Re: [RFC PATCH] docs/memory-barriers.txt: Rewrite "KERNEL I/O BARRIER EFFECTS" section

From: Arnd Bergmann
Date: Tue Feb 19 2019 - 06:32:11 EST

On Tue, Feb 19, 2019 at 11:27 AM Thomas Petazzoni
<thomas.petazzoni@xxxxxxxxxxx> wrote:
> On Mon, 18 Feb 2019 21:37:25 +0100
> Arnd Bergmann <arnd@xxxxxxxx> wrote:
> > Ah, it seems we actually do that on 32-bit ARM, at least on one platform,
> > see 6a02734d420f ("ARM: mvebu: map PCI I/O regions strongly ordered")
> > and prior commits.
> Yes, some Marvell Armada 32-bit platforms have an errata that require
> the PCI MEM and PCI I/O regions to be mapped strongly ordered.
> BTW, this requirement prevents us from using the pci_remap_iospace()
> API from drivers/pci, because it assumes page attributes of
> pgprot_device(PAGE_KERNEL). That's why we're still using the
> ARM-specific pci_ioremap_io() function.


> > > I would still prefer to document the weaker semantics as the portable
> > > interface, unless there are portable drivers relying on this today (which
> > > would imply that it's widely supported by other architectures).
> >
> > I don't know of any portable driver that actually relies on it, but
> > that's mainly because there are very few portable drivers that
> > use inb()/outb() in the first place. How many of those require
> > the non-posted behavior I don't know
> >
> > Adding Thomas, Gregory and Russell to Cc, as they were involved
> > in the discussion that led to the 32-bit change, maybe they are
> > aware of a specific example.
> I'm just arriving in the middle of this thread, and I'm not sure to
> understand what is the question. If the question is whether PCI I/O is
> really used in practice, then I've never seen it be used with Marvell
> platforms (but I'm also not aware of all PCIe devices people are
> using). I personally have a hacked-up version of the e1000e driver
> that intentionally does some PCI I/O accesses, that I use as a way to
> validate that PCI I/O support is minimally working, but that's it.

The main question is whether we know of a portable (not specific
to a single architecture) driver for PCIe (or PCI, but probably not ISA)
that uses outb/outw/outl in a way that relies on the on-posted
behavior of those, compared to posted writeb/writew/writel that are
only required to complete before another I/O operation on the same
device but whose completion is not ordered with regard to the rest of
the system.

I think an example of this would be a driver using outb() to disable
an interrupt, and then relying on the the interrupt no longer happening
after the outb().