From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020529 Description of problem: Notebook/laptop screen detection is frustrating. On Dell Inspiron 4000 notebook laptop, the installation process does not identify the screen, the user must guess [ordinary VGA for instance], firstboot fails "no screens", the recovery diagnostics are hard to interpret [which choice gets the right answer, and why wasn't that choice selected automatically before?], and the only sure way to fix it is to hand edit /etc/X11/XF86Config. Anaconda knows hundreds of obscure CRT monitors, yet a popular laptop cannot be installed successfully because "ddc probe failed". However, XFree86 successfully determines the model number of the screen (IBM ITXG76E2) with its ddc prober [with no help from /etc/X11/XF86Config]. Please, anaconda, learn as much as XFree86 about laptop screen detection, and the proper Horiz/Vert frequencies for the 100 most popular screens. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.Fresh install Phoebe on Dell Inspiron 4000 notebook/laptop. 2.Accept default bog-standard VGA when screen cannot be identified during anaconda's X11 setup. 3.Boot. Actual Results: firstboot fails "no screens". Recovery dialogs make no sense; in particular, the default answer just goes into a loop. /var/log/XFree86.0.log shows that XFree86 knows the panel model number: ----- (--) R128(0): Chipset: "ATI Rage 128 Mobility LF (AGP)" (ChipID = 0x4c46) (--) R128(0): Linear framebuffer at 0xe8000000 (--) R128(0): MMIO registers at 0xfdffc000 (II) R128(0): Video RAM override, using 8192 kB instead of 8192 kB (**) R128(0): VideoRAM: 8192 kByte (128-bit SDR SGRAM 1:1) (**) R128(0): Using flat panel for display (II) R128(0): Panel size: 1024x768 (II) R128(0): Panel ID: IBM ITXG76E2 (II) R128(0): Panel Type: Color, Single, TFT (II) R128(0): Panel Interface: LVDS (II) R128(0): PLL parameters: rf=2700 rd=60 min=12500 max=25000; xclk=10500 (II) Loading sub module "ddc" (II) LoadModule: "ddc" (II) Loading /usr/X11R6/lib/modules/libddc.a (II) Module ddc: vendor="The XFree86 Project" compiled for 4.2.99.2, module version = 1.0.0 ABI class: XFree86 Video Driver, version 0.6 ----- Expected Results: screen detected and Horiz/Vert frequencies selected correctly during installation. Additional info:
If I had a Dell Inspiron 4000 to test with, I might have a chance of fixing this bug. But sadly we get no desktop/laptop samples from Dell, so it's quite difficult to sort out issues that are hardware specific without access to the hardware. To be honest, the lack of hardware is maddening.
I'm willing to help, even in strange and wonderful ways. (This shouldn't be at the top of your list [it's not at the top of mine, either], but it might be productive and fun.) For instance, I'll use dosemu to run something that you provide, and send you the results. Or, we ought to be able to setup so that you can use a shell running on the Dell under Phoebe, with X11 display+keyboard+mouse across the 'net to your desktop. In other words: think BIG, and we just might be able to provide you enough of an environment to make signficant progress.
Take a look at bug #80402. He has an Inspiron 4000, and I made some changes that made detection work on his machine. Try redhat-config-xfree86-0.7.1-3 out of Rawhide and see if that works for you.
I obtained redhat-config-xfree86-0.7.1-3.noarch.rpm from rpmfind.net ("Rawhide for i386, 3 days old or less"; ftp directly from redhat was impossibly slow), and did "rpm --upgrade" on the already-installed Phoebe on Dell Inspiron 4000. Then I got back to single-user mode text console (from default GNOME2 graphical desktop) via "telinit 3", and ran /usr/bin/redhat-config-xfree86 as root. All four of the probe results were "none". The Xserver started, and the graphical dialog had "Unprobed monitor" with correct ATI Rage Mobility r128 card. The H+V frequncies were those of the old existing /etc/X11/XFree86Config file. Retaining the old frequencies is an improvement over some versions within the last year that just reverted to bog-standard VGA (ignoring the old XFree86Config) when monitor could not be probed. The only difference in the resulting XFree86Config file was the comment saying that redhat-config-xfree86 was the author instead of pyxconf. Then I deleted the XFree86Config file and ran redhat-config-xfree86 again, hoping to simulate some aspects of a fresh install. This time two consecutive attempts to start an Xserver failed, no graphics ever appeared, and it exited to the shell with no XFree86Config file being written. So this looks to me like a failure to detect the display panel, and a failure to recover (not even VGA). Then I restored my backup of XFree86Config, and ran again. This time I went to the Advanced tab, selected Dell 1024X Laptop display panel, and tried to OK, but it hung. The pointer moved, but clicking anywhere, including the OK and Cancel buttons for both the upper and lower dialogs, had no effect. I recovered by Ctrl-Alt-F2 to a text console, then "kill -TERM" to the xconf.py process. This looks to me like another failure. So, I would rate this as not much progress. Maybe I didn't start in the right place?
Try upgrading to rhpl-0.77-1 and redhat-config-xfree86-0.7.2-2 out of Rawhide and see if that changes anything. Some code was added th rhpl over the weekend that should improve detection of laptop LCDs that can't be probed via ddcprobe. Is the behavior any better?
Just now I retrieved rhpl-0.77-1 from ftp://rawhide.redhat.com/pub/redhat/linux/rawhide/i386/RedHat/RPMS/ but the latest redhat-config-xfree86 was redhat-config-xfree86-0.7.1-6.noarch.rpm 69 KB 01/17/2003 06:48:00 PM which lags the 0.7.2-2 that you cite. Neither rpmfind.net nor www.alltheweb.com could find 0.7.2-2. Is there a different place to look, or should I just wait until tomorrow and look again?
I think wait until tomorrow.
I installed these 3 .rpms from today's Rawhide: redhat-config-mouse-1.0.3-4.noarch.rpm redhat-config-xfree86-0.7.2-2.noarch.rpm rhpl-0.78-1.i386.rpm Then I removed /etc/X11/XF86Config and ran redhat-config-xfree86. First I got syntax error because /usr/share/redhat-config-xfree86/xconf.py line 356 was missing the closing right parenthesis ')'. I added that and reran: ----- # redhat-config-xfree86 * ddcprobe returned bogus values: ID: None Name: None HorizSync: None VertSync: None Couldn't start X server, trying with a fresh configuration * ddcprobe returned bogus values: ID: None Name: None HorizSync: None VertSync: None Traceback (most recent call last): File "/usr/share/redhat-config-xfree86/xconf.py", line 356, in ? print _("Trying with card:", current_card.getCardData()["NAME"]) TypeError: gettext() takes exactly 2 arguments (3 given) [root@host X11]# ----- So I consider this "no progress" because it doesn't generate XF86Config at all.
Oops. I forgot to put the closing parenthesis on that line, and I think you put it in the wrong place. It should have looked like: print _("Trying with card:"), current_card.getCardData()["NAME"] Please try with redhat-config-xfree86-0.7.2-3. Sorry about the mistake.
redhat-config-xfree86-0.7.2-3 still does not produce a XF86Config: ----- # cd /etc/X11; cp XF86Config XF86Config.bak # rm XF86Config # redhat-config-xfree86 * ddcprobe returned bogus values: ID: None Name: None HorizSync: None VertSync: None Couldn't start X server, trying with a fresh configuration * ddcprobe returned bogus values: ID: None Name: None HorizSync: None VertSync: None Trying with card: ATI Rage 128 Mobility Error, cannot start X server. # -----
Try running 'redhat-config-xfree86 --set-hsync="31.5-90.0" --set-vsync="59.0-75.0"' Does that work? This problem happens because we can't probe the LCD. This also seems to affect the Inspiron 8000 (bug #81436).
Yes, probably that works. Exact methodology: fresh install original Phobe .isos using TEXT install. Change unprobed monitor freqs during install to ranges indicated above. Boot with extra parameter "single". rpm -Uvh with files: redhat-config-mouse-1.0.4-3.noarch.rpm redhat-config-xfree86-0.7.3-2.noarch.rpm redhat-artwork-0.65-1.i386.rpm Exit single-user mode, fall into firstboot, it works.
I suspect that this bug is a dupe of bug #81436. If you install using the graphical installer, what monitor is selected in the monitor selection screen? If that monitor isn't correct, you should be able to select you monitor from the list. If your monitor isn't available in the list, you should be able to click on the "Generic" tab and then select "Generic Laptop Display Panel 1024x768". Does that allow firstboot to come up correctly?
Default graphical install of Workstation selects "Unprobed monitor > Unprobed monitor" [two lines, indented, with same text]. Selecting "Dell > Dell 1024X Laptop Display Panel" gives Horiz 31.5-48.5 KHz and Vert 59.0-75.0 Hz; this worked for firstboot. There is a small amount of confusion because "1024X" is not a model number, but instead means "native resolution of 1024 pixels in horizontal direction." Selecting "Generic > Generic Laptop Display Panel 1024 x 768" gives Horiz 31.5-48.5 KHz and Vert 40-70 Hz; this would work because XFree86 [on "Dell 1024X" selection] chooses a modeline that gives 60Hz refresh.
Ok, I've modified the entries in MonitorsDB so that they read: Dell; Dell 1024x768 Laptop Display Panel; 0; 31.5-48.5; 59.0-75.0; 1 instead of: Dell; Dell 1024X Laptop Display Panel; 0; 31.5-48.5; 59.0-75.0; 1 That should resolve the confusion about which item to pick from the monitor list. This change should be in hwdata-0.74-1. I think that's about all I can do for this release. For users with these laptop screens, we're going to need them to pick the correct monitor during install or they're going to have problems. We can't probe the monitor correctly because the screen does not provide the necessary ddcprobe information. However, if bug #81436 gets fixed so that the r128 driver allows 800x600 resolutions to work, then this shouldn't matter. Please open a new bug report against gpm for the "-R" behavior causing problems with X.
Nevermind my comments about gpm. I was looking at the wrong bug report. Sorry.
Thank you. The changed entry of Dell; Dell 1024x768 Laptop Display Panel; 0; 31.5-48.5; 59.0-75.0; 1 helps significantly.
There is a stack of 64 bugs that have been in Modified state for a long period of time. I am closing these as Rawhide now. If you find that the issue is not fixed, please reopen this report.