Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #85015 > unrolled thread

Bug#1076372: linux-image-6.12.1+debian+tj: Force MDTS patch

Back to article view | Back to linux.debian.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Bug#1076372: linux-image-6.12.1+debian+tj: Force MDTS patch Stefan <debian@simg.de> - 2025-01-03 15:40 +0100
    Bug#1076372: linux-image-6.12.1+debian+tj: Force MDTS patch Stefan <debian@simg.de> - 2025-03-24 17:00 +0100
      Bug#1076372: linux-image-6.12.1+debian+tj: Force MDTS patch Salvatore Bonaccorso <carnil@debian.org> - 2025-03-24 21:10 +0100
        Bug#1076372: linux-image-6.12.1+debian+tj: Force MDTS patch Stefan <debian@simg.de> - 2025-03-28 00:10 +0100

#85015 — Bug#1076372: linux-image-6.12.1+debian+tj: Force MDTS patch

Hi,

Am 19.12.24 um 14:05 schrieb Tj:
> Package: linux-image-6.12.1+debian+tj
> Followup-For: Bug #1076372
> X-Debbugs-Cc: tj.iam.tj@proton.me
>
> This patch ought to enable forcing MDTS for *all* NVME devices. Untested
> but applies on master and back to v6.3.7 at least, and builds for
> master.

Almost a perfect hit. The bug is triggered by the line after the patch,
where `ctrl->max_hw_sectors` is set to another value than defined by
MDTS. That behavior was introduced in 6.3.7 and is closely related to
the patch you mentioned earlier.

The data corruption disappears, if the `ctrl->max_hw_sectors` is set
according to MDTS (no need to alter that setting), see comment #4 at
https://bugzilla.kernel.org/show_bug.cgi?id=219609 for details.

Regards Stefan

[toc] | [next] | [standalone]

#86591

Hi,

the issue seems to be fixed in Firmware (BIOS) Version 4.10

Regards Stefan

m 21.03.25 um 20:00 schrieb Salvatore Bonaccorso:
> Control: tags -1 + moreinfo
>
> Hi Stefan,
>
> We once again looked at your bugs in our weekly kernel team meeting
> (cf.
> https://meetbot.debian.net/debian-kernel/2025/debian-kernel.2025-03-19-20.00.html)
> and the lengthy upstream discussion is not clear to us with the
> results.
>
> Is it the case that the BIOS update was tested and that it fixed all
> of your reported regressions (aka. #1076372 and #1090717 can be
> considered not a src:linux issue after all, and can be closed)?
>
> Is this understanding correct? Can you give us a status update?
>
> Regards,
> Salvatore
>

[toc] | [prev] | [next] | [standalone]

#86595

Hi Stefan,

On Mon, Mar 24, 2025 at 04:50:31PM +0100, Stefan wrote:
> Hi,
> 
> the issue seems to be fixed in Firmware (BIOS) Version 4.10

Only the "Longsys/Kingston NVMe drive" one or as well "Lexar NM790
NVMe drive" one?

Regards,
Salvatore

[toc] | [prev] | [next] | [standalone]

#86645

Hi Salvatore,

the read errors in 1st M.2 socket, which were easy to reproduce and
which occurred with many different SSD's, are fixed.

The write errors that occurred with the Kingston SSD in 2nd Socked on
rear side were either caused by a hardware failure, are fixed or won't
be fixed, because these errors are (almost) impossible to reproduce.
(Furthermore the Kingston SSD moved to another PC.)

Regards Stefan

Am 24.03.25 um 20:59 schrieb Salvatore Bonaccorso:
> Hi Stefan,
>
> On Mon, Mar 24, 2025 at 04:50:31PM +0100, Stefan wrote:
>> Hi,
>>
>> the issue seems to be fixed in Firmware (BIOS) Version 4.10
>
> Only the "Longsys/Kingston NVMe drive" one or as well "Lexar NM790
> NVMe drive" one?
>
> Regards,
> Salvatore
>

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web