Invalid argument

Technical issues/questions of an intermediate or advanced nature.
User avatar
Ed_P
Contributor
Contributor
Posts: 3207
Joined: 06 Feb 2013, 22:12
Distribution: Cinnamon 3.2.2 64-bit ISO
Location: Western NY, USA

Re: Invalid argument

Post#31 by Ed_P » 19 Jan 2017, 07:44

I'm gonna wait for 3.2.3. :D
Ed

biotec
White ninja
White ninja
Posts: 23
Joined: 23 Jan 2014, 23:50
Distribution: slackware
Location: Oviedo

Re: Invalid argument

Post#32 by biotec » 22 Jan 2017, 14:50

Hi!

I've run in this (stupid?) bug as well. In my case, I found it when trying to build a package using slackbuild a script that had 'cp -a ...' in it. Since slackbuild scripts are set to stop with ANY error, I couldn't build the package and it made me to investigate too. I'll summarize my findings, in part similar to those described in this thread.

1) Simply "mkdir aa; cp -a aa bb" already shows the error, no need to do anything more complicated.
2) In spite of the error report, It looks like permissions are correctly copied (I can't find any permission difference).
3) If the destination is on a non-aufs filesystem, NO error is shown, whatever is the originating filesystem.
4) If the destination is on a aufs filesystem, the error message IS shown, whatever is the originating filesystem.
5) any 4.9.x kernel shows the situation, but kernels from earlier series do not.
6) going back to porteus 3.2.1 does not solve the situation (with kernel 4.9.x).
7) replacing the coreutils in 001-core.xzm with the latest slackware version (8.26) does not solve the situation either in my hands (using kernel 4.9.4 and 4.9.5). This is disapointing after reading what was reported by ncmprhnsbl.

At this stage, I tend to think it is somehow related to how aufs behave in kernel 4.9, but I do not know the intrincacies of kernels and aufs to go further. Maybe I could try changing the aufs config, but this kind of trial and error is painful (each kernel compilation takes more than 90 min in my computer).

User avatar
ncmprhnsbl
DEV Team
DEV Team
Posts: 880
Joined: 20 Mar 2012, 03:42
Distribution: 3.2.2-64bit xfce/openbox
Location: australia
Contact:

Re: Invalid argument

Post#33 by ncmprhnsbl » 22 Jan 2017, 23:03

good stuff.. :good:
biotec wrote:At this stage, I tend to think it is somehow related to how aufs behave in kernel 4.9
i'm tending this way too, based on other non-aufs systems not displaying ths error..
but whether its something in the aufs patch or some other kernel config is also beyond my present abilities...
Forum Rules : http://forum.porteus.org/viewtopic.php?f=35&t=44

User avatar
brokenman
Site Admin
Site Admin
Posts: 5503
Joined: 27 Dec 2010, 03:50
Distribution: Porteus v3.2rcX all desktops
Location: Brazil
Contact:

Re: Invalid argument

Post#34 by brokenman » 23 Jan 2017, 03:03

Yes. This behaviour started with the latest aufs patch. Not much we can do at this point.
How do i become super user?
Wear your underpants on the outside and put on a cape.

User avatar
Ed_P
Contributor
Contributor
Posts: 3207
Joined: 06 Feb 2013, 22:12
Distribution: Cinnamon 3.2.2 64-bit ISO
Location: Western NY, USA

Re: Invalid argument

Post#35 by Ed_P » 23 Jan 2017, 07:34

How are other Slackware distros handling the problem?
Ed

biotec
White ninja
White ninja
Posts: 23
Joined: 23 Jan 2014, 23:50
Distribution: slackware
Location: Oviedo

Re: Invalid argument

Post#36 by biotec » 23 Jan 2017, 08:34

Good new!

This problem was also reported on the aufs forum from last friday, and yesterday, the aufs developer sent a patch to the forum:

https://sourceforge.net/p/aufs/mailman/ ... /35619254/

I applied the patch on top of a previously compilation kernel tree (v.4.9.5) and recompiled ... and the problem seems gone. :Yahoo!:

As the patch only modifies one file, it is not necessary to do a compilation from scratch if you already had a full compilation tree, as I did, and recompilation only took 3 minutes. Only bzImage is modified (modules remain untouched).

Go to the link above to get the patch named a.patch.bz2 that is attached to the relevant post lnked above. At the time of writting, aufs4-standalone doesn't yet incorporate this patch, so if you compile a kernel 4.9.x, you must apply a.patch.bz2 after applying aufs4-standalone.

Regards to all.

User avatar
Ed_P
Contributor
Contributor
Posts: 3207
Joined: 06 Feb 2013, 22:12
Distribution: Cinnamon 3.2.2 64-bit ISO
Location: Western NY, USA

Re: Invalid argument

Post#37 by Ed_P » 23 Jan 2017, 16:23

Can the patch be made into an .xzm module? For us ISO booting users. :)
Ed

User avatar
Blaze
DEV Team
DEV Team
Posts: 1425
Joined: 28 Dec 2010, 11:31
Distribution: ⟰ Porteus 3.2 Cinnamon x86_64
Location: ☭ Russian Federation, Lipetsk region, Dankov
Contact:

Re: Invalid argument

Post#38 by Blaze » 23 Jan 2017, 17:14

Hi biotec.
Can you share your /porteus/base/000-kernel.xzm and /boot/syslinux/vmlinuz?
Thanks.
Linux porteus 4.13.3-porteus #1 SMP PREEMPT Sat Sep 23 18:22:13 x86_64 Intel(R) Core(TM) i5-6600K CPU @ 3.50GHz GenuineIntel GNU/Linux
MS-7A12 » [AMD/ATI] Tobago PRO [Radeon R7 360 / R9 360 OEM] (rev 81) » Vengeance LPX 16GB DDR4 K2 3200MHz C16

Bogomips
Full of knowledge
Full of knowledge
Posts: 2560
Joined: 25 Jun 2014, 15:21
Distribution: 3.2.2 Cinnamon & KDE5
Location: London

Re: Invalid argument

Post#39 by Bogomips » 23 Jan 2017, 19:52

Ed_P wrote:Can the patch be made into an .xzm module? For us ISO booting users. :)
From what I've gathered so far, kernel 4.9.0 just has to be rebuilt/(relinked?) with the patched module, so most likely all that would be needed would be the replacement of vmlinuz.
Linux porteus 4.4.0-porteus #3 SMP PREEMPT Sat Jan 23 07:01:55 UTC 2016 i686 AMD Sempron(tm) 140 Processor AuthenticAMD GNU/Linux
NVIDIA Corporation C61 [GeForce 6150SE nForce 430] (rev a2) MemTotal: 901760 kB MemFree: 66752 kB

biotec
White ninja
White ninja
Posts: 23
Joined: 23 Jan 2014, 23:50
Distribution: slackware
Location: Oviedo

Re: Invalid argument

Post#40 by biotec » 23 Jan 2017, 22:52

As Bogomips states, only vmlinuz/bzImage is modified by the patch. However, I am compiling my own kernels with different settings that fit my needs, not Neko kernels, so my kernel cannot be used alone as a replacement. Neko work releasing kernels is of first order and of great benefit for the Porteus comunity, so I prefer not to polute the forum with a non-compatible kernel. He surely will release a patched kernel soon (once you have a compilation tree, applying the patch and recompiling takes about 5 minutes).

For those compiling their own kernels: let's say you have downloaded a.patch from the link in my previous mail, and let's you decompressed and saved it in your home directory, and let's say you have the kernel compilation tree at the usual place:

cd /usr/src/linux
patch -p1 < ~/a.patch
make

and in about 3 minutes, you have the correct bzImage 8)

Voilà.

biotec
White ninja
White ninja
Posts: 23
Joined: 23 Jan 2014, 23:50
Distribution: slackware
Location: Oviedo

Re: Invalid argument

Post#41 by biotec » 23 Jan 2017, 23:00

I must update myself: as expected, Neko has already updated his work! :beer:

User avatar
brokenman
Site Admin
Site Admin
Posts: 5503
Joined: 27 Dec 2010, 03:50
Distribution: Porteus v3.2rcX all desktops
Location: Brazil
Contact:

Re: Invalid argument

Post#42 by brokenman » 24 Jan 2017, 02:16

Thanks for the info Biotec.

Junjiro will update the aufs tree within several days. I will release an update for Porteus after this. Others may update and test with Neko's latest kernel before then. Good job!
How do i become super user?
Wear your underpants on the outside and put on a cape.

neko
DEV Team
DEV Team
Posts: 930
Joined: 09 Feb 2013, 09:55
Distribution: APorteus-FVWM-ja-x86_64.iso
Location: japan

Re: Invalid argument

Post#43 by neko » 24 Jan 2017, 04:17

@biotec
Thank you very much for your information.
I updated new v4.9.5 one that was adapted by the additional patch.

Excuse me for delay responce.
I cleanup a kernel build work directory everytimes in order to keep space in my poor PC.
So it took 3 hours to rebuild kernel (32Bit & 64Bit).

Thanks.

Post Reply