The complete sentence is:
oldest non-working kernel is 6.3.7 (package linux-image-6.3.0-1-amd64),
6.3.5 (latest version of package package linux-image-6.3.0-0-amd64) works.
Am 26.07.24 um 10:45 schrieb Stefan:
Hi,
oldest non-working kernel is 6.3.7 (package linux-image-6.3.0-1-amd64),
6.3.5 (latest version of package package linux-image-6.3.0-0-amd64)
The file corruption does not occur if the file is still in buffer.
Furthermore the corruption do not occur during reading, i.e. damaged
files stay damaged after reboot.
In kernels 6.3.* to early 6.5.* another strange error occurs when nvme
module is loaded: "Device not ready; aborting initialisation. CSTS=0x0."
Waiting a while and reloading nvme module helps. Sometimes several tries
are required. See the attached screenshot.
Regards Stefan
Am 24.07.24 um 17:53 schrieb Diederik de Haas:
Control: found -1 6.9.7-1~bpo12+1
On Wednesday, 24 July 2024 17:24:44 CEST Stefan wrote:
I ran a few other tests:
1. tried package "linux-image-6.1.0-22-amd64": works
2. tried package "linux-image-6.9.7+bpo-amd64": does not work
Via https://snapshot.debian.org/package/linux-signed-amd64/ you can find
older (bpo) kernels then 6.5.10+1~bpo12+1 (where it was initially filed
against) and it's useful to know what the last kernel version after 6.1
was where it worked properly and the first one where it broke.
