Page 1 of 1

Why does 'xwin' sometimes cause troubles for 'X' and 'startx' does not?

Posted: Thu Dec 07, 2023 10:39 am
by ozboomer

Hi, again...

I'm trying to use some of the 'more modern' Puppies on a certain PC that has 2 monitors running through an old NVIDIA card (an EVGA GeForce 6200 PCI card, from about 2004).

However, like someone has found before me, starting 'X' is a bit of a problem.

'X' works 'out of the box' (that is, with no savefile) across both monitors under Xenial64 7.5 (64 bit)... as does Bookworm64 (64 bit)... and I can get Slacko 6.3.2 (32 bit) to work with a bit of xrandr manipulation, together with an old NVIDIA video driver... but FossaPup64 (as FossaPup64 and F96-CE), as well as most any of the 64-bit Puppies (like Bionic, Slacko64.. and a couple of others) run to the command line... but when 'xwin' starts, the display is all black, with an underbar cursor at the top left corner of the first screen... and I need to reboot.

When possible, if I boot to a command prompt, and run /usr/bin/startx, it seems like everything works Ok... but if I try to run /usr/bin/xwin, I get the dreaded underbar cursor problem.

Going through the available options in 'xorgwizard' and going through all the reboots doesn't change the behaviour.

I'm just wondering if anyone has some understanding of the differences between 'startx' and 'xwin'... or maybe I'll just have to do the 'forensic examination' of both scripts to try and work out why one allows 'X' to work while the other doesn't... Ugh!

It might be a bit academic, as I'm likely going to change over the 'troublesome' PC by this time next year... but I'm still trying (arguably pointlessly) to create one savefile for a given Puppy variant to make maintaining 'Puppy' easier across a couple of PCs.

Always appreciate any thoughts folks might be able to pass on....


Re: Why does 'xwin' sometimes cause troubles for 'X' and 'startx' does not?

Posted: Thu Dec 07, 2023 1:50 pm
by rockedge

@ozboomer
Sounds like swapping the fdrv SFS or changing the kernel will fix it. This must be a case of missing firmware in the kernel package.

It appears that the firmware and kernel have the needed graphic card drivers in Xenial-7.5 but not in the newer Puppy's kernel package.

Have you tried to use the same kernel that works in Xenial in F96-CE_4?


Re: Why does 'xwin' sometimes cause troubles for 'X' and 'startx' does not?

Posted: Thu Dec 07, 2023 10:31 pm
by ozboomer
rockedge wrote: Thu Dec 07, 2023 1:50 pm

Sounds like swapping the fdrv SFS or changing the kernel will fix it. This must be a case of missing firmware in the kernel package.
[ . . . ]
Have you tried to use the same kernel that works in Xenial in F96-CE_4?

Huh.. Never even thought of that.

Tried swapping-in vmlinuz from Xenial64 7.5 and it booted... but the 'depmod' load failed.. but couldn't find anything in the boot logs about it... but I'm not expert in understanding them.. and 'xwin' worked Ok... but I couldn't get the higher resolution/dual displays working (both screens showed the same content and the resolution couldn't be changed from 1024x768)... but then, maybe I need to install the 32-bit driver... if it'll work - I didn't think I had to with 'Xenial'... but I'm not sure.

So, maybe there's a bit more still to try along that path... Fanx!


Re: Why does 'xwin' sometimes cause troubles for 'X' and 'startx' does not?

Posted: Thu Dec 07, 2023 11:50 pm
by rockedge

@ozboomer,
when changing the vmlinuz also change the zdrv SFS (has the modules) and the fdrv SFS (this depends on the kernel you want to use) that goes along with the vmlinuz. There might not be a fdrv SFS depending on the kernel version. The kernel in Xenial64-7.5 has 2 components.

Screenshot(74).jpg
Screenshot(74).jpg (17.62 KiB) Viewed 349 times

For a simple test swap in Xenial64-7.5's kernel package into a F96-CE_4 which might work correctly.

  1. Copy zdrv_xenialpup64_7.5.sfs and rename it to zdrv_fossapup64_9.6.sfs

  2. Copy the renamed zdrv and the vmlinuz from Xenial64-7.5 into F96-CE_4 to replace those kernel files that are present.

  3. Rename the F96-CE_4 fdrv_fossapup64_9.6.sfs to No-fdrv_fossapup64_9.6.sfs

Then try to boot the F96-CE_4


Re: Why does 'xwin' sometimes cause troubles for 'X' and 'startx' does not?

Posted: Fri Dec 08, 2023 1:02 am
by ozboomer
rockedge wrote: Thu Dec 07, 2023 11:50 pm

For a simple test swap in Xenial64-7.5's kernel package into a F96-CE_4 which might work correctly.

  1. Copy zdrv_xenialpup64_7.5.sfs and rename it to zdrv_fossapup64_9.6.sfs

  2. Copy the renamed zdrv and the vmlinuz from Xenial64-7.5 into F96-CE_4 to replace those kernel files that are present.

  3. Rename the F96-CE_4 fdrv_fossapup64_9.6.sfs to No-fdrv_fossapup64_9.6.sfs

This is one of the reasons I gravitated more to 'Fossa' (and the other 'more recent' Puppies) where the 'hardware(firmware) dependent' components are largely separated from the 'distro' per se... for just this sort of situation :)

Will give this a try in a lil' bit... Ta..


Re: Why does 'xwin' sometimes cause troubles for 'X' and 'startx' does not?

Posted: Fri Dec 08, 2023 5:56 am
by ozboomer

I tried the suggested changes... and tried a 'no X' boot... and that went through Ok... but the message 'Making the filesystem usable' showed 'failed'. Doing a 'normal' boot, the filesystem 'failed' message was displayed again... and as it called 'xwin', the 'underbar cursor' was displayed and the PC was again hung. As I said, I'm not great on the log files.. so maybe others can 'decode' something from the logs:

logs.zip
(16.78 KiB) Downloaded 11 times

I'm sure I didn't mix-up the files (which is admittedly, easy to do)... I also removed the ydrv_* as well as leaving it around.. and it made no difference..

Again, to roll out my most-often used quote, "'tis a puzzlement"