DebianDog - Jessie (21 June 2017)

A home for all kinds of Puppy related projects
Post Reply
Message
Author
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#101 Post by saintless »

Hi, Fred.

I think you will be happy to know I have P-III machine loading modules without error from NTFS and ext4 partitions using live-boot-2 and 3 :)
I can't say what is the problem yet (maybe hdd driver issue or similar). I will test boot with older kernel on the machine with ntfs-3g errors in the next days. Maybe some module in kernel 3.16 (or Jessie firmware) doesn't work well with some HDD type.

Edit: I guess the new way mount command works in combination with driver bug is the reason for this ntfs-3g error. Still downgrading mount deb seems the safe and universal solution for me but I like to test more to confirm.

Edit2: Do you know if porteus-boot has different module probe/load on boot compared to live-boot?

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#102 Post by fredx181 »

Hi Toni,
I think you will be happy to know I have P-III machine loading modules without error from NTFS and ext4 partitions using live-boot-2 and 3 Smile
I can't say what is the problem yet (maybe hdd driver issue or similar). I will test boot with older kernel on the machine with ntfs-3g errors in the next days.
Ah, there's hope! :)
Edit2: Do you know if porteus-boot has different module probe/load on boot compared to live-boot?
Yes, probably very different, as you know, sfs made it possible to make porteus-boot work with Debian kernel.
The way he did was loading all modules from /lib/modules/<kernel-version> from initrd1.xz
This increased the RAM usage to much so I made selection, but very much on the safe side (still a lot).
It's in linuxrc:

Code: Select all

echo $i"loading kernel modules..."
   for MODULE in mbcache aufs exportfs ext4 fat vfat fuse isofs nls_cp437 nls_iso8859-1 nls_utf8 reiserfs squashfs xfs libata ahci libahci sata_sil24 pdc_adma sata_qstor sata_sx4 ata_piix sata_mv sata_nv sata_promise sata_sil sata_sis sata_svw sata_uli sata_via sata_vsc pata_ali pata_amd pata_artop pata_atiixp pata_atp867x pata_cmd64x pata_cs5520 pata_cs5530 pata_cs5535 pata_cs5536 pata_efar pata_hpt366 pata_hpt37x pata_it8213 pata_it821x pata_jmicron pata_marvell pata_netcell pata_ns87415 pata_oldpiix pata_pdc2027x pata_pdc202xx_old pata_rdc pata_sc1200 pata_sch pata_serverworks pata_sil680 pata_sis pata_triflex pata_via pata_isapnp pata_mpiix pata_ns87410 pata_opti pata_rz1000 ata_generic loop cdrom hid usbhid mptscsih mptspi mptsas tifm_core cb710 mmc_block mmc_core sdhci wbsd tifm_sd cb710-mmc via-sdmmc vub300 sdhci-pltfm scsi_mod scsi_transport_spi xts scsi_transport_sas sd_mod sr_mod usb-common usbcore ehci-hcd ehci-pci ohci-hcd uhci-hcd xhci-hcd usb-storage dm_mod; do
      modprobe $MODULE 2>/dev/null
I think you are on the right track with this.

Fred
mcewanw
Posts: 3169
Joined: Thu 16 Aug 2007, 10:48
Contact:

#103 Post by mcewanw »

saintless wrote:Hi, William.
Even the first preview iso Light-Debian-Core-Live-CD-Squeeze was with SMP kernel. Infact you, William, were the reason to build no-PAE version :)
Yes, I know... :-) When forcepae kernel line worked for me this time, I was really surprised (and happy). Gave what used to be my main work horse a new lease of life in terms of my supporting it.

saintless wrote: Why there is no person interested enough from DebianDog 64-bit version to suggest help and support for 64-bit version?
...
Toni
Yes, that's a good point of course. As far as 64bit capable machines are concerned I just suddenly found myself with one and then three... for $50 each, and with 2GB RAM. I am still running 32bit DD on them, but my partner wants that 'viber' comms program which only has 64bit Linux version - otherwise I wouldn't even be thinking about it. Truth is, I've never been interested in actually building any new distribution, preferring to build or maintain small system apps and leaving distribution building to others. There is never enough time in the day though admittedly I have been slowing down in terms of my own contributions. Who knows, perhaps I will have a go at it sometime, if no one else jumps in, but don't hold your breathe cos, it not being 'my thing', I think it is unlikely...

William
github mcewanw
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#104 Post by saintless »

Thanks, Fred!
I will do some testing in the next days but it is driver bug with sure.
With liquorix kernel loading modules from ext4 and ntfs works from the machine I have the problem using official kernel 3.16.
To make the test easier for you here is separate kernel module I made earlier for testing SID. It has all boot methods initrd files:
028-kernel-4.0-1.dmz.2-liquorix-686-pae.tar.gz (sorry, you have to use forcepae) ;)
I'm not sure if we can fix this with the default jessie kernel but we know the reason for the error now.

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#105 Post by fredx181 »

Hi Toni,

Tested sfs-load booted with the liquorix kernel you provided and indeed it works without problems.
Btw, with Debian kernel, loading sfs from ext2 or ext3 also doesn't work (not only from NTFS and ext4 as you mentioned), I think it's with all filesytem types.

Fred
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#106 Post by saintless »

fredx181 wrote:Btw, with Debian kernel, loading sfs from ext2 or ext3 also doesn't work (not only from NTFS and ext4 as you mentioned), I think it's with all filesytem types.
Module loading works from ext3 for me without changing the kernel or downgrading mount, Fred. I will test ext2 and vfat when I have time. For the moment I confirm the same error only loading from NTFS or ext4 partition.
My guess is the error depends on different hardware because of some driver bug and works without problem on some hardware like mine P-III machine.
But since porteus-boot works well for you for all partitions maybe only some different or extra module needs to be loaded on boot for live-boot. Testing will show.

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#107 Post by fredx181 »

Hi Toni,
Module loading works from ext3 for me without changing the kernel or downgrading mount, Fred.
For info: I tested on usb stick fresh created second partition ext2 as well as ext3, from both sfs-load failed.
(with live-boot 2 and 3, debian kernel, no downgrading mount)

Fred
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#108 Post by fredx181 »

Hi Toni,

Perhaps a little step forward, I managed to create the same problem with porteus-boot :)
By adding ntfs to the list of modules to load in linuxrc, and making new initrd1.xz with that change, sfs-load fails to activate with the same error.

Edit: Just guessing:
I was wondering why accessing NTFS drive works with porteus-boot while it's not in modprobe list at boot from linuxrc. (in initrd1.xz how it is now)
Inside initrd1.xz there's /usr/bin/ntfs-3g and /usr/lib/libntfs-3g.so, that could take care of that, instead of the kernel module (ntfs.ko).
Edit2: And my NTFS partition is not mounted by ntfs-3g, but by fuse (in case ntfs.ko is not loaded and boot partition is ext4 with porteus-boot):

Code: Select all

root@jessie:~# mount | grep /dev/sda1
/dev/sda1 on /mnt/sda1 type fuseblk (rw,relatime,user_id=0,group_id=0,allow_other,blksize=4096)
Edit2b: Same as above with live-boot, maybe pup-volume-monitor handles it this way.
I'll do some more tests for different cases.
Edit3:Doing "modprobe -r ntfs" makes it work with live-boot:

Code: Select all

root@jessie:~# loadmodule -a /media/sda2/02-isotools-xorriso-isolinux.sfs
sfs
ntfs-3g: Failed to access volume 'aufs': No such file or directory
......
......
News, support and information:  http://tuxera.com
ERROR: Module '02-isotools-xorriso-isolinux.sfs' activate failed
root@jessie:~# modprobe -r ntfs
root@jessie:~# loadmodule -a /media/sda2/02-isotools-xorriso-isolinux.sfs
sfs
SUCCESS: Module '02-isotools-xorriso-isolinux.sfs' activated
ACTION: updating mimeinfo cache
Getting closer now!

Fred
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#109 Post by saintless »

Hi, Fred.

Nice job! All I had time to do was loading all modules from your porteus-boot list from /etc/modules with live-boot and it didn't solve the problem. BTW this module is missing and can't be loaded: sdhci-pltfm. Maybe it is not included in official Jessie kernel 3.16.

I think if we have to mod initrd.img it is better to keep downgraded mount as fix. I don't want to edit initrd.img. It is better to keep it untouched (created with update-initramfs command).
But if adding or blacklisting a module inside 01-filesystem.squashfs can fix the problem without downgrading mount I do not mind such fix.

Still think downgrading mount is easy, safe and reversable solution. We can upgrade it again anytime by installing official jessie deb. But removing the changes from the main module or from initrd.img is more difficult and we need to test such change properly before accept it as better fix.
What do you think, Fred?

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#110 Post by fredx181 »

A little late, but finally yad is part of official Debian repository for stretch and sid:

https://packages.debian.org/stretch/yad

Fred
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#111 Post by saintless »

fredx181 wrote:A little late, but finally yad is part of official Debian repository for stretch and sid:
I know, I had strange looking remasterdog with some warnings upgrading SID few days ago. Unfortunately we will have to keep older yad or you will have to mod all yad scripts (and packages) to work with latest yad for SID (and Jessie if yad gets there too).

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#112 Post by fredx181 »

Hi Toni,
Still think downgrading mount is easy, safe and reversable solution. We can upgrade it again anytime by installing official jessie deb. But removing the changes from the main module or from initrd.img is more difficult and we need to test such change properly before accept it as better fix.
Yes, I think at this moment also the mount downgrade is best, maybe blacklisting ntfs has consequence that we cannot know about yet.

Edit:
I know, I had strange looking remasterdog with some warnings upgrading SID few days ago. Unfortunately we will have to keep older yad or you will have to mod all yad scripts (and packages) to work with latest yad for SID (and Jessie if yad gets there too).
Yes, we'll have to keep older yad untill I find the courage :) to make changes to (hopefully) some or most of the scripts depending on yad.


Fred
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#113 Post by saintless »

Hi, Fred.
fredx181 wrote:...maybe blacklisting ntfs has consequence that we cannot know about yet.
I think it will be a problem to blacklist ntfs. Read Nooby's post and the next answer from me. What I did to make it boot from ntfs partition was installing ntfs-3g and rebuilding initrd.img with ntfs-3g included.
I'm not sure now if the bug is in mount script or in new ntfs-3g version but since downgrading mount works we are fine.

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#114 Post by fredx181 »

Hi Toni,

I think I made another small mistake with openbox version:
A few weeks after jessie became stable I experimented with dist-upgrade to Sid and got distorted graphics caused by the new xserver-xorg-video-intel package. (see screenshot)
My intel video card:

Code: Select all

root@jessie:~# lspci |grep VGA
00:02.0 VGA compatible controller: Intel Corporation 82852/855GM Integrated Graphics Device (rev 02)
So I searched for solution and found it by adding a config file for xorg:
/usr/share/X11/xorg.conf.d/20-intel.conf, with content:

Code: Select all

Section "Device"
   Identifier  "Intel Graphics"
   Driver      "intel"
   Option      "AccelMethod"  "uxa"
EndSection
Ok this solved it for Sid, but... I added it in Jessie version, thinking it wouldn't do any harm.
Now I fould that it does harm because when trying out today DD openbox_xfce on a Dell latitude laptop (with Nvidia video card), X won't start, and removing the 20-intel.conf file solves it.

So in short, /usr/share/X11/xorg.conf.d/20-intel.conf should be removed from Jessie-openbox_xfce.

Can you docoment this in some way?

Fred
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#115 Post by saintless »

Hi, Fred.

Added as fix number 1 here.
I will change it if you like to add something more.

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#116 Post by fredx181 »

Toni wrote:Added as fix number 1 here.
Thanks Toni, clear info!

Fred
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#117 Post by fredx181 »

Hi Toni, All,

isotools deb upgrade to version 1.0.1, see Edit here:
http://murga-linux.com/puppy/viewtopic. ... 481#849481

Toni, here's yad version 0.28.1, I compiled and packaged.(remove dummy .tar)
The difference from the stretch/sid version is that it's compiled with gtk2.
gtk2 is still the default, btw, when building.
Some applications will look a little different with this newer yad, e.g text wrapped, but not as terrible as the gtk3 build from stretch/sid (maybe it's me, but I can't make displays look right with this gtk3 build, seems like height is ignored in a lot of cases, and gets excessive height)

Edit: Re-attached yad .deb package, this one should work on both, wheezy and jessie (didn't test with Sid)

Fred
Last edited by fredx181 on Fri 12 Jun 2015, 20:23, edited 1 time in total.
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#118 Post by saintless »

Hi, Fred.

Both isotools and yad available from dd-jessie repository (section Extra and Updated). Added changes to the description line in isotools deb.
I'm not sure we need to replace yad with newer version. It is not available in Jessie yet. If you think it should be removed from our repository I will add it in no-repo-packages section.
BTW new yad doesn't work in Wheezy.
Wheezy repository updated with isotools (I had to add isolinux from Jessie also in dd-wheezy-repo - it breaks syslinux but installing syslinux will auto-remove isolinux and isotools and the opposite).

Toni
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#119 Post by fredx181 »

Hi Toni,
I'm not sure we need to replace yad with newer version. It is not available in Jessie yet. If you think it should be removed from our repository I will add it in no-repo-packages section.
BTW new yad doesn't work in Wheezy.
See my Edit in previous post for yad 0.28 working on wheezy and jessie.
I think adding it in no-repo-packages is just fine, the older yad is still ok for our needs, just wanted to share the latest yad build with gtk2.
Wheezy repository updated with isotools (I had to add isolinux from Jessie also in dd-wheezy-repo - it breaks syslinux but installing syslinux will auto-remove isolinux and isotools and the opposite).
Ok, thanks, if it's still problematic I can make special wheezy version for isotools.

Fred
User avatar
saintless
Posts: 3862
Joined: Sat 11 Jun 2011, 13:43
Location: Bulgaria

#120 Post by saintless »

Thanks, Fred!
New yad works in SID also. Wheezy and Jessie repository updated.

Toni
Post Reply