The Latest Kernel Release.

enorbet · 10-10-2019, 12:51 PM

Quote:

Originally Posted by abga

Thanks for the warning with the further development of the "beloved" Module Versioning. And I thought they'll scrap it for good.
Honestly, I provided the hints with the symlink because you were compiling a development kernel and considered that maybe they forgot to create it in the modules_install make target. I couldn't imagine that they changed the format of Module.symvers ... well, boredom should be catalogued as a psychiatric disorder.

Boredom? What's that? There isn't time in one lifetime just to read all the books I'd like to read... all the places I'd like to see... all the girls I'd like to.... ummm... meet, yeah that's it, meet. Although at my age all I can do is run to the end of the chain and bark

3rensho · 10-11-2019, 12:47 PM

4.19.79 and related have been released.

cwizardone · 10-11-2019, 03:34 PM

Kernel updates 5.3.6, 4.19.79 and 4.14.149 are now available at

https://www.kernel.org/

The change logs,

https://cdn.kernel.org/pub/linux/ker...hangeLog-5.3.6

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.19.79

https://cdn.kernel.org/pub/linux/ker...geLog-4.14.149

And, as previously mentioned, the 5.2.yy series is now EOL.
https://cdn.kernel.org/pub/linux/ker...angeLog-5.2.21

Andypoo · 10-12-2019, 11:43 AM

Just upgraded a btrfs system from 4.19.73 to 4.19.78 on Friday to be greeted by a rootfs kernel panic on one of my servers.

Thought at first I had botched the upgrade with 4.19.79 being released at the same time, but have now gone in via IPMI and upgraded to 4.19.79 and the same panic occurs.

Does anybody know where to get older kernels these days (4.19.73 - 4.19.77) so I can try to narrow down and assume I have not had a momentary mental lapse?

I am using the standard huge kernel without an initrd.

Thanks,
Andypoo.

ponce · 10-12-2019, 11:48 AM

Quote:

Originally Posted by Andypoo

Does anybody know where to get older kernels these days (4.19.73 - 4.19.77) so I can try to narrow down and assume I have not had a momentary mental lapse?

http://slackware.uk/cumulative/slack...slackware64/a/

kgha · 10-12-2019, 11:50 AM

Quote:

Originally Posted by Andypoo

Does anybody know where to get older kernels these days (4.19.73 - 4.19.77) so I can try to narrow down and assume I have not had a momentary mental lapse?

You can find them here: http://slackware.uk/cumulative/slack...slackware64/a/

Ooops - great minds think alike, but some think faster

Andypoo · 10-12-2019, 12:31 PM

Thanks ponce and kgha. Seems it wasn't anything silly that I did. Just did a quick check through these kernels and discovered:

4.19.73 OK
4.19.74 OK
4.19.75 OK
4.19.76 PANIC
4.19.77 PANIC
4.19.78 PANIC
4.19.79 PANIC

Now looking into the changes. There were no btrfs-specific changes between 4.19.75 and 4.19.76 that I can see at first glance from kernel.org.

EDIT: No Slackware config changes between 4.19.75 and 4.19.76 so it must be something upstream.

Andypoo.

Andypoo · 10-12-2019, 01:43 PM

For the curious, there is a nasty regression in F2FS code that leads to some (non-F2FS) filesystems no longer being mounted.

This is the commit that broke 4.19.76:

https://git.kernel.org/pub/scm/linux...64199aadefb493

And for the background into why it broke, there was a patch in 5.3 about this from July:

https://git.kernel.org/pub/scm/linux...e994158b6f8eb3

I e-mailed the maintainers. Does anybody know the proper way to report major regressions in longterm? This one seems pretty bad as people may find it out the same way I did (on a server kernel upgrade).

EDIT: Issue is present on 4.14, 4.19 and 5.2 currently.

Andrew.

blancamolinos · 10-12-2019, 05:25 PM

This is a demonstration of the disastrous patch policy of LTS kernels. The people who maintain them do not even verify that the kernel can boot. Hopefully they will check that the kernel with the patches compiles without errors, but who knows. It is better to follow the stable kernel than this crap of LTS kernels.

Andypoo · 10-12-2019, 05:32 PM

Quote:

Originally Posted by blancamolinos

This is a demonstration of the disastrous patch policy of LTS kernels. The people who maintain them do not even verify that the kernel can boot. Hopefully they will check that the kernel with the patches compiles without errors, but who knows. It is better to follow the stable kernel than this crap of LTS kernels.

To be fair, they did probably boot for most people. (I couldn't find any cases personally of people yelling about this yet besides me).

But it did strike me as odd how this commit made it to LTS (two LTS infact, on different days!). I think they were trying to deal with a separate F2FS issue at the time, but this patch seems independent from that and is essentially some refactoring to standardise behaviour for the future, with the resulting issue potentially affecting a much wider audience (since it is now no longer about F2FS but other filesystems).

Andypoo.

Lysander666 · 10-12-2019, 05:38 PM

For what it's worth, I'm running 4.19.76 with no issues. I don't update with every new kernel point release though, but when it's convenient for me.

rogan · 10-12-2019, 06:12 PM

One of my machines are using btrfs and haven't had any problems sofar on any kernel (up to 5.4.0-rc2).
I use btrfs tools in the initrd with the generic kernel (add -B when making it). The huge kernel however remains untested. The root file system is single-device on a sata ssd.

tramtrist · 10-12-2019, 06:43 PM

Disappointing

rogan · 10-12-2019, 06:50 PM

I just fired up the btrfs machine and installed the 4.19.79 slackware huge kernel.
It indeed panics on root device access. The generic works as usual (with the btrfs tools).
My other kernels on this machine have only btrfs support compiled in.

rogan · 10-12-2019, 06:53 PM

It seems as an issue present on kernels with support for many file systems compiled in...
I'll try a huge without f2fs tomorrow.