Yesterday I read a report of ext4 file system corruption with the 4.19.x kernel,

http://lkml.iu.edu/hypermail/linux/k...1.3/02611.html

And, this morning it is being reported in the media,

https://www.phoronix.com/scan.php?pa....19-Corruption

There is a difference of opinion as to if this is true and, if so, the cause,

http://lkml.iu.edu/hypermail/linux/k...1.3/03195.html

Edit in: FWIW, I have not had any problems with 4.19.3/4/5 with -current and the recent Nvidia "Long lived" or "short lived" drivers.

4 members found this post helpful.

I did - possibly (would have to investigate, but ext4 died).

Slackware-current

SSD, SATA-attached

I have two SSD disks (SATA), the one that failed was unencrypted, with four primary partitions only
( I have encrypted partition on the second SSD - but this one thankfully is ok)

yes I am using discard
e.g.
/dev/sda1 / ext4 discard,noatime,errors=remount-ro 1 1

- ok will turn discard off

1 members found this post helpful.

In preparation for hopefully a new stable release in the next month or two ( a guy can hope) it was time to test the latest kernel 4.19.y. Happily I can report that 4.19.5 does build under 14.2 and I've been running for last 24 hours without issues. I used the config file from dusks 4.19 /config directory as a starter config, executed make oldconfig, then make menuconfig, changed a few things for the emachine Pentium D 820 (older processor), saved, and then executed the standard remaining steps to build a kernel. Currently running as well as the last 4.4.164 kernel from last week. Using straight nouveau with an NVIDIA 8400GS card. Stable Slackware64 with llvm on ext4 SATA drive, no multilib, only QT5 and vlc from AlienBob. Seems quite stable. Now if I could just get the crazy USB 058f:6377 Alcor Micro Corp. AU6375 4-LUN card internal reader to actually function correctly (ie allow read/write and respond in less than 15 seconds to card insertions) my system would be really stable. Cheers, BrianA_MN

2 members found this post helpful.

Duplicate

Kernel updates 4.19.6, 4.14.85, 4.9.142, 4.4.166 and 3.18.128 are now available at,

https://www.kernel.org/

The change logs,

https://cdn.kernel.org/pub/linux/ker...angeLog-4.19.6

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.14.85

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.9.142

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.4.166

https://cdn.kernel.org/pub/linux/ker...geLog-3.18.128


Dave's Unofficial Slackbuilt Kernel update for the 4.4.x series
can be found at, https://dusk.idlemoor.tk

3 members found this post helpful.

4.19.6 x 64 running fine here

1 members found this post helpful.

Ditto, with -current and the Nvidia-415.18 "short lived" driver.

1 members found this post helpful.

The newest "mainline" kernel, version 4.20-rc5, has been been made available for testing.

The tarball, https://git.kernel.org/torvalds/t/linux-4.20-rc5.tar.gz

The 4.20.0 kernel maybe released by Christmas, maybe not. Read all about it in
Mr. Torvalds' announcement, http://lkml.iu.edu/hypermail/linux/k...2.0/00502.html

It looks like the STIBP problem will be resolved with the release of the 4.19.7 kernel update due out late tomorrow or early Thursday, depending on your time zone.

http://lkml.iu.edu/hypermail/linux/k...2.0/02166.html

Curious, why not official 4.19 config directly from PV? You can find it here:

http://ftp.osuosl.org/pub/slackware/...ernel-configs/

1 members found this post helpful.

It is now being reported this corruption problem is not from within the EXT4 code.

The story is here, https://www.phoronix.com/scan.php?pa...ssue-Likely-MQ

1 members found this post helpful.

Because that config is for -current Slackware. Dusk's is for stable (at this time 14.2). I didn't want to use the 4.19.5 config from PV in case it had any changes implemented specifically for packages of -current or newer CPU and hardware. Dusk's version was specifically for "stable" x86_64. I did do a diff -y of the two files ans say PV's has some hardware specific changes for current. I also note that in the -current changelog that PV has implemented some specific kernel config changes for -current binaries (like newer gcc and glibc) and new hardware than I'm NOT using. I did not want to chase ghosts or functionality introduced for newer hardware and binary packages than stable requires. It would be interesting to try a kernel build with PV's config against 14.2, but I wanted a more sure config. So why didn't I just use Dusk's kernel binary builds? Because I know how to build for my specific installation with stable's gcc and glibc. May be the kernel building experts know more about which config to start with and I'm willing to learn from them. My decisions were to take a cautious approach, and that worked. HTH Cheers BrianA_MN

1 members found this post helpful.

Slackware-current users didn't notice the bug because of

3 members found this post helpful.

No it isn't (and that makes no difference anyway). When I've got time, I'll replace my amateur config with Pat's

3 members found this post helpful.

Kernel updates 4.19.7, 4.14.86 and 4.9.143 are now available at,

https://www.kernel.org/

The change logs,

https://cdn.kernel.org/pub/linux/ker...angeLog-4.19.7

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.14.86

https://cdn.kernel.org/pub/linux/ker...ngeLog-4.9.143

1 members found this post helpful.