Re: [PATCH v1 for-next] spi: mchp-pci1xxxx: Updated memcpy implementation for x64 and bcm2711 processors

[Mark Brown]

On Wed, Mar 19, 2025 at 01:58:38PM +0000, Rengarajan.S@xxxxxxxxxxxxx wrote:
> Hi Mark,
> 
> Thanks for reviewing the patch and apologies for delayed response.
> 
> On Mon, 2025-02-24 at 14:30 +0000, Mark Brown wrote:
> > [Some people who received this message don't often get email from
> > broonie@xxxxxxxxxx. Learn why this is important at 
> > https://aka.ms/LearnAboutSenderIdentification ]
> > 
> > EXTERNAL EMAIL: Do not click links or open attachments unless you
> > know the content is safe

You have not quoted any context from the message you're replying to so I
don't really know what you're talking about.

> I went through several patches related to similar issues, and mostsuggest handling it on a SoC basis. The reasoning is that a system
> may have an affected PCIe root complex while still having other
> devices in the SoC that can, or even require, 64-bit accesses.

> The following are some of the patches that I had looked into:
> https://lore.kernel.org/lkml/20210226140305.26356-2-nsaenzjulienne@xxxxxxx/T/#u

That's a adding a generic 64bit-mmio-broken property - that's an example
of something that's not quirking off the SoC compatible.  Doesn't seem
to have reached mainline though.

> https://lore.kernel.org/linux-arm-kernel/c188698ca0de3ed6c56a0cf7880e1578aa753077.camel@xxxxxxx/T/#u

> Can you please suggest any alternate methods that we could use to
> handle this in a more generic manner instead of making it Soc-specific.

That thread seems to be going down a similar direction - adding a
generic quirk that the accesses are broken.  Both these threads seem to
be suggesting something like what I was thinking of, you've got a
generic DT property or some other indication that the device can't use
64 bit accesses on this platform.

Attachment: signature.asc
Description: PGP signature