Re: ohci: USB devices only detected at boot time

Alan Stern wrote:
> On Thu, 15 Feb 2007, Raphael Assenat wrote:
> 
> 
>>>Okay.  This says that RHSC works correctly while the controller is 
>>>running, but it doesn't work when the controller is suspended.
>>
>>hmm. So that's a bug in the chip? Would preventing this specific 
>>controller from being suspended be an acceptable solution? What are 
>>the implication of not suspending the controller when all the ports 
>>are free?
>>
>>lspci entry for this controller:
>>0000:00:01.4 0c03: 1283:8152 (rev 07)
>>
>>
>>
>>>Post a copy of the contents of /sys/class/usb_host/usb_host2/registers
>>>while no devices are plugged in, and then again after you plug in a device
>>>and nothing happens.
>>
>>Here. Apparently nothing changes...
>>
** removed register dumps **
> 
> 
> Wow, that's really buggy!  Not only is there no interrupt request, the
> port status shows that the chip doesn't even know you plugged in the
> device.
> 
> Clearly the chip's wakeup function is completely broken.  The only way you
> can get it to work is to prevent the controller from ever being suspended.  
> You can do that putting these lines in a system startup file:
> 
> 	echo disabled >/sys/bus/usb/devices/usb2/power/wakeup
> 	echo -n 0 >/sys/bus/usb/devices/usb2/power/state
> 
> Note that you will need to keep the CONFIG_PM_SYSFS_DEPRECATED option for 
> that second line to work properly.
That would certainly work, but I dont like much the idea of relying 
on a deprecated option. Wont it disappear in the future?

How about adding an OHCI_QUIRK_NO_RH_SUSPEND flag in struct ohci_hcd 
  and using it to disable suspend for this chip? The flag can be set 
automatically in ohci-pci.c. The only thing I'm not sure about is 
the appropriate place to handle this flag in ohci-hub.c....

A patch that does this is attached to this email. Please comment.

Best regards,
-- 
Raphael Assenat
8D Technologies Inc.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Disabled root hub suspend for the buggy it8152 ohci controller.

Signed-off-by: Raphael Assenat <[email protected]>

--- linux-2.6.20/drivers/usb/host/ohci-pci.c	2007-02-04 13:44:54.000000000 -0500
+++ linux-2.6.20-8d/drivers/usb/host/ohci-pci.c	2007-02-19 11:32:45.000000000 -0500
@@ -94,6 +94,13 @@
 				"enabled Compaq ZFMicro chipset quirk\n");
 		}
 
+		else if (pdev->vendor == PCI_VENDOR_ID_ITE
+				&& pdev->device == 0x8152) {
+			ohci->flags |= OHCI_QUIRK_NO_RH_SUSPEND;
+			ohci_dbg (ohci,
+				"workaround: disabling root hub suspend for ITE8152\n");
+		}
+	
 		/* RWC may not be set for add-in PCI cards, since boot
 		 * firmware probably ignored them.  This transfers PCI
 		 * PM wakeup capabilities (once the PCI layer is fixed).
--- linux-2.6.20/drivers/usb/host/ohci.h	2007-02-04 13:44:54.000000000 -0500
+++ linux-2.6.20-8d/drivers/usb/host/ohci.h	2007-02-19 11:39:07.000000000 -0500
@@ -396,6 +396,7 @@
 #define	OHCI_QUIRK_INITRESET	0x04			/* SiS, OPTi, ... */
 #define	OHCI_BIG_ENDIAN		0x08			/* big endian HC */
 #define	OHCI_QUIRK_ZFMICRO	0x10			/* Compaq ZFMicro chipset*/
+#define	OHCI_QUIRK_NO_RH_SUSPEND	0x20		
 	// there are also chip quirks/bugs in init logic
 
 };
--- linux-2.6.20/drivers/usb/host/ohci-hub.c	2007-02-04 13:44:54.000000000 -0500
+++ linux-2.6.20-8d/drivers/usb/host/ohci-hub.c	2007-02-19 11:33:19.000000000 -0500
@@ -350,7 +350,8 @@
 
 		/* if no devices have been attached for one second, autostop */
 		} else {
-			if (changed || any_connected) {
+			if (changed || any_connected || 
+					ohci->flags & OHCI_QUIRK_NO_RH_SUSPEND) {
 				ohci->autostop = 0;
 				ohci->next_statechange = jiffies +
 						STATECHANGE_DELAY;