[GIT PULL] ata fixes for 6.16-rc4

[GIT PULL] ata fixes for 6.16-rc4

[Niklas Cassel]

Linus,

The following changes since commit a7b3b77fd111d49f8e25624e4ea1046322a57baf:

  ata: ahci: Disallow LPM for Asus B550-F motherboard (2025-06-17 11:11:39 +0200)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux tags/ata-6.16-rc4

for you to fetch changes up to 3e0809b1664b9dc650d9dbca9a2d3ac690d4f661:

  ata: ahci: Use correct DMI identifier for ASUSPRO-D840SA LPM quirk (2025-06-25 15:17:57 +0200)

----------------------------------------------------------------
ata fix for 6.16-rc4

 - Use the correct DMI identifier for ASUSPRO-D840SA LPM quirk such that
   the quirk actually gets applied (me)

----------------------------------------------------------------
Niklas Cassel (1):
      ata: ahci: Use correct DMI identifier for ASUSPRO-D840SA LPM quirk

 drivers/ata/ahci.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Re: [GIT PULL] ata fixes for 6.16-rc4

[pr-tracker-bot]

The pull request you sent on Fri, 27 Jun 2025 12:14:31 +0200:

> git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux tags/ata-6.16-rc4

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/9c7331f150549d3c724051aa11c5dcb51bf3aa99

Thank you!

-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html

Re: [GIT PULL] ata fixes for 6.16-rc4

[Linus Torvalds]

On Fri, 27 Jun 2025 at 03:14, Niklas Cassel <cassel@kernel.org> wrote:
>
>  - Use the correct DMI identifier for ASUSPRO-D840SA LPM quirk such that
>    the quirk actually gets applied (me)

I've obviously pulled this, but it made me think (once again) that our
DMI matching routines are incredibly odd and hard to use.

Maybe "hard to use" isn't the right phrase: it's just that they are
pretty inflexible and you have to write out the exact right
incantation.

Yes, we have that whole "exact vs relaxed" string comparison thing,
and default to a relaxed comparison (and almost everybody uses that).
The difference there is "strcmp" vs "strstr".

But it still requires the exact slot.

So it's still incredibly inflexible with the whole "oh, you wanted to
compare the *BOARD* name, not the BIOS vendor name or the system
vendor name".

Yes, in some situations you probably do want the specificity, but I do
suspect we could have been much more relaxed, and have some kind of
"board / bios / vendor - who cares?" model of DMI matching (and same
for the whole "name vs version").

Because I strongly suspect that if you had two strings like "ASUSTeK
COMPUTER INC." and "ASUSPRO D840MB_M840SA", and you just asked for the
sane match, it would all just work without having to specify that it's
DMI_SYS_VENDOR / DMI_PRODUCT_NAME.

So the whole "you need to specify the exact DMI slot to match" does
seem counter-productive.

That said, it's probably much too late to worry about this, and
obviously the whole baroque exact slot matching model mostly does
work.

Except when the slot gets copy-pasted from the entry above, which used
a different one because one vendor filled in "version" instead of
"product name".

            Linus