[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20251217151952.GG31492@ziepe.ca>
Date: Wed, 17 Dec 2025 11:19:52 -0400
From: Jason Gunthorpe <jgg@...pe.ca>
To: Manivannan Sadhasivam <mani@...nel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@....qualcomm.com>,
Bjorn Helgaas <bhelgaas@...gle.com>, linux-pci@...r.kernel.org,
linux-kernel@...r.kernel.org, iommu@...ts.linux.dev,
Naresh Kamboju <naresh.kamboju@...aro.org>,
Pavankumar Kondeti <quic_pkondeti@...cinc.com>,
Xingang Wang <wangxingang5@...wei.com>,
Marek Szyprowski <m.szyprowski@...sung.com>,
Robin Murphy <robin.murphy@....com>
Subject: Re: [PATCH v2 3/4] PCI: Disable ACS SV capability for the broken IDT
switches
On Tue, Dec 09, 2025 at 08:20:39PM +0900, Manivannan Sadhasivam wrote:
> On Tue, Dec 02, 2025 at 03:15:33PM -0400, Jason Gunthorpe wrote:
> > On Tue, Dec 02, 2025 at 07:52:50PM +0530, Manivannan Sadhasivam wrote:
> > > @@ -544,6 +544,7 @@ struct pci_dev {
> > > #endif
> > > u16 acs_cap; /* ACS Capability offset */
> > > u16 acs_capabilities; /* ACS Capabilities */
> > > + u16 acs_broken_cap; /* Broken ACS Capabilities */
> >
> > Why do we need this? Have the quirk function accep tthe
> > acs_capabilities from the register and return the value to program
> > into struct pci_dev ?
> >
>
> We dont have any quirk levels between pci_acs_init() and pci_acs_enable() that
> will allow us to modify pci_dev::acs_capabilities in the quirk function. Hence,
> I came up with one more member to pass the broken caps.
Call the quirk function directly from the ACS path? We have things
like that already for ACS?
Jason
Powered by blists - more mailing lists