Something's wrong on the kernels ide/atapi interface for cd-rom's. I
got a Asus CD-S520/A and it seems that I can't sweep from the beginning
to the end of the CD without getting a sense error. Commands like:
dd if=/dev/scd0 of=/file.iso
dd if=/dev/hdc of=/file.iso
give
---
Jun 28 19:28:33 deneb kernel: I/O error: dev 0b:00, sector 1159912
Jun 28 19:28:33 deneb kernel: I/O error: dev 0b:00, sector 1160064
(note)
the sectors do change if I change de medium, but on the same medium the
falty sectores are always the same. All the cd's are in top conditions
without a scratch and clean.
---
(with and without the scsi emulation) are pointless. On my former HP
9100c cdr/rw writer and Phillips 40x reader this didn't happen. What is
also curious is that original brand cd's (like a game or the maxim's cd
data catalog) is red 100% by this Asus cd reader. I thought is as not
an error with the interface but with cdrecord but after testing a cd
burned with nero (windows os) the same happened. If I change brands the
same happens too (tested TDK and Sony).
As a final note it seems that when mounted the cd can be accessed
perfectly (or so it seems).
I would write a patch/workaround myself but it seems that I can't
understand this behaviour.
Hoping that this report was useful.. best regards, Marco Ferra.
--- Debian GNU/Linux 3.0r1 ---
My configuration:
Motherboard: ASUS CUV4X
cat /proc/version
Linux version 2.4.18 (root@deneb) (gcc version 2.95.4 20011002 (Debian
prerelease)) #1 Mon Jun 23 00:27:00 WEST 2003
cat /proc/version
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 8
model name : Pentium III (Coppermine)
stepping : 3
cpu MHz : 701.626
cache size : 256 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca
cmov pat pse36 mmx fxsr sse
bogomips : 1399.19
cat /proc/modules
ipt_state 576 2
ipt_mac 640 0 (unused)
ipt_limit 960 0 (unused)
ipt_LOG 3072 0 (unused)
ip_conntrack_irc 2400 0 (unused)
ip_conntrack_ftp 3136 0 (unused)
ip_conntrack 12884 3 [ipt_state ip_conntrack_irc
ip_conntrack_ftp]
mga 102480 1
agpgart 12448 3
(note: scsi emulation is directly on the kernel image)
cat /proc/ioports
0000-001f : dma1
0020-003f : pic1
0040-005f : timer
0060-006f : keyboard
0080-008f : dma page reg
00a0-00bf : pic2
00c0-00df : dma2
00f0-00ff : fpu
0170-0177 : ide1
01f0-01f7 : ide0
02f8-02ff : serial(set)
0376-0376 : ide1
0378-037a : parport0
03c0-03df : vga+
03f6-03f6 : ide0
03f8-03ff : serial(set)
0cf8-0cff : PCI conf1
b400-b43f : Ensoniq ES1371 [AudioPCI-97]
b400-b43f : es1371
b800-b8ff : Realtek Semiconductor Co., Ltd. RTL-8139
b800-b8ff : 8139too
d000-d01f : VIA Technologies, Inc. UHCI USB (#2)
d000-d01f : usb-uhci
d400-d41f : VIA Technologies, Inc. UHCI USB
d400-d41f : usb-uhci
d800-d80f : VIA Technologies, Inc. Bus Master IDE
d800-d807 : ide0
d808-d80f : ide1
e400-e4ff : VIA Technologies, Inc. VT82C686 [Apollo Super ACPI]
e800-e80f : VIA Technologies, Inc. VT82C686 [Apollo Super ACPI]
cat /proc/iomem
00000000-0009fbff : System RAM
0009fc00-0009ffff : reserved
000a0000-000bffff : Video RAM area
000c0000-000c7fff : Video ROM
000f0000-000fffff : System ROM
00100000-13febfff : System RAM
00100000-002343e9 : Kernel code
002343ea-00292b77 : Kernel data
13fec000-13feefff : ACPI Tables
13fef000-13ffefff : reserved
13fff000-13ffffff : ACPI Non-volatile Storage
f8000000-f80000ff : Realtek Semiconductor Co., Ltd. RTL-8139
f8000000-f80000ff : 8139too
f8800000-f9efffff : PCI Bus #01
f8800000-f8ffffff : Matrox Graphics, Inc. MGA G400 AGP
f9000000-f9003fff : Matrox Graphics, Inc. MGA G400 AGP
f9f00000-fbffffff : PCI Bus #01
fa000000-fbffffff : Matrox Graphics, Inc. MGA G400 AGP
fc000000-fdffffff : VIA Technologies, Inc. VT8605 [ProSavage PM133]
ffff0000-ffffffff : reserved
cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: ASUS Model: CD-S520/A Rev: 2.0K
Type: CD-ROM ANSI SCSI revision: 02
On Sat, 28 Jun 2003, Marco Ferra wrote:
> Something's wrong on the kernels ide/atapi interface for cd-rom's. I
> got a Asus CD-S520/A and it seems that I can't sweep from the
> beginning to the end of the CD without getting a sense error.
> Commands like:
>
> dd if=/dev/scd0 of=/file.iso
> dd if=/dev/hdc of=/file.iso
>
> give
>
> ---
>
> Jun 28 19:28:33 deneb kernel: I/O error: dev 0b:00, sector 1159912
> Jun 28 19:28:33 deneb kernel: I/O error: dev 0b:00, sector 1160064
>
> (note) the sectors do change if I change de medium, but on the same
> medium the falty sectores are always the same. All the cd's are in
> top conditions without a scratch and clean.
Does this happen in the middle of a CD or always at the end? If tha
latter is the case have a look at cdrecords README.verify. Then you are
bitten by the TAO readahead bug. Try to burn a CD with -raw96r.
regards
Markus
Although it seems that all the cd's were recorded at TAO mode the read
error does happen randomly at the middle of the cd's. Or I think that
is on the middle. I will test this right away.
Nevertheless it doesn't justify the same error with a cd burned with nero.
Best regards
Marco
It is _always_ at the end. -raw96r doesn't seem to exist but reading
the README.verify file the -pad argument option was mentioned. Learning
from the manpage it seems that this option can be used to correct this
situation. I will get a blank cd tomorrow to try it. Thanks a lot.
Tell me just more one thing using the data contained in the cd's
recorded this way is bad? Or can it be used normally? (I always used
them and seemed OK).
Best regards
Marco
On Sat, 28 Jun 2003, Marco Ferra wrote:
> It is _always_ at the end. -raw96r doesn't seem to exist
Then you must have an old, or as Joerg would put it, ancient version.
> but reading the README.verify file the -pad argument option was
> mentioned. Learning from the manpage it seems that this option can be
> used to correct this situation.
I don't think so. -pad is normally only for audio cds, IIRC. You could
also try -dao, which will work as long as your burners firmware isn't
b0rked, as it was with my acer 2010.
> I will get a blank cd tomorrow to try it. Thanks a lot. Tell me just
> more one thing using the data contained in the cd's recorded this way
> is bad? Or can it be used normally? (I always used them and seemed
> OK).
The data is all good. And as lond as you don't use readcd or dd or
something similiar you won't realize the bug.
regards
Markus
You were right once more. I was using the cdrecord version 1.10 shiped
with Debian 3.0 (woody). After compiling the latest
cdrtools-2.00.3.tar.gz and grasping the manpage I could see the -raw96r
option. Now the final test is to burn the cd. But I'm sure that all be
fine.
Once again I can't thank you enough. This 'strange' error have been killing
my brain cells all day.
My best regards, Marco.