Re: [PATCH 2/2] PCI: Workaround ACS hw bug for IDT switch

[Alex Williamson <alex.williamson@redhat.com>]

On Wed, 11 Apr 2018 12:06:21 -0400
James Puthukattukaran <james.puthukattukaran@oracle.com> wrote:

> This patch implements the hw workaround found in the IDT switch.
> 
> The IDT switch incorrectly flags an ACS source violation on a read config
> request to an end point device on the completion (IDT 89H32H8G3-YC,
> errata #36) even though the PCI Express spec states that completions are
> never affected by ACS source violation (PCI Spec 3.1, Section 6.12.1.1). 
> Here's
> the specific copy of the errata text
> 
> "Item #36 - Downstream port applies ACS Source Validation to Completions
> Section 6.12.1.1 of the PCI Express Base Specification 3.1 states
> that completions are never affected
> by ACS Source Validation. However, completions received by a
> downstream port of the PCIe switch from a device that has not yet
> captured a PCIe bus number are incorrectly dropped by ACS source
> validation by the switch downstream port.
> 
> Workaround: Issue a CfgWr1 to the downstream device before issuing
> the first CfgRd1 to the device.
> This allows the downstream device to capture its bus number; ACS
> source validation no longer stops
> completions from being forwarded by the downstream port. It has been
> observed that Microsoft Windows implements this workaround already;
> however, some versions of Linux and other operating systems may not. "
> 
> The suggested workaround by IDT is to issue a configuration write to the
> downstream device before issuing the first config read. This allows the
> downstream device to capture its bus number, thus avoiding the ACS
> violation on the completion. In order to make sure that the device is ready
> for config accesses, we do what is currently done in making config reads
> till it succeeds and then do the config write as specified by the errata.
> However, to avoid hitting the errata issue when doing config reads, we
> disable ACS SV around this process.
> 
> The patch does the following -
> 
> 1. Disable ACS source violation if enabled.
> 2. Wait for config space access to become available by reading vendor id
> 3. Do a config write to the end point (errata workaround)
> 4. Enable ACS source validation (if it was enabled to begin with)


It seems like the suggested workaround is to precede the config read
with a config write.  Why do we need to involve toggling ACS SV?  Does
this avoid a race that the device might only see the read and not the
write and therefore still trigger the SV violation?  Per my reply on
1/2, I don't think there's a well defined calling convention for the
quirk as proposed.

> --
> 
>   drivers/pci/quirks.c | 38 ++++++++++++++++++++++++++++++++++++++
>   1 file changed, 38 insertions(+)

There's no sign-off here, the above diffstat should be below the triple
dash, not above.  Please see Documentation/SubmittingPatches.  Thanks,

Alex

> ---
> 
> diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
> index bf423e3..932e0ff 100644
> --- a/drivers/pci/quirks.c
> +++ b/drivers/pci/quirks.c
> @@ -4738,13 +4738,51 @@ static void quirk_gpu_hda(struct pci_dev *hda)
>   DECLARE_PCI_FIXUP_CLASS_FINAL(PCI_VENDOR_ID_NVIDIA, PCI_ANY_ID,
>   			      PCI_CLASS_MULTIMEDIA_HD_AUDIO, 8, quirk_gpu_hda);
> 
> +static int pci_idt_acs_quirk(struct pci_bus *bus, int devfn, int enable,
> +				bool found)
> +{
> +	int pos;
> +	u16 cap;
> +	u16 ctrl;
> +	int retval;
> +	struct pci_dev *dev = bus->self;
> +
> +
> +	/* Write 0 to the devfn device under the PCIE switch (bus->self)
> +	 * as part of forcing the devfn number to latch with the device below
> +	 */
> +	if (found)
> +		pci_bus_write_config_word(bus, devfn, PCI_VENDOR_ID, 0);
> +
> +
> +	/* Enable/disable ACS SV feature (based on enable flag) */
> +	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ACS);
> +	if (!pos)
> +		return -ENODEV;
> +
> +	pci_read_config_word(dev, pos + PCI_ACS_CAP, &cap);
> +
> +	if (!(cap & PCI_ACS_SV))
> +		return -ENODEV;
> 
> +	pci_read_config_word(dev, pos + PCI_ACS_CTRL, &ctrl);
> +
> +	retval = !!(ctrl & cap & PCI_ACS_SV);
> +	if (enable)
> +		ctrl |= (cap & PCI_ACS_SV);
> +	else
> +		ctrl &= ~(cap & PCI_ACS_SV);
> +
> +	pci_write_config_word(dev, pos + PCI_ACS_CTRL, ctrl);
> +	return retval;
> +}
> 
>   static const struct pci_dev_acs_quirk{
>   	u16 vendor;
>   	u16 device;
>   	int (*acs_quirk)(struct pci_bus *bus, int devfn, int enable, bool found);
>   } pci_dev_acs_quirks[] = {
> +	{ PCI_VENDOR_ID_IDT, 0x80b5, pci_idt_acs_quirk},
>   	{0}
>   };
>