Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Status of affs support in the kernel and affstools

19 views
Skip to first unread message

John Paul Adrian Glaubitz

unread,
Jun 13, 2023, 2:50:04 AM6/13/23
to
Hello!

I was wondering what the current status of the affs driver in the kernel is,
did it receive the fixes that Michael(?) was working on? I vaguely remember
there were some issues with larger filesystems or so and that someone was
working on fixing these issues.

Also, I noticed that Roman Zippel actually created an affstools package [1]
back in the day and I wonder whether anyone has any experience with it. In
particular, I'm wondering what »different mkfs.affs« Artem Bityutskiy is
talking about here [2]. Is there an alternative tool for creating and checking
affs filesystems?

If we package it for Debian, it would be possible to create affs partitions
from debian-installer which would be useful for anyone wanting to install
Debian on an Amiga without a Workbench installation on the harddisk.

Thanks,
Adrian

> [1] https://www.xs4all.nl/~zippel/affstools-0.1a.tar.gz
> [2] https://marc.info/?l=linux-m68k&m=133888043312812&w=2

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer
`. `' Physicist
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

John Paul Adrian Glaubitz

unread,
Jun 13, 2023, 3:10:03 AM6/13/23
to
Hi Artem!

On Tue, 2023-06-13 at 09:53 +0300, Artem Bityutskiy wrote:
> I apologize, that message was sent 10 years ago and I do not remember details
> any longer.

No worries, thanks for the quick reply!

> What I do remember is that I was doing some cross-filesystem changes and
> struggled to test AFFS. At that time it felt that AFFS was abandoned. But I
> remember I managed to create an AFFS partition, mount it in Linux, and test my
> changes.

I see. I'm just wondering what other mkfs.affs command you might have found that
you were using there. Maybe Geert or anyone else on this list can remember.

Adrian

Stefan Reinauer

unread,
Jun 13, 2023, 3:20:04 AM6/13/23
to
I have found amitools from Christian Vogelgsang to be a good tool for creating and manipulating affs. 

Artem Bityutskiy

unread,
Jun 13, 2023, 3:20:04 AM6/13/23
to
On Tue, 2023-06-13 at 08:47 +0200, John Paul Adrian Glaubitz wrote:
> Also, I noticed that Roman Zippel actually created an affstools package [1]
> back in the day and I wonder whether anyone has any experience with it. In
> particular, I'm wondering what »different mkfs.affs« Artem Bityutskiy is
> talking about here [2]. Is there an alternative tool for creating and checking
> affs filesystems?

Hi Adrian,

I apologize, that message was sent 10 years ago and I do not remember details
any longer.

What I do remember is that I was doing some cross-filesystem changes and
struggled to test AFFS. At that time it felt that AFFS was abandoned. But I
remember I managed to create an AFFS partition, mount it in Linux, and test my
changes.

Thanks,
Artem.

John Paul Adrian Glaubitz

unread,
Jun 13, 2023, 3:30:03 AM6/13/23
to
Hi Martin!

On Tue, 2023-06-13 at 09:20 +0200, Martin Steigerwald wrote:
> John Paul Adrian Glaubitz - 13.06.23, 08:47:11 CEST:
> > I was wondering what the current status of the affs driver in the
> > kernel is, did it receive the fixes that Michael(?) was working on? I
> > vaguely remember there were some issues with larger filesystems or so
> > and that someone was working on fixing these issues.
> (...)
> I will reply to the other thread and ask again. Maybe it was just a last
> question / answer round missing. Or maybe someone else can confirm that
> the patch indeed went in.

I greatly appreciate that. Thanks a lot for taking care of this!

Adrian

Martin Steigerwald

unread,
Jun 13, 2023, 3:30:03 AM6/13/23
to
John Paul Adrian Glaubitz - 13.06.23, 08:47:11 CEST:
> I was wondering what the current status of the affs driver in the
> kernel is, did it receive the fixes that Michael(?) was working on? I
> vaguely remember there were some issues with larger filesystems or so
> and that someone was working on fixing these issues.

Well I just searched for the large size fixing patches discussed on
linux-m68k mailing list under subject titles "Amiga RDB partition
support fixes":

While I found a steady stream of fixes for affs, including related to
Amiga permission bit handling, quite recent fixes, even in 2023, I did
not found that size fixing patch.

I recalled a discussion where Michael Schmitz talked to Jens Axboe after
I asked whether the patch got in on 2023-08-21. I thought back then it
would have went in, but maybe the last question asked by Jens Axboe did
not get answered?

At least I did not found the patch in git history. Maybe I made a
mistake during search.

I will reply to the other thread and ask again. Maybe it was just a last
question / answer round missing. Or maybe someone else can confirm that
the patch indeed went in.

Best,
--
Martin

John Paul Adrian Glaubitz

unread,
Jun 13, 2023, 3:30:03 AM6/13/23
to
Hi Stefan!

On Tue, 2023-06-13 at 00:15 -0700, Stefan Reinauer wrote:
> I have found amitools from Christian Vogelgsang to be a good tool
> for creating and manipulating affs.

Looks interesting [1] but it doesn't seem to contain any tool for
creating and checking AFFS filesystems.

Adrian

> [1] https://github.com/cnvogelg/amitools

Stefan Reinauer

unread,
Jun 13, 2023, 3:40:04 AM6/13/23
to
Check out the format section... it will create filesystems:  https://amitools.readthedocs.io/en/latest/tools/xdftool.html

I haven't tried it directly on a drive as opposed to an image but there's a good likelihood it might just work.

Michael Schmitz

unread,
Jun 13, 2023, 4:00:03 AM6/13/23
to
Hi Martin,

Am 13.06.2023 um 19:20 schrieb Martin Steigerwald:
> John Paul Adrian Glaubitz - 13.06.23, 08:47:11 CEST:
>> I was wondering what the current status of the affs driver in the
>> kernel is, did it receive the fixes that Michael(?) was working on? I
>> vaguely remember there were some issues with larger filesystems or so
>> and that someone was working on fixing these issues.
>
> Well I just searched for the large size fixing patches discussed on
> linux-m68k mailing list under subject titles "Amiga RDB partition
> support fixes":
>
> While I found a steady stream of fixes for affs, including related to
> Amiga permission bit handling, quite recent fixes, even in 2023, I did
> not found that size fixing patch.
>
> I recalled a discussion where Michael Schmitz talked to Jens Axboe after
> I asked whether the patch got in on 2023-08-21. I thought back then it
> would have went in, but maybe the last question asked by Jens Axboe did
> not get answered?

I can't remember leaving a question of Jens' unanswered - I can go back
and have another look at the thread though.

I admit I've lost enthusiasm for endless follow-up when a maintainer
just won't take a patch. Chances are I've got something wrong and it's
possible the patch does no longer apply cleanly.

If someone else wants to take that patch and give it a fresh try, I'm OK
with that.

> At least I did not found the patch in git history. Maybe I made a
> mistake during search.

No, it never went in as far as I know.

Cheers,

Michael

Martin Steigerwald

unread,
Jun 13, 2023, 7:00:04 AM6/13/23
to
Thanks for your reply, Michael. And all your efforts to get the patch in.

Michael Schmitz - 13.06.23, 09:58:09 CEST:
> Am 13.06.2023 um 19:20 schrieb Martin Steigerwald:
> > John Paul Adrian Glaubitz - 13.06.23, 08:47:11 CEST:
> >> I was wondering what the current status of the affs driver in the
> >> kernel is, did it receive the fixes that Michael(?) was working on?
> >> I
> >> vaguely remember there were some issues with larger filesystems or
> >> so
> >> and that someone was working on fixing these issues.
> >
> > Well I just searched for the large size fixing patches discussed on
> > linux-m68k mailing list under subject titles "Amiga RDB partition
> > support fixes":
> >
> > While I found a steady stream of fixes for affs, including related
> > to
> > Amiga permission bit handling, quite recent fixes, even in 2023, I
> > did not found that size fixing patch.
> >
> > I recalled a discussion where Michael Schmitz talked to Jens Axboe
> > after I asked whether the patch got in on 2023-08-21. I thought
> > back then it would have went in, but maybe the last question asked
> > by Jens Axboe did not get answered?
>
> I can't remember leaving a question of Jens' unanswered - I can go
> back and have another look at the thread though.

Well I was just assuming that cause as I looked at thread display in my
mail appliation, there was a mail from Jens with a question without an
answer mail to it. But it was also just about backporting to Linux
stable. I would be happy to see it accepted for vanilla/master already.

> I admit I've lost enthusiasm for endless follow-up when a maintainer
> just won't take a patch. Chances are I've got something wrong and it's
> possible the patch does no longer apply cleanly.

I understand. Well let's see whether Jens Axboe answers. From my
experience regarding flexible I/O tester he normal accepts reasonable
patches. But I'd also consider the patch submitted as reasonable.

So maybe he just forgot to put it in.

> If someone else wants to take that patch and give it a fresh try, I'm
> OK with that.

I see.

Well I will first wait a few days whether Jens replies.

> > At least I did not found the patch in git history. Maybe I made a
> > mistake during search.

> No, it never went in as far as I know.

Would really be nice to have this settled.

Thanks,
--
Martin

Martin Steigerwald

unread,
Jun 20, 2023, 5:30:03 PM6/20/23
to
Hi John, hi Artem, hi Michael.

John Paul Adrian Glaubitz - 13.06.23, 09:22:52 CEST:
> On Tue, 2023-06-13 at 09:20 +0200, Martin Steigerwald wrote:
> > John Paul Adrian Glaubitz - 13.06.23, 08:47:11 CEST:
> > > I was wondering what the current status of the affs driver in the
> > > kernel is, did it receive the fixes that Michael(?) was working
> > > on? I
> > > vaguely remember there were some issues with larger filesystems or
> > > so
> > > and that someone was working on fixing these issues.
> >
> > (...)
> > I will reply to the other thread and ask again. Maybe it was just a
> > last question / answer round missing. Or maybe someone else can
> > confirm that the patch indeed went in.
>
> I greatly appreciate that. Thanks a lot for taking care of this!

The Amiga RDB large partition fixes went in now.

Thanks to everyone involved! Special thanks to Michael for going through
13 iterations!

https://lore.kernel.org/linux-m68k/10e23e7d-d2b0-f540...@gmail.com/T/#m0eea1b4b33ff8d96ebfe35546bce3a67256ee166

---------- Weitergeleitete Nachricht ----------

Betreff: Re: [PATCH v13 0/3] Amiga RDB partition support fixes
Datum: Dienstag, 20. Juni 2023, 22:28:34 CEST
Von: Jens Axboe <ax...@kernel.dk>
An: linux...@vger.kernel.org, Michael Schmitz <schmi...@gmail.com>
Kopie: linux...@vger.kernel.org, ge...@linux-m68k.org, h...@lst.de,
mar...@lichtvoll.de, fth...@linux-m68k.org



On Wed, 21 Jun 2023 08:17:22 +1200, Michael Schmitz wrote:
> another last version of this patch series, only change
> in this version is in the patch 2 subject (to better
> reflect what it's about), and adding Geert's review
> tag to said patch.
>
> I hope I've crossed all i's and dotted all t's now...
>
> [...]

Applied, thanks!

[1/3] block: fix signed int overflow in Amiga partition support
commit: fc3d092c6bb48d5865fec15ed5b333c12f36288c
[2/3] block: change all __u32 annotations to __be32 in affs_hardblocks.h
commit: 95a55437dc49fb3342c82e61f5472a71c63d9ed0
[3/3] block: add overflow checks for Amiga partition support
commit: b6f3f28f604ba3de4724ad82bea6adb1300c0b5f

Best regards,
--
Jens Axboe




-------------------------------------------------------------

Thanks,
--
Martin
0 new messages