Re: [PATCH v3] PCI: Workaround wrong flags completions for IDT switch

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 06/14/2017 07:32 AM, Bjorn Helgaas wrote:
On Tue, Jun 13, 2017 at 07:19:57PM -0400, James Puthukattukaran wrote:
On Jun 13, 2017, at 6:14 PM, Bjorn Helgaas <helgaas@xxxxxxxxxx> wrote:
On Tue, Jun 13, 2017 at 02:30:55PM -0400, james puthukattukaran wrote:
On 6/13/2017 1:00 PM, Yinghai Lu wrote:
On Mon, Jun 12, 2017 at 2:48 PM, Bjorn Helgaas <helgaas@xxxxxxxxxx> wrote:
On Fri, Jun 09, 2017 at 04:16:17PM -0700, Yinghai Lu wrote:
From: James Puthukattukaran <james.puthukattukaran@xxxxxxxxxx>

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).
Can you include a URL where this erratum is published?  If not, can
you include the actual erratum text here?
Here's 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.
This doesn't mention anything about disabling ACS.  Issuing a
config write to devices downstream of an IDT bridge sounds simpler
than what this patch does.  Why don't you do that?
The issue is how will we know is the config write succeeds if the
device is not ready? I thought it was simpler to disable acs for the
sake of the read and when we know that the device is ready ( returns
vendor id from read), it's ready for subsequent config write.
If that's a problem, it sounds like the errata text is wrong or at
least incomplete.  If disabling ACS SV is required, the errata text
should mention it.

But I don't think it is a problem.  Per PCIe r3.1, sec 2.3.2, if a
Root Complex receives a CRS completion for a Configuration Write, it
must re-issue the request.

Or did you actually try that and find that it didn't work?
I tried the write and it did not work. This could be because the root port gave up after getting a few CRS responses? I know that the IDT switch takes some time to be ready even to respond with CRS response (firmware initialization, etc). The spec does not say how often the root complex would retry the request and how long. So, I put in a 1 second delay before the write and the write seemed to latch/work. The real issue here is there's no programmatic way (that I know of) to check if the write failed.

I thought instead of putting delays (which might vary for different devices and might not work in certain cases), the sure way of knowing the device exists and is responding to config requests.

--James



[Index of Archives]     [DMA Engine]     [Linux Coverity]     [Linux USB]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Greybus]

  Powered by Linux