Tecra/Protege stops booting at Xvesa and Xorg (SOLVED)

Please post any bugs you have found
Message
Author
User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

#21 Post by rerwin »

The 8514 was an early IBM display type, which apparently was 1024x768. It is included in what you filter out, but is the only 1024x768 entry the problem adapters put out to ddcprobe. So the simple fix is to not filter it out.

But there is no 800x600 "timing:" entry from the MagicGraph adapter, even though one of the Porteges maxes at 800x600. So that needs to be added the way I did it. But there is another approach we might take.

My ThinkPad that has a similar adapter yields no "timing:" entries. The consequence seems to be that the wizard concludes it cannot determine the appropriate resolutions, so it presents a list of all possibilities. Maybe that is how we should handle the problem cases, instead of trying to fill in the gaps.

As far as contacting the Xorg/Xvesa developers, I have assumed that the Puppy regulars would have more clout with them than some stranger (me). You would certainly be able to speak with them more credibly.

(I am away from home so I can't get deeply into things this week. I can try the developers next week, if necessary.)

dalderton
Posts: 177
Joined: Sun 22 Apr 2007, 08:33

#22 Post by dalderton »

Richard I worked out how to do the remaster properly,there is a /tmp/etc/ X11 that i had to put the problem-adapters in before burning the remastered CD.I did that successfully but it just hung as the original Puppy unleashed did after the flash of probe hardware.nothing then the probe hardware message and a cursor. I tried the various workarounds but nothing worked. I realise that you are away so I will keep an eye on this thread . My problem is solved but it would be nice if it happened as a matter of course from the original CD Thanks again Regards Dennis

User avatar
Dougal
Posts: 2502
Joined: Wed 19 Oct 2005, 13:06
Location: Hell more grotesque than any medieval woodcut

#23 Post by Dougal »

rerwin wrote:My ThinkPad that has a similar adapter yields no "timing:" entries. The consequence seems to be that the wizard concludes it cannot determine the appropriate resolutions, so it presents a list of all possibilities. Maybe that is how we should handle the problem cases, instead of trying to fill in the gaps.
You mean go for the "your screen could not be detected" dialog? It could be worth trying -- I'll have to look into it.
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind

DaveR
Posts: 28
Joined: Wed 30 May 2007, 15:08

problem with Satellite 1800/1805

#24 Post by DaveR »

I have a similar problem.

After reading these posts I tried this:
F2 then choose C to boot from CD
command "puppy pfix=nox acpi=off"
(because allowing normal boot hangs system as below)
all OK to #_ prompt
(pen drive is sda1)
copied problem-adapters > /etc/X11
copied xorgwizard-tecra-trident > /usr/sbin
ran ddcprobe to get line "oem:Trident CYBER 8620"
added that to "problem-adapters"
ran xorgwizard-tecra-trident
one line of pink? text appears and disappears too rapidly to read
system hangs - [ctrl]c will NOT break out
[alt]f2 will not open text mode
[ctrl][alt]del will NOT break out
cycle power

Any suggestions would be most welcome.

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

#25 Post by rerwin »

Sorry to hear dalderton and DaveR are having problems with the fix.

Dennis, since the wizard looks for the problem-adapters file in /etc/X11, you need to make sure it is there before you run the wizard. You can do that with your remaster by booting up to the "probing-hang", then pressing ctrl-c to get the prompt, then entering:
cp /tmp/etc/X11/problem-adapters /etc/X11/

Then run xorgwizard-tecra-trident (or whatever you might have renamed it to).

If it still doesn't work, try booting with pfix=ram,nox

Dave, thanks for trying this fix script. Your description sounds like you did what is needed. But verify that the problem-adapters file looks right following your addition of your Trident adapter. The new text must be on its own line, not an extension of an existing line. I expect the file to read (excluding the initial comment line):

Code: Select all

oem: MagicGraph 256 AV 44K PRELIMINARY
oem:Trident Cyber 8820
oem:Trident CYBER 8620
Be sure you typed it exactly as ddcprobe shows it, respecting case.

If it still doesn't work, and if you did an "enter" at the end of the last line of problem-adapters, try removing the line-feed by backspacing from the end of the file until the I-bar cursor follows the "8620" and is not on its own line.

BTW, When I get the hang, I always see the pink "probing" line, which then disappears and then reappears with the blinking cursor below it. When the fix works, I see a very brief display of a line I assume to be the initial "probing" line, then the "probing" line a bit longer, then the dialog for choosing a video mode. How much of that actually happens for you?

If you can, would you post the output from ddcprobe? At least the "timing:" lines. Thanks.

It is possible that the problem with your adapter is something else than the Xvesa call in the script. But let's be sure we have the fix correct before giving up.

youb
Posts: 6
Joined: Fri 16 Nov 2007, 02:40
Location: France

#26 Post by youb »

Hi,
I'm new here, and want to try Puppy Linux 3.01 on my Laptop : Toshiba Portege 3110ct with Trident Cyber9525DVD graphic adapter. The screen can show 8000x600 16bit max if I'm not mistaken. I'm using loadlin to run Linux from a win98 partition.

I have the same problem, and apparently the fix doesn't work (I think I'm doing all this correctly, I'm a newbie but followed the instructions...)
I also tried other proposed fixes on this forum with no luck.

I follow the instructions here, hit Ctrl+c to get to prompt, then copy the files and run the script. It shows the pink "probing" message, then hangs with a nice black screen... I tried copying the line from ddcprobe in the problem-adapter file with no success.

I noticed something strange when I run ddcprobe, it shows me :
vbe: VESA 2.0 detected
oem: Copyright 1997 TRIDENT MICROSYSTEMS INC. (the line I copied)
memory: 2560kb
mode: 640x480x16m
... lists other modes up to ...
mode: 640x480x256
Segmentation fault


This last line seems a bit funky to me... but i'm mot an expert.
Any help would be great ! I tried to give as much info here... please tell me if you need more.

Mat

dalderton
Posts: 177
Joined: Sun 22 Apr 2007, 08:33

#27 Post by dalderton »

Richard I have no problem getting the program to run it was just that I was trying to make a CD that would load to my laptop as a normal Puppy CD does.I can modify it when it hangs after your original help but I was trying to make it automatic.So what I thought was that if I used your xorgwizard file and added the problem-adapters file to /etc/X11 the above automatic load would occur.This did not happen.However Puppy 3.0.1 is happily running on my hard drive in the laptop from your original fix for me. Regards Dennis

DaveR
Posts: 28
Joined: Wed 30 May 2007, 15:08

Toshiba Satellite 1805-S203

#28 Post by DaveR »

Thanks for the quick come-back rewin.

This time I let 'mp' copy the line from the output of 'ddcprobe' and paste it into my new '/etc/X11/problem-adapters' file. I verified that there was no line feed at the end of the line.

ran 'xorgwizard-tecra-trident' - this time I could see that the pink text said something about probing video. After that line flashes on the screen it remains completely black without anything displayed. Same result as before - complete lock-up.

FYI there are at least 2 other threads involving the same problem - in the How To section (http://www.murga-linux.com/puppy/viewtopic.php?t=23050) and in the House Training Bugs section (http://www.murga-linux.com/puppy/viewtopic.php?t=22873)

Here is the output from 'ddcprobe':
vbe: VESA 2.0 detected.
oem: Trident CYBER 8620
memory: 8192kb
mode: 1280x1024x64k
mode: 1280x1024x32k
mode: 1024x768x16m
mode: 1280x1024x256
mode: 640x480x16m
mode: 800x600x16m
mode: 1024x768x32k
mode: 1024x768x64k
mode: 800x600x32k
mode: 800x600x64k
mode: 640x480x32k
mode: 640x480x64k
mode: 1024x768x256
mode: 1280x1024x16
mode: 320x200x32k
mode: 320x200x64k
mode: 320x200x16m
mode: 640x400x256
mode: 640x480x256
mode: 800x600x256
mode: 1024x768x16
mode: 132x25 (text)
mode: 132x43 (text)
mode: 132x60 (text)
mode: 80x60 (text)
mode: 800x600x16
edid:
edid: 1 3
id: 5082
eisa: TOS5082
serial: 00000000
manufacture: 0 1990
input: analog signal.
screensize: 27 21
gamma: 1.000000
dpms: RGB, no active off, suspend, standby
timing: 720x400@70 Hz (VGA 640x400, IBM)
timing: 640x480@75 Hz (VESA)
timing: 800x600@72 Hz (VESA)
timing: 1024x768@87 Hz Interlaced (8514A)
ctiming: 1280x1024@60
ctiming: 1600x1200@60
dtiming: 1024x768@0
monitorname: TOSHIBA Inte
monitorname: rnal 1024x76
monitorname: 8 Panel

==================================

For what it's worth these are the relevant lines from Ubuntu 7.10's xorg.conf:
Section "Device"
Identifier "Trident Microsystems CyberBlade/i1"
Driver "trident"
BusID "PCI:1:0:0"
EndSection

Section "Monitor"
Identifier "Generic Monitor"
Option "DPMS"
HorizSync 28-51
VertRefresh 43-60
EndSection

Section "Screen"
Identifier "Default Screen"
Device "Trident Microsystems CyberBlade/i1"
Monitor "Generic Monitor"
DefaultDepth 24
SubSection "Display"
Depth 1
Modes "1024x768"
EndSubSection
SubSection "Display"
Depth 4
Modes "1024x768"
EndSubSection
SubSection "Display"
Depth 8
Modes "1024x768"
EndSubSection
SubSection "Display"
Depth 15
Modes "1024x768"
EndSubSection
SubSection "Display"
Depth 16
Modes "1024x768"
EndSubSection
SubSection "Display"
Depth 24
Modes "1024x768"
EndSubSection
EndSection

dalderton
Posts: 177
Joined: Sun 22 Apr 2007, 08:33

#29 Post by dalderton »

Richard: I have got the problem organized. if I use your xorgwizard-tecra file with the modified if grep -q 'oem' line (479 ) and use it in the remastered cd after "chmod a+q xorgwizard-tecra" and burn it the cd then operates as you said it should i.e. gets to the pink probe line flashes then a pause then the line comes back momentarily then the choice of resolutions with 1024*768*16 highlighted and proceeds normally.Bingo
This means that the CD can be used on my Toshiba TE 2000 Laptop as a live CD or loaded to the hard disc without any glitches. I wonder if other laptops with similar problems can use this system without further modifications from that original fix.Is the problem-adapters file really necessary?Mine is fine without it and my trident bit is not mentioned on my xorgwizard. Anyway all is well with me Thank You Regards Dennis
EDIT This fix should work for DaveR as his adapter is apparently the same as mine D

DaveR
Posts: 28
Joined: Wed 30 May 2007, 15:08

Satellite 1805S203

#30 Post by DaveR »

dalderton, thanks for thinking about my problem but I don't understand your post. Is the "if [..." line quoted below what you want me to modify? How should it read?

#For some Tecra (8000) and Portege laptops, avoid Xvesa hang by substituting an Xmodes list known to work; also enable 800x600 and 1024x768 by adding "timing: .. (VESA)" lines for them in the ddcprobe listing. #rerwin
if [ -f /etc/X11/problem-adapters ] && grep -q -f /etc/X11/problem-adapters /tmp/ddcprobe.txt;then
echo -e "640x480x16\n640x480x24\n800x600x16\n800x600x24\n1024x768x16\n1024x768x24" > /tmp/Xmodes.txt

Dave

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

New version of fixed xorgwizard

#31 Post by rerwin »

Dennis & Dave, please do not proagate the workaround that gets Dennis' laptop to work; it is not appropriate for general use because all video adapters would be treated the same.

I am attaching a new version of the wizard that should fix everyone's problem that has been reported. Instead of trying to substitute for the failing Xvesa results, it is consistent with the current design that lets the user choose a video mode if the wizard cannot determine it. (That's how my ThinkPads are handled, even though they do not hang like the Tecras/Porteges.) The "problem-adapters" file is no longer used, so the "remaster mess" is avoided.

In addition to checking for specific adapters, it looks for segment faults and failure to find the needed info from ddcprobe; that should take care of Mat's case. For Dennis, I included two tests for the 8820 because the "oem:" is always followed by a space; Dennis, please check your ddcprobe output and tell me which format is correct (with or without the space).

Please, would each of you try the new version, xorgwizard2, and give me feedback. Thanks.

(After copying/extracting xorgwizard2 to /usr/sbin, rename it to xorgwizard to have it run during bootup. Otherwise, when bootup hangs, ctrl-c out of the hang and enter xorgwizard2, then xwin.)
Attachments
xorgwizard2.tar.gz
File contains both a normal version, xorgwizard2, and a debug version for troubleshooting, xorgwizard2d.
Expect to see a long list of possible video modes to choose from..
(32.5 KiB) Downloaded 960 times
Last edited by rerwin on Mon 19 Nov 2007, 16:39, edited 5 times in total.

DaveR
Posts: 28
Joined: Wed 30 May 2007, 15:08

Satellite 1805 S203

#32 Post by DaveR »

rewin, before I forget it my "oem:" is followed by 1 space - oem: Trident CYBER 8620
I'll give the new wizard a try in a little while
Dave

DaveR
Posts: 28
Joined: Wed 30 May 2007, 15:08

Satellite 1805 S203

#33 Post by DaveR »

Still no luck ;-(

edited the problem-adapters file to include my "oem: Trident CYBER 8620" line
boot from CD
command "puppy pfix=nox"
all OK to #_ prompt
plug-in pen drive then mount /dev/sda1 /mnt/flash
ls /mnt/flash
cp /mnt/flash/proble~2 /etc/X11/problem-adapters
cp /mnt/flash/xorgwi~2 /usr/sbin/xorgwizard2
ran xorgwizard2
"Puppy Video Wizard" window opens - choose Xorg - Enter
one line of pink text appears "Probing video hardware ..."
system hangs with empty black screen and no cursor - wait a minute
[ctrl]c will NOT break out
[ctrl][alt]del will NOT break out
cycle power to quit

youb
Posts: 6
Joined: Fri 16 Nov 2007, 02:40
Location: France

#34 Post by youb »

Thanks for your time on that problem... no luck either..
The new script his behaving the same on my laptop as on Dave's .
Pink "probing", then hangs with black screen.
Ctrl+Alt+Del produces a little reaction from my HDD activity led...

dalderton
Posts: 177
Joined: Sun 22 Apr 2007, 08:33

#35 Post by dalderton »

Richard I tried it on mine and it did the same hang as the original xorgwizard. Regards Dennis

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

Let's work with the latest version -- xorgwizard2

#36 Post by rerwin »

I am sorry you are having trouble testing this fix. So let's all work with the latest version, which is simpler than the others, so less room for errors.

First, I'll explain the logic -- what it checks for -- and then how I test, step by step. Here is the code that should run for your cases (EDIT (11/19/07): corrected to match latest version):

Code: Select all

#For some Tecra and Portege laptops (and any others generating a segment fault or edidfail in ddcprobe), avoid Xvesa/other hang by skipping detection and letting the user choose the appropriate video mode.   #rerwin begin
if grep -q -F -x 'edidfail
oem: MagicGraph 256 AV 44K PRELIMINARY
oem: Trident CYBER 8820
oem: Trident CYBER 8620' /tmp/ddcprobe.txt;then
 HORIZSYNC=""
 VERTREFRESH=""
else   #rerwin end
. . .  [logic that hangs]
fi   #rerwin
. . .
#if ddcprobe also failed, have to ask monitor type...
if [ "$HORIZSYNC" = "" -o "$VERTREFRESH" = "" ];then #"" if monitor not plug-n-play.
[Displays the long list of all possible safe video modes (A through Z).]
The "grep" statement includes all of the strings (between the ticks) it considers indicative of a problem adapter. Please be sure that your ddcprobe output contains one of those exact strings; I took them from your postings, so please tell me of any typos that crept in.

Given that the code is so simple, if it finds a match it will display the dialog. Therefore the hanging suggests that it is not finding a match.

Now, here are my test steps, avoiding any complications, on my Tecra8000 that has no HD:
1. Extract xorgwizard2 from xorgwizard2.tar.gz to the top level of a USB flash drive, by Menu>Utilities>Xarchive archiver, on some other puppy system.

2. In the subject laptop, insert the flash drive, boot up from the Puppy 3.01 CD, and enter "puppy pfix=ram,nox" at the boot prompt.

3. Select a keyboard layout and <enter>.

4. Enter "mount /dev/sda1 /mnt/flash -t vfat" to mount the flash drive and use long filenames (-t vfat).

5. Enter "cp /mnt/flash/xorgwizard2 /usr/sbin/" to copy the script.

6. Enter "umount /mnt/flash" to unmount the flash drive and avoid any problems caused by its being already mounted.

7. Enter "xorgwizard2" to run the script.

8. Select "Xorg" in the resultant dialog.

9. Wait for the two pink "Probing" lines to come and go -- The wizard dialog beginning with "Automatic probing ... was unsuccessful" should be displayed.

10. Select the second (B) line if your screen resolution is 800x600; the third (C) line if 1024x768, then OK, and OK again.

11. At the resultant prompt, enter "xwin" to start the desktop.
If it still hangs, boot again the same way but at the after-bootup prompt (before step 4 above) enter "ddcprobe" and find the line we are trying to match, to be sure the text is exactly what the script is looking for. For youb's case I rely on ddcprobe including a "Segmentation fault" line in its output, since the "oem:" line is too vague -- it is a copyright notice and does not specify a particular adapter model; so be sure I got the upper/lower cases right, for an exact match. To verify there is only the one xorgwizard2 in the system, enter "which xorgwizard2", that should show either nothing or "/usr/sbin"; anything else suggests the wrong file is being run.

If you do get it to work you can rename the script (in /usr/sbin) to "xorgwizard" to have it run automatically at bootup.

I need to verify it works so I can hand it off to dougal for integration into his version and a puppy release. Thanks for sticking with me on this.
Richard
Last edited by rerwin on Mon 19 Nov 2007, 16:40, edited 3 times in total.

dalderton
Posts: 177
Joined: Sun 22 Apr 2007, 08:33

#37 Post by dalderton »

I will sit this one out for a bit as I dont have a flash drive.My option is to replace the original xorgwizard with your modified one and burn it as a remastered live or installable cd.This modified xorgwizard does not work when used in this way. Regards Dennis

youb
Posts: 6
Joined: Fri 16 Nov 2007, 02:40
Location: France

#38 Post by youb »

Added pfix=ram,nox to loadlin's options
Boot linux, at prompt :
mkdir /mnt/hda3
mount /dev/hda3 /mnt/hda3 -t vfat (my C: win drive, where I put the downloaded file)
cp /mnt/hda3/xo/xorgwizard2 /usr/sbin
umount /mnt/hda3
xorgwizard2 -- same problem again.
I checked my ddcprobe output and "Segmentation fault" is written ok.

Isn't it normal since I am using the same xorgwizard2 as yesterday ?

Thanks, Mat

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

I need your assistance; I do not have your laptops to test.

#39 Post by rerwin »

Dennis, please don't sit this out. I cannot resolve the dilemma without you. Your remaster technique should now work with the fix, since the file in /etc/X11 is no longer involved. (Go ahead and rename the script to xorgwizard for your remaster, so it will be invoked automatically.)

I want the fix to be as generally applicable as we can make it, which means getting it to detect your all of particular video adapters by matching something in the ddcprobe output.

Just one thing to make clear, I have built the fix on the Puppy 3.01 version of the wizard; so be sure you are all testing with 3.01.

If testing with 3.01 doesn't work, I will have to add debug code to the fix to try to find where the problem is.

My only other suspicion is that maybe the forum is squeezing some space-characters out of the oem: and segment-fault lines in your posts. So if you notice beginning- or trailing spaces or multiple adjacent spaces in the text we are trying to match, please tell me. (And, Dennis, does your ddcprobe "oem:" line have a space after the colon?)

One more time ... Thanks.
Richard

youb
Posts: 6
Joined: Fri 16 Nov 2007, 02:40
Location: France

#40 Post by youb »

I'm using 3.01, and the "Segmentation fault" is the same in the output and the xorgwizard2 file (I will double check). No problem for further testing... Thanks, Mat
EDIT: I've been browsing through the forum... and I read a post about the minimum amount of RAM needed to run Puppy... I have 64mb, could it be the cause of my problem ? (I use loadlin, therefore if I'm not mistaken it loads everything in RAM and I don't use any swap.)

Post Reply