tags 926202 + fixed-upstream We (at Kinvolk) hit the very same bug (fault_state(0x5862)!) and created a way to reproduce with fio to have heavy IO. All the info is here: https://github.com/flatcar-linux/Flatcar/issues/12
As it's shown in the link, this was fixed in https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=9e034c61951acdc1378d936875641fcb61b6a7a3 and 4.19.65 is the first stable release to include it. Also, note there are more fixes in later stable release, but our repro doesn't hit this as of 4.19.65 The bug you reference in Ubuntu is similar but is not the same, IIUC. The fault state hex code returned there is different (not 0x5862), so this probably indicates a different controller error. Debian linux-image-4.19.0-6-amd64 seems to incoporate this commit/fix in version 4.19.67-1. In backports (I see the report is for backports), this seems to be fixed in linux-image-4.19.0-0.bpo.6-amd64 version 4.19.67-2~bpo9+1. Hope this helps
