On 3/11/25 4:00 PM, Jens Axboe wrote:

On 3/11/25 8:59 AM, Milan Broz wrote:

...

On 3/11/25 4:03 AM, Shinichiro Kawasaki wrote:

...

I created fix candidate patches to address the blktests nvme/039 failure [1]. This may work for the failures Ondrej and Milan observe too, hopefully.

Hi,

I quickly tried to run the test with todays' mainline git and mentioned two patches:   https://lkml.kernel.org/linux-block/20250311024144.1762333-2-shinichiro.kawa...   https://lkml.kernel.org/linux-block/20250311024144.1762333-3-shinichiro.kawa... and it looks like our SED Opal tests are fixed, no errors or warnings, thanks!

...

Jens, Alan, could you take a look in the patches and see if they make sense?

Please, fix it before the 6.14 final, this could cause serious data corruption, at least on systems using SED Opal drives.

Fix is already queued up.

Hi Jens,

it seems the bug was "successfully" backported to stable 6.13.4+ and I do not see any fix in the stable queue yet.

We were hit by it today, as all locked ranges for Opal devices now returning empty data instead of failing read.

Please could you check what need to be backported to 6.13 stable?

IMO these patches should be backported: https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commi... https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commi...

Thanks, Milan