagnostic tool, not
> claim the bug didn't exist because the option was available.
Ok.
> > Anyway, i am at your disposition if you need this bug investigated, and may
> > look into this myself also as time permits.
>
> Basically I need help finding out what VT restoration logic is being used,
> and why the bug submitter sees his text VTs corrected if he shuts down X
> from the same VT that started it, but not otherwise.
Ah.
Friendly,
Sven Luther
Reply to: