[PATCH] Au1200 OHCI/EHCI fixes

[PATCH] Au1200 OHCI/EHCI fixes

[Manuel Lauss]

Hello,

This little patch fixes compileproblems in Au1000
OHCI/EHCI code. Run-tested on custom Au1200 Platform.

Signed-off-by: Manuel Lauss <mano@roarinelk.homelinux.net>


diff -Naurp linux-2.6.17.7-base/drivers/usb/host/ehci-au1xxx.c linux-2.6.17.7-work/drivers/usb/host/ehci-au1xxx.c
--- linux-2.6.17.7-base/drivers/usb/host/ehci-au1xxx.c	2006-08-02 08:12:42.378136000 +0200
+++ linux-2.6.17.7-work/drivers/usb/host/ehci-au1xxx.c	2006-08-09 13:57:54.576811000 +0200
@@ -111,9 +111,9 @@ int usb_ehci_au1xxx_probe(const struct h
 
 	/* Au1200 AB USB does not support coherent memory */
 	if (!(read_c0_prid() & 0xff)) {
-		pr_info("%s: this is chip revision AB!\n", dev->dev.name);
+		pr_info("%s: this is chip revision AB!\n", dev->dev.driver->name);
 		pr_info("%s: update your board or re-configure the kernel\n",
-			dev->dev.name);
+			dev->dev.driver->name);
 		return -ENODEV;
 	}
 #endif
diff -Naurp linux-2.6.17.7-base/drivers/usb/host/ohci-au1xxx.c linux-2.6.17.7-work/drivers/usb/host/ohci-au1xxx.c
--- linux-2.6.17.7-base/drivers/usb/host/ohci-au1xxx.c	2006-08-02 08:12:42.398136000 +0200
+++ linux-2.6.17.7-work/drivers/usb/host/ohci-au1xxx.c	2006-08-10 08:46:56.430581000 +0200
@@ -84,6 +84,11 @@ static void au1xxx_start_ohc(struct plat
 	au_writel(USBH_ENABLE_INIT, USB_HOST_CONFIG);
 	udelay(1000);
 
+	/* wait for reset complete (read register twice; see au1500 errata) */
+	while (au_readl(USB_HOST_CONFIG),
+		!(au_readl(USB_HOST_CONFIG) & USBH_ENABLE_RD))
+		udelay(1000);
+
 #else   /* Au1200 */
 
 	/* write HW defaults again in case Yamon cleared them */
@@ -101,11 +106,6 @@ static void au1xxx_start_ohc(struct plat
 
 #endif  /* Au1200 */
 
-	/* wait for reset complete (read register twice; see au1500 errata) */
-	while (au_readl(USB_HOST_CONFIG),
-		!(au_readl(USB_HOST_CONFIG) & USBH_ENABLE_RD))
-		udelay(1000);
-
 	printk(KERN_DEBUG __FILE__
 	": Clock to USB host has been enabled \n");
 }
@@ -157,9 +157,9 @@ static int usb_ohci_au1xxx_probe(const s
 	/* Au1200 AB USB does not support coherent memory */
 	if (!(read_c0_prid() & 0xff)) {
 		pr_info("%s: this is chip revision AB !!\n",
-			dev->dev.name);
+			dev->dev.driver->name);
 		pr_info("%s: update your board or re-configure the kernel\n",
-			dev->dev.name);
+			dev->dev.driver->name);
 		return -ENODEV;
 	}
 #endif

Re: [PATCH] Au1200 OHCI/EHCI fixes

[Daniel Mack]

Hi,

On Aug 10, 2006, at 8:53 AM, Manuel Lauss wrote:

> This little patch fixes compileproblems in Au1000
> OHCI/EHCI code. Run-tested on custom Au1200 Platform.

This has already been fixed - a similar patch went upstream to 2.6.18- 
rc3.
Did you check out the latest git?

Daniel

Re: [PATCH] Au1200 OHCI/EHCI fixes

[Manuel Lauss]

On Thu, Aug 10, 2006 at 03:27:43PM +0200, Daniel Mack wrote:

> On Aug 10, 2006, at 8:53 AM, Manuel Lauss wrote:
> 
> >This little patch fixes compileproblems in Au1000
> >OHCI/EHCI code. Run-tested on custom Au1200 Platform.
> 
> This has already been fixed - a similar patch went upstream to 2.6.18- 
> rc3.
> Did you check out the latest git?

I searched the mips ML prior to sending.
Guess I'll have to have a talk about git with
the firewall guy...

Sorry for the noise.

-- 
 ml.

Re: [PATCH] Au1200 OHCI/EHCI fixes

[Ralf Baechle]

On Thu, Aug 10, 2006 at 04:19:14PM +0200, Manuel Lauss wrote:

> I searched the mips ML prior to sending.
> Guess I'll have to have a talk about git with
> the firewall guy...

My favorite firewalling product is selling at unbeatably low prices
under the name patch cable ;-)

Short of such a high-performance firewalling enhancment you can also use
the rsync protocol or the http protocol with git.  Both are significantly
less efficient than the native git protocol so should only be used if
plan A that is the git protocol fails.

  Ralf

Re: [PATCH] Au1200 OHCI/EHCI fixes

[Ricardo Mendoza]

On 8/10/06, Daniel Mack <daniel@caiaq.de> wrote:

> This has already been fixed - a similar patch went upstream to 2.6.18-
> rc3.
> Did you check out the latest git?

Actually I think a problem stood alive, there is a parent-less #endif
in -rc4 that screwed up compile. I think it has been there since -rc2.

     Ricardo

Re: [PATCH] Au1200 OHCI/EHCI fixes

[Yoichi Yuasa]

Hi,

On Fri, 11 Aug 2006 22:45:28 -0400
"Ricardo Mendoza" <mendoza.ricardo@gmail.com> wrote:

> On 8/10/06, Daniel Mack <daniel@caiaq.de> wrote:
> 
> > This has already been fixed - a similar patch went upstream to 2.6.18-
> > rc3.
> > Did you check out the latest git?
> 
> Actually I think a problem stood alive, there is a parent-less #endif
> in -rc4 that screwed up compile. I think it has been there since -rc2.

Try this git.
Some fixes are included in it after rc4.
Of course, #endif fix also.

git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/linux-2.6.git/

Yoichi