Slacko32/64 700 BETA 2

A home for all kinds of Puppy related projects
Message
Author
User avatar
James C
Posts: 6618
Joined: Thu 26 Mar 2009, 05:12
Location: Kentucky

#31 Post by James C »

gyro wrote:@James C,

I notice most of your kernel lines have a pdev1=sdb1, but 1 shown doesn't.
If you use a kernel line that doesn't specify a pdev1 (or pupsfs) it will go looking on all partitions.

So what kernel line did you use?

Got error messages, what were they?

gyro

Used the one shown, only changed the title.

Error messages were something about ext4 errors on sdb7,sdb8 and sdb9 which are now occupied by AntiX, Trisquel and Devuan and not being able to find the slacko64 sfs on sdb9.

http://www.murga-linux.com/puppy/viewto ... 447#920447

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#32 Post by gyro »

@James C,
Sorry, I didn't pay close enoough attention.
Yes, the one for pup15.
Nothing unusual there, so still got nothing.
gyro

User avatar
James C
Posts: 6618
Joined: Thu 26 Mar 2009, 05:12
Location: Kentucky

#33 Post by James C »

gyro wrote:@James C,
Sorry, I didn't pay close enoough attention.
Yes, the one for pup15.
Nothing unusual there, so still got nothing.
gyro

I took a quick,quick glance in the initrd,didn't notice anything obvious and just switched to the initrd out of 6.3.0 since I knew it worked.

jlst

#34 Post by jlst »

I can't reproduce the issue. Maybe some debug code can shed light on the situation. Maybe implement pfix=debug and restore syslogd and klogd for that pfix (output goes to /tmp)..

jlst

#35 Post by jlst »

"debugsave" script

If you're having problems booting puppy, this can help
save debug info to a partition for further analysis,
assuming a kernel panic does not occur.

Just type:

debugsave

and follow the instructions
Last edited by jlst on Tue 30 Aug 2016, 15:09, edited 6 times in total.

User avatar
smokey01
Posts: 2813
Joined: Sat 30 Dec 2006, 23:15
Location: South Australia :-(
Contact:

#36 Post by smokey01 »

Gyro I compiled zarfy which works fine but the changes are not persistent. As soon as I run zarfy again the monitors are swapped back but on next restart all reverts to original wrong way round.

jlst

#37 Post by jlst »

nice app smokey01. it's really old.

apparently there's no switch to make it apply settings, so i was searching
for a config file and found it. i identified some values usable by xrandr.

usage:

run zarfy
change settings
click on OK (maybe Apply first)
restart X/reboot/etc

(and it will apply basic settings)
Last edited by jlst on Tue 30 Aug 2016, 15:55, edited 1 time in total.

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#38 Post by gyro »

smokey01 wrote:Gyro I compiled zarfy which works fine but the changes are not persistent. As soon as I run zarfy again the monitors are swapped back but on next restart all reverts to original wrong way round.
Thats why I put a little script in /root/Startup.
gyro

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#39 Post by gyro »

@boot fails with relsase initrd.gz folk.

Just to start on the same page, could you please do test booting with the following configuration:
Do a non uefi boot of a frugal install on an ext partition containing only these 4 files, initrd.gz, puppy_slacko64_6.9.5.sfs, vmlinuz, zdrv_slacko64_6.9.5.sfs.
With these boot paramaters "pmedia=atahd", "pfix=fsckp", and appropriate "pdev1=" and "psubdir=".

Note: the pfix=fsckp, this tells the new init to fsck every ext partition before just before mounting it. The important thing for testing is that it displays the output of e2fsck on the console, so it should be obvious if the "pdev1=" parameter is not being honoured.

We need to try to reduce the variation, in order to pinpoint the significant difference.

Oh, by all means use the debug version produced by jlist.

If this does not provide a resolution, then I can produce a series of init scripts that sinply stop and display significant variables, (then reboot with ctrl+alt+del, it's ok no partitions are mounted when it stops).
The first version would do this very early to establish that the boot parameters have been correctly interpreted, and then subquent versions move this further down the code, until we find something that is not right.
Quite tedious, but it's pretty much the method I used to build the thing.

gyro

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#40 Post by gyro »

I'm working on an init which has several breakout points, controled by "pdebug=n" boot parameter.
"pdebug=0" will dropout very early, "debug=1" will dropout later, and so on...
So downloading a single initrd.gz will provide many testing points.
Even just seeing which debug point you get to before a problem occurs, will give us a starting clue.
gyro

User avatar
01micko
Posts: 8741
Joined: Sat 11 Oct 2008, 13:39
Location: qld
Contact:

#41 Post by 01micko »

Thanks @gyro for joining the discussion.

Hopefully this weekend I can build a 32 bit version with the latest 'rationalise' at that time and further the discussion/findings/fixes in a new thread.

Cheers!
Puppy Linux Blog - contact me for access

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#42 Post by gyro »

I have atached a debgug init for this slacko64, as init.gz.
It's also available as an updated initrd.gz from http://www.fishprogs.software/puppy/slacko64/initrd.gz.
This init accepts "pdebug=" boot parameters.
"pdebug=0" drops out to the console just after intepreting boot parameters.
"pdebug=3" drops out just before getting the no puppy...sfs message.
1,2 are in between. There is no "pdebug=4".
At each pdebug point it displays a message showing significant init variables at that point. The different pdebug points show different messages.
At the console you can run commands like "mount" to confirm that no partitions are still mounted.
Use ctrl+alt+del keys to reboot.

How to use:
Start by booting with "pdebug=0", and then "pdebug=1" etc.. noting the last pdebug point that was successful.
Please report what was the last successful pdebug point number and the contents of the line above showing the init variables.

Edit: This version has been replaced, see later post

gyro
Last edited by gyro on Tue 30 Aug 2016, 16:44, edited 1 time in total.

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#43 Post by gyro »

An example of successful boots using pdebug.

grub4dos kernel line:

Code: Select all

kernel /puppy/slacko64/vmlinuz libata.noacpi=1 pmedia=atahd pupsfs=49baa82d-8c69 psubdir=/puppy/slacko64 pfix=fsckp pdebug=0
produces this variables display line:

Code: Select all

PMEDIA=atahd PUPSYS=49baa82d-8c69 PSUBDIR=/puppy/slacko64 pfix=fsckp
Using pdebug=1 produces:

Code: Select all

PUPSYS= P_BP_ID=49baa82d-8c69
Using pdebug=2 produces:

Code: Select all

ONE_PART=sdb4 ONE_FN=/puppy/slacko64/puppy_slacko64_6.9.5.sfs LOOK_PUP= LOOK_SAVE=
Using pdebug=3 produces:

Code: Select all

ONE_PART=sdb4 ONE_TRY_FN=/puppy/slacko64/puppy_slacko64_6.9.5.sfs PDRV=sdb4,ext4,/puppy/slacko64/puppy_slacko64_6.9.5.sfs
And pdebug=4 results in boot to desktop.

gyro

jlst

#44 Post by jlst »

haha silly me. zarfy is awesome. it should be included in all puppies. it's really tiny.

so it has the -l option, at first thought it didn't work.

so i tested with 2 monitors and now i think puppy remembers that last settings and will apply them everytime X starts...
Attachments
xinitrc__zarfy-1.0.pet
(3.11 KiB) Downloaded 406 times

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#45 Post by gyro »

A new version of an init script with debug support is attached.
An updated initrd.gz is available at http://www.fishprogs.software/puppy/slacko64/initrd.gz.

This a a slightly refined version, in that the debug messages are always written if any value of pdebug is set. i.e. if you boot with "pdebug=y" all debug messages will be appended to bootinit.log.
If "pdebug=" is set to 0,1,2,3 init will drop out as per previous version.

So, for initial testing boot "pdebug=y". If the boot works, there's a few extra lines in "/initrd/tmp/bootinit.log". If it fails, all the messages should appear on the screen following the "bootinit.log" line.

Example debug lines from a successful boot:

Code: Select all

0: PMEDIA=atahd PDRV=49baa82d-8c69 PSUBDIR=/puppy/slacko64 pfix=fsckp
1: PDRV= P_BP_ID=49baa82d-8c69 P_BP_FN=
2: ONE_PART=sdb4 ONE_FN=/puppy/slacko64/puppy_slacko64_6.9.5.sfs ONE_MP=/mnt/pdrv
3: ONE_PART=sdb4 ONE_TRY_FN=/puppy/slacko64/puppy_slacko64_6.9.5.sfs PDRV=sdb4,ext4,/puppy/slacko64/puppy_slacko64_6.9.5.sfs
Edit: Attachemnt deleted, superceeded by later version, see later post.

gyro
Last edited by gyro on Wed 31 Aug 2016, 11:23, edited 1 time in total.

User avatar
mavrothal
Posts: 3096
Joined: Mon 24 Aug 2009, 18:23

#46 Post by mavrothal »

Couldn't help it and played a bit more with it. :D
I often boot from a CD/Image and have savefiles/folder in a partition.
So far puppies could find the savefile automatically.
Not anymore.
You now need to define psave (and realize that you need to "tab" to pass any kernel arguments).

Puppy boot arguments where there mostly to help when boot without user input was failing or speed up/customize things.
Would be nice to remain user/nooby friendly and not follow fatdog, debiandog, porteus, tinycore (all of which I very much like) and other more "involved" mini/frugal OSs.

BTW if anyone needs the broadcom wl driver for 4.1.31 x86_64, is here
== [url=http://www.catb.org/esr/faqs/smart-questions.html]Here is how to solve your[/url] [url=https://www.chiark.greenend.org.uk/~sgtatham/bugs.html]Linux problems fast[/url] ==

User avatar
smokey01
Posts: 2813
Joined: Sat 30 Dec 2006, 23:15
Location: South Australia :-(
Contact:

#47 Post by smokey01 »

I can confirm the zarfy -l switch works. When zarfy is run it creates a config file in /root./zarfy The -l switch forces it to be read. By placing "zarfy -l" in .xinitrc it works fine. I guess placing a script with the contents "zarfy -l" in /root/startup would work equally as well but you you might see the screens change as scripts in startup are run after .xinitrc, I think.

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#48 Post by gyro »

mavrothal wrote:Couldn't help it and played a bit more with it. :D
I often boot from a CD/Image and have savefiles/folder in a partition.
So far puppies could find the savefile automatically.
Not anymore.
It should automatically find them in the root of the partition. On a cd boot It will look for /slacko64save... (since PSUBDIR is empty.)
mavrothal wrote:(and realize that you need to "tab" to pass any kernel arguments).
I don't understand what this means.

gyro

gyro
Posts: 1798
Joined: Tue 28 Oct 2008, 21:35
Location: Brisbane, Australia

#49 Post by gyro »

A new debug version of the init script is attached.
A new debug version of the initrd.gz has been uploaded to the same place as the previous version.

The only change is the inclusion of debug messages inside the search loop. So if a search is done, it can easily be seen which partitions it looked at.

Edit: Attachment deleted, replaced by new version, see later post.

gyro
Last edited by gyro on Wed 31 Aug 2016, 17:28, edited 2 times in total.

User avatar
mavrothal
Posts: 3096
Joined: Mon 24 Aug 2009, 18:23

#50 Post by mavrothal »

gyro wrote:
mavrothal wrote:Couldn't help it and played a bit more with it. :D
I often boot from a CD/Image and have savefiles/folder in a partition.
So far puppies could find the savefile automatically.
Not anymore.
It should automatically find them in the root of the partition. On a cd boot It will look for /slacko64save... (since PSUBDIR is empty.)
The savefolder is at the root of the partition. The partition is detected according to the logs but it does not load.
Good to know that is a bug rather than design change, though.
gyro wrote:
mavrothal wrote:(and realize that you need to "tab" to pass any kernel arguments).
I don't understand what this means.

gyro
Not an init issue.
It has to do with how you pass arguments now with the UEFI/Grub bootloader.
== [url=http://www.catb.org/esr/faqs/smart-questions.html]Here is how to solve your[/url] [url=https://www.chiark.greenend.org.uk/~sgtatham/bugs.html]Linux problems fast[/url] ==

Post Reply