Red Hat Bugzilla – Bug 15813
cdrom install not starting X on C&T 65555
Last modified: 2007-04-18 12:28:05 EDT
Upgrading from 6.2, the installer starts TUI without asking
(did not happen in 6.2; card is C&T 65555, DDC probe is ok in 6.2, checked now)
This defect is considered MUST-FIX for Winston Gold-release
Did you catch any messages on why it may fail? Do you have a PS/2 mouse?
RC1 but not 6.2: "found video card controller unknown" on Alt-F3, then goes TUI.
That particular message is ok - it means it didn't find a card that requires the
agpgart module to be loaded.
Does Xconfigurator properly setup your X display?
Does it "click in" to X before going to text mode? See any X server messages
To msf: how do I know that? Xconfigurator still hasn't had a chance to come up at this stage in install...
To msw: no X messages, nada, nil, null, zilch. No screen flashes either.
how much system ram do you have? cat /proc/meminfo from VC2
Do you really need the "cat" or is it enough if I tell you? It's a standard Toshiba Satellite 320cds, with 32MB.
Well, I'd like to know tha the kernel is detecting the amount of ram properly...
Ok, checked: kernel detects all 32MB and has 9MB free when the shell on VC-2 is available.
BTW, another test machine I have has 128MBm, the kernel sees only 64MB (due to the VIA chipset, I guess) and X starts up fine on install.
On a booted system, run:
python -c "from kudzu import *; probe(CLASS_VIDEO, BUS_UNSPEC, PROBE_ALL)"
and send the output, please.
python -c "from kudzu import *; print probe(CLASS_VIDEO, BUS_UNSPEC, PROBE_ALL)"
[(None, 'Card:Chips & Technologies CT65555', 'C&T|F65555 HiQVPro')]
Ugh, there is no reason why this shouldn't work. This is a laptop?
Yes, it's a Toshiba Satellite 320cds.
Could you install it in text mode and send me the X configuration it writes out,
as well as any working X configuration you can come up with?
TUI RC1 install: detects card as CT65520, Xconfigurator loops when using custom monitor (so went for similar monitor type), X setup ok.
Created attachment 2919 [details]
Heavily customized XF86Config for 6.2
Created attachment 2920 [details]
X config generated by RC1 installer
Created attachment 2921 [details]
X config generated by me with Xconfigurator in RC1 post-installation checkup
Still there in RC2. After the installation, Xconfigurator generates a valid setup.
Added pbrown because apparently Xconfigurator setup works, but anaconda does
not. Needs to have a unified solution in the future.
This should be fixed with the installer that is currently in CVS. The new installer tries to use the frame buffer if possible, and this will allow the installer to
work on many cards that didn't work in the past. I'm resolving this bug, but please reopen if the problem persists in future versions.
Verified as resolved.