On 11/16/2015 01:30 PM, Stephen Warren wrote:
On 11/13/2015 10:58 AM, Andrew Bresticker wrote:
On Fri, Nov 13, 2015 at 8:32 AM, Thierry Reding
<thierry.reding@xxxxxxxxx> wrote:
On Wed, Nov 04, 2015 at 01:59:51PM -0700, Stephen Warren wrote:
On 11/04/2015 10:11 AM, Thierry Reding wrote:
From: Thierry Reding <treding@xxxxxxxxxx>
Extend the binding to cover the set of feature found in Tegra210.
diff --git
a/Documentation/devicetree/bindings/phy/nvidia,tegra-xusb-padctl.txt b/Documentation/devicetree/bindings/phy/nvidia,tegra-xusb-padctl.txt
+For Tegra210, the list of valid PHY nodes is given below:
+- utmi: utmi-0, utmi-1, utmi-2, utmi-3
+ - functions: "snps", "xusb", "uart"
+- hsic: hsic-0, hsic-1
+ - functions: "snps", "xusb"
+- pcie: pcie-0, pcie-1, pcie-2, pcie-3, pcie-4, pcie-5, pcie-6
+ - functions: "pcie-x1", "usb3-ss", "pcie-x4"
+- sata: sata-0
+ - functions: "usb3-ss", "sata"
usb2-bias also needs to be present.
I'm not sure about this. All of the driver code that I've looked deals
with the usb2-bias pad internally. As far as I can tell, this pad needs
to be configured to whatever any of the other pads is configured for. I
think that means if any of the UTMI pads is configured for XUSB then the
usb2-bias pad must also be configured for XUSB. Which would also imply
that if one of the UTMI pads is configured for XUSB, all of them must be
configured for XUSB.
I was told by hardware engineers at NVIDIA that (at least on
Tegra124/Tegra132) the usb2-bias pad must be configured in the
XUSB_PADCTL register space if UTMI pad 0 is muxed to XUSB. If UTMI
pad 0 is muxed to SNPS, then the usb2-bias pad is configured in the
USB register space (base 0x7d000000). You may want to follow up
internally to confirm this. If it's true, that could make things here
a bit nastier, especially if we want to support configurations where
some pads are muxed to XUSB while others are muxed to SNPS.
Hmm. I've certainly successfully tested a configuration where UTMI pad 0
was handled by the SNPS controller and other pads by the XUSB controller
*and* where I set the usb2-bias "pad"'s muxing and configuration via the
XUSB PADCTL module. In that case, I /had/ to configure usb2-bias via
XUSB PADCTL or the other XUSB pads didn't work. However, perhaps that
was because the XUSB controller driver probed before the SNPS driver;
perhaps if they'd probed the other way around and the SNPS driver
configured the bias pad, then everything would have worked without
configuring the bias pad via XUSB PADCTL.
I suppose I'll have to start another internal thread to get the full
details, and differentiate between "recommended" and "supported" and
"must" vs. "can"/"should".
I've discussed this with a HW engineer, and apparently this rule is
simply a recommendation, with the acknowledgement that everything will
work perfectly fine if we ignore it.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
