Can't boot Salix -Openbox-14.1 on AMD K6-2 machines

[coyotl]

A little weird perhaps that the UBCD will not boot on this hardware? Would think they used low level code
for this that would boot on anything?
If there is a DOS or Freedos based repair CD and that will not boot either them maybe the hw must be faulty?

http://www.hirensbootcd.org/download/

Hiren's boot CD -if one of the old versions of this could be found and it does not work it must be hw fault.

[junqueman]

A little weird perhaps that the UBCD will not boot on this hardware? Would think they used low level code
for this that would boot on anything?
If there is a DOS or Freedos based repair CD and that will not boot either them maybe the hw must be faulty?

http://www.hirensbootcd.org/download/

Hiren's boot CD -if one of the old versions of this could be found and it does not work it must be hw fault.

I do appreciate the thought, but all other indications are the hardware is functional. I now have 2 optical drives in the AMD K6-III box: a DVD read-only, and a CD R/W. Both drives behave the same for each CD I've tried. (Obviously I can only try to boot DVDs in the DVD drive.) And you've probably seen the list of distros that I've had no trouble booting from CD. A Windows 98 CD boots without a problem as well. Also, the other machine, the AMD K6-2, has exactly the same behavior in terms of which CDs will and will not boot including the UBCD. As westms had said, one might then conclude there is a problem with the CDs. However, they all boot OK on a newer machine with a Pentium 4 CPU, or in a VM on a contemporary Macbook. So from all the distros I've attempted, I've pretty much concluded at this point there is a compatibility issue with certain versions (IIRC 4.06, 4.07, at least) of Isolinux, and also probably certain versions of the kernel, with either the BIOS, or CPUs, or both, in the AMD K6 machines.

I have another data point: I was almost successful in running and installing antiX-16-b5-386-full, the latest (as of a couple of weeks ago, anyway) beta of antiX, from the live CD. I was able to get it to boot to a text-mode installer on the AMD K6-III. (I did not try the K6-2 but I'm quite confident at this point that the results would be the same). After installation it wouldn't boot, but I was able to make edits to /boot/grub/menu.lst so that it would, but again, only into a text/terminal mode. The same live CD booted to the GUI desktop on a VM, and also installed fine in the VM (i.e., boots successfully to the desktop without any edits to menu.lst).

[westms]

A little weird perhaps that the UBCD will not boot on this hardware? Would think they used low level code
for this that would boot on anything?
If there is a DOS or Freedos based repair CD and that will not boot either them maybe the hw must be faulty?

http://www.hirensbootcd.org/download/

Hiren's boot CD -if one of the old versions of this could be found and it does not work it must be hw fault.

I do appreciate the thought, but all other indications are the hardware is functional. I now have 2 optical drives in the AMD K6-III box: a DVD read-only, and a CD R/W. Both drives behave the same for each CD I've tried. (Obviously I can only try to boot DVDs in the DVD drive.) And you've probably seen the list of distros that I've had no trouble booting from CD. A Windows 98 CD boots without a problem as well. Also, the other machine, the AMD K6-2, has exactly the same behavior in terms of which CDs will and will not boot including the UBCD. As westms had said, one might then conclude there is a problem with the CDs. However, they all boot OK on a newer machine with a Pentium 4 CPU, or in a VM on a contemporary Macbook. So from all the distros I've attempted, I've pretty much concluded at this point there is a compatibility issue with certain versions (IIRC 4.06, 4.07, at least) of Isolinux, and also probably certain versions of the kernel, with either the BIOS, or CPUs, or both, in the AMD K6 machines.

I have another data point: I was almost successful in running and installing antiX-16-b5-386-full, the latest (as of a couple of weeks ago, anyway) beta of antiX, from the live CD. I was able to get it to boot to a text-mode installer on the AMD K6-III. (I did not try the K6-2 but I'm quite confident at this point that the results would be the same). After installation it wouldn't boot, but I was able to make edits to /boot/grub/menu.lst so that it would, but again, only into a text/terminal mode. The same live CD booted to the GUI desktop on a VM, and also installed fine in the VM (i.e., boots successfully to the desktop without any edits to menu.lst).

One must examine all causes and rule out one by one. The drive is therefore not defective and the CDs are written correctly.

Either, the bootloader software requires the functionality of the feature identifier CMOV, but the CPU is not built with it, or the BIOS is faulty. The former seems more likely to me.

With Salix Xfce 14.2beta1 (32-bit, i586), gapan has provided a version, that could run on the CPUs AMD K6-2 and AMD K6-3D+ when all other non-supported CPU features were also off. An experiment with this beta distribution would bring clarity. Better you wait until the final release, you save one DVD.

Windows 98 does not require the questionable CPU features, because that runs on much older computers.
The same goes for antiX-16-b5-386-full, I think. The 386 suggests that it is running on the same CPU, so, theoretically, on any newer CPU architectures. Since you can edit /boot/grub/menu.lst, the operating system is running yet. What happens then? Can you see how the X server is trying out unsuccessfully? Or are error messages visible, because of missing video driver. What messages are in /var/log/Xorg.0.log? Probably missing a suitable graphics driver, which you can possibly obtain. Perhaps the kernel is not configured for this graphics.

Can not say much detailed about Macbook Pro. A CMOV problem does not exist. The Macbook has a newer BIOS, of course, so called "Open Firmware" (like the PC BIOS). The program flow conditions at boot time are very different than in an about fifteen year old computer.

Also, I can not say what effects the virtual machine has. Never used KVM. ISOLINUX needs some functions of a PC BIOS. Is this emulated by the virtual machine?

Booting on a different hardware in VM to test out ISOLINUX or BIOS problems, is in my opinion quite worthless. You might use the CD to boot directly on the Macbook, to see if it works. But, this "Open Firmware" is UEFI firmware and probably works in UEFI native mode, then you need to switch to a BIOS-like mode. I would leave it alone.

coyotl could perhaps say, for which CPU architectures UBCD is actually suitable. I could not find anything.

[junqueman]

Either, the bootloader software requires the functionality of the feature identifier CMOV, but the CPU is not built with it, or the BIOS is faulty. The former seems more likely to me.

Seems logical to me.

With Salix Xfce 14.2beta1 (32-bit, i586), gapan has provided a version, that could run on the CPUs AMD K6-2 and AMD K6-3D+ when all other non-supported CPU features were also off. An experiment with this beta distribution would bring clarity. Better you wait until the final release, you save one DVD.

OK. BTW I've had success in booting and installing Salix 14.1 Fluxbox (as well as Openbox) on a third system - this one uses an AMD K6-3D CPU. However, the X server aborts before the GUI can launch, and I get a command-line login. I will post the warnings and errors from the resulting Xorg.0.log in a separate post.

Windows 98 does not require the questionable CPU features, because that runs on much older computers.
The same goes for antiX-16-b5-386-full, I think. The 386 suggests that it is running on the same CPU, so, theoretically, on any newer CPU architectures. Since you can edit /boot/grub/menu.lst, the operating system is running yet. What happens then? Can you see how the X server is trying out unsuccessfully? Or are error messages visible, because of missing video driver. What messages are in /var/log/Xorg.0.log? Probably missing a suitable graphics driver, which you can possibly obtain. Perhaps the kernel is not configured for this graphics.

I actually made the edits to /boot/grub/menu.lst from the graphical text editor running from the Wary Puppy live CD, but I'm sure I could have done it from command-line in AntiX. What happens is I get a command-line login with no obvious error messages preceding (similar to Salix on the "third" machine mentioned above). Here is a list of all the warnings and errors from /var/log/Xorg.0.log (only the warnings and errors are shown, with a few exceptions because they look like they may be related to the errors and therefore may show useful information):

[119.970] (WW) The directory "usr/share/fonts/X11/cyrillic" does not exist
[120.308] (WW) The directory "usr/share/fonts/X11/75dpi/" does not exist
[120.313] (WW) The directory "usr/share/fonts/X11/75dpi/" does not exist
[124.222] (WW) Falling back to old probe method for modesetting
[124.222] (EE) open /dev/dri/card0: No such file or directory
[124.223] (WW) Falling back to old probe method for fbdev
[124.541] (WW) Falling back to old probe method for vesa
[127.523] [drm] failed to load kernel module "savage"
[127.523] (EE) [drm] drm Open failed
[127.523] (EE) SAVAGE(0): [drm] DRI ScreenInit failed. Disabling DRI
[127.523] (EE) SAVAGE(0): DRI isn't enabled
[127.523] (--) SAVAGE(0): Chose mode 11a @ 60 Hz
[128.078] (EE) AIGLX: reverting to software rendering
[140.242] (EE) XKB: Could not invoke xkbcomp
[140.242] (EE) XKB: Couldn't compile keymap
[140.242] (EE) XKB: Failed to load keymap. Loading default keycap instead.
[140.268] (EE) XKB: Could not invoke xkbcomp
[140.269] (EE) XKB: Couldn't compile keymap
[140.269] (EE) XKB: Failed to compile keymap
[140.269] (EE) Keyboard initialization failed. This could be a missing or incorrect setup of xkeyboard-config.
[140.269] (EE)
Fatal Server Error:
[140.269] (EE) Failed to activate core devices. (EE)
[140.269] (EE)
Please consult the X.Org Foundation support at http://wiki.x.org for help.
[140.269] (EE) Please also check the log file at "/var/log/xorg.0.log" for additional information.
[140.269] (EE)
[140.710] (EE) Server terminated with error (1). Closing log file.

Also, I can not say what effects the virtual machine has. Never used KVM. ISOLINUX needs some functions of a PC BIOS. Is this emulated by the virtual machine?

Well, since the VMs so far have successfully run every distro I've tried on them, I'd have to conclude "yes." For my VMs I use Virtualbox for Mac OS X.

Booting on a different hardware in VM to test out ISOLINUX or BIOS problems, is in my opinion quite worthless. You might use the CD to boot directly on the Macbook, to see if it works. But, this "Open Firmware" is UEFI firmware and probably works in UEFI native mode, then you need to switch to a BIOS-like mode. I would leave it alone.

Yeah, I've been doing this just to test that the CDs (or DVDs) are OK, and to explore some of the cool distros that I haven't been able to get running on the HOH

[junqueman]

I mentioned in my preceding post that I was able to complete installation of Salix 14.1 (Fluxbox) on a third, previously untried, machine of similar vintage to the other 2 machines I had been using. This machine uses an AMD K6-3D 350 MHz CPU. I was able to boot from the CD using the huge.s kernel (which could not be successfully done with the other 2 machines). I did not install LILO, because it has given me problems previously, so I was going to install grub using the instructions at viewtopic.php?p=16546. So after installation I re-booted from the Salix installation CD by entering this at the "boot:" prompt: huge.s root=/dev/sda1 rdinit= ro.
All was good until I got a dialog box telling me "Failed to start the X server (your graphical interface). It is likely that it is not set up correctly. Would you like to view the X server output to diagnose the problem?"

To which I replied "Yes". This is the output that was displayed:

Initializing built-in extension DR12
Loading extension GLX
(EE)
Fatal server error:
(EE) no screens found (EE)
(EE)
Please consult the The X.Org Foundation support at http://wiki.x.org for help
(EE) Please also check the log file at "/car/log/Xorg.0.log for additional information.
(EE)
(EE) Server terminated with error (1) Closing log file.

Another dialog box appeared asking "Would you like to view the detailed X server output as well?" Here are the warnings and errors (only) from that:

[92.494] (WW) The directory "/usr/share/fonts/local" does not exist.
[92.494] (WW) The directory "/usr/share/fonts/OTF" does not exist.
[92.498] (WW) The directory "/usr/share/fonts/CID" does not exist.
[92.525] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[92.578] (WW) Warning, couldn't open module sis
[92.578] (EE) Failed to load module "sis" (module does not exist, 0)
[92.607] (WW) Warning, couldn't open module fbdev
[92.608] (EE) Failed to load module "fbdev" (module doesn't exist, 0)
[92.617] (WW) Falling back to old probe method for modesetting
[92.618] (EE) open /dev/dri/card0: No such file or directory
[92.650] (EE) screen(s) found, but none have a usable configuration
[92.650] (EE)
Fatal server error:
[92.651] (EE) no screens found (EE)
[92.651] (EE)
Please consult the X.org Foundation support http://wiki.x.org for help
[92.652] (EE) Please also check the log file at "/var/log/Xorg.0.log"
[92.652] (EE)
[92.659] (EE) Server terminated with error (1). Closing log file

After viewing the complete file I get the following:
darkstar login [87.997317] Disabling IRQ #10
[91.156874] Disabling IRQ #10
[92.306708] Disabling IRQ #10
gdm-binary[680] Warning: Failed to start X server several times in a short time period; disabling display :0
Welcome to Linux 3.10.17 (tty1)
darkstar login:

...at which point I can log in to a command-line terminal with the user name and password I had set up during the Salix 14.1 installation. I did so and viewed the "/var/log/Xorg.0.log" file, which appeared to be exactly the same as the file that was previously displayed and from which I had copied the errors (above).

It appears that I am (at least) missing a video driver, but I don't know how to install it even if I was able to locate it. The video chip in this machine is SiS 6326. Any assistance in resolving this would be greatly appreciated. So close and yet so far!

As a side note, surprisingly AntiX-16-b5_386.full boots and installs with graphics! in this machine, but exhibits some other weirdness - for example, the Control Centre screen won't come up, and the web browser crashes on just about any page I try except for the home page. We probably shouldn't explore these issues in this forum, though - it's not a Salix topic.

[westms]

I mentioned in my preceding post that I was able to complete installation of Salix 14.1 (Fluxbox) on a third, previously untried, machine of similar vintage to the other 2 machines I had been using. This machine uses an AMD K6-3D 350 MHz CPU. I was able to boot from the CD using the huge.s kernel (which could not be successfully done with the other 2 machines). I did not install LILO, because it has given me problems previously, so I was going to install grub using the instructions at viewtopic.php?p=16546. So after installation I re-booted from the Salix installation CD by entering this at the "boot:" prompt: huge.s root=/dev/sda1 rdinit= ro.
All was good until I got a dialog box telling me "Failed to start the X server (your graphical interface). It is likely that it is not set up correctly. Would you like to view the X server output to diagnose the problem?"

To which I replied "Yes". This is the output that was displayed:

Initializing built-in extension DR12
Loading extension GLX
(EE)
Fatal server error:
(EE) no screens found (EE)
(EE)
Please consult the The X.Org Foundation support at http://wiki.x.org for help
(EE) Please also check the log file at "/car/log/Xorg.0.log for additional information.
(EE)
(EE) Server terminated with error (1) Closing log file.

Another dialog box appeared asking "Would you like to view the detailed X server output as well?" Here are the warnings and errors (only) from that:

Yes, a driver could not be found or is not installed.
Try:

Code: Select all

sudo slapt-get --search xf86-video-sis

If packet xf86-video-sis is not installed, install it and reboot. (man slapt-get)

If installed, do a ls /usr/lib/xorg/modules/drivers and look for the files:
sis_drv.la
sis_drv.so

If existing, then it becomes more complicated: Add a file with a name like 50-device-sis6326.conf to /etc/X11/xorg.conf.d
The content could be:

Code: Select all

Section "Device"
	Identifier  "SiS 6326 VGA Display Adapter"
	Driver      "sis"
	Option      "NoAccel"   "1"
#	Option      "DRI"       "1"
EndSection

and reboot. Can not tell if this is good enough. Maybe you still need a Monitor and a Screen section, too. But for a first try...

Disabling IRQ #10

The IRQ 10 could be used to synchronize write operations into the display memory in graphics mode, with the vertical retrace.

And to this

I did not install LILO, because it has given me problems previously, [...]

I want to say: This is not a good idea, because if you look for an error, then you do not change many things at once. Which problem with LILO might have been that?

Added: It can not hurt to read this: man sis

[junqueman]

Yes, a driver could not be found or is not installed.
Try:

Code: Select all

sudo slapt-get --search xf86-video-sis

If packet xf86-video-sis is not installed, install it and reboot. (man slapt-get)

If installed, do a ls /usr/lib/xorg/modules/drivers and look for the files:
sis_drv.la
sis_drv.so

If existing, then it becomes more complicated: Add a file with a name like 50-device-sis6326.conf to /etc/X11/xorg.conf.d
The content could be:

Code: Select all

Section "Device"
	Identifier  "SiS 6326 VGA Display Adapter"
	Driver      "sis"
	Option      "NoAccel"   "1"
#	Option      "DRI"       "1"
EndSection

and reboot. Can not tell if this is good enough. Maybe you still need a Monitor and a Screen section, too. But for a first try...

Thanks once again, westms, for the assistance. The sis driver was not installed, so I installed it per your directions. After reboot, this got me a little farther. The graphical login screen attempted to come up and was visible for an instant, then it immediately disappeared. This happened several times, and then the following text was displayed: "The display has been shut down about 6 times in the last 90 seconds. It is likely that something bad is going on. Waiting for 2 minutes before trying again on display :0." 2 minutes was enough for me to verify from command line that the sis_dv.la and sis_dv.so files were now present in directory /usr/lib/xorg/modules/drivers (as they should be), but not long enough for me to create a new file per your suggestion in /etc/X11/xorg.conf.d (I need to refresh my memory on how to use an editor that works without graphics, and I'm not a fast typist). Is there a way I can abort the X server from the command line to give me time to try things?

And to this

I did not install LILO, because it has given me problems previously, [...]

I want to say: This is not a good idea, because if you look for an error, then you do not change many things at once. Which problem with LILO might have been that?

I had previously attempted a standard Salix 14.1 installation with the LILO boatloader, on another hard disk, and although there were no suspicious messages during the install process, it would not boot from the hard disk after installation. Then, I think (but I'm not sure at this point) this is the same disk on which I could no longer read, or create, a partition table from Linux (using gparted, for example). However I could create and format partitions from DOS with fdisk, and read and write from Windows 98. Basically the disk became useless for my Linux experiments. Again, at this point I am not positive it was the same hard drive on which I had attempted to install Salix 14.1 with LILO, but was reluctant to take the chance, especially since LILO wouldn't boot Salix anyway in my previous attempt.

[westms]

Thanks once again, westms, for the assistance. The sis driver was not installed, so I installed it per your directions. After reboot, this got me a little farther. The graphical login screen attempted to come up and was visible for an instant, then it immediately disappeared. This happened several times, and then the following text was displayed: "The display has been shut down about 6 times in the last 90 seconds. It is likely that something bad is going on. Waiting for 2 minutes before trying again on display :0." 2 minutes was enough for me to verify from command line that the sis_dv.la and sis_dv.so files were now present in directory /usr/lib/xorg/modules/drivers (as they should be), but not long enough for me to create a new file per your suggestion in /etc/X11/xorg.conf.d (I need to refresh my memory on how to use an editor that works without graphics, and I'm not a fast typist). Is there a way I can abort the X server from the command line to give me time to try things?

No idea what you've done there. I wrote, you could install the package. If the package is installed, the files in question are present. The program call slapt-get --filelist xf86-video-sis should show the files sis_drv.la and sis_drv.so .

Of course, you can terminate the X server, kill him. Or reboot up in runlevel 1 or 3, then you can work in peace.
And read the manpage: man sis.

Editors for terminal: nano, vi/vim/elvis, joe, jed/emacs. nano is already installed. For Real Men: ed.

[junqueman]

No idea what you've done there. I wrote, you could install the package. If the package is installed, the files in question are present. The program call slapt-get --filelist xf86-video-sis should show the files sis_drv.la and sis_drv.so .

Yes; I believe I did install the package. "sudo slapt-get --install f86-video-sis" After installation of the package, which appeared to go flawlessly, I just navigated to the directory with the driver files to make sure they were present.

Of course, you can terminate the X server, kill him. Or reboot up in runlevel 1 or 3, then you can work in peace.

With the interruptions every two minutes I had a hard time figuring out which running process was the one needing killing. As far as booting up in run level 1 or 3, I don't know what that means, or how to make that happen. Remember, I'm a noob.

And read the manpage: man sis.

I did. It looks like the same information I had previously reviewed at ftp://www.x.org/pub/X11R7.5/doc/man/man4/sis.4.html

Editors for terminal: nano, vi/vim/elvis, joe, jed/emacs. nano is already installed. For Real Men: ed.

Using nano I created the file you suggested, invoking the sis driver with the "NoAccel" option. I am now greeted with a stable login screen and desktop. Progress! There are some things that don't work properly; the most annoying one is that Firefox appears not to launch, or crashes before any window appears. Another problem is that when in LX terminal I cannot see the text I have typed until after I hit "Enter." Thirdly, the display doesn't refresh properly after drawing a window - horizontal and vertical lines are left on the "wallpaper" and in parts of open windows. But this is farther than I've gotten with Salix on this vintage of hardware to date.

I would like to identify what is different in this hardware which allows "huge.s" to boot whereas the other AMD K6-III would not. However, the inxi command isn't recognized.

[westms]

As far as booting up in run level 1 or 3, I don't know what that means, or how to make that happen. Remember, I'm a noob.

On bootloader command line, e.g. for runlevel 3: The_Kernel_Name 3

A noob who uses Virtualbox but does not know what runlevels are.

Using nano I created the file you suggested, invoking the sis driver with the "NoAccel" option. I am now greeted with a stable login screen and desktop. Progress! There are some things that don't work properly;

the most annoying one is that Firefox appears not to launch, or crashes before any window appears.

Then this Firefox is not for your AMD K6.

Another problem is that when in LX terminal I cannot see the text I have typed until after I hit "Enter."

Looks like no echo. A mis-configuration, perhaps. Correct it or try other terminal emulations.

Thirdly, the display doesn't refresh properly after drawing a window - horizontal and vertical lines are left on the "wallpaper" and in parts of open windows. But this is farther than I've gotten with Salix on this vintage of hardware to date.

To /etc/X11/xorg.conf.d add the files:

50-serverflags.conf

Code: Select all

Section "ServerFlags"
	Option 		"DontZap"		"off"
#	Option		"DontZoom"
	Option		"BlankTime"      "0"
	Option		"StandbyTime"    "0"
	Option		"SuspendTime"    "0"
	Option		"OffTime"        "0"
EndSection

50-files.conf

Code: Select all

Section "Files"
	ModulePath   "/usr/lib/xorg/modules"

	FontPath     "/usr/share/fonts/local"
#	FontPath     "/usr/share/fonts/TTF"    # default
	FontPath     "/usr/share/fonts/OTF"
#	FontPath     "/usr/share/fonts/Type1"  # default
#	FontPath     "/usr/share/fonts/misc"   # default
	FontPath     "/usr/share/fonts/CID"
#	FontPath     "/usr/share/fonts/75dpi/:unscaled"   # default
#	FontPath     "/usr/share/fonts/100dpi/:unscaled"  # default
#	FontPath     "/usr/share/fonts/75dpi"     # default
#	FontPath     "/usr/share/fonts/100dpi"    # default
#	FontPath     "/usr/share/fonts/cyrillic"  # default

#	RGBPath      "/usr/share/X11/rgb"         # rgb.txt # obsolete
EndSection

50-module.conf

Code: Select all

Section "Module"
	Load        "dbe"  	# Double buffer extension  # default
	Load        "extmod" # Okay, as long DGA extension is disabled

	SubSection  "extmod"
		Option    "omit xfree86-dga"  # don't initialise the DGA extension
	EndSubSection

	Load        "type1"
	Load        "freetype"
	Load        "glx"  # default
EndSection

50-extensions.conf

Code: Select all

Section "Extensions"
	Option	"Composite"          "Enable"  # (RENDER acceleration)
EndSection

Perhaps it helps. And don't be stupid and type all this in. Copy it to files on a full working system and move the files to the HOH. There are more files to be add. But for the moment...

I would like to identify what is different in this hardware which allows "huge.s" to boot whereas the other AMD K6-III would not. However, the inxi command isn't recognized.

With the older kernels, on both machines:
cat /proc/cpuinfo