2.6.1-love3 "Watch the love!"

Message

OneOfOne · Post by **OneOfOne** » Tue Jan 13, 2004 2:30 am

antisthenes wrote:I know how to use loop-devices. I already use the loop-AES fix, but I don't need AES or any sort of encryption. All I want is my loopback devices, and without the loop-AES module replacing the normal one, I get something like:
Code: Select all
ioctl: LOOP_SET_FD: Invalid argument
(can't remember the exact error).

I want the loop devices to work out of the box, and that's what I was asking.

Thanks, though.

something got really really messed up in the loop.c code, that's why loop-aes works (it makes its own loop module).

peace

charlieg · Post by **charlieg** » Tue Jan 13, 2004 3:20 am

jassi wrote:Hey just a suggestion.. wud the following patches be good enough to get implemented in the love sources ??
Internet Speedy Patch

They are not kernel source patches, they are scripts to demonstrate a technique. This is something to be implemented outside of kernel space - ie in user space.

i_hate_your_os · Post by **i_hate_your_os** » Tue Jan 13, 2004 3:26 am

antisthenes wrote:Does anyone know if loop devices work on this one? I'm forced to use a random thing I got off loop-aes.sf.net to do it, because otherwise I get ioctl errors. Someone know?

Just tried it. Still no "love". BUT, upgraded util-linux to 2.12-r4 and, guess what?

All better.

[EDIT]

ack! cancel that... I messed up the kernels in /boot. Even with util-linux-2.12-r4, no loopy love... will have to try mm2 sometime and see if it works there...
[/EDIT]

jrz · Post by **jrz** » Tue Jan 13, 2004 5:03 am

hi, when i try to emerge it keeps trying to download the patch file even though i already have it

PrakashP · Post by **PrakashP** » Tue Jan 13, 2004 5:14 am

Delete the digest file.

jrz · Post by **jrz** » Tue Jan 13, 2004 6:10 am

doh!

thx

sawanv · Post by **sawanv** » Tue Jan 13, 2004 11:55 am

Hey man...you guys rock!! This is a very good effort and kudos to ya'all!! Love sources forever!!

But do manage to get SOME sleep.....

Sawan

ed0n · Post by **ed0n** » Tue Jan 13, 2004 1:14 pm

Wedge_ wrote:
ed0n wrote:Can anyone tell me why when I am trying to rmmod a module its segfaulting
It is segfaulting with 2.6.1-love1/2/3 and with module-init-tools-0.9.15_pre4/
module-init-tools-3.0_pre5.
Does anything appear in dmesg? Do you have module unloading support in the kernel? (sorry, had to ask )

Wedge_ :

first question: no I think.
second question: yes .

Proton · Post by **Proton** » Tue Jan 13, 2004 1:41 pm

How's the supermount situation that packetios reported back there? I'm thinking of upgrading, but without it working I'd rather wait... Does it really oops?

BTW, I've noticed there's now version 2.0.3b of the supermount-ng patch, but I can't see what's new because sorceforge is down for maintenance...

sp1nm0nkey · Post by **sp1nm0nkey** » Wed Jan 14, 2004 12:31 am

Yes, supermount is seriously b0rked.

Code: Select all

Unable to handle kernel NULL pointer dereference at virtual address 00000000
 printing eip:
d19c16f3
*pde = 00000000
Oops: 0000 [#3]
CPU:    0
EIP:    0060:[<d19c16f3>]    Tainted: PF  VLI
EFLAGS: 00010297
EIP is at parse_options+0xf0/0x1e3 [supermount]
eax: 00000000   ebx: cef27017   ecx: 00000005   edx: d19c25a3
esi: ce3d5e7c   edi: c87f9f00   ebp: c87fac00   esp: ce3d5e6c
ds: 007b   es: 007b   ss: 0068
Process mount (pid: 5663, threadinfo=ce3d4000 task=c1e7a080)
Stack: cef27017 d19c47a0 ce3d5e7c c87fac00 cef27004 cef27016 cff732d0 000000d0
       c0154452 c0386c6c 00000042 cff6bc68 c87fac00 c87fac00 fffffff4 c87f9f00
       d19c18fd 00000000 c87fac00 d19c45c0 fffffff4 c87fac00 00000000 fffffff4
Call Trace:
 [<c0154452>] alloc_super+0xc2/0x19d
 [<d19c18fd>] supermount_read_super+0x3b/0x1e3 [supermount]
 [<c0155136>] get_sb_nodev+0x64/0x92
 [<c01552b9>] do_kern_mount+0x8e/0x100
 [<d19c18c2>] supermount_read_super+0x0/0x1e3 [supermount]
 [<c016807f>] do_add_mount+0x90/0x17b
 [<c01683a9>] do_mount+0x141/0x19c
 [<c01681fe>] copy_mount_options+0x94/0xfe
 [<c0168738>] sys_mount+0xa0/0xe2
 [<c036ea4b>] syscall_call+0x7/0xb

Code: 04 e8 51 6d 84 ee 85 c0 89 c2 2e 0f 85 62 ff ff ff 8b 44 24 0c 89 07 e9 61  ff ff ff c7 07 01 00 00 00 e9 56 ff ff ff 8b 44 24 44 <80> 38 00 74 1f 89 04 24  e8 b9 fd ff ff 31 d2 85 c0 89 47 0c 0f

This is with the options "dev=/dev/cdroms/cdrom0,--,users" . Also, whenever I use bash to go /mnt/cdro<tab><tab> it freezes for a few minutes, and this is with sysctl -w dev.cdroms.autoclose=0. When the fs finally mounts and I access it, the directory appears empty (YES, there is a cd-rom in the drive

) So, yes, this is a problem.

Evil Dark Archon · Post by **Evil Dark Archon** » Wed Jan 14, 2004 6:00 am

i hope supermount gets fixed in the next love revision, luckily i found the bug before i changed my fstab, so i will avoid supermount until this bug gets fixed.

c0bblers · Post by **c0bblers** » Wed Jan 14, 2004 6:21 am

Hi,
Here's a problem that had me pulling much hair with love-sources so I thought I'd share....Packet writing. I've been trying this on and off for ages but never got it working, but when having another go with love-sources I hit a new problem. It kernel ooped on me when mounting the pktcdvd device. Long story short, packet writing doesn't like any IO scheduler other then the anticipatory scheduler, dunno why I'm not an expert. Setting the scheduler for the cdrw drive to anticipatory fixed this problem. As these two features are both included in the sources, perhaps this should be highlighted as an issue? Or is this a quirk specific to my system? Anyone else tried packet writing?

BTW The problem I'd always had with my Acer 20x10x40 drive was that it couldn't read udf files properly, never mind write the things....files kept getting the ends cuts off them. Turned out the be the drives firmware, so now after umteen million attempts I appear to have packet writing working *crosses fingers*.

Cheers,
James

sawanv · Post by **sawanv** » Wed Jan 14, 2004 7:24 am

Hey guys I got packet writing working!!!

Not that I did anything personally

...just followed the instructions at packet-cd.sourceforge.net and it worked fine.

Of course this was after upgrading to 2.6.1-love3. At present it only works on cdrws'...which is a pity but I am sure that will be fixed soon.

@c0bblers:

Check the forum thread for 2.6.1-rc2-love.....Wedge gave me instructions on how to set it up for love sources. I have a LITE-ON CDRW and it worked fine for me.

Lemme know...

Sawan

c0bblers · Post by **c0bblers** » Wed Jan 14, 2004 9:23 am

Hi,
I have packet writing working (though haven't tested it that much), I was just relating my experiences with the io scheduler problem I ran into. I've also added the pktsetup command to local.start and local.stop and added an entry to fstab. Mounting a cdrw is now two mouse clicks away with Nautilus....nice

Cheers,
James

sawanv · Post by **sawanv** » Wed Jan 14, 2004 9:40 am

@cobblers

Yeah I added it in fstab too.....was wondering how to rig pktsetup and you gave me a hint...local.start and stop. So thats what those scripts are for.. ...

But I am not sure that is the right place for them.....suppose you setup during boot and copy some files and then you want to burn some stuff (on an ordinary cd i.e. not cdrw). Then you will probably encounter some problem until you tear down the interface via "pktsetup -d".

Will experiment and let you know.

Setting the scheduler for the cdrw drive to anticipatory fixed this problem.

How the heck do you do that?? I thought there was just one scheduler and once it was set (via elevator= at boot), thats it. Can you set different schedulers for different kinds of operations?

Sawan

Tiggar · Post by **Tiggar** » Wed Jan 14, 2004 10:02 am

Hello,

thanks for these patches. The new kernel works fine. But now I have some serious problems. And I want to know if it possible that this relates to the knew kernel.

1st. couldn't compile dc-gui and fluxbox
2nd. old version of fluxbox works, but gnome 2.4 doesn't do his job, complete systems freezes after logging in.

I appologize for the bad word constellation and hope you can help me.
Thank you,

Tiggar

PrakashP · Post by **PrakashP** » Wed Jan 14, 2004 11:17 am

mm3 is out.

ett_gramse_nap · Post by **ett_gramse_nap** » Wed Jan 14, 2004 11:58 am

PrakashKC wrote:mm3 is out.

Just wonder how long until next love-release...

c0bblers · Post by **c0bblers** » Wed Jan 14, 2004 12:08 pm

sawanv wrote: Then you will probably encounter some problem until you tear down the interface via "pktsetup -d".

Works ok for me.....maybe its not the best way to do it.....I might fiddle around with a script of some sort if I can be bothered....

sawanv wrote:
Setting the scheduler for the cdrw drive to anticipatory fixed this problem.
How the heck do you do that?? I thought there was just one scheduler and once it was set (via elevator= at boot), thats it. Can you set different schedulers for different kinds of operations?

You can change the scheduler individually for each drive via sysfs thanks to Nick Piggin's runtime selectable I/O scheduler...for example, I (now) leave the default scheduler on (i.e. I don't pass anything at boot) and switch on cfq scheduling for my hard disk in local.start

echo cfq > /sys/block/hda/queue/io_scheduler

Wedge_ · Post by **Wedge_** » Wed Jan 14, 2004 1:05 pm

c0bblers wrote:It kernel ooped on me when mounting the pktcdvd device. Long story short, packet writing doesn't like any IO scheduler other then the anticipatory scheduler, dunno why I'm not an expert. Setting the scheduler for the cdrw drive to anticipatory fixed this problem. As these two features are both included in the sources, perhaps this should be highlighted as an issue? Or is this a quirk specific to my system? Anyone else tried packet writing?

Ah, that's what was doing it. Thanks, I'd run into the same problem a couple of times. I'd guess it must be something wrong with the packet-writing code rather than CFQ, but I don't know for sure. At least it's nice and easy to change the scheduler

jay · Post by **jay** » Wed Jan 14, 2004 3:38 pm

How about hacking prism54 cvs drivers and swsusp2 (software suspending next generation) into love sources? Those patches apply cleanly.

Wedge_ · Post by **Wedge_** » Wed Jan 14, 2004 4:05 pm

I think the prism54 drivers are on the todo list for the next release.

SPY_jmr1 · Post by **SPY_jmr1** » Wed Jan 14, 2004 4:07 pm

I'm just double checking, but...

The next release (2.6.1-love4?) will be jumped to mm3, correct?

I only ask, because the mm3 changelog indicates a few keyboard patches/fixes for my laptop...

Spy

Wedge_ · Post by **Wedge_** » Wed Jan 14, 2004 4:26 pm

I would imagine so, the releases are usually pretty up to date

mondauge · Post by **mondauge** » Wed Jan 14, 2004 4:34 pm

Hi, I have successfully compiled and booted the current love-sources. They are working great so far, but I have encountered a big problem: I can neither mount cds nor dvds. My system is keeping me telling this:

Code: Select all

unable to mount: wrong filesystem type, illegal options, superblock of /dev/cdroms/cdrom0 damaged or too many filesystems mounted

(the above message is translated because I have set LANG to de_DE@euro)

dmesg reports the following when mount a cd in my DVD-ROM:

Code: Select all

hdc: packet command error: status=0x51 { DriveReady SeekComplete Error }
hdc: packet command error: error=0xb0
hdc: request sense failure: status=0x51 { DriveReady SeekComplete Error }
hdc: request sense failure: error=0xb0LastFailedSense 0x0b
Unable to identify CD-ROM format.

so, does anyone have a clue how to fix this? I didn't have these problems with development-sources-2.6.0 and I would regret if I had to switch back

mondauge

2.6.1-love3 "Watch the love!"

Packet writing WORKS!!!!!

some emerging problems after switching to 2.6.1-love3