Slackware 15.0 ext2 file system

[Thomas Gibson]

Is anybody else having problems reading ext2 file drives with latest
kernel 5.13? All OK with kernel 5.12.
My system crashes if I try to read from an ext2 drive. The error message
is extremely long ending up displaying cpu registers.
Slackware 15.0 is on an ext4 drive.

[Eric Pozharski]

Post by Thomas Gibson
Is anybody else having problems reading ext2 file drives with latest
kernel 5.13? All OK with kernel 5.12.

5.13.2 is fine too. Complained about year 2038 though. Also, might be
relevant, no trigger happy helpers here.

Post by Thomas Gibson
My system crashes if I try to read from an ext2 drive. The error
message is extremely long ending up displaying cpu registers.

I suspect that would be tail of CPU-dump. Actual kernel panic is way
above. And with buffer-scroll being killed for good you're in the dark
now. I have two ideas how to work around this (one more ridiculous then
other), I hope those are in vain.

*CUT*

[Eric Pozharski]

Post by Eric Pozharski

Post by Thomas Gibson
Is anybody else having problems reading ext2 file drives with latest
kernel 5.13? All OK with kernel 5.12.

5.13.2 is fine too. Complained about year 2038 though. Also, might
be relevant, no trigger happy helpers here.

Upgraded to 5.14.9 -- still nothing, ext2 mounts fine (granted no
helpers). Thus whatever it was is either fixed or isn't ext2 problem
per se.

*CUT*

[Thomas Gibson]

Post by Eric Pozharski

Post by Eric Pozharski

Post by Thomas Gibson
Is anybody else having problems reading ext2 file drives with latest
kernel 5.13? All OK with kernel 5.12.

5.13.2 is fine too. Complained about year 2038 though. Also, might
be relevant, no trigger happy helpers here.

Upgraded to 5.14.9 -- still nothing, ext2 mounts fine (granted no
helpers). Thus whatever it was is either fixed or isn't ext2 problem
per se.
*CUT*

I have recently tested ext2 file systems on kernel 5.14.11 and you are
correct, the problem has been corrected.

[John McCue]

Post by Thomas Gibson
Is anybody else having problems reading ext2 file drives with latest
kernel 5.13? All OK with kernel 5.12.
My system crashes if I try to read from an ext2 drive. The error message
is extremely long ending up displaying cpu registers.
Slackware 15.0 is on an ext4 drive.

When was it formatted ? IIRC a kernel change was made a
few releases ago which rolled it into ext4.

So, maybe to an fsck should be done in case something
changed.

[Thomas Gibson]

Post by John McCue

Post by Thomas Gibson
Is anybody else having problems reading ext2 file drives with latest
kernel 5.13? All OK with kernel 5.12.
My system crashes if I try to read from an ext2 drive. The error message
is extremely long ending up displaying cpu registers.
Slackware 15.0 is on an ext4 drive.

When was it formatted ? IIRC a kernel change was made a
few releases ago which rolled it into ext4.
So, maybe to an fsck should be done in case something
changed.

Thanks for responding. I can create a new ext2 file system but it still
cpu dumps if I try to read from it. I tried e2fsck on an existing file
system to no avail.

[Robert Komar]

I'm running slackware off of an SD card with an ext2 root filesystem.
Since the upgrade to kernel 5.13, I have not been able to boot the
system. It panics as it is loading the initrd. I assumed that I
was missing some modules for the new kernel, but now I think I am
probably hitting the same ext2 problem. It's hard to know because
I can't scroll back up to the start of the panic log, as others
have also pointed out. Bah!

Rob Komar

[Thomas Gibson]

Post by Robert Komar
I'm running slackware off of an SD card with an ext2 root filesystem.
Since the upgrade to kernel 5.13, I have not been able to boot the
system. It panics as it is loading the initrd. I assumed that I
was missing some modules for the new kernel, but now I think I am
probably hitting the same ext2 problem. It's hard to know because
I can't scroll back up to the start of the panic log, as others
have also pointed out. Bah!
Rob Komar

Thank you for responding. I was starting to think I was the only one
having problems with this new kernel.

[noel]

Post by Robert Komar
I'm running slackware off of an SD card with an ext2 root filesystem.
Since the upgrade to kernel 5.13, I have not been able to boot the
system. It panics as it is loading the initrd. I assumed that I was
missing some modules for the new kernel, but now I think I am probably
hitting the same ext2 problem. It's hard to know because I can't scroll
back up to the start of the panic log, as others have also pointed out.
Bah!
Rob Komar

mount the filesystem on a PC, chroot and build a kernel with everything
in - ie: your not using an init image you are using full kernel.
or you could chroot and remake your init image first.

as for not scrolling back, you have that dumbfuck so called genius
torvalds to thank for that. had a boot issue once on a pentium, if I had
scrol back I'd see the obvious error and correct it, but no scrollback
meant hours and hours and hours of hair pulling - eventually grabbed a
video camera and recorded the boot up, replayed the recording and saw the
error, fixes in 15 seconds, i had and still have plenty of choice words
for torvalds

[John Forkosh]

...I have not been able to boot the system. It panics as it is
loading the initrd. It's hard to know [why] because I can't scroll
back up to the start of the panic log, as others have also pointed out.
Rob Komar

had a boot issue once on a pentium, if I had scroll back I'd see
the obvious error and correct it, but no scrollback meant hours
and hours and hours of hair pulling - eventually grabbed a
video camera and recorded the boot up, replayed the recording and saw
the error, fixes in 15 seconds

What a great idea. I've had an ongoing similar problem, but it never
occurred to me to record the screen "offline" (so to speak).
You think that'll work using a smartphone video (actually, I guess
I'll be trying it regardless)?

Any possibility of an "online" (so to speak) recording of the screen?
I guess you'd need an already-booted-and-running system for that,
and then there's probably an application to do that somewhere (I don't
offhand know of any such thing, but I imagine it exists somewheres).
But I can't think of any way to do that while booting from an
installation medium. Is there maybe some way to do that???
Or maybe some simple change to initrd that would amount to a
"*.* -debuglogfile" line in syslog.conf???

[Henrik Carlqvist]

Post by John Forkosh
Any possibility of an "online" (so to speak) recording of the screen?

It would be possible to boot the kernel with the console set to a serial
port and have a serial cable connected to another machine logging all
data to the serial port.

regards Henrik

[John Forkosh]

Post by Henrik Carlqvist

Post by John Forkosh
Any possibility of an "online" (so to speak) recording of the screen?

It would be possible to boot the kernel with the console set to a serial
port and have a serial cable connected to another machine logging all
data to the serial port.
regards Henrik

Thanks for the additional good idea (assuming some usb-to-serial driver
or physical adapter is available since I no longer have any pc's with
an actual serial, or parallel for that matter, port).
It also occurred to me to plug the hdmi cable from the pc into the
hdmi input of a vcr or vdr recorder. And since I happen to already have
an hdmi splitter (one input, three outputs), that may be the most
straightforward approach.

[Ted Heise]

On Tue, 3 Aug 2021 04:12:40 -0000 (UTC),

Post by John Forkosh

...I have not been able to boot the system. It panics as it
is loading the initrd. It's hard to know [why] because I
can't scroll back up to the start of the panic log, as others
have also pointed out. Rob Komar

had a boot issue once on a pentium, if I had scroll back I'd
see the obvious error and correct it, but no scrollback meant
hours and hours and hours of hair pulling - eventually grabbed
a video camera and recorded the boot up, replayed the
recording and saw the error, fixes in 15 seconds

What a great idea. I've had an ongoing similar problem, but it
never occurred to me to record the screen "offline" (so to
speak). You think that'll work using a smartphone video
(actually, I guess I'll be trying it regardless)?

Some 20 years ago I used a video recorder to capture EKG traces in
a clinical study, rather than try to figure out how to get the
data out of the units. A most practical approach.

I would expect the smartphone video to work okay. Resolution may
not be optimal, but probably good enough. I used my iPhone SE to
record the video from an old VCR as it played on a tv screen. Not
even close to HD ( :), but surprisingly good.

[Ralph Spitzner]

noel wrote on 8/2/21 4:36 PM:
[...]

Post by noel
as for not scrolling back, you have that dumbfuck so called genius
torvalds to thank for that. had a boot issue once on a pentium, if I had
scrol back I'd see the obvious error and correct it, but no scrollback

[...]
I imagine it's a bit hard to do screen scrolling when the
keyboard and video drivers (and all others) are dead/ haltet.
would you prefer a windows like:
KERNEL PANIC
ERROR #DEADBEEF ?

-rasp

[Robert Komar]

Post by Ralph Spitzner
[...]

Post by noel
as for not scrolling back, you have that dumbfuck so called genius
torvalds to thank for that. had a boot issue once on a pentium, if I had
scrol back I'd see the obvious error and correct it, but no scrollback

[...]
I imagine it's a bit hard to do screen scrolling when the
keyboard and video drivers (and all others) are dead/ haltet.
KERNEL PANIC
ERROR #DEADBEEF ?
-rasp

I think most of us would prefer the scrolling that was there for almost
thirty years. It wasn't perfect, but it was better than nothing.

Back to the original topic, I have slackware-current running in a VM
as well. The root filesystem is not ext2, so I can boot without the
initrd. In there, I can create an ext2 filesystem in a file, mount
that, write to it, read from it,... So, it looks like ext2 works
fine in that system. Perhaps my problem on the netbook is really
with the initrd as I first thought, and the OP has something broken
in his system. I can't duplicate his ext2 problems within the VM.

Rob Komar

[noel]

Post by Robert Komar

[...]

Post by noel
as for not scrolling back, you have that dumbfuck so called genius
torvalds to thank for that. had a boot issue once on a pentium, if I
had scrol back I'd see the obvious error and correct it, but no
scrollback

[...]
I imagine it's a bit hard to do screen scrolling when the keyboard and
video drivers (and all others) are dead/ haltet.
KERNEL PANIC ERROR #DEADBEEF ?
-rasp

I think most of us would prefer the scrolling that was there for almost
thirty years. It wasn't perfect, but it was better than nothing.
Back to the original topic, I have slackware-current running in a VM as
well. The root filesystem is not ext2, so I can boot without the
initrd. In there, I can create an ext2 filesystem in a file, mount
that, write to it, read from it,... So, it looks like ext2 works fine
in that system. Perhaps my problem on the netbook is really with the
initrd as I first thought, and the OP has something broken in his
system. I can't duplicate his ext2 problems within the VM.
Rob Komar

current kernels dont like ext2, plain and simple

I pulled an ol toshiba usb drive out from storage since I knew it was
ext2, nope, stack errors and killed at every first attempt...

converted it to ext4,

tune2fs -O extents,uninit_bg,dir_index,has_journal /dev/sdf1

and voila! no errors, so its just ext2 theyve rooted up

way to go linus...

(im not gonna sift through 10K trolls on LKML to see if its been brought
up there yet.)

(I didnt see mr spitzers post, I didnt miss anything, my troll filters
seem to be working well)

[Thomas Gibson]

Post by noel

Post by Robert Komar

[...]

Post by noel
as for not scrolling back, you have that dumbfuck so called genius
torvalds to thank for that. had a boot issue once on a pentium, if I
had scrol back I'd see the obvious error and correct it, but no
scrollback

[...]
I imagine it's a bit hard to do screen scrolling when the keyboard and
video drivers (and all others) are dead/ haltet.
KERNEL PANIC ERROR #DEADBEEF ?
-rasp

I think most of us would prefer the scrolling that was there for almost
thirty years. It wasn't perfect, but it was better than nothing.
Back to the original topic, I have slackware-current running in a VM as
well. The root filesystem is not ext2, so I can boot without the
initrd. In there, I can create an ext2 filesystem in a file, mount
that, write to it, read from it,... So, it looks like ext2 works fine
in that system. Perhaps my problem on the netbook is really with the
initrd as I first thought, and the OP has something broken in his
system. I can't duplicate his ext2 problems within the VM.
Rob Komar

current kernels dont like ext2, plain and simple
I pulled an ol toshiba usb drive out from storage since I knew it was
ext2, nope, stack errors and killed at every first attempt...
converted it to ext4,
tune2fs -O extents,uninit_bg,dir_index,has_journal /dev/sdf1
and voila! no errors, so its just ext2 theyve rooted up
way to go linus...
(im not gonna sift through 10K trolls on LKML to see if its been brought
up there yet.)
(I didnt see mr spitzers post, I didnt miss anything, my troll filters
seem to be working well)

Thanks for this. I was beginning to think it was just me. I have
reverted back to kernel 5.12.16-smp and all is OK.

[noel]

Post by Thomas Gibson

Post by noel

Post by Robert Komar

[...]

Post by noel
as for not scrolling back, you have that dumbfuck so called genius
torvalds to thank for that. had a boot issue once on a pentium, if I
had scrol back I'd see the obvious error and correct it, but no
scrollback

[...]
I imagine it's a bit hard to do screen scrolling when the keyboard
and video drivers (and all others) are dead/ haltet.
KERNEL PANIC ERROR #DEADBEEF ?
-rasp

I think most of us would prefer the scrolling that was there for
almost thirty years. It wasn't perfect, but it was better than
nothing.
Back to the original topic, I have slackware-current running in a VM
as well. The root filesystem is not ext2, so I can boot without the
initrd. In there, I can create an ext2 filesystem in a file, mount
that, write to it, read from it,... So, it looks like ext2 works fine
in that system. Perhaps my problem on the netbook is really with the
initrd as I first thought, and the OP has something broken in his
system. I can't duplicate his ext2 problems within the VM.
Rob Komar

current kernels dont like ext2, plain and simple
I pulled an ol toshiba usb drive out from storage since I knew it was
ext2, nope, stack errors and killed at every first attempt...
converted it to ext4,
tune2fs -O extents,uninit_bg,dir_index,has_journal /dev/sdf1
and voila! no errors, so its just ext2 theyve rooted up
way to go linus...
(im not gonna sift through 10K trolls on LKML to see if its been
brought up there yet.)
(I didnt see mr spitzers post, I didnt miss anything, my troll filters
seem to be working well)

Thanks for this. I was beginning to think it was just me. I have
reverted back to kernel 5.12.16-smp and all is OK.

Is there any reason youve stuck to ext2?

I admit it took me until just last year to move from ext3 to ext4, my
earlier attempts to try ext4 were very many years ago when it was new,
too new, it was unstable, quotas did not work at all, but its matured
over the years, I'm running production servers on it and not missed a
beat.

[Jimmy Johnson]

Post by noel
Is there any reason youve stuck to ext2?

I've been wondering the same thing?

I test linux, so when I moved from ext3 to ext4 in 2009 I expected
people to start moving with new linux installs, it's a better file system.

Now when 15 gets released the move will be to the newest linux EXT file
system BTRFS. BTRFS is a even better EXT file system I've just been
waiting for it to get well supported.

[Aragorn]

Post by Jimmy Johnson

Post by noel
Is there any reason youve stuck to ext2?

I've been wondering the same thing?
I test linux, so when I moved from ext3 to ext4 in 2009 I expected
people to start moving with new linux installs, it's a better file system.
Now when 15 gets released the move will be to the newest linux EXT
file system BTRFS. BTRFS is a even better EXT file system I've just
been waiting for it to get well supported.

btrfs is not a filesystem in the ext family. It is developed at
Oracle, and if anything, it is much more akin to ZFS, albeit that the
RAID-5/6 functionality of btrfs is still not usable. The on-disk
format has however already long been stable and it is quite resilient.

I've been using btrfs here on my production machine for over two years
already — I do use ext4 for /boot because GRUB doesn't like btrfs
(although there is a patched version that does) — and I haven't had any
problems with it so far. It didn't even budge after two unexpected power
failures. Earlier I had been using XFS, with which I wasn't so lucky
under those very same circumstances.

Some of the advantages — for more information, check the Wikipedia page
or the man page...:

° Copy-on-write.

° Supports transparent inline compression and will autodetect the
compression type and compression factor.

° Autodetects whether it's running off of an SSD, and if so, will
automatically enable performance optimizations for SSDs.

° Snapshots.

° Subvolumes. Think of them as separate partitions, except that the
free disk pace is shared among all the subvolumes. Subvolumes are
however not regarded as separate block devices.

° Multiple root directories per filesystem, so that you can mount a
different subvolume as the root volume of the partition.

° Has an auto-defragmentation mount option — should only be used on
spinning rust disks, of course.

° Supports swap files for those who don't want to use a swap partition.

° Supports RAID-0 and RAID-1.

° Data AND metadata journaling. The journal is always replayed after
an unclean shutdown, even for read-only filesystems.

° Allows for rebalancing the trees after lots of moving files between
individual subvolumes.

° Excellent documentation.

[Jimmy Johnson]

Post by Aragorn
btrfs is not a filesystem in the ext family.

The main btrfs features include: "Extent based file storage".
https://www.kernel.org/doc/html/latest/filesystems/btrfs.html

[Aragorn]

Post by Jimmy Johnson

Post by Aragorn
btrfs is not a filesystem in the ext family.

The main btrfs features include: "Extent based file storage".
https://www.kernel.org/doc/html/latest/filesystems/btrfs.html

Yes, but that has nothing to do with the ext, ext2, ext3 or ext4
filesystems.

ext — the original Linux filesystem — was essentially just the Minix
filesystem. ext2 was a different kind of filesystem — it was more
advanced than the Minix fileystem.

ext3 is simply ext2 with journaling added. ext4 is ext3 with hashed
trees enabled by default — they are available but disabled by default
in ext3 — and with extent-based storage (which neither ext, ext2 or
ext3 have).

The "ext" in the name of ext, ext2, ext3 and ext4 has nothing to
do with the concept of extent-based storage, but instead stands for
"extended filesystem".

btrfs is an entirely different type of filesystem — the fact that it
uses extents is irrelevant. ZFS, XFS, JFS and (the non-canonical)
reiser4 are also extent-based, but none of them have anything to do
with ext, ext2, ext3 or ext4 either.

[Thomas Gibson]

Post by noel

Post by Thomas Gibson

Post by noel

Post by Robert Komar
Back to the original topic, I have slackware-current running in a VM
as well. The root filesystem is not ext2, so I can boot without the
initrd. In there, I can create an ext2 filesystem in a file, mount
that, write to it, read from it,... So, it looks like ext2 works fine
in that system. Perhaps my problem on the netbook is really with the
initrd as I first thought, and the OP has something broken in his
system. I can't duplicate his ext2 problems within the VM.
Rob Komar

current kernels dont like ext2, plain and simple
I pulled an ol toshiba usb drive out from storage since I knew it was
ext2, nope, stack errors and killed at every first attempt...
converted it to ext4,
tune2fs -O extents,uninit_bg,dir_index,has_journal /dev/sdf1
and voila! no errors, so its just ext2 theyve rooted up
way to go linus...
(im not gonna sift through 10K trolls on LKML to see if its been
brought up there yet.)

Thanks for this. I was beginning to think it was just me. I have
reverted back to kernel 5.12.16-smp and all is OK.

Is there any reason youve stuck to ext2?
I admit it took me until just last year to move from ext3 to ext4, my
earlier attempts to try ext4 were very many years ago when it was new,
too new, it was unstable, quotas did not work at all, but its matured
over the years, I'm running production servers on it and not missed a
beat.

No is the sort answer. For some reason I stuck with ext2 file systems
from Slackware 3. I did not know you could convert from ext2 to ext4
without a lot of rebuilding. Thanks for that. I have Slakware 15.0
running on ext3 and Slackware 14.2 on ext2. I will convert soon.
Thanks again

[Jimmy Johnson]

I have Slakware 15.0 running on ext3

That's fine with me, but why when slackware 15 is being designed to use
btrfs.

[Chris Elvidge]

Post by Jimmy Johnson

I have Slakware 15.0 running on ext3

That's fine with me, but why when slackware 15 is being designed to use
btrfs.

Wo says it was designed to use btrfs?
My Slack15 installation uses ext4.

[Jimmy Johnson]

Post by Chris Elvidge

Post by Jimmy Johnson
That's fine with me, but why when slackware 15 is being designed to
use btrfs.

Wo says it was designed to use btrfs?

I said, it's being designed to use btrfs. :)

Post by Chris Elvidge
My Slack15 installation uses ext4.

And so do I, but when 15 is released it will be installed on btrfs.

[John Forkosh]

Post by Jimmy Johnson

Post by Chris Elvidge
My Slack15 installation uses ext4.

And so do I, but when 15 is released it will be installed on btrfs.

What's the future of lilo with respect to all this?
Is grub (or whatever) going to replace it as standard with slackware?

[Jimmy Johnson]

Post by John Forkosh

Post by Jimmy Johnson

Post by Chris Elvidge
My Slack15 installation uses ext4.

And so do I, but when 15 is released it will be installed on btrfs.

What's the future of lilo with respect to all this?
Is grub (or whatever) going to replace it as standard with slackware?

That I don't know, I use my own boot loader so I don't think about it
much, but I will be surprised if there is a major change in the
installer. Do you think there will be a RC for us to test?

[noel]

Post by Jimmy Johnson

Post by Chris Elvidge

Post by Jimmy Johnson
That's fine with me, but why when slackware 15 is being designed to
use btrfs.

Wo says it was designed to use btrfs?

I said, it's being designed to use btrfs. :)

Post by Chris Elvidge
My Slack15 installation uses ext4.

And so do I, but when 15 is released it will be installed on btrfs.

If that was teh case though -current would already, long, been using it,
it isnt, last -current fresh install I made was back in December, it
chose ext4.

[noel]

Post by Chris Elvidge

Post by Jimmy Johnson

I have Slakware 15.0 running on ext3

That's fine with me, but why when slackware 15 is being designed to use
btrfs.

Wo says it was designed to use btrfs?
My Slack15 installation uses ext4.

reiserwas fantastic on raid, unfortunately raid or no raid, it did not
like hard powerfails, dataloss was a huge problem, like ZFS used to

[Chris Elvidge]

Post by Thomas Gibson

Post by noel

Post by Thomas Gibson

Post by noel

Post by Robert Komar
Back to the original topic, I have slackware-current running in a VM
as well. The root filesystem is not ext2, so I can boot without the
initrd. In there, I can create an ext2 filesystem in a file, mount
that, write to it, read from it,... So, it looks like ext2 works fine
in that system. Perhaps my problem on the netbook is really with the
initrd as I first thought, and the OP has something broken in his
system. I can't duplicate his ext2 problems within the VM.
Rob Komar

current kernels dont like ext2, plain and simple
I pulled an ol toshiba usb drive out from storage since I knew it was
ext2, nope, stack errors and killed at every first attempt...
converted it to ext4,
tune2fs -O extents,uninit_bg,dir_index,has_journal /dev/sdf1
and voila! no errors, so its just ext2 theyve rooted up
way to go linus...
(im not gonna sift through 10K trolls on LKML to see if its been
brought up there yet.)

Thanks for this. I was beginning to think it was just me. I have
reverted back to kernel 5.12.16-smp and all is OK.

Is there any reason youve stuck to ext2?
I admit it took me until just last year to move from ext3 to ext4, my
earlier attempts to try ext4 were very many years ago when it was new,
too new, it was unstable, quotas did not work at all, but its matured
over the years, I'm running production servers on it and not missed a
beat.

No is the sort answer. For some reason I stuck with ext2 file systems
from Slackware 3. I did not know you could convert from ext2 to ext4
without a lot of rebuilding. Thanks for that. I have Slakware 15.0
running on ext3 and Slackware 14.2 on ext2. I will convert soon.
Thanks again

On my first Slackware (around 3) I used reiserfs; and used it's recovery
features.

[Henrik Carlqvist]

Post by noel
Is there any reason youve stuck to ext2?

I was not the original poster, but still I might give some such reasons
even though I have not tried Slackware-current and suffered any problems
with ext2.

Ext2 as opposed to ext3 and ext4 does not have journaling. Usually
journaling is a good thing and I use ext4 for my everyday use. However,
when I do backups to hard drives which are intended for "write once,
maybe read a few times" that journaling is not to much use. As such I
have during several years stuck to ext2 for those backups.

I would not call it a show-stopper, I will still have some older
Slackware installations capable of reading my backups, but it would be a
major disadvantage if the kernel in upcoming Slackware 15 totally crashes
when trying to read such a backup.

regards Henrik

[Aragorn]

Post by Henrik Carlqvist

Post by noel
Is there any reason youve stuck to ext2?

I was not the original poster, but still I might give some such
reasons even though I have not tried Slackware-current and suffered
any problems with ext2.
Ext2 as opposed to ext3 and ext4 does not have journaling. Usually
journaling is a good thing and I use ext4 for my everyday use.
However, when I do backups to hard drives which are intended for
"write once, maybe read a few times" that journaling is not to much
use. As such I have during several years stuck to ext2 for those
backups.
I would not call it a show-stopper, I will still have some older
Slackware installations capable of reading my backups, but it would
be a major disadvantage if the kernel in upcoming Slackware 15
totally crashes when trying to read such a backup.

The ext4 driver in the kernel is perfectly capable of handling ext2 and
ext3, and is now the preferred driver for those two filesystems, even.
The old driver is still available too, but it's deprecated.

[Jimmy Johnson]

Post by Henrik Carlqvist

Post by noel
Is there any reason youve stuck to ext2?

I was not the original poster, but still I might give some such reasons
even though I have not tried Slackware-current and suffered any problems
with ext2.
Ext2 as opposed to ext3 and ext4 does not have journaling. Usually
journaling is a good thing and I use ext4 for my everyday use. However,
when I do backups to hard drives which are intended for "write once,
maybe read a few times" that journaling is not to much use. As such I
have during several years stuck to ext2 for those backups.
I would not call it a show-stopper, I will still have some older
Slackware installations capable of reading my backups, but it would be a
major disadvantage if the kernel in upcoming Slackware 15 totally crashes
when trying to read such a backup.
regards Henrik

ext4 is working great with the new kernels, quiet snappy and ext4 is
totally backwards compatible with ext2, what I would suggest is at the
first signs of trouble to go to /etc/fstab and change ext2 to ext4 and
everything will be the same, meta data and all.

[noel]

Post by Aragorn
ext4 is
totally backwards compatible with ext2,

unless your using a later kernel :)
I have no doubt its nothing that Pat has done, its a patch linus or greg
have authorised.

[Robert Komar]

Post by noel
Is there any reason youve stuck to ext2?
I admit it took me until just last year to move from ext3 to ext4, my
earlier attempts to try ext4 were very many years ago when it was new,
too new, it was unstable, quotas did not work at all, but its matured
over the years, I'm running production servers on it and not missed a
beat.

I'm not the OP, but as I said earlier, I'm running slackware on a
netbook from an SD card. I use ext2 to reduce wear on the card.
If I upgraded to ext4, I would want to disable journalling and
any other feature that does a lot of extra writing to the device.
I'm not sure if that is possible, and it wasn't necessary until
the latest kernel upgrade.

Cheers,
Rob Komar

[Chris Elvidge]

Post by Robert Komar

Post by noel
Is there any reason youve stuck to ext2?
I admit it took me until just last year to move from ext3 to ext4, my
earlier attempts to try ext4 were very many years ago when it was new,
too new, it was unstable, quotas did not work at all, but its matured
over the years, I'm running production servers on it and not missed a
beat.

I'm not the OP, but as I said earlier, I'm running slackware on a
netbook from an SD card. I use ext2 to reduce wear on the card.
If I upgraded to ext4, I would want to disable journalling and
any other feature that does a lot of extra writing to the device.
I'm not sure if that is possible, and it wasn't necessary until
the latest kernel upgrade.
Cheers,
Rob Komar

https://foxutech.com/how-to-disable-enable-journaling/ for example?

i.e. use -O ^has_journal

[Robert Komar]

Post by noel

Post by Robert Komar

[...]

Post by noel
as for not scrolling back, you have that dumbfuck so called genius
torvalds to thank for that. had a boot issue once on a pentium, if I
had scrol back I'd see the obvious error and correct it, but no
scrollback

[...]
I imagine it's a bit hard to do screen scrolling when the keyboard and
video drivers (and all others) are dead/ haltet.
KERNEL PANIC ERROR #DEADBEEF ?
-rasp

I think most of us would prefer the scrolling that was there for almost
thirty years. It wasn't perfect, but it was better than nothing.
Back to the original topic, I have slackware-current running in a VM as
well. The root filesystem is not ext2, so I can boot without the
initrd. In there, I can create an ext2 filesystem in a file, mount
that, write to it, read from it,... So, it looks like ext2 works fine
in that system. Perhaps my problem on the netbook is really with the
initrd as I first thought, and the OP has something broken in his
system. I can't duplicate his ext2 problems within the VM.
Rob Komar

current kernels dont like ext2, plain and simple
I pulled an ol toshiba usb drive out from storage since I knew it was
ext2, nope, stack errors and killed at every first attempt...
converted it to ext4,
tune2fs -O extents,uninit_bg,dir_index,has_journal /dev/sdf1
and voila! no errors, so its just ext2 theyve rooted up
way to go linus...
(im not gonna sift through 10K trolls on LKML to see if its been brought
up there yet.)
(I didnt see mr spitzers post, I didnt miss anything, my troll filters
seem to be working well)

Thanks for the tune2fs tip. It was a bit more work for me, though, so
I thought I'd give an update here in case it helps anyone.

Besides modifying the root filesystem with tune2fs (but with
^has_journal to disable journaling), I modified /etc/fstab to change the
root file system type from ext2 to ext4. I also added the "noatime"
option to reduce writing on the SD card. I had to modify my mkinitrd
script to remove the "-f ext2" option I was using. I also had to modify
the /etc/lilo.conf file and add "rootfstype=ext4" to the "append"
parameter and re-run lilo. Without the last change, the root filesystem
would only be attempted to be mounted as ext2 and would fail because of
unrecognized options. Perhaps autodetection of the filesystem type is
not working properly at that stage, because adding "rootfstype=ext4" to
the kernel boot options seems like it shouldn't be necessary.

The mkinitrd stuff was problematic because it needed the system to be
booted to run. I had to make that change to the script and rebuild the
initrd images before starting to make changes to the root filesystem.
Fortunately I could re-run lilo on the SD card from my desktop, so I was
able to fix that even when the system would not boot.

Cheers,
Rob Komar