[PATCH v3] PCI: dw-rockchip: Enable async probe by default

[Lukas Wunner]

On Tue, Mar 24, 2026 at 08:44:22PM -0700, Dmitry Torokhov wrote:
> I think the bigger question is why PCI does something different from
> every other bus? Why doesn't it rely on driver core to bind devices to
> driver?

This was introduced in 2013 with commit 4f535093cf8f ("PCI: Put pci_dev
in device tree as early as possible").  Corresponding patch submission:

https://lore.kernel.org/all/1358803252-20639-12-git-send-email-yinghai@kernel.org/

The rationale in the commit message for performing device registration
earlier is that devices would otherwise not be found by for_each_pci_dev().
I don't know why that was a problem at the time.  The patch is the last
one in the series, so there's no other patch succeeding and thus
depending on it, which could provide a clue why this was necessary.

Needless to say, changing the behavior now risks regressions.

The author of the patch is apparently no longer around.  Last activity
on the mailing list was in October 2017.  Access to kernel.org infra
persists though, so maybe he's still reachable:

https://git.kernel.org/pub/scm/docs/kernel/pgpkeys.git/plain/keys/A7C92F7A39BD11DA.asc

I'm not defending the patch, merely trying to explain the history
as far as it is possible.

Thanks,

Lukas