https://bugzilla.kernel.org/show_bug.cgi?id=200753 Martin Steigerwald (Martin@xxxxxxxxxxxx) changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |Martin@xxxxxxxxxxxx --- Comment #12 from Martin Steigerwald (Martin@xxxxxxxxxxxx) --- Interesting. That may explain why on doing a "blkdiscard" on the partition of an Ext4 filesystem mounted as / on an SSD does not seem to show anything in "dmesg -w" started before the "blkdiscard" command. I always was quite a bit puzzled about this. This happened with a running Plasma desktop which attempts to write things quite regularly. When accessing a command not yet in cache I do receive "error in format of binary file" (roughly translated from German language). But still nothing in kernel log ring buffer. This is on some Fujitsu workstation with Intel CPU + chipset. I assume the kernel uses libata AHCI for accessing it. No access to such a workstation at the moment. I made a test with creating an 1 exibyte large XFS filesystem on a large sparse file on a way smaller XFS filesystem and on writing to the 1 exibyte large filesystem beyond the free space left in the underlying XFS gave error messages like "lost buffer write" in kernel log. I never tried the "blkdiscard" thing with XFS so far tough. -- You are receiving this mail because: You are watching the assignee of the bug.