Patchwork [V1] USB: Add support for Xilinx USB host controller

login
register
mail settings
Submitter Julie Zhu
Date Sept. 21, 2009, 7:38 p.m.
Message ID <20090921193821.CB9DF1620046@mail78-sin.bigfish.com>
Download mbox | patch
Permalink /patch/34016/
State Superseded
Headers show

Comments

Julie Zhu - Sept. 21, 2009, 7:38 p.m.
Add bus glue driver for Xilinx USB host controller. The controller can be
configured as HS only or HS/FS hybrid. The driver uses the device tree file
to configure the driver according to the setting in the hardware system.

This driver has been tested with usbtest using the NET2280 PCI card.

Signed-off-by: Julie Zhu <julie.zhu@xilinx.com>
Signed-off-by: John Linn <john.linn@xilinx.com>

---

V1 - Updated per comments from Benjamin Herrenschmidt
     <benh@kernel.crashing.org> and Grant Likely <grant.likely@secretlab.ca>.
---
 Documentation/powerpc/dts-bindings/xilinx.txt |   11 +
 drivers/usb/host/Kconfig                      |   15 +-
 drivers/usb/host/ehci-hcd.c                   |    5 +
 drivers/usb/host/ehci-xilinx-of.c             |  302 +++++++++++++++++++++++++
 4 files changed, 331 insertions(+), 2 deletions(-)
 create mode 100644 drivers/usb/host/ehci-xilinx-of.c
Grant Likely - Sept. 21, 2009, 8:47 p.m.
On Mon, Sep 21, 2009 at 12:38 PM, Julie Zhu <julie.zhu@xilinx.com> wrote:
> Add bus glue driver for Xilinx USB host controller. The controller can be
> configured as HS only or HS/FS hybrid. The driver uses the device tree file
> to configure the driver according to the setting in the hardware system.

Looks good, one minor nit....

> +static struct of_device_id ehci_hcd_xilinx_of_match[] = {
> +               {.compatible = "xlnx,xps-usb-host-1.00.a",},
> +               {.compatible = "xlnx,xps-usb-host-1.01.a",},
> +               {.compatible = "xlnx,xps-usb-host-1.02.a",},
> +       {},
> +};

Ideally, 1.01 and 1.02 device tree nodes would also claim
compatibility with the original 1.00.a version (assuming of course
that it really is register level compatible) making the 2nd and 3rd
lines redundant, but that's just a minor nit and this doesn't hurt
anything.  Do what you prefer.

Acked-by: Grant Likely <grant.likely@secretlab.ca>

Cheers,
g.
Benjamin Herrenschmidt - Sept. 21, 2009, 8:49 p.m.
On Mon, 2009-09-21 at 13:38 -0600, Julie Zhu wrote:

> +	iounmap(hcd->regs);
> +err_ioremap:
> +	irq_dispose_mapping(irq);

You missed that one too :-)

I'm actually considering making irq_dispose_mapping() a private
API and rename it to something like __irq_dispose_mapping() :-)

I think the only reasonable use case for it is when re-allocating
MSIs to avoid using the whole virq space.

Cheers,
Ben.
Anton Vorontsov - Sept. 21, 2009, 10:56 p.m.
On Tue, Sep 22, 2009 at 06:49:40AM +1000, Benjamin Herrenschmidt wrote:
> On Mon, 2009-09-21 at 13:38 -0600, Julie Zhu wrote:
> 
> > +	iounmap(hcd->regs);
> > +err_ioremap:
> > +	irq_dispose_mapping(irq);
> 
> You missed that one too :-)
> 
> I'm actually considering making irq_dispose_mapping() a private
> API and rename it to something like __irq_dispose_mapping() :-)

I.e. turning irq_dispose_mapping() into nop, and __-prefixed
(private) variant would actually dispose irq mapping?

The nop part is important, because otherwise having no cleanup
variant of an allocation/mapping routine feels odd and might raise
questions by non-ppc people (e.g. subsystem maintainers when they
see no cleanup path for such a call).

And if we'll want to refcount irq mappings in the future, some
drivers will be ready without modifications, i.e. the ones that
got irq_dispose_mapping() wrong:

$ git grep irq_dispose drivers/ | wc -l
42

Thanks,
Benjamin Herrenschmidt - Sept. 21, 2009, 11:43 p.m.
On Tue, 2009-09-22 at 02:56 +0400, Anton Vorontsov wrote:
> On Tue, Sep 22, 2009 at 06:49:40AM +1000, Benjamin Herrenschmidt wrote:
> > On Mon, 2009-09-21 at 13:38 -0600, Julie Zhu wrote:
> > 
> > > +	iounmap(hcd->regs);
> > > +err_ioremap:
> > > +	irq_dispose_mapping(irq);
> > 
> > You missed that one too :-)
> > 
> > I'm actually considering making irq_dispose_mapping() a private
> > API and rename it to something like __irq_dispose_mapping() :-)
> 
> I.e. turning irq_dispose_mapping() into nop, and __-prefixed
> (private) variant would actually dispose irq mapping?
> 
> The nop part is important, because otherwise having no cleanup
> variant of an allocation/mapping routine feels odd and might raise
> questions by non-ppc people (e.g. subsystem maintainers when they
> see no cleanup path for such a call).
> 
> And if we'll want to refcount irq mappings in the future, some
> drivers will be ready without modifications, i.e. the ones that
> got irq_dispose_mapping() wrong:

Yup. I agree. I'll look into it

Ben.

Patch

diff --git a/Documentation/powerpc/dts-bindings/xilinx.txt b/Documentation/powerpc/dts-bindings/xilinx.txt
index 80339fe..ea68046 100644
--- a/Documentation/powerpc/dts-bindings/xilinx.txt
+++ b/Documentation/powerpc/dts-bindings/xilinx.txt
@@ -292,4 +292,15 @@ 
        - reg-offset : A value of 3 is required
        - reg-shift : A value of 2 is required
 
+      vii) Xilinx USB Host controller
+
+      The Xilinx USB host controller is EHCI compatible but with a different
+      base address for the EHCI registers, and it is always a big-endian
+      USB Host controller. The hardware can be configured as high speed only,
+      or high speed/full speed hybrid.
+
+      Required properties:
+      - xlnx,support-usb-fs: A value 0 means the core is built as high speed
+                             only. A value 1 means the core also supports
+                             full speed devices.
 
diff --git a/drivers/usb/host/Kconfig b/drivers/usb/host/Kconfig
index 1a920c7..95e4689 100644
--- a/drivers/usb/host/Kconfig
+++ b/drivers/usb/host/Kconfig
@@ -90,14 +90,25 @@  config USB_EHCI_TT_NEWSCHED
 
 config USB_EHCI_BIG_ENDIAN_MMIO
 	bool
-	depends on USB_EHCI_HCD && (PPC_CELLEB || PPC_PS3 || 440EPX || ARCH_IXP4XX)
+	depends on USB_EHCI_HCD && (PPC_CELLEB || PPC_PS3 || 440EPX || ARCH_IXP4XX || XPS_USB_HCD_XILINX)
 	default y
 
 config USB_EHCI_BIG_ENDIAN_DESC
 	bool
-	depends on USB_EHCI_HCD && (440EPX || ARCH_IXP4XX)
+	depends on USB_EHCI_HCD && (440EPX || ARCH_IXP4XX || XPS_USB_HCD_XILINX)
 	default y
 
+config XPS_USB_HCD_XILINX
+	bool "Use Xilinx usb host EHCI controller core"
+	depends on USB_EHCI_HCD && (PPC32 || MICROBLAZE)
+	select USB_EHCI_BIG_ENDIAN_DESC
+	select USB_EHCI_BIG_ENDIAN_MMIO
+	---help---
+		Xilinx xps USB host controller core is EHCI compilant and has
+		transaction translator built-in. It can be configured to either
+		support both high speed and full speed devices, or high speed
+		devices only.
+
 config USB_EHCI_FSL
 	bool "Support for Freescale on-chip EHCI USB controller"
 	depends on USB_EHCI_HCD && FSL_SOC
diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c
index 11c627c..7672624 100644
--- a/drivers/usb/host/ehci-hcd.c
+++ b/drivers/usb/host/ehci-hcd.c
@@ -1107,6 +1107,11 @@  MODULE_LICENSE ("GPL");
 #define OF_PLATFORM_DRIVER	ehci_hcd_ppc_of_driver
 #endif
 
+#ifdef CONFIG_XPS_USB_HCD_XILINX
+#include "ehci-xilinx-of.c"
+#define OF_PLATFORM_DRIVER	ehci_hcd_xilinx_of_driver
+#endif
+
 #ifdef CONFIG_PLAT_ORION
 #include "ehci-orion.c"
 #define	PLATFORM_DRIVER		ehci_orion_driver
diff --git a/drivers/usb/host/ehci-xilinx-of.c b/drivers/usb/host/ehci-xilinx-of.c
new file mode 100644
index 0000000..f1ae855
--- /dev/null
+++ b/drivers/usb/host/ehci-xilinx-of.c
@@ -0,0 +1,302 @@ 
+/*
+ * EHCI HCD (Host Controller Driver) for USB.
+ *
+ * Bus Glue for Xilinx EHCI core on the of_platform bus
+ *
+ * Copyright (c) 2009 Xilinx, Inc.
+ *
+ * Based on "ehci-ppc-of.c" by Valentine Barshak <vbarshak@ru.mvista.com>
+ * and "ehci-ppc-soc.c" by Stefan Roese <sr@denx.de>
+ * and "ohci-ppc-of.c" by Sylvain Munaut <tnt@246tNt.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; either version 2 of the License, or (at your
+ * option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+ * or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+ * for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+#include <linux/signal.h>
+
+#include <linux/of.h>
+#include <linux/of_platform.h>
+
+/**
+ * ehci_xilinx_of_setup - Initialize the device for ehci_reset()
+ * @hcd:	Pointer to the usb_hcd device to which the host controller bound
+ *
+ * called during probe() after chip reset completes.
+ */
+static int ehci_xilinx_of_setup(struct usb_hcd *hcd)
+{
+	struct ehci_hcd	*ehci = hcd_to_ehci(hcd);
+	int		retval;
+
+	retval = ehci_halt(ehci);
+	if (retval)
+		return retval;
+
+	retval = ehci_init(hcd);
+	if (retval)
+		return retval;
+
+	ehci->sbrn = 0x20;
+
+	return ehci_reset(ehci);
+}
+
+/**
+ * ehci_xilinx_port_handed_over - hand the port out if failed to enable it
+ * @hcd:	Pointer to the usb_hcd device to which the host controller bound
+ * @portnum:Port number to which the device is attached.
+ *
+ * This function is used as a place to tell the user that the Xilinx USB host
+ * controller does support LS devices. And in an HS only configuration, it
+ * does not support FS devices either. It is hoped that this can help a
+ * confused user.
+ *
+ * There are cases when the host controller fails to enable the port due to,
+ * for example, insufficient power that can be supplied to the device from
+ * the USB bus. In those cases, the messages printed here are not helpful.
+ */
+static int ehci_xilinx_port_handed_over(struct usb_hcd *hcd, int portnum)
+{
+	dev_warn(hcd->self.controller, "port %d cannot be enabled\n", portnum);
+	if (hcd->has_tt) {
+		dev_warn(hcd->self.controller,
+			"Maybe you have connected a low speed device?\n");
+
+		dev_warn(hcd->self.controller,
+			"We do not support low speed devices\n");
+	} else {
+		dev_warn(hcd->self.controller,
+			"Maybe your device is not a high speed device?\n");
+		dev_warn(hcd->self.controller,
+			"The USB host controller does not support full speed "
+			"nor low speed devices\n");
+		dev_warn(hcd->self.controller,
+			"You can reconfigure the host controller to have "
+			"full speed support\n");
+	}
+
+	return 0;
+}
+
+
+static const struct hc_driver ehci_xilinx_of_hc_driver = {
+	.description		= hcd_name,
+	.product_desc		= "OF EHCI",
+	.hcd_priv_size		= sizeof(struct ehci_hcd),
+
+	/*
+	 * generic hardware linkage
+	 */
+	.irq			= ehci_irq,
+	.flags			= HCD_MEMORY | HCD_USB2,
+
+	/*
+	 * basic lifecycle operations
+	 */
+	.reset			= ehci_xilinx_of_setup,
+	.start			= ehci_run,
+	.stop			= ehci_stop,
+	.shutdown		= ehci_shutdown,
+
+	/*
+	 * managing i/o requests and associated device resources
+	 */
+	.urb_enqueue		= ehci_urb_enqueue,
+	.urb_dequeue		= ehci_urb_dequeue,
+	.endpoint_disable	= ehci_endpoint_disable,
+
+	/*
+	 * scheduling support
+	 */
+	.get_frame_number	= ehci_get_frame,
+
+	/*
+	 * root hub support
+	 */
+	.hub_status_data	= ehci_hub_status_data,
+	.hub_control		= ehci_hub_control,
+#ifdef	CONFIG_PM
+	.bus_suspend		= ehci_bus_suspend,
+	.bus_resume		= ehci_bus_resume,
+#endif
+	.relinquish_port	= NULL,
+	.port_handed_over	= ehci_xilinx_port_handed_over,
+
+	.clear_tt_buffer_complete = ehci_clear_tt_buffer_complete,
+};
+
+/**
+ * ehci_hcd_xilinx_of_probe - Probe method for the USB host controller
+ * @op:		pointer to the of_device to which the host controller bound
+ * @match:	pointer to of_device_id structure, not used
+ *
+ * This function requests resources and sets up appropriate properties for the
+ * host controller. Because the Xilinx USB host controller can be configured
+ * as HS only or HS/FS only, it checks the configuration in the device tree
+ * entry, and sets an appropriate value for hcd->has_tt.
+ */
+static int __devinit
+ehci_hcd_xilinx_of_probe(struct of_device *op, const struct of_device_id *match)
+{
+	struct device_node *dn = op->node;
+	struct usb_hcd *hcd;
+	struct ehci_hcd	*ehci;
+	struct resource res;
+	int irq;
+	int rv;
+	int *value;
+
+	if (usb_disabled())
+		return -ENODEV;
+
+	dev_dbg(&op->dev, "initializing XILINX-OF USB Controller\n");
+
+	rv = of_address_to_resource(dn, 0, &res);
+	if (rv)
+		return rv;
+
+	hcd = usb_create_hcd(&ehci_xilinx_of_hc_driver, &op->dev,
+				"XILINX-OF USB");
+	if (!hcd)
+		return -ENOMEM;
+
+	hcd->rsrc_start = res.start;
+	hcd->rsrc_len = res.end - res.start + 1;
+
+	if (!request_mem_region(hcd->rsrc_start, hcd->rsrc_len, hcd_name)) {
+		printk(KERN_ERR __FILE__ ": request_mem_region failed\n");
+		rv = -EBUSY;
+		goto err_rmr;
+	}
+
+	irq = irq_of_parse_and_map(dn, 0);
+	if (irq == NO_IRQ) {
+		printk(KERN_ERR __FILE__ ": irq_of_parse_and_map failed\n");
+		rv = -EBUSY;
+		goto err_irq;
+	}
+
+	hcd->regs = ioremap(hcd->rsrc_start, hcd->rsrc_len);
+	if (!hcd->regs) {
+		printk(KERN_ERR __FILE__ ": ioremap failed\n");
+		rv = -ENOMEM;
+		goto err_ioremap;
+	}
+
+	ehci = hcd_to_ehci(hcd);
+
+	/* This core always has big-endian register interface and uses
+	 * big-endian memory descriptors.
+	 */
+	ehci->big_endian_mmio = 1;
+	ehci->big_endian_desc = 1;
+
+	/* Check whether the FS support option is selected in the hardware.
+	 */
+	value = (int *)of_get_property(dn, "xlnx,support-usb-fs", NULL);
+	if (value && (*value == 1)) {
+		ehci_dbg(ehci, "USB host controller supports FS devices\n");
+		hcd->has_tt = 1;
+	} else {
+		ehci_dbg(ehci,
+			"USB host controller is HS only\n");
+		hcd->has_tt = 0;
+	}
+
+	/* Debug registers are at the first 0x100 region
+	 */
+	ehci->caps = hcd->regs + 0x100;
+	ehci->regs = hcd->regs + 0x100 +
+			HC_LENGTH(ehci_readl(ehci, &ehci->caps->hc_capbase));
+
+	/* cache this readonly data; minimize chip reads */
+	ehci->hcs_params = ehci_readl(ehci, &ehci->caps->hcs_params);
+
+	rv = usb_add_hcd(hcd, irq, 0);
+	if (rv == 0)
+		return 0;
+
+	iounmap(hcd->regs);
+err_ioremap:
+	irq_dispose_mapping(irq);
+err_irq:
+	release_mem_region(hcd->rsrc_start, hcd->rsrc_len);
+err_rmr:
+	usb_put_hcd(hcd);
+
+	return rv;
+}
+
+/**
+ * ehci_hcd_xilinx_of_remove - shutdown hcd and release resources
+ * @op:		pointer to of_device structure that is to be removed
+ *
+ * Remove the hcd structure, and release resources that has been requested
+ * during probe.
+ */
+static int ehci_hcd_xilinx_of_remove(struct of_device *op)
+{
+	struct usb_hcd *hcd = dev_get_drvdata(&op->dev);
+	dev_set_drvdata(&op->dev, NULL);
+
+	dev_dbg(&op->dev, "stopping XILINX-OF USB Controller\n");
+
+	usb_remove_hcd(hcd);
+
+	iounmap(hcd->regs);
+	release_mem_region(hcd->rsrc_start, hcd->rsrc_len);
+
+	usb_put_hcd(hcd);
+
+	return 0;
+}
+
+/**
+ * ehci_hcd_xilinx_of_shutdown - shutdown the hcd
+ * @op:		pointer to of_device structure that is to be removed
+ *
+ * Properly shutdown the hcd, call driver's shutdown routine.
+ */
+static int ehci_hcd_xilinx_of_shutdown(struct of_device *op)
+{
+	struct usb_hcd *hcd = dev_get_drvdata(&op->dev);
+
+	if (hcd->driver->shutdown)
+		hcd->driver->shutdown(hcd);
+
+	return 0;
+}
+
+
+static struct of_device_id ehci_hcd_xilinx_of_match[] = {
+		{.compatible = "xlnx,xps-usb-host-1.00.a",},
+		{.compatible = "xlnx,xps-usb-host-1.01.a",},
+		{.compatible = "xlnx,xps-usb-host-1.02.a",},
+	{},
+};
+MODULE_DEVICE_TABLE(of, ehci_hcd_xilinx_of_match);
+
+static struct of_platform_driver ehci_hcd_xilinx_of_driver = {
+	.name		= "xilinx-of-ehci",
+	.match_table	= ehci_hcd_xilinx_of_match,
+	.probe		= ehci_hcd_xilinx_of_probe,
+	.remove		= ehci_hcd_xilinx_of_remove,
+	.shutdown	= ehci_hcd_xilinx_of_shutdown,
+	.driver		= {
+		.name	= "xilinx-of-ehci",
+		.owner	= THIS_MODULE,
+	},
+};