[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250214154415.ns5li3ywlxawmzz6@thinkpad>
Date: Fri, 14 Feb 2025 21:14:15 +0530
From: Manivannan Sadhasivam <mani@...nel.org>
To: Janne Grunau <j@...nau.net>
Cc: Bjorn Helgaas <helgaas@...nel.org>,
Alyssa Rosenzweig <alyssa@...enzweig.io>,
Lorenzo Pieralisi <lpieralisi@...nel.org>,
Krzysztof Wilczyński <kw@...ux.com>,
Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>,
Rob Herring <robh@...nel.org>, Bjorn Helgaas <bhelgaas@...gle.com>,
linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] PCI: apple: Add depends on PAGE_SIZE_16KB
On Tue, Feb 11, 2025 at 11:00:19PM +0100, Janne Grunau wrote:
> On Tue, Feb 11, 2025 at 08:56:04PM +0100, Janne Grunau wrote:
> > On Tue, Feb 11, 2025 at 12:38:59PM -0600, Bjorn Helgaas wrote:
> > > On Tue, Feb 11, 2025 at 01:03:52PM -0500, Alyssa Rosenzweig wrote:
> > > > From: Janne Grunau <j@...nau.net>
> > > >
> > > > The iommu on Apple's M1 and M2 supports only a page size of 16kB and is
> > > > mandatory for PCIe devices. Mismatched page sizes will render devices
> > > > useless due to non-working DMA. While the iommu prints a warning in this
> > > > scenario, it seems a common and hard to debug problem, so prevent it at
> > > > build-time.
> > >
> > > Can we include a sample iommu warning here to help people debug this
> > > problem?
> >
> > I don't remember and it might have changed in the meantime due to iommu
> > subsystem changes. What currently happens is that
> > apple_dart_attach_dev_identity() fails with -EINVAL. I can't say whether
> > that results in a failure to probe now. I'll test and report back.
>
> Using a kernel with 4K page size It results now in following warning per
> PCI device:
>
> | ------------[ cut here ]------------
> | WARNING: CPU: 7 PID: 260 at drivers/iommu/iommu.c:2979 iommu_setup_default_domain+0x348/0x590
This should be added to the patch description.
- Mani
--
மணிவண்ணன் சதாசிவம்
Powered by blists - more mailing lists