I have made a new version of the packet patch available here: http://w1.894.telia.com/~u89404340/patches/packet/packet-2.4.18-pre4-2.patch... Changes since the 2.4.18-pre2 version: packet-2.4.18-pre4-2: - Don't do unnecessary work in pkt_release_dev and pkt_open. Unless it is the first open/last close, very little needs to be done. - Don't flush the cache on last close if the device was only opened for reading. - When opening/closing the packet block device, call blkdev_get and blkdev_put on the underlying cdrom device. This makes drive locking work automatically, and also gives better handling if the tray is open or the drive is not ready when you try to mount a filesystem. packet-2.4.18-pre4: - Fixed kernel hangs after I/O errors. - Make sure not to overflow the "read speed" value in pkt_set_speed(). packet-2.4.18-pre3: - Made it possible to set up/tear down the pktcdvd device even if there is no disc in the drive. - Don't lock the drive door at setup time, only keep the door locked while the device is opened for read or write. -- Peter Osterlund - petero2@telia.com http://w1.894.telia.com/~u89404340
Hello Peter.
Yesterday i have download your new version of packet writing patch
(packet-2.4.18-pre4-2.patch.bz2) and have applied it to kernel source tree
(pure 2.4.17) with no rejects.
Have successfully compile kernel (UDF driver is built in, pktcdvd layer is
accessable as loadable module) with UDF write support.
Reboot ..
rm /dev/pktcdvd0
mknod /dev/pktcdvd0
/./sbin/insmod /lib/modules/2.4.17/kernel/drivers/block/pktcdvd.o
pktsetup /dev/pktcdvd0 /dev/cdrom
# Try to format CDRW disk with UDF fs.
cdrwtool -d /dev/cdrom -q
--------- cdrwtool responce ----------
using device /dev/cdrom
1312KB internal buffer
setting write speed to 12X
Settings for /dev/cdrom:
Fixed packets, size 32
Mode-2 disc
I'm going to do a quick setup of /dev/cdrom. The disc is going to be blanked and formatted with
one big track. All data on the device will be lost!! Press CTRL-C to cancel now.
Initiating quick disc blank
Disc capacity is 275744 blocks (551488KB/538MB)
Formatting track
Writing UDF structures to disc
7200 (EET) /-120
--------- end ---------
Kernel messages from /var/log/messages :
Jan 23 14:49:19 my kernel: pktcdvd: v0.0.2k 09/05/2001 Jens Axboe (axboe@suse.de)
Jan 23 14:50:20 my kernel: pktcdvd: writer hdc sucessfully registered
As you can see formatting with cdrwtool takes about 20 minutes and not finished
(CDRW LED backs to normal light (yellow); cdrwtool stop responding and lock the CDRW drive).
On reboot i also have see next messages on screen (interrupt loss):
Jan 23 15:30:22 my kernel: hdc: lost interrupt
Jan 23 15:31:03 my last message repeated 4 times
Jan 23 15:32:04 my last message repeated 6 times
....
After reboot Windows DirectCD (Linux can't process disk) recognize filesystem disk of CDRW disk as
.. CDFS.
The same situation with packet-2.4.17.patch for 2.4.17 kernel.
Whats up ?????
I don't wanna use Windows anymore to format CDRW disk .. but
inspite of that DirectCD support for my drive is too expensive (~200 Mbytes) this program
has stability and usability features (for example quick disk UDF formatting takes ~ 3 minutes).
Mail list related to linux-udf project (which contain udf tools) seems not worked.
I use:
software: RedHat 7.2 (patched kernel 2.4.17, gcc-2.95-3) .
hardware: IDE Mitsumi CDRW 4805 (4x8x32), AMD K6-2 300 Mhz.
Please help me.
Thanks.
Regards,
Sergiy Kudryk.
P.S. I attach kernel configuration file.
--- Peter Osterlund
I have made a new version of the packet patch available here:
http://w1.894.telia.com/~u89404340/patches/packet/packet-2.4.18-pre4-2.patch...
Changes since the 2.4.18-pre2 version:
packet-2.4.18-pre4-2:
- Don't do unnecessary work in pkt_release_dev and pkt_open. Unless it is the first open/last close, very little needs to be done.
- Don't flush the cache on last close if the device was only opened for reading.
- When opening/closing the packet block device, call blkdev_get and blkdev_put on the underlying cdrom device. This makes drive locking work automatically, and also gives better handling if the tray is open or the drive is not ready when you try to mount a filesystem.
packet-2.4.18-pre4:
- Fixed kernel hangs after I/O errors.
- Make sure not to overflow the "read speed" value in pkt_set_speed().
packet-2.4.18-pre3:
- Made it possible to set up/tear down the pktcdvd device even if there is no disc in the drive.
- Don't lock the drive door at setup time, only keep the door locked while the device is opened for read or write.
-- Peter Osterlund - petero2@telia.com http://w1.894.telia.com/~u89404340
__________________________________________________ Do You Yahoo!? Send FREE video emails in Yahoo! Mail! http://promo.yahoo.com/videomail/
On Wed, 23 Jan 2002, Sergiy Kudryk wrote:
As you can see formatting with cdrwtool takes about 20 minutes and not finished (CDRW LED backs to normal light (yellow); cdrwtool stop responding and lock the CDRW drive).
On reboot i also have see next messages on screen (interrupt loss):
Jan 23 15:30:22 my kernel: hdc: lost interrupt Jan 23 15:31:03 my last message repeated 4 times Jan 23 15:32:04 my last message repeated 6 times ....
A wild guess. If DMA is enabled on the CDRW, maybe turning it off will help: hdparm -d 0 /dev/hdc
After reboot Windows DirectCD (Linux can't process disk) recognize filesystem disk of CDRW disk as .. CDFS.
What does this mean? Was the disc correctly formatted even though cdrwtool hang?
I don't wanna use Windows anymore to format CDRW disk .. but inspite of that DirectCD support for my drive is too expensive (~200 Mbytes)
Are you referring to the fact that UDF formatted discs can not hold as much data as ISO9660 discs? This is true for linux as well.
this program has stability and usability features (for example quick disk UDF formatting takes ~ 3 minutes).
Does anyone know how the quick formatting in DirectCD works? Will a quick-formatted disc work with the current linux packet writing module, or is special support needed?
P.S. I attach kernel configuration file.
You have scsi emulation compiled as a module, but don't use it in the above example. I think someone reported that his problems were solved when he started to use scsi emulation. -- Peter Osterlund - petero2@telia.com http://w1.894.telia.com/~u89404340
Peter Osterlund wrote:
On Wed, 23 Jan 2002, Sergiy Kudryk wrote:
As you can see formatting with cdrwtool takes about 20 minutes and not finished (CDRW LED backs to normal light (yellow); cdrwtool stop responding and lock the CDRW drive).
On reboot i also have see next messages on screen (interrupt loss):
Jan 23 15:30:22 my kernel: hdc: lost interrupt Jan 23 15:31:03 my last message repeated 4 times Jan 23 15:32:04 my last message repeated 6 times ....
A wild guess. If DMA is enabled on the CDRW, maybe turning it off will help: hdparm -d 0 /dev/hdc
After reboot Windows DirectCD (Linux can't process disk) recognize filesystem disk of CDRW disk as .. CDFS.
What does this mean? Was the disc correctly formatted even though cdrwtool hang?
I don't wanna use Windows anymore to format CDRW disk .. but inspite of that DirectCD support for my drive is too expensive (~200 Mbytes)
Are you referring to the fact that UDF formatted discs can not hold as much data as ISO9660 discs? This is true for linux as well.
this program has stability and usability features (for example quick disk UDF formatting takes ~ 3 minutes).
Does anyone know how the quick formatting in DirectCD works? Will a quick-formatted disc work with the current linux packet writing module, or is special support needed?
I'd guess they'll format only part of the track and format the rest on demand. I think I've spotted something like this in the UDF specs saying it is possible. Moreover, I've creaded a 160MB fs, put it on the RW, played with it a bit (put some files on it) and linux was reporting 43MB free. When I took this disc to a windows friend, his DirectCD claimed there is ~390MB free on the disk, so it seems it is that way. BTW: I was also testing that 512sec/packet on his machine and he was reading it without problems, although windows refused to write on that disc. Nenik
this program has stability and usability features (for example quick disk UDF formatting takes ~ 3 minutes).
Does anyone know how the quick formatting in DirectCD works? Will a quick-formatted disc work with the current linux packet writing module, or is special support needed?
Well, some drives support background formating. (Which is required for CD-MRW)
I'd guess they'll format only part of the track and format the rest on demand. I think I've spotted something like this in the UDF specs saying it is possible.
Moreover, I've creaded a 160MB fs, put it on the RW, played with it a bit (put some files on it) and linux was reporting 43MB free. When I took this disc to a windows friend, his DirectCD claimed there is ~390MB free on the disk, so it seems it is that way.
You can also extend the track (which would of course require futzing with the UDF structures as well)
BTW: I was also testing that 512sec/packet on his machine and he was reading it without problems, although windows refused to write on that disc.
Wonder if it correctly parses the sparing table *heh* (Reading is (theoretically) transparent no matter what the packet size) Ben
Ben Fennema wrote:
BTW: I was also testing that 512sec/packet on his machine and he was reading
it without problems, although windows refused to write on that disc.
Wonder if it correctly parses the sparing table *heh* (Reading is (theoretically) transparent no matter what the packet size)
Hmm, it can't be that transparent as I can read this RW in my writer but not in my CDROM (which handles 32spp discs without problems). And I've even tried to tune mkudf for this (it stores packet size somewhere on the disc).
Ben
-- To unsubscribe, e-mail: packet-writing-unsubscribe@suse.com For additional commands, e-mail: packet-writing-help@suse.com
On Thu, Jan 24, 2002 at 08:18:35AM +0100, Petr Nejedly wrote:
Ben Fennema wrote:
BTW: I was also testing that 512sec/packet on his machine and he was reading
it without problems, although windows refused to write on that disc.
Wonder if it correctly parses the sparing table *heh* (Reading is (theoretically) transparent no matter what the packet size)
Hmm, it can't be that transparent as I can read this RW in my writer but not in my CDROM (which handles 32spp discs without problems). And I've even tried to tune mkudf for this (it stores packet size somewhere on the disc).
The Linux UDF code does sector translations if it figures out the CDROM drive doesn't do fixed packet reads, so if your CDROM test is the ability to read fixed packet written CDRW's, it might not be accurate. (but it only handles 32spp packets). Of course, I've seen writters get it wrong when reading, so ya never know.. Ben
Hello I can't get stable UDF writing results using IDE CDRW drive with SCSI emulation. I plays with different mount options: sync, async, atime, noatime, .. but no result. Also tries enable/disable UDF write caching .. Here is error messages from /var/log/messages: Jan 27 20:37:55 my kernel: Ignoring read error on sector:49768 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49772 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49776 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49780 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49784 Jan 27 20:37:55 my kernel: I/O error: dev 0b:00, sector 58352 Jan 27 20:37:55 my kernel: I/O error: dev 0b:00, sector 58352 Jan 27 20:37:55 my kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000034 Jan 27 20:37:55 my kernel: printing eip: Jan 27 20:37:55 my kernel: c0173c07 Jan 27 20:37:55 my kernel: *pde = 00000000 Jan 27 20:37:55 my kernel: Oops: 0000 Jan 27 20:37:55 my kernel: CPU: 0 Jan 27 20:37:55 my kernel: EIP: 0010:[udf_add_entry+2407/3044] Not tainted Jan 27 20:37:55 my kernel: EIP: 0010:[<c0173c07>] Not tainted Jan 27 20:37:55 my kernel: EFLAGS: 00010246 Jan 27 20:37:55 my kernel: eax: 00000000 ebx: c1ac2460 ecx: 000000d8 edx: 000000b0 Jan 27 20:37:55 my kernel: esi: c1ac2460 edi: c4811f50 ebp: c4811ec4 esp: c4811b48 Jan 27 20:37:55 my kernel: ds: 0018 es: 0018 ss: 0018 Jan 27 20:37:55 my kernel: Process kdeinit (pid: 1241, stackpage=c4811000) Jan 27 20:37:55 my kernel: Stack: c1ac2460 cb9e67c0 c4811f50 c4811ba0 00000000 c4811ba4 c4811ba8 00000000 Jan 27 20:37:55 my kernel: 00000000 00000000 00000000 00000000 00000028 0000002c 00000000 00000000 Jan 27 20:37:55 my kernel: 00000000 ceee3000 00002f5c 00000054 00000000 00000000 00000000 00000000 Jan 27 20:37:55 my kernel: Call Trace: [udf_add_aext+1051/1248] [getblk+24/64] [bread+34/108] [udf_tread+52/56] [udf_fileident_read+546/1016] Jan 27 20:37:55 my kernel: Call Trace: [<c0171827>] [<c012fd3c>] [<c012ff2a>] [<c0179e9c>] [<c0179a3e>] Jan 27 20:37:55 my kernel: [getblk+24/64] [bread+34/108] [udf_find_entry+726/1204] [udf_release_data+14/20] [udf_find_entry+1186/1204] [udf_bitmap_new_block+180/1676] Jan 27 20:37:55 my kernel: [<c012fd3c>] [<c012ff2a>] [<c0173002>] [<c017a56a>] [<c01731ce>] [<c016be84>] Jan 27 20:37:55 my kernel: [udf_bitmap_new_block+1597/1676] [udf_get_pblock+109/128] [udf_new_inode+612/652] [udf_mkdir+116/568] [cached_lookup+14/80] [vfs_mkdir+116/172] Jan 27 20:37:55 my kernel: [<c016c40d>] [<c0175b7d>] [<c016e278>] [<c0174214>] [<c01360a2>] [<c0137584>] Jan 27 20:37:55 my kernel: [sys_mkdir+137/208] [system_call+51/64] Jan 27 20:37:55 my kernel: [<c0137645>] [<c0106d33>] Jan 27 20:37:55 my kernel: Jan 27 20:37:55 my kernel: Code: 8b 40 34 01 c2 89 95 c4 fc ff ff e9 52 01 00 00 8b 4d 10 89 Jan 27 20:37:55 my kernel: I/O error: dev 0b:00, sector 49792 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49792 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49796 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49800 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49804 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49808 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49812 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49816 Jan 27 20:37:55 my kernel: Ignoring read error on sector:49820 Jan 27 20:37:55 my kernel: I/O error: dev 0b:00, sector 49828 How achieve stable writing to CDRW disc ? Does anybody get good results with IDE CDRW drive ? I use kernel 2.4.17 with patch 2.4.18-pre4-2. Thanks. Regards, Sergiy Kudryk __________________________________________________ Do You Yahoo!? Great stuff seeking new owners in Yahoo! Auctions! http://auctions.yahoo.com
Hello Peter
Thanks a lot !
--- Peter Osterlund
On Wed, 23 Jan 2002, Sergiy Kudryk wrote:
As you can see formatting with cdrwtool takes about 20 minutes and not finished (CDRW LED backs to normal light (yellow); cdrwtool stop responding and lock the CDRW drive).
On reboot i also have see next messages on screen (interrupt loss):
Jan 23 15:30:22 my kernel: hdc: lost interrupt Jan 23 15:31:03 my last message repeated 4 times Jan 23 15:32:04 my last message repeated 6 times ....
A wild guess. If DMA is enabled on the CDRW, maybe turning it off will help: hdparm -d 0 /dev/hdc
Unfortunally turning off DMA on /dev/hdc doesn't help - same trouble with formatting.
After reboot Windows DirectCD (Linux can't process disk) recognize filesystem disk of CDRW disk as .. CDFS.
What does this mean? Was the disc correctly formatted even though cdrwtool hang?
No, not correctly. DirectCD showed disc as CDFS with capacity ~700 Mb without free space.
I don't wanna use Windows anymore to format CDRW disk .. but inspite of that DirectCD support for my drive is too expensive (~200 Mbytes)
Are you referring to the fact that UDF formatted discs can not hold as much data as ISO9660 discs? This is true for linux as well.
Sorry i mean DirectCD 5.0 and Easy CD Creator space usage on harddrive after installing in c:\program files\adaptec .. I know that UDF allow to use only ~ 530 Mbytes on CDRW disc.
this program has stability and usability features (for example quick disk UDF formatting takes ~ 3 minutes).
Does anyone know how the quick formatting in DirectCD works? Will a quick-formatted disc work with the current linux packet writing module, or is special support needed?
Quick formatting (i always use this method) with DirectCD work perfectly (takes 1-3 minutes), and Linux UDF read/write support layer works with DirectCD formatted discs without problems. No special support is needed.
P.S. I attach kernel configuration file.
You have scsi emulation compiled as a module, but don't use it in the above example. I think someone reported that his problems were solved when he started to use scsi emulation.
Thanks. SCSI emulation really helps me to solve formatting (with cdrwtool) problem: My success story: --- Add SCSI emulation parameter (i compile ide-scsi support as module) to kernel line (/boot/grub/grub.conf): --- ... kernel .... ro root=/dev/hda4 hdc=ide-scsi ... --- Add "new" scsi device: --- # mknod /dev/sr0 b 11 0 --- Change line (cdrom -> sr0; for readonly support) in /etc/fstab --- ... /dev/sr0 /mnt/cdrom udf,iso9660 noauto,user,ro,iocharset=koi8-u 0 0 ... # reboot --- Check SCSI devices --- # cdrecord -scanbus Cdrecord 1.10 (i686-pc-linux-gnu) Copyright (C) 1995-2001 J�rg SchillinLinux sg driver version: 3.1.22 Using libscg version 'schily-0.5' scsibus0: 0,0,0 0) 'MITSUMI ' 'CR-48X5TE ' '1.8A' Removable CD- 0,1,0 1) * 0,2,0 2) * 0,3,0 3) * 0,4,0 4) * 0,5,0 5) * 0,6,0 6) * 0,7,0 7) * --- Plug pktcdvd module --- # /./sbin/insmod /./lib/modules/2.4.17/kernel/drivers/block/pktcdvd.o --- View active modules --- # /./sbin/lsmod Module Size Used by pktcdvd 16480 0 (unused) nls_koi8-u 3888 0 (autoclean) sr_mod 11872 0 (autoclean) sg 27840 0 (autoclean) tulip 36912 1 ide-scsi 7376 0 md 42976 0 (unused) sb 7344 0 sb_lib 32272 0 [sb] uart401 6032 0 [sb_lib] sound 52320 0 [sb_lib uart401] soundcore 3472 5 [sb_lib sound] unix 13696 123 (autoclean) --- Setup CDRW --- # pktsetup /dev/pktcdvd0 /dev/sr0 --- Format CDRW disc with UDF ------- # cdrwtool -d /dev/sr0 -q using device /dev/sr0 1312KB internal buffer setting write speed to 12x Settings for /dev/sr0: Fixed packets, size 32 Mode-2 disc I'm going to do a quick setup of /dev/sr0. The disc is going to be blanked and formatted with one big track. All data on the device will be lost!! Press CTRL-C to cancel now. Initiating quick disc blank Disc capacity is 275744 blocks (551488KB/538MB) Formatting track Writing UDF structures to disc 7200 (EET)/-120 Quick setup complete! --- Last line important - she shows that disc is correctly formatted. Then i try to write (with/without SCSI emulation) to CDRW disc .. and after some transations CDRW drive locks with i/o errors (/var/log/messages): Jan 24 19:29:23 my kernel: Ignoring read error on sector:10196 Jan 24 19:29:23 my kernel: I/O error: dev 0b:00, sector 10208 Jan 24 19:29:23 my kernel: Ignoring read error on sector:10208 Jan 24 19:29:23 my kernel: Ignoring read error on sector:10212 Jan 24 19:29:23 my kernel: Ignoring read error on sector:10216 Jan 24 19:29:23 my kernel: Ignoring read error on sector:10220 .... So today i have two questions: 1. How to achieve stable UDF writing ? 2. Where i can find documentation for udf tools (from udf-0.9.5 package) ?
-- Peter Osterlund - petero2@telia.com http://w1.894.telia.com/~u89404340
Thanks ! Regards, Sergiy Kudryk __________________________________________________ Do You Yahoo!? Great stuff seeking new owners in Yahoo! Auctions! http://auctions.yahoo.com
I have TEAC-W524E, but can't format CDRW - linux-2.4.18-pre4 - packet-2.4.18-pre4 ----------cut-------- cdrwtool -d /dev/hdc -q using device /dev/hdc 1575KB internal buffer setting write speed to 12x Settings for /dev/hdc: Fixed packets, size 32 Mode-2 disc I'm going to do a quick setup of /dev/hdc. The disc is going to be blanked and formatted with one big track. All data on the device will be lost!! Press CTRL-C to cancel now. Initiating quick disc blank Disc capacity is 295264 blocks (590528KB/576MB) Formatting track Writing UDF structures to disc 10800 (MSK)/-180 wait_cmd: Input/output error Command failed: 2a 00 00 00 00 00 00 00 20 00 00 00 - sense 0e.4c.19 Error writing packet 0 (ffffffff) -------cut syslog--------------- Jan 24 12:11:04 gambrinus kernel: hdc: irq timeout: status=0xd0 { Busy } Jan 24 12:11:04 gambrinus kernel: hdc: ATAPI reset complete -------------------------------- P.S. In SCSI emulation for cdrecord write disks is Ok! -- Thanks
participants (5)
-
Ben Fennema -
Peter Osterlund -
Petr Nejedly -
Serge Khmel -
Sergiy Kudryk