Hello all,
I'm unable to use particular video cd under self-compiled kernel
2.6.18. Reading first 8 megabytes goes well, and then - bulk of errors
and nothing. The same cd works perfectly under windows xp sp2 on the
same machine - system can mount it, read its contents and play it.
BUGGIE MONSTER WARNING: This optical drive is somehow.... special :D
Have you ever erased dvd-r or dvd+r medium? Well, now I have done it
too.... First, I put a dvd recorded half a year ago. It worked fine.
Once. Second time, the system was just unable to read anything from
the cd. I tried it on another machine - winxp said the dvd is empty,
but was unable to record anything there of course. Then, after some
study, I discovered each sector of the disc can be read one time, max.
two times - after this its contents is lost. What is interesting
hovewer, it does not actually happen all the time - sometimes
everything works just perfect, sometimes dvd is destroyed. Especially
when recording a dvd (what works always fine) I try to read it it gets
'erased' - nothing such was observed on windows, but I can't tell it
does not happen ever - just haven't investigated.
This cd is however bought, so I suspect it is not recorded - system
fails tests on the original, and on its copy (which has been
successfully made on another machine with nero). The symptomes are
always the same. However, on windows it just works fine.
The machine is AMD Athlon 64 3200+ 1GB RAM. The magic drive is LG
GSA-H20L (it's IDE). I've upgraded the firmware from 1.00 to 1.01,
because I thought it could help - it didn't.
Attachments - .config and two dmesgs. The first is what happens when I
try to use xine - first seconds (8mbytes) works fine, and then this
(note - the kernel was tainted then by ati's fglrx driver - that's why
i've done the second test). The second test is done without xwindows
(kernel is not tained though) - i've tried dd if=/dev/hda of=/dev/null
- that's how i now the first 8 megs work fine. It includes the startup
messages too. You can see I have reloaded cdrom driver with debug=1
there, hoping it will expose some usefull info - it seems it didn't
however.
I will be gratefull for your responses - especially to the vcd case,
because it is very repetitive (note - the problem is with this
particular vcd - i have other that works fine). The dvd mysterious is
I think too mysterious to investigate it without some..... ide packet
dump?
--
wixor
Ar Llu, 2006-10-16 am 20:53 +0200, ysgrifennodd wixor:
> (kernel is not tained though) - i've tried dd if=/dev/hda of=/dev/null
> - that's how i now the first 8 megs work fine. It includes the startup
> messages too. You can see I have reloaded cdrom driver with debug=1
> there, hoping it will expose some usefull info - it seems it didn't
> however.
Multimedia sectors on a VCD are not readable by dd or the filesystem
layer, you must use a specific vcd player app so the dd test fail is
expected. The xine one is more interesting. Is Xine being fooled by the
fact its a multisession disk of some form perhaps >?
On 10/17/06, Alan Cox <[email protected]> wrote:
> Multimedia sectors on a VCD are not readable by dd or the filesystem
> layer, you must use a specific vcd player app so the dd test fail is
> expected.
Hmm.... so this is more like audio CD with separate tracks and no data
track than like film on DVD? OK then, but how does xine tell kernel
"give me video tracks"? Is there some magic switch_to_video ioctl on
cdrom devices or what? And should the dd test fail with kernel errors
or just with dd saying 'the device is empty'? And why even hal doesn't
recognize the disc as video cd, but recognizes audio cds? (ok, it does
not recognize all video cds I have, but only this one is unplayable,
so it is probably lack of support, just asking to make sure).
>The xine one is more interesting. Is Xine being fooled by the
>fact its a multisession disk of some form perhaps >?
The xine does not display any errors. It just hangs and I have to kill
(-TERM is enough) it. Well..... i have no idea, really. The original
CD is bought, so i suspect it is classical old-style traditional disc
with no multi sessions. It is compatibile with VCD2.0 standard as far
as i remember, but I might be wrong as well. The copy of it I made is
made by nero express & windows xp on other machine, so I belive it is
bitwise exact (which is propably wrong, but... :D). The both original
and copy behave in the same way. Is there any way to dump the disc
structure?
One more - winxp shows files and directories on this linux-unplayable
vcd. Is it like there no directories in fact, and all this contents is
just file virtualization of the vcd tracks (but why the hell would be
autorun.inf and some executable !! file there?) or there is some data
track there, but it is non-standard or not supported? I can provide
you with nero image of the disc i have used during copying
(450mbytes), but i'm not broadband, and you propably also have some
better things to transmit, so....
Thanks for replies
--
wixor
Ar Maw, 2006-10-17 am 20:16 +0200, ysgrifennodd wixor:
> Hmm.... so this is more like audio CD with separate tracks and no data
> track than like film on DVD? OK then, but how does xine tell kernel
Yes. To make video fit the mpeg video stream is encoded without all the
error correction. There may or may not be multiple tracks, each is a
chapter in the film basically.
> "give me video tracks"? Is there some magic switch_to_video ioctl on
> cdrom devices or what? And should the dd test fail with kernel errors
There are ioctls for getting sectors in non data format (just like for
reading audio bitstreams to play them).
Now where it all gets weirder is that some forms of VCD (especially the
ones for philips short lived interactive stuff) have an ISO file system
on them but where sector numbers in the file system for video blocks
point to blocks that are not 2K data blocks but mpeg blocks that the
file system layer can't handle, so a VCD disk can appear mountable and
the like.
Alan
Am Mittwoch 18 Oktober 2006 00:38 schrieb Alan Cox:
> Now where it all gets weirder is that some forms of VCD (especially the
> ones for philips short lived interactive stuff) have an ISO file system
> on them but where sector numbers in the file system for video blocks
> point to blocks that are not 2K data blocks but mpeg blocks that the
> file system layer can't handle, so a VCD disk can appear mountable and
> the like.
Perhaps I confuse things with SVCDs, but the ISO part *should* actually be
mountable as it could contain valid data. I remeber authorinhg such disks
(ie, puttins some stuff into the ISO part), which worked nicely, ie. it plays
fine in standalone and eg. in Windows I could also access data on the ISO fs.
PS: Could you please check your clock. My mailer gets confused as your mails
seem to come 20 minutes from the future. ;-)
Cheers,
--
(?= =?)
//\ Prakash Punnoor /\\
V_/ \_V
On Tue, Oct 17, 2006 at 11:38:52PM +0100, Alan Cox wrote:
> Now where it all gets weirder is that some forms of VCD (especially the
> ones for philips short lived interactive stuff) have an ISO file system
> on them but where sector numbers in the file system for video blocks
> point to blocks that are not 2K data blocks but mpeg blocks that the
> file system layer can't handle, so a VCD disk can appear mountable and
> the like.
PSX1 CDs are often like that too. Video files are in mode2 sectors
but the filesystem is ISO9660. It's a mix of mjpeg-like and adpcm
instead of mpeg though.
Makes the FIBMAP restrictions annoying too.
OG.
On 10/18/06, Alan Cox <[email protected]> wrote:
> Now where it all gets weirder is that some forms of VCD (especially the
> ones for philips short lived interactive stuff) have an ISO file system
> on them but where sector numbers in the file system for video blocks
> point to blocks that are not 2K data blocks but mpeg blocks that the
> file system layer can't handle, so a VCD disk can appear mountable and
> the like.
OK, but this is still mountable only from windows....... Is this ISO
filesystem hidden somewhere, or what? And that still does not explain
errors from xine, and the 8 megs that i actually can read using dd.
All after all - even if this disc would contain totally unsupported
tracks, with absolutly weird data, kernel should recognize it and
report something like:
cdrom: there are no tracks on hda i can recognize
(or something like that). If the errors do happen, it means kernel
thinks he can read the data, and he actually can't, yes? Is here
anything I can do to improve support of this disc under linux, or is
this just another hell-knows-what thing, the kernel implementation is
ok, and this is only some m$-dontated extension that prevents us from
accessing this disc? Even if it is, shouldn't it be implemented if it
is possible?
Thanks
--
wixor
Ar Iau, 2006-10-19 am 18:35 +0200, ysgrifennodd wixor:
> OK, but this is still mountable only from windows....... Is this ISO
> filesystem hidden somewhere, or what? And that still does not explain
It isn't hidden it should be in the first session
> errors from xine, and the 8 megs that i actually can read using dd.
> All after all - even if this disc would contain totally unsupported
> tracks, with absolutly weird data, kernel should recognize it and
> report something like:
The different types aren't per track but per sector.
> this just another hell-knows-what thing, the kernel implementation is
> ok, and this is only some m$-dontated extension that prevents us from
> accessing this disc? Even if it is, shouldn't it be implemented if it
> is possible?
Actually all the VCD stuff is a mix of the Chinese government and
Philips. The specs are available although pricy.
The kernel provides the infrastructure for arbitary CD handling, but
provides only the file system support in kernel. It can't (and doesn't
appear to need to) handle VCD in kernel it provides the tools to Xine
and friends instead.
Alan
On 10/19/06, Alan Cox <[email protected]> wrote:
> It isn't hidden it should be in the first session
So tell me then, where is it if windows can see it, but linux can't?
> The kernel provides the infrastructure for arbitary CD handling, but
> provides only the file system support in kernel. It can't (and doesn't
> appear to need to) handle VCD in kernel it provides the tools to Xine
> and friends instead.
Well, no doubt it is only m$ who can have vcd player in the kernel -
the point is the CD does NOT get read, neither by xine nor by any
other known to me tool. I asked here because the kernel is the only
source of any messages addressing the unplayability of the CD - all
the others just hang - so it looks like a kernel issue, doesn't it?
And as far as I understand you, you are trying to tell me it is
everything OK with the kernel, and the problem lays somewhere else,
yes?
wixor
Ar Iau, 2006-10-19 am 19:39 +0200, ysgrifennodd wixor:
> the point is the CD does NOT get read, neither by xine nor by any
> other known to me tool. I asked here because the kernel is the only
> source of any messages addressing the unplayability of the CD - all
> the others just hang - so it looks like a kernel issue, doesn't it?
No. It looks to me like an application problem from what has been posted
so far. Its hard to be sure however. Can you try 2.6.18.1 which fixed
one or two problems with CD handling that might be enough to have
confused Xine.
OK, I'm sorry for troubling you - it is indeed a xine problem - i've
ripped the whole vcd with vcdxrip, and even the executable file - just
everything, and it worked fine. Thanks for paying attention.
--
wixor
20 Eki 2006 Cum 16:05 tarihinde, Alan Cox şunları yazmıştı:
> No. It looks to me like an application problem from what has been posted
> so far. Its hard to be sure however. Can you try 2.6.18.1 which fixed
> one or two problems with CD handling that might be enough to have
> confused Xine.
With 2.6.18.1 and with 10 different VCD's (all of them are original copy and
they work with my VCD player without a problem) dmesg flooded with following;
hdb: command error: status=0x51 { DriveReady SeekComplete Error }
hdb: command error: error=0x50 { LastFailedSense=0x05 }
ide: failed opcode was: unknown
ATAPI device hdb:
Error: Illegal request -- (Sense key=0x05)
Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
The failed "Read 10" packet command was:
"28 00 00 01 5e ec 00 00 02 00 00 00 00 00 00 00 "
end_request: I/O error, dev hdb, sector 359344
Buffer I/O error on device hdb, logical block 44918
hdb: command error: status=0x51 { DriveReady SeekComplete Error }
hdb: command error: error=0x50 { LastFailedSense=0x05 }
ide: failed opcode was: unknown
ATAPI device hdb:
Error: Illegal request -- (Sense key=0x05)
Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
The failed "Read 10" packet command was:
"28 00 00 01 5e ee 00 00 02 00 00 00 00 00 00 00 "
end_request: I/O error, dev hdb, sector 359352
Buffer I/O error on device hdb, logical block 44919
hdb: command error: status=0x51 { DriveReady SeekComplete Error }
hdb: command error: error=0x50 { LastFailedSense=0x05 }
ide: failed opcode was: unknown
ATAPI device hdb:
Error: Illegal request -- (Sense key=0x05)
Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
The failed "Read 10" packet command was:
"28 00 00 01 5e ec 00 00 02 00 00 00 00 00 00 00 "
end_request: I/O error, dev hdb, sector 359344
Buffer I/O error on device hdb, logical block 44918
Cheers
--
S.Çağlar Onur <[email protected]>
http://cekirdek.pardus.org.tr/~caglar/
Linux is like living in a teepee. No Windows, no Gates and an Apache in house!
Hi,
I am having problems reading VCDs under various Linux distros (Fedora Core 6, openSUSE 10.2, Slackware 11 with the 2.6 kernel), and while searching Google for a solution I found that this problem has been mentioned on the LKML list too. (http://lkml.org/lkml/2006/10/29/95)
I didn't see any responses after the post linked to above, so I'd like to add that I too get this problem and that I've tried with various VCDs and players. In previous versions of these distros I could just mount the VCD and copy the *.DAT files across; but in the current versions I can't even mount! dmesg gets flooded with errors such as the below:
hdc: command error: status=0x51 { DriveReady SeekComplete Error }
hdc: command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
ide: failed opcode was: unknown
ATAPI device hdc:
Error: Illegal request -- (Sense key=0x05)
Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
The failed "Read 10" packet command was:
"28 00 00 00 73 f2 00 00 01 00 00 00 00 00 00 00 "
end_request: I/O error, dev hdc, sector 118728
Buffer I/O error on device hdc, logical block 29682
hdc: command error: status=0x51 { DriveReady SeekComplete Error }
hdc: command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
ide: failed opcode was: unknown
ATAPI device hdc:
Error: Illegal request -- (Sense key=0x05)
Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
The failed "Read 10" packet command was:
"28 00 00 00 73 f3 00 00 03 00 00 00 00 00 00 00 "
end_request: I/O error, dev hdc, sector 118732
Buffer I/O error on device hdc, logical block 29683
Buffer I/O error on device hdc, logical block 29684
Buffer I/O error on device hdc, logical block 29685
hdc: command error: status=0x51 { DriveReady SeekComplete Error }
hdc: command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
ide: failed opcode was: unknown
ATAPI device hdc:
Error: Illegal request -- (Sense key=0x05)
Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
The failed "Read 10" packet command was:
"28 00 00 00 73 f2 00 00 02 00 00 00 00 00 00 00 "
end_request: I/O error, dev hdc, sector 118728
Buffer I/O error on device hdc, logical block 29682
Buffer I/O error on device hdc, logical block 29683
hdc: command error: status=0x51 { DriveReady SeekComplete Error }
hdc: command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
ide: failed opcode was: unknown
ATAPI device hdc:
Error: Illegal request -- (Sense key=0x05)
Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
The failed "Read 10" packet command was:
"28 00 00 00 73 f2 00 00 02 00 00 00 00 00 00 00 "
end_request: I/O error, dev hdc, sector 118728
Buffer I/O error on device hdc, logical block 29682
Buffer I/O error on device hdc, logical block 29683
hdc: command error: status=0x51 { DriveReady SeekComplete Error }
hdc: command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
Another link on the net (by the same poster as the LKML link above) mentioned that this problem appears *after* kernel 2.6.16. So I downgraded my Slackware kernel to 2.6.16 and sure enough the problem goes away. I haven't tried with the 2.6.19 kernel, so I can't confirm if that solves the problem or not.
Thanks,
Rakhesh
ps. Am not subscribed to this list. Please cc me any replies etc. Thanks.
____________________________________________________________________________________
Have a burning question?
Go to http://www.Answers.yahoo.com and get answers from real people who know.
09 Ara 2006 Cts 08:06 tarihinde, Rakhesh Sasidharan şunları yazmıştı:
> Hi,
>
> I am having problems reading VCDs under various Linux distros (Fedora Core
> 6, openSUSE 10.2, Slackware 11 with the 2.6 kernel), and while searching
> Google for a solution I found that this problem has been mentioned on the
> LKML list too. (http://lkml.org/lkml/2006/10/29/95)
See http://bugme.osdl.org/show_bug.cgi?id=7460
/ismail
Thanks. Found that after I posted to the list; have put my comments too there.
----- Original Message ----
From: Ismail Donmez <[email protected]>
To: Rakhesh Sasidharan <[email protected]>
Cc: [email protected]
Sent: Saturday, December 9, 2006 12:14:28 PM
Subject: Re: VCD not readable under 2.6.18
09 Ara 2006 Cts 08:06 tarihinde, Rakhesh Sasidharan ?unlar? yazm??t?:
> Hi,
>
> I am having problems reading VCDs under various Linux distros (Fedora Core
> 6, openSUSE 10.2, Slackware 11 with the 2.6 kernel), and while searching
> Google for a solution I found that this problem has been mentioned on the
> LKML list too. (http://lkml.org/lkml/2006/10/29/95)
See http://bugme.osdl.org/show_bug.cgi?id=7460
/ismail
____________________________________________________________________________________
Yahoo! Music Unlimited
Access over 1 million songs.
http://music.yahoo.com/unlimited
> I didn't see any responses after the post linked to above, so I'd like to add that I too get this problem and that I've tried with various VCDs and players. In previous versions of these distros I could just mount the VCD and copy the *.DAT files across; but in the current versions I can't even mount! dmesg gets flooded with errors such as the below:
>
> hdc: command error: status=0x51 { DriveReady SeekComplete Error }
> hdc: command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
> ide: failed opcode was: unknown
> ATAPI device hdc:
> Error: Illegal request -- (Sense key=0x05)
> Illegal mode for this track or incompatible medium -- (asc=0x64, ascq=0x00)
> The failed "Read 10" packet command was:
> "28 00 00 00 73 f2 00 00 01 00 00 00 00 00 00 00 "
Your system tried to read a Video data block. The usual cure for this
problem is to remove Gnome, or at least kill all the Gnome stuff and flip
to init level 3 then mount the cd from the command line.
The kernel is correctly reporting it was asked to do something stupid.
Older kernels would fail to report many of these errors due to a bug in
the ide-cd reporting logic.
Slackware 11 (2.6.18 kernel) didn't have GNOME installed. So if it is a GNOME thing, how come Slackware too was giving the same error?
Also, if previously the kernel wasn't reporting errors when asked to do stuff like this with VCDs, how come it was still able to mount the VCDs and allow playback/ copying of the files in it? Its ok if the kernel reports more errors than before, as long as it allows playback/ copying ...
Regards,
Rakhesh
ps. Please do cc me on replies to thread.
----- Original Message ----
From: Alan <[email protected]>
To: Rakhesh Sasidharan <[email protected]>
Cc: [email protected]; [email protected]
Sent: Saturday, December 9, 2006 5:21:20 PM
Subject: Re: VCD not readable under 2.6.18
Your system tried to read a Video data block. The usual cure for this
problem is to remove Gnome, or at least kill all the Gnome stuff and flip
to init level 3 then mount the cd from the command line.
The kernel is correctly reporting it was asked to do something stupid.
Older kernels would fail to report many of these errors due to a bug in
the ide-cd reporting logic.
____________________________________________________________________________________
Any questions? Get answers on any topic at http://www.Answers.yahoo.com. Try it now.
09 Ara 2006 Cts 15:21 tarihinde, Alan şunları yazmıştı:
> > I didn't see any responses after the post linked to above, so I'd like to
> > add that I too get this problem and that I've tried with various VCDs and
> > players. In previous versions of these distros I could just mount the VCD
> > and copy the *.DAT files across; but in the current versions I can't even
> > mount! dmesg gets flooded with errors such as the below:
> >
> >
> >
> > hdc: command error: status=0x51 { DriveReady SeekComplete Error }
> > hdc: command error: error=0x54 { AbortedCommand LastFailedSense=0x05 }
> > ide: failed opcode was: unknown
> > ATAPI device hdc:
> > Error: Illegal request -- (Sense key=0x05)
> > Illegal mode for this track or incompatible medium -- (asc=0x64,
> > ascq=0x00) The failed "Read 10" packet command was:
> > "28 00 00 00 73 f2 00 00 01 00 00 00 00 00 00 00 "
>
> Your system tried to read a Video data block. The usual cure for this
> problem is to remove Gnome, or at least kill all the Gnome stuff and flip
> to init level 3 then mount the cd from the command line.
Well my bet is xine-lib is buggy somehow as I can reproduce this bug with
kaffeine ( KDE media player ).
Regards,
ismail
> Well my bet is xine-lib is buggy somehow as I can reproduce this bug with
> kaffeine ( KDE media player ).
Under what situations - and does it play the disc despite the messages ?
09 Ara 2006 Cts 16:36 tarihinde, Alan şunları yazmıştı:
> > Well my bet is xine-lib is buggy somehow as I can reproduce this bug with
> > kaffeine ( KDE media player ).
>
> Under what situations - and does it play the disc despite the messages ?
Just pop in the VCD and try to play, Kaffeine either crashes or doesn't play
at all.
Regards,
ismail
09 Ara 2006 Cts 16:15 tarihinde, Ismail Donmez şunları yazmıştı:
> Well my bet is xine-lib is buggy somehow as I can reproduce this bug with
> kaffeine ( KDE media player ).
Same symptoms occur with mplayer also, dmesg flooded with warnings, what about
hal or KDE's cdpolling backend?
Cheers
--
S.Çağlar Onur <[email protected]>
http://cekirdek.pardus.org.tr/~caglar/
Linux is like living in a teepee. No Windows, no Gates and an Apache in house!
Infact, just inserting a CD is enough. No need for a media player to try and access the files. :)
The backend must be polling and trying to mount the disc upon insertion. Kernel 2.6.16 and before did that fine, but kernel 2.6.17 and above don't and give error messages. Which explains why downgrading the kernel solves the problem. (If it were a HAL or KDE/ GNOME problem then shouldn't downgrading the kernel *not* help?) Just thinking aloud ...
----- Original Message ----
From: S.?a?lar Onur <[email protected]>
To: Ismail Donmez <[email protected]>
Cc: Alan <[email protected]>; Rakhesh Sasidharan <[email protected]>; [email protected]; [email protected]
Sent: Saturday, December 9, 2006 8:09:05 PM
Subject: Re: VCD not readable under 2.6.18
09 Ara 2006 Cts 16:15 tarihinde, Ismail Donmez ?unlar? yazm??t?:
> Well my bet is xine-lib is buggy somehow as I can reproduce this bug with
> kaffeine ( KDE media player ).
Same symptoms occur with mplayer also, dmesg flooded with warnings, what about
hal or KDE's cdpolling backend?
Cheers
--
S.?a?lar Onur <[email protected]>
http://cekirdek.pardus.org.tr/~caglar/
Linux is like living in a teepee. No Windows, no Gates and an Apache in house!
____________________________________________________________________________________
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com
Rakhesh Sasidharan wrote:
> Infact, just inserting a CD is enough. No need for a media player to try and access the files. :)
>
> The backend must be polling and trying to mount the disc upon insertion. Kernel 2.6.16 and before did that fine, but kernel 2.6.17 and above don't and give error messages. Which explains why downgrading the kernel solves the problem. (If it were a HAL or KDE/ GNOME problem then shouldn't downgrading the kernel *not* help?) Just thinking aloud ...
I believe the problem is not that the kernel is providing an error, but
that it is providing an error which is taken as "DIDN'T WORK THIS TIME,
BUT TRY AGAIN" rather than a permanent error. Asking every
media-consious application to be rewritten is perhaps not the best
solution, either return another error, or return what application expect
(non-error but no data??)
Changes which break a large number of applications are probably ill-advised.
>
> ----- Original Message ----
> From: S.?a?lar Onur <[email protected]>
> To: Ismail Donmez <[email protected]>
> Cc: Alan <[email protected]>; Rakhesh Sasidharan <[email protected]>; [email protected]; [email protected]
> Sent: Saturday, December 9, 2006 8:09:05 PM
> Subject: Re: VCD not readable under 2.6.18
>
> 09 Ara 2006 Cts 16:15 tarihinde, Ismail Donmez ?unlar? yazm??t?:
>> Well my bet is xine-lib is buggy somehow as I can reproduce this bug with
>> kaffeine ( KDE media player ).
>
> Same symptoms occur with mplayer also, dmesg flooded with warnings, what about
> hal or KDE's cdpolling backend?
>
> Cheers
09 Ara 2006 Cts 19:23 tarihinde, Rakhesh Sasidharan ?unlar? yazm??t?:
> Infact, just inserting a CD is enough. No need for a media player to try
> and access the files. :)
>
> The backend must be polling and trying to mount the disc upon insertion.
> Kernel 2.6.16 and before did that fine, but kernel 2.6.17 and above don't
> and give error messages. Which explains why downgrading the kernel solves
> the problem. (If it were a HAL or KDE/ GNOME problem then shouldn't
> downgrading the kernel *not* help?) Just thinking aloud ...
But i cannot reproduce the problem that way, in my case dmesg flooded as soon
as somebody trying to _access_ to VCD. I disabled hal and closed KDE to test
and that problem no longer reproducible for me. So its really seems a
userspace problem and i think all of them (KDE's cdpolling backend, hal,
mplayer and xine-lib) has problems with kernels >= 2.6.17
--
S.?a?lar Onur <[email protected]>
http://cekirdek.pardus.org.tr/~caglar/
Linux is like living in a teepee. No Windows, no Gates and an Apache in house!
On Sat, 9 Dec 2006 09:23:32 -0800 (PST)
Rakhesh Sasidharan <[email protected]> wrote:
> Infact, just inserting a CD is enough. No need for a media player to try and access the files. :)
>
> The backend must be polling and trying to mount the disc upon insertion. Kernel 2.6.16 and before did that fine, but kernel 2.6.17 and above don't and give error messages. Which explains why downgrading the kernel solves the problem. (If it were a HAL or KDE/ GNOME problem then shouldn't downgrading the kernel *not* help?) Just thinking aloud ...
The old kernel erroneously failed to report errors in some cases so the
answer to that bit is a definite - no -. That side is a desktop problem.
The fact people are saying that in addition vcd players are breaking is a
bit more mysterious.
> BUT TRY AGAIN" rather than a permanent error. Asking every
> media-consious application to be rewritten is perhaps not the best
> solution, either return another error, or return what application expect
> (non-error but no data??)
Unfortunately nobody changed the behaviour just fixed bugs that were
hiding existing bad behaviour. Please take the polling mess up with the
HAL and KDE developers.
Alan
Oh, VCD players are breaking coz they are trying to mount the device and then access files from it. Mounting itself fails, and so the VCD players cant read any files ... that's what I think.
----- Original Message ----
From: Alan <[email protected]>
To: Rakhesh Sasidharan <[email protected]>
Cc: [email protected]; [email protected]; Ismail Donmez <[email protected]>; [email protected]
Sent: Sunday, December 10, 2006 4:44:42 AM
Subject: Re: VCD not readable under 2.6.18
On Sat, 9 Dec 2006 09:23:32 -0800 (PST)
Rakhesh Sasidharan <[email protected]> wrote:
> Infact, just inserting a CD is enough. No need for a media player to try and access the files. :)
>
> The backend must be polling and trying to mount the disc upon insertion. Kernel 2.6.16 and before did that fine, but kernel 2.6.17 and above don't and give error messages. Which explains why downgrading the kernel solves the problem. (If it were a HAL or KDE/ GNOME problem then shouldn't downgrading the kernel *not* help?) Just thinking aloud ...
The old kernel erroneously failed to report errors in some cases so the
answer to that bit is a definite - no -. That side is a desktop problem.
The fact people are saying that in addition vcd players are breaking is a
bit more mysterious.
____________________________________________________________________________________
Need a quick answer? Get one in minutes from people who know.
Ask your question on http://www.Answers.yahoo.com
Well yeah, same here. In your case HAL/ KDE was trying to access the VCD. In my case, I was trying to access the VCD by mounting it.
Had the error been one on Fedora/ openSUSE or distros like that, I would have blamed it on the desktop side. But in Slackware, I wasn't running KDE, and it doesn't even have GNOME or HAL -- so how then can it be one of those? I was trying to do a simple mount of the VCD (as I've always done over these years, to copy the .dat movie files across) -- and that itself failed!
----- Original Message ----
From: S.?a?lar Onur <[email protected]>
To: Rakhesh Sasidharan <[email protected]>
Cc: Ismail Donmez <[email protected]>; Alan <[email protected]>; [email protected]
Sent: Sunday, December 10, 2006 3:19:14 AM
Subject: Re: VCD not readable under 2.6.18
But i cannot reproduce the problem that way, in my case dmesg flooded as soon
as somebody trying to _access_ to VCD. I disabled hal and closed KDE to test
and that problem no longer reproducible for me. So its really seems a
userspace problem and i think all of them (KDE's cdpolling backend, hal,
mplayer and xine-lib) has problems with kernels >= 2.6.17
--
S.?a?lar Onur <[email protected]>
http://cekirdek.pardus.org.tr/~caglar/
Linux is like living in a teepee. No Windows, no Gates and an Apache in house!
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Just dropping a note that I booted into single user mode today and tried mounting the VCD -- and it worked! So I guess, yeah, the problem must be something to do with HAL etc. Some change to the kernel probably breaks something in HAL (just a guess) .......... whatever, am stuck now coz I can't play VCDs until this problem is fixed! Heh!
Rakhesh
----- Original Message ----
From: Alan <[email protected]>
To: Bill Davidsen <[email protected]>
Cc: Rakhesh Sasidharan <[email protected]>; [email protected]
Sent: Sunday, December 10, 2006 4:50:32 AM
Subject: Re: VCD not readable under 2.6.18
> BUT TRY AGAIN" rather than a permanent error. Asking every
> media-consious application to be rewritten is perhaps not the best
> solution, either return another error, or return what application expect
> (non-error but no data??)
Unfortunately nobody changed the behaviour just fixed bugs that were
hiding existing bad behaviour. Please take the polling mess up with the
HAL and KDE developers.
Alan
____________________________________________________________________________________
Yahoo! Music Unlimited
Access over 1 million songs.
http://music.yahoo.com/unlimited
Hi again!
I realize what I am going to post below is OT to this list, but I thought I'd mention it here anyways just in case some other user is having a similar problem and happens to follow this thread in the hope of finding a solution. I've tested this on Fedora Core 6 and openSUSE 10.2, both running GNOME, and it works.
As I mentioned in my previous mail, this VCDs not readable problem is not a kernel one. The newer kernels are more vocal about the errors, and refuse to mount the VCDs, but the problem is one of HAL-kernel interaction. To verify that, boot into single user mode, and its possible to mount VCDs without any problems.
As a temporary work-around for those who wish to watch VCDs and are unable to coz of this problem, all one needs to do is disable HAL from polling the CD/ DVD drive. I stumbled upon this idea by doing a "ps ax |grep hald" and noticing that there's an entry for "hald-addon-storage: polling /dev/hdc". I killed this process, and bingo! I could mount VCDs perfectly fine! Ofcourse, killing this process has the side effect that none of your CDs are automatically mounted any more (till the next reboot that is).
If you wish to permanently stop polling of the CD/ DVD drive, then you have to edit the HAL policy file for storage devices (located at "/usr/share/hal/fdi/policy/10osvendor/20-storage-methods.fdi" in the distros I mentioned above). In my case, this is how the first few lines of that file looked like:
<?xml version="1.0" encoding="UTF-8"?>
<deviceinfo version="0.2">
<device>
<match key="info.udi" string="/org/freedesktop/Hal/devices/computer">
<append key="info.callouts.add" type="strlist">hal-storage-cleanup-all-mo$
</match>
<match key="storage.media_check_enabled" bool="true">
<append key="info.addons" type="strlist">hald-addon-storage</append>
</match>
The key here is the part that goes as 'key="storage.media_check_enabled" bool="true"'. This is what tells HAL to keep polling the drives. What we want to do now is disable polling for CD/DVD drives, and so we add something like this *before* the above stuff:
<!-- Temporary hack to solve my VCD problem -->
<device>
<match key="storage.bus" string="ide">
<!-- <match key="storage.model" string="TSSTcorpCD/DVDW TS-L532M"> -->
<match key="block.device" string="/dev/hdc">
<merge key="storage.media_check_enabled" type="bool">false</merge>
</match>
<!-- </match> -->
</match>
</device>
In my case, the CD/ DVD drive is /dev/hdc. Which is why I am matching on that and changing its polling to false. Its also possible to match based on the model etc etc (in my case I've commented that out with <!-- to -->).
This is how the first few lines of the storage policies file looks, after the changes:
<?xml version="1.0" encoding="UTF-8"?>
<deviceinfo version="0.2">
<!-- Temporary hack to solve my VCD problem -->
<device>
<match key="storage.bus" string="ide">
<!-- <match key="storage.model" string="TSSTcorpCD/DVDW TS-L532M"> -->
<match key="block.device" string="/dev/hdc">
<merge key="storage.media_check_enabled" type="bool">false</merge>
</match>
<!-- </match> -->
</match>
</device>
<device>
<match key="info.udi" string="/org/freedesktop/Hal/devices/computer">
<append key="info.callouts.add" type="strlist">hal-storage-cleanup-all-mo$
</match>
<match key="storage.media_check_enabled" bool="true">
<append key="info.addons" type="strlist">hald-addon-storage</append>
</match>
After making the changes, restart HAL, and the problem is solved! :)
I figure readers of the LKML probably know all this stuff well; but I didn't, and it took me a bit of effort to figure what to do, and that's why I am posting it all here. A special mention to this link -- http://ubuntuforums.org/archive/index.php/t-64388.html -- from where I got an idea of where the policy files are located and what I could do.
Guess that's all for now.
Thanks,
Rakhesh
ps. Any replies, please cc me. I am not subscribed to this list.
----- Original Message ----
From: Alan <[email protected]>
To: Rakhesh Sasidharan <[email protected]>
Cc: [email protected]; [email protected]; Ismail Donmez <[email protected]>; [email protected]
Sent: Sunday, December 10, 2006 4:44:42 AM
Subject: Re: VCD not readable under 2.6.18
On Sat, 9 Dec 2006 09:23:32 -0800 (PST)
Rakhesh Sasidharan <[email protected]> wrote:
> Infact, just inserting a CD is enough. No need for a media player to try and access the files. :)
>
> The backend must be polling and trying to mount the disc upon insertion. Kernel 2.6.16 and before did that fine, but kernel 2.6.17 and above don't and give error messages. Which explains why downgrading the kernel solves the problem. (If it were a HAL or KDE/ GNOME problem then shouldn't downgrading the kernel *not* help?) Just thinking aloud ...
The old kernel erroneously failed to report errors in some cases so the
answer to that bit is a definite - no -. That side is a desktop problem.
The fact people are saying that in addition vcd players are breaking is a
bit more mysterious.
____________________________________________________________________________________
Any questions? Get answers on any topic at http://www.Answers.yahoo.com. Try it now.