https://bugzilla.kernel.org/show_bug.cgi?id=217769 --- Comment #12 from Mariusz Gronczewski (xani666@xxxxxxxxx) --- The system was never marked as dirty so fsck (which fixed the problem) wasn't happening on boot. So yeah, once I upgraded kernel if it failed the first time it failed on every reboot after till manually fscking or downgrading kernel. On the one machine where it failed after the boot (reading of certain files triggered it), that too was repeatable Maybe on "corruption of in-memory data" driver should also mark FS as dirty? After all if data is corrupted from actual memory error (and not like here, from loading bad data) there is nonzero chance some of that data ended up being written to disk -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
