Re: [PATCH] PCI: Fix endianness issues in pci_bus_read_config()

From: Hans Zhang
Date: Wed Aug 13 2025 - 03:53:04 EST




On 2025/8/13 15:47, Niklas Schnelle wrote:
On Tue, 2025-08-12 at 22:44 +0800, Hans Zhang wrote:

On 2025/8/4 18:09, Gerd Bayer wrote:
On Mon, 2025-08-04 at 11:06 +0800, Hans Zhang wrote:

On 2025/8/1 19:30, Gerd Bayer wrote:
On Fri, 2025-08-01 at 16:24 +0530, Manivannan Sadhasivam wrote:

<--- snip --->



Dear all,

According to the issue mentioned by Lukas and Mani. Gerd has already
been tested on the s390. I have tested it on the RK3588 and it works
fine. RK3588 uses Synopsys' PCIe IP, that is, the DWC driver. Our
company's is based on Cadence's PCIe 4.0 IP, and the test function is
normal. All the platforms I tested were based on ARM.

The following is the patch based on the capability-search branch. May I
ask everyone, do you have any more questions?

Gerd, if there's no problem, I'll add your Tested-by label.

Before you add that I'd like to re-test with the "final" patch.

Branch:
https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/log/?h=capability-search

Patch:

<--- snip --->

Please bear with me while I'm working on that.


Dear Gerd,

May I ask if there is any update?



I plan to submit the v15 version of my series based on v6.17-rc1.
The modification method is like the previous comment:
https://lore.kernel.org/linux-pci/06012cc6-824d-4a7d-85c9-9995ec915382@xxxxxxx/

Best regards,
Hans


Hi Hans,

Gerd is currently out so I just gave the patch you provided against
capability-search-v14 a try on s390. Didn't see any issues with the
previously broken device probing. As I understand it Bjorn asked you to
send a complete v15 and then for people to test that. I like that
approach and would prefer to provide a Tested-by for v15 rather than
via a patch on top. Gerd should be back next week too. Does that work
for you?


Hi Niklas,

Ok, no problem. I'm starting to prepare the patch for V15 now.

Best regards,
Hans