This is a persistent problem since FreeBSD 6.x. DVD RAM formatted media with FAT or UFS as filesystems aren't a problem at all, but UDF (diverent versions) isn't readable in FreeBSD. mount_udf /dev/acd0 /cdrom results in mount_udf: /dev/acd0: Invalid argument I tried it with different version of UDF and different operating systems (Debian Etch/Windows XP and Mac OS X Leopard) but to no avail in FreeBSD. Maybe the udfdump output of such a media is of some use: UDF volume sets marked alive : Volume set `0939115C UDF Volume Set` (1 volume) with 8193 partitions Partition number 8192 at device `/dev/cd0' session 0 from sector 272(+0) for 2236160 sectors Primary volume `UDF Volume` (part 1/1) created by implementator `*Microsoft Windows' contains logical volume `` mapping 0 on 8192 as direct recording data metadata retrieving logical volume dependencies Logical volume `0939115C UDF Volume Set`:`UDF Volume`:`` integrity TAG: descriptor 9, serial_num 61994 at sector 36, crc length 118 bytes Logical volume integrity descriptor Timestamp (2008 01 26 at 17:01:55.01.17.36) Integrity type closed Next integrety sequence at 0 for 0 bytes Next free unique file ID 139 Length of implementation use area 46 bytes Number of partitions 1 Partition 0 : 335924 blocks free space out of 2236160 blocks Implemenator Id `*Apple Mac OS X UDF FS` (MacOS [03 01 02 00 00 00 ]) (flags=0) Number of files 48 Number of directories 6 Minimum readversion UDFv 201 Minimum writeversion UDFv 201 Maximum writeversion UDFv 201 marked closed at (2008 01 26 at 17:01:55.01.17.36) TAG: descriptor 8, serial_num 61994 at sector 37, crc length 496 bytes Terminating descriptor supporting tables For partition mapping 0->8192 fileset(s) Fileset descriptor extent at sector 0 within partion 0 for 2048 bytes TAG: descriptor 256, serial_num 61994 at sector 0, crc length 496 bytes Fileset descriptor Timestamp (2007 08 04 at 09:57:18.06.07.00) Interchange level 3 Max interchange level 3 Charset lists 1 Max charset lists 1 Fileset number 0 Fileset descriptor number 0 Logical volume id `` Fileset id `UDF Volume Set` Copyright file id `` Abstract file id `` DomainId `*OSTA UDF Compliant` (UDFv 201; no write protect ) (flags=0) Rootdir ICB found at sector 153 within partion space 0 for 2048 bytes Next extend for fileset at sector 0 within partion space 0 for 0 bytes Streamdir ICB found at sector 0 within partion space 0 for 0 bytes TAG: descriptor 266, serial_num 61994 at sector 153, crc length 796 bytes Extended file entry ICB Prior direct entries recorded (excl.) 0 ICB Strategy type 4 ICB Strategy type flags 0 0 ICB Maximum number of entries (non strat 4) 1 ICB indirect entries/depth 0 ICB File type 4 ICB Parent ICB in logical block 0 of mapped partition 0 ICB Flags 3 File/directory information using : internal in the ICB File flags : Uid -1 Gid -1 Permissions 3def Link count 3 Record format 0 Record display attributes 0 Record length 0 Information length 596 Object size 596 Logical blocks recorded 0 Access time (2008 01 26 at 17:01:52.37.77.10) Modification time (2008 01 26 at 17:01:52.37.77.14) Creation time (2007 08 04 at 10:57:18.06.07.00) Attribute time (2007 08 04 at 11:52:25.75.25.17) Checkpoint 1 Extended attributes ICB at at sector 0 within partion space 0 for 0 bytes Streamdir ICB at at sector 0 within partion space 0 for 0 bytes Implementation `*Apple Mac OS X UDF FS` (MacOS [03 01 02 00 00 00 ]) (flags=0) UniqueID 0 Length of extended attribute area 0 Length of allocation descriptors 596 Allocation descriptors : [ blob at internal blob here for 596 bytes ] checking writable filesets used/freed space tables For partition mapping 0->8192 Unalloced space bitmap TAG: descriptor 264, serial_num 61994 at sector 16, crc length 8 bytes Space bitmap Number of bits 2236160 Number of bytes 279520 Marked parts at : [00000014 - 00000015] [00000155 - 00000157] [00000172 - 00000172] [00000181 - 00000183] [00000226 - 00000226] [00000242 - 00000243] [00000250 - 00000251] [00000267 - 00000268] [00000280 - 00000280] [00000282 - 00000287] [00000289 - 00000290] [00000294 - 00000303] [00000305 - 00000311] [00000316 - 00022367] [01922330 - 02236160] Free space found on this partition [00000014 - 00000015] [00000155 - 00000157] [00000172 - 00000172] [00000181 - 00000183] [00000226 - 00000226] [00000242 - 00000243] [00000250 - 00000251] [00000267 - 00000268] [00000280 - 00000280] [00000282 - 00000287] [00000289 - 00000290] [00000294 - 00000303] [00000305 - 00000311] [00000316 - 00022367] [01922330 - 02097147] [02097148 - 02236159] Physical partition's unallocated space : 687972352 UDF volume sets marked alive : Volume set `0939115C UDF Volume Set` (1 volume) with 8193 partitions Partition number 8192 at device `/dev/cd0' session 0 from sector 272(+0) for 2236160 sectors Primary volume `UDF Volume` (part 1/1) created by implementator `*Microsoft Windows' contains logical volume `` mapping 0 on 8192 as direct recording data metadata Resulting list of alive sets : UDF volume sets marked alive : Volume set `0939115C UDF Volume Set` (1 volume) with 8193 partitions Partition number 8192 at device `/dev/cd0' session 0 from sector 272(+0) for 2236160 sectors Primary volume `UDF Volume` (part 1/1) created by implementator `*Microsoft Windows' contains logical volume `` mapping 0 on 8192 as direct recording data metadata Closing discs Disc access statistics sector reads 148 (296 Kbyte) sector written 0 (0 Kbyte) switches 0 Fix: Not a real fix but an option: I can mount UDF formatted DVD RAM media with sysutils/udfclient (option fusefs): udfclientfs /dev/cd0 /cdrom and read/copy the content. How-To-Repeat: mount_udf /dev/acd0 /cdrom results in mount_udf: /dev/acd0: Invalid argument
Responsible Changed From-To: freebsd-i386->freebsd-bugs This does not sound i386-specific. I wonder if it is related to kern/120967?
CC: Mark Linimon: no, it's not related. As a maintainer of sysutils/udfclient port and a guy who spent a lot of time on in-kernel UDF FS, I can tell that this problem is caused by FreeBSD fs/udf supporting only older UDF standards. Newer UDF features (>=2.00), like extended file entries, are not supported. So, I am not sure how to qualify this PR: a bug, a feature request, or ... One thing is for sure: FreeBSD UDF fs doesn't get attention that it actually needs. -- Andriy Gapon
Responsible Changed From-To: freebsd-bugs->scottl Assign to udf author Scott Long, who seems to be collecting udf PRs.
Oliver, if you still have that disk around and have an opportunity to check it with FreeBSD, could you please try the patch attached? The patch is against RELENG_7 but it should apply cleanly to not so distant releases/branches. This patch contains larger fixes for various UDF problems and in addition it also contains very basic support for UDF 2.X. Technically speaking it adds support for Extended File Entry, but not anything else. Just in case, to apply the patch you need the following: 1. FreeBSD sources :-) 2. cd /usr/src 3. patch -p1 < /path/to/the/patch 4. cd sys/modules/udf 5. make obj && make depend && make && make install 6. kldload udf The above assumes that you use udf as module. If you have it built into kernel you then instead of 5 and 6 you would need to build the kernel and reboot. -- Andriy Gapon
freebsd is still missing udf 2.x support. if wikipedia is correct the latest revision is udf 2.60 from 2005. reinoud zandijk has taken scotts udf 1.x work and extended it to support udf 2.60. [1] the code is in the netbsd (and also openbsd) tree. maybe somebody is willing to backport it to freebsd. cheers. alex [1] http://mail-index.netbsd.org/current-users/2006/02/02/0027.html -- Alexander Best
on 19/07/2010 05:07 Alexander Best said the following: > freebsd is still missing udf 2.x support. if wikipedia is correct the > latest revision is udf 2.60 from 2005. > > reinoud zandijk has taken scotts udf 1.x work and extended it to > support udf 2.60. [1] the code is in the netbsd (and also openbsd) > tree. maybe somebody is willing to backport it to freebsd. Yeah. http://www.freebsd.org/projects/ideas/ideas.html#p-portudf -- Andriy Gapon
In the header of this bug report the tag: Importance currently reads: Normal Affects Only Me This is incorrect. This issue affects others. I am presently using PCBSD10.3 and this exact issue is still present. I have also noted from google that it has affected several others. Anyone who wants to mount a modern UDF CD/DVD ROM will not be able to on FreeBSD (at least with version 10.3). I suspect correcting this tag could have an impact on getting some attention to this problem.
mount_udf is clearly a problem, i cant mount or read any kind of new dvd. something like this seems to me important because it heavily impacts your workflow if things like this dont work. we now have bhyve and this requires using newer kinds of media, making iso-images and other stuff.
This is still a problem. I've been debating for years to switch to FreeBSD as I really enjoy the BSD license over GPL. With XFCE4 FreeBSD works fairly well. However, I installed virtualbox so I could install Windows 7 in a VM for things that I will not be able to use natively. I am unable to mount my Windows 7 DVD because of this.
Hi. mount_udf and the udf kernel code was originally my baby, but I did some very naive things in it that worked for mainstream DVD-ROM media but were not fully compliant with the spec. I can work on this for a few evenings, but I don't have a way to easily create a DVD-RAM test case. Would it work to do a 'dd' image of your test disk and make it available to me to download? That might help to accelerate my efforts on this.
Hi Scott, Glad to see a response. Was not expecting one! The DVD is simply the Windows 7 install DVD, but you could achieve the same result with the freely downloadable Windows 10 DVD and burning it. The Windows 10 ISO is unfortunately downloaded by a tool so I don't know how well it work on FreeBSD with Wine: https://www.microsoft.com/en-us/software-download/windows10. If it helps, I could make a stripped down Windows 7 ISO that removes the large WIM files and upload that somewhere or email you a link. Thanks, Frank
Is dd equal to illegal disk copy tools? You need to copy your disk exactly equal the way it is if you want to do any testing. This means i have to make an illegal copy for you, can we do this? Rights could be an problem, i could help you otherwise by testing it myself on my own machine. There is many things i can test, i will give you some information what protocol is required on those disks. You can work on this.
Hi! This would be nice to have. Wanted to create fat32 Windows 10 installer USB, but couldn't mount downloaded udf iso with these symptoms.
Keyword: patch or patch-ready – in lieu of summary line prefix: [patch] * bulk change for the keyword * summary lines may be edited manually (not in bulk). Keyword descriptions and search interface: <https://bugs.freebsd.org/bugzilla/describekeywords.cgi>
Triage: * severity reduced, to the norm for a non-security bug report * …
With bugmeister hat, return to pool.