Working `fsck.udf` badly needed

[ehem]

To me this seems the one super-duper urgent need for wide usage of the UDF filesystem, a working fsck program. Without this and improperly removed USB key results in unrepairable filesystem damage. Worse, damage like that accumulates and over time you end up with a major problem.

Personally, I'd rate this as urgent. udftools has template files, they merely need to be filled in...

[pali]

Hi! I know that fsck tools is important. Currently @argorain is working on it, see WIP pull request: https://github.com/pali/udftools/pull/7

[argorain]

Hi, that is correct, I am working on it. I have a very little time for that, but it is slowly moving forward.

[callegar]

Can you please provide some information on the current status of the udf fsck?

• the pull request mentions it being a beta version, is it expected to be safe enough to use or at least to try?
• looks like the repo from which the pull request originates stopped seeing commits shortly after the pull request was made. Can someone please detail if this is just about limited time resources to invest, of if the project is in wait for more feedback, etc.?

[argorain]

Hi, I am author of fsck for udf but I have no time to work on it now. There are some issues on ARM platform what was raised later as requirement from @pali plus some other stuff. Overall I think it is safe to try it but I am not sure about production use. There were some nasty issues on large drives so it definitely needs testing and feedback but I am out of time, at least for now.

[callegar]

@argorain thank you for the prompt answer and for the effort in giving us the beta. I'll give it a try, then, truly hoping that you'll be able to recover work on it soon. I just wanted to make sure that there was nothing holding you back, such as lack of feedback from testers on systems different from yours.

[argorain]

No problem. Please keep in mind it is really an beta so run it rather on image or copy of your drive rather that live data. And if you encounter anything odd, please let me know. I am not saying I'll work on it anytime soon but always once in a while I am going thru those reports.

[Siltaar]

As a matter of encouragement here is an article from 2013 that already states that the world is in the starting blocks for using UDF as USB-key filesystem as soon as an fsck.udf will be working on GNU+Linux :-)

https://tanguy.ortolo.eu/blog/article93/usb-udf

[ghost]

Sorry to "move the knife in the wound" (it's a french expression) But it is a shame that M$ has a working fsck.udf and not "us" ;) And I have no problems writing with win10 on a UDF partition whereas in Linux seems like there is an issue with some big files (when using steam). Yes I am my self guilty of not opening a bug report about that .

Cheers freind(s).

[ht990332]

Sorry to "move the knife in the wound" (it's a french expression) But it is a shame that M$ has a working fsck.udf and not "us" ;) And I have no problems writing with win10 on a UDF partition whereas in Linux seems like there is an issue with some big files (when using steam). Yes I am my self guilty of not opening a bug report about that .

Cheers freind(s).

Where can I get the Microsoft fsck.udf please?

[ghost]

I have not searched for the binary, it is a little complicated in windows, I had to dig with the command tasklist or taskslist I think to pick up the name of the "desktop windows terminal client" which is mstsc.exe for example ( a while ago)

You can do the same, I am not under a windows right at the moment, so I cant do it .

So many times when I have issues, I plug the disk on a machine with windows and right click on it, then select "tools" ... then "check and repair",

It takes hours on my 500G ssd ... but it works .

ps: I really like UDF, since it is native on most OSs, and is written independently from any OS provider ... and perfs are just fine.

Cheers freind(s).

[ht990332]

I have not searched for the binary, it is a little complicated in windows, I had to dig with the command tasklist or taskslist I think to pick up the name of the "desktop windows terminal client" which is mstsc.exe for example ( a while ago)

You can do the same, I am not under a windows right at the moment, so I cant do it .

So many times when I have issues, I plug the disk on a machine with windows and right click on it, then select "tools" ... then "check and repair",

It takes hours on my 500G ssd ... but it works .

ps: I really like UDF, since it is native on most OSs, and is written independently from any OS provider ... and perfs are just fine.

Cheers freind(s).

That is not fsck.udf then. It's a native windows utility. Your comment was misleading as it led me to believe Microsoft wrote a Linux fsck.udf utility. Please don't spread inaccurate information to us users.

[ghost]

I missleaded folks like you, that do not understand .

fsck.udf == FileSystemChecK for UDF .

So you dont know that a "M$-Windows" window is actually a program ?

It is a GUI for the microsoft udf fsck utility .

M$ writing utilities for Linux users ? even my "grandma" laughed at that.

Cheers and xanax

Le 07/09/2020 16:11, Hussam Al-Tayeb a écrit :

I have not searched for the binary, it is a little complicated in windows, I had to dig with the command tasklist or taskslist I think to pick up the name of the "desktop windows terminal client" which is mstsc.exe for example ( a while ago)

You can do the same, I am not under a windows right at the moment, so I cant do it .

So many times when I have issues, I plug the disk on a machine with windows and right click on it, then select "tools" ... then "check and repair",

It takes hours on my 500G ssd ... but it works .

ps: I really like UDF, since it is native on most OSs, and is written independently from any OS provider ... and perfs are just fine.

Cheers freind(s).

That is not fsck.udf then. It's a native windows utility. Your comment was misleading as it let me to believe Microsoft wrote a Linux fsck.udf utility. Please don't spread inaccurate information to us users.

-- You are receiving this because you commented. Reply to this email directly, view it on GitHub [1], or unsubscribe [2].

Links:

[1] https://github.com/pali/udftools/issues/21#issuecomment-688351210 [2] https://github.com/notifications/unsubscribe-auth/AEZ2YGQNTWJRJDDEC24HIB3SETSXXANCNFSM4FN4QPUA

[ht990332]

1. Microsoft's utility is called chkdsk and not fsck. It has no GUI. The GUI you see is a management console plugin.
2. Microsoft writes utilities for Linux all the time, including a process viewer, their MSSQL server, (a recent MR for Mesa), etc.. etc.. I can go on till tomorrow..
3. A Microsoft employee is the co-maintainer of the Linux kernel stable branches and Microsoft contributes a substantial amount of patches to the mainline kernel. In other words, Microsoft co-maintain the upstream LTS Linux kernels alongside greg k-h. Without Microsoft, LTS releases would barely be 2 years instead of the current 6 years.
4. Microsoft ships a Linux kernel as part of WSL2.
5. Microsoft was the largest contributor to open source projects https://www.infoworld.com/article/3253948/who-really-contributes-to-open-source.html
6. Microsoft owns, github, the website we are posting on now.

Your grandma needs to re-evaluate her position on Microsoft ;)

[ghost]

Please stop being as ridiculous as you are,

because I almost lost my grandma, I read out loud you message the her, and she was eating !

she started LoLing like a youngster and shocked so bad ... the nurse just got in... be quite little boy.

Le 07/09/2020 17:03, Hussam Al-Tayeb a écrit :

• Microsoft's utility is called chkdsk and not fsck.
• Microsoft writes utilities for Linux all the time, including a process viewer, their MSSQL server, (a recent MR for Mesa), etc.. etc.. I can go on till tomorrow..
• A Microsoft employee is the co-maintainer of the Linux kernel stable branches and Microsoft contributes a substantial amount of patches to the mainline kernel. In other words, Microsoft co-maintain the upstream LTS Linux kernels alongside greg k-h.
• Microsoft ships a Linux kernel as part of WSL2.
• Microsoft was the largest contributor to open source projects https://www.infoworld.com/article/3253948/who-really-contributes-to-open-source.html
• Microsoft owns, github, the website we are posting on now.

Your grandma needs to re-evaluate her position on Microsoft ;)

-- You are receiving this because you commented. Reply to this email directly, view it on GitHub [1], or unsubscribe [2].

Links:

[1] https://github.com/pali/udftools/issues/21#issuecomment-688380380 [2] https://github.com/notifications/unsubscribe-auth/AEZ2YGT6MCTDNOTNH73UCK3SETY5JANCNFSM4FN4QPUA

[pali]

Hello @hussamT @remi75! Independently of which opinion I have on this, I would like if you do not spam this issue. Feel free to move this discussion to other place. I want to have this issue tracker relevant and technically orientated. If you have anything which could help and improve UDF checking tool for Linux, you are free put arguments here. But discussion who and how contribute to Linux world is relevant with topic "udftools does not provide stable checking tool yet".

EDIT: I really do not like any moderation, but this discussion is now off-topic. I marked last comments via github button "off-topic", hopefully they are not deleted and just packed. (I really do not want to delete any comments).

[pali]

To move discussion back...

But it is a shame that M$ has a working fsck.udf and not "us" ;)

@argorain did some testing of MS UDF check disk tool and results were that it broke damaged UDF filesystems even more.

@smagnani did another tests and results were that it cannot handle and detect native 4K disks (correctly).

So I would not say it is "working" corectly. But maybe something changed... who knows.

[ghost]

a) well when did he do the tests ?

I have used it again this week with zero issues: a 500G ssd disk

over USB 3.1 ... every-time linux-steam trashes it, I hook it up on tha windows and it get fixed. it is a fact, fyi: I use linux and windows since the early 90' ... so trust me please, I have trashed / fixed ALOT of disks in my life and that with many FSs and OSs.

b) well, since M$ is reading this thread, you are free to file a bug report with them :)

Cheers and xanax .

Le 07/09/2020 18:40, pali a écrit :

To move discussion back...

But it is a shame that M$ has a working fsck.udf and not "us" ;)

@argorain [1] did some testing of MS UDF check disk tool and results were that it broke damaged UDF filesystems even more.

@smagnani [2] did another tests and results were that it cannot handle and detect native 4K disks (correctly).

So I would not say it is "working" corectly. But maybe something changed... who knows.

-- You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub [3], or unsubscribe [4].

Links:

[1] https://github.com/argorain [2] https://github.com/smagnani [3] https://github.com/pali/udftools/issues/21#issuecomment-688428962 [4] https://github.com/notifications/unsubscribe-auth/AEZ2YGTBPG2WQ6K3TAPFRMLSEUEGPANCNFSM4FN4QPUA

[ht990332]

Hello @hussamT @remi75! Independently of which opinion I have on this, I would like if you do not spam this issue. Feel free to move this discussion to other place. I want to have this issue tracker relevant and technically orientated. If you have anything which could help and improve UDF checking tool for Linux, you are free put arguments here. But discussion who and how contribute to Linux world is relevant with topic "udftools does not provide stable checking tool yet".

EDIT: I really do not like any moderation, but this discussion is now off-topic. I marked last comments via github button "off-topic", hopefully they are not deleted and just packed. (I really do not want to delete any comments).

Sure. Thank you for the feedback but out of courtesy, shouldn't you also be moderating false information? I was led to believe there was a fsck.udf binary from another vendor that did not truly exist. Also the regarding the "M$" labeling, shouldn't that be moderated as well? Not all Linux software developers, technicians, and system administrators reflect a toxic view towards developers from a competitive operation system. Some of us who perform technology work at engineering firms for a living try as much as possible to portray a positive, professional, and diplomatic image that shies away from such terminology (especially towards the competition).

[ghost]

Listen mister man, I Might have made one mistake, I didnt write "fsck.udf" , instead I wrote fsck.udf ,

Do you understand that ? It is a provocation from you to pretend to not have understood that , otherwise,

I urge you to change carrier and go to a m$Donald, there they have a strong policy about trademarks.

Also, If you "do not like moderation" , get out of your office, and go have a more intelligent chit-chat around the corner about the quality of the hot-dogs across the street.

Cheers and sleeping-pills

Le 07/09/2020 21:00, Hussam Al-Tayeb a écrit :

Hello @hussamT [1] @remi75 [2]! Independently of which opinion I have on this, I would like if you do not spam this issue. Feel free to move this discussion to other place. I want to have this issue tracker relevant and technically orientated. If you have anything which could help and improve UDF checking tool for Linux, you are free put arguments here. But discussion who and how contribute to Linux world is relevant with topic "udftools does not provide stable checking tool yet".

EDIT: I really do not like any moderation, but this discussion is now off-topic. I marked last comments via github button "off-topic", hopefully they are not deleted and just packed. (I really do not want to delete any comments).

Sure. Thank you for the feedback but out of courtesy, shouldn't you also be moderating false information? I was led to believe there was a fsck.udf binary from another vendor that did not truly exist. Also the regarding the "M$" labeling, shouldn't that be moderated as well? Not all Linux software developers, technicians, and system administrators reflect a toxic view towards developers from a competitive operation system. Some of us who perform technology work at engineering firms for a living try as much as possible to portray a positive, professional, and diplomatic image that shies away from such terminology.

-- You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub [3], or unsubscribe [4].

Links:

[1] https://github.com/hussamT [2] https://github.com/remi75 [3] https://github.com/pali/udftools/issues/21#issuecomment-688476208 [4] https://github.com/notifications/unsubscribe-auth/AEZ2YGV7UT6VRS3FE54EPULSEUUWXANCNFSM4FN4QPUA

[smagnani]

Two suggestions for moving the ball forward on a functional fsck:

1. Come to some agreement on the most important use case. For me this is UDF 2.01 on a flash disk, used for interchange with Windows. This reduces the scope to something that might be manageable.

B. Collect some examples of corruption to determine which are most common and therefore most important to fix. Examples would be hugely helpful, both to udffsck development and regression testing, as well as possibly pointing to driver issues (based on trends and analysis/speculation on how various types of corruption could have occurred). For this I have in mind a program and/or script that would scan a partition and gather up all the blocks with UDF metadata into a tarfile, which could then be used to reconstruct an image of the filesystem (without any file content, of course).

[Randrianasulu]

there was link to udf_test archive, archive was downloaded and looked into, not tried to compile it on Android/termux yet.. https://github.com/gmerlino/format-udf/wiki/fsck-tools-for-UDF

[pali]

udf_test is still available to download from moved Philips Parther Area website and seems that registration is not needed anymore.

[Randrianasulu]

@pali there is now fsck_udf from netbsd (also can create udf 2.50 fs on hd)

https://github.com/NetBSD/src/commit/c146cfcba0b33269459f26b6772e16b01da952d9

"Newfs_udf and makefs now also support formatting of UDF 2.50 with a metadata partition."

I tried to make those (makefs/fsck) compile on Linux (see my repo) but they sadly do not work as intended (some descriptors are off, and fsck can segfault on just created test image) - strangely there is no segfault on termux...

[hdante]

Hello, any updates on fsck.udf ?

[argorain]

Hello, not from my side. It is not forgotten but I didn't had time to finish merging and to be frank, not really will to do that. But hey, maybe at some point in future I'll get back to it. Vojtech

pá 26. 4. 2024 v 3:00 odesílatel Henrique @.***> napsal:

Hello, any updates on fsck.udf ?

— Reply to this email directly, view it on GitHub https://github.com/pali/udftools/issues/21#issuecomment-2078413924, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABZTSJMVDZWBH5ROXEXQMILY7GRKTAVCNFSM4FN4QPUKU5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TEMBXHA2DCMZZGI2A . You are receiving this because you were mentioned.Message ID: @.***>

[hdante]

Hello, ok, no problem, but if you don't want to implement it anymore and still think it's important and should be maintained, may I suggest you that you post a public request for help ? This should be in contrast to the recently seen xz backdoor, where the lack of publicity allowed bad actors to hijack the slowly evolving project.

[argorain]

Fair enough but I am not maintainer though, Pali is. Are you volunteering (wink, wink)? Jokes aside, recovery part of what I wrote is not really good but I still think integrity  checking is pretty solid so maybe some refactoring and merging at least that might be worth considering. Odesláno z mého zařízení Galaxy -------- Původní zpráva --------Od: Henrique @.> Datum: 26.04.24 16:34 (GMT+01:00) Komu: pali/udftools @.> Cc: Vojtech Vladyka @.>, Mention @.> Předmět: Re: [pali/udftools] Working fsck.udf badly needed (#21) Hello, ok, no problem, but if you don't want to implement it anymore and still think it's important and should be maintained, may I suggest you that you post a public request for help ? This should be in contrast to the recently seen xz backdoor, where the lack of publicity allowed bad actors to hijack the slowly evolving project.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>

[hdante]

No, I can't, I have a chronic illness and don't have free time. Less important than immediately finishing it is urgently announcing that help is needed.

[hdante]

More important

[smagnani]

FYI. when I was working on a UDF-centric product several years ago I did try to make incremental improvements to Vojtech's work. You may find my fork (or at least some of the changesets) useful.

https://github.com/smagnani/udftools.git

I do use this udffsck to check integrity periodically and to make simple repairs such as closing a LVID left open by a surprise dismount.

Unfortunately I too no longer have time to devote to this project.

Best Regards, Steve