Skip to contentSkip to conversation

BugReport (test6): USB (ACPI), SWSUSP, E100

DISCUSSION Thomas Winkler 3 participants 4 messages Sep 29, 2003 – Oct 4, 2003

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hello,

I sent a similar report for 2.6.0test5 already (see:
http://linux.derkeiler.com/Mailing-Lists/Kernel/2003-09/3017.html).
Back than I was told to wait for the next kernel (test6) and check again. So here are the results with test6:

- USB - still dead
This notebook has some IRQ routing problems and requires ACPI to get it right. When using older 2.4 Kernels I always had to apply the ACPI patches to get USB working. With 2.6-test3 the IRQ routing / ACPI stuff worked out of the box. With test5 and test6 USB is not working.
The following of dmesg seems to be interesting (full dmesg at the end of the mail):

drivers/usb/host/uhci-hcd.c: USB Universal Host Controller Interface driver v2.1
uhci-hcd 0000:00:1f.2: UHCI Host Controller
irq 9: nobody cared!

Call Trace:
 [<c010ae0a>] __report_bad_irq+0x2a/0x90
 [...]
 [<c01072a9>] kernel_thread_helper+0x5/0xc

handlers:
[<c0204f3f>] (acpi_irq+0x0/0x16)
Disabling IRQ #9

- SWSUSP
In contrast du test5 there now is a /proc/acpi/sleep file again. But an

echo 4 > /proc/acpi/sleep shows no effect at all. SWSUSP is enabled in the kernel (full .config at the end of the mail).

echo 3 > /proc/acpi/sleep produces the following output and then the prompt returns again:
Stopping tasks: ==================
stopping tasks failed (1 tasks remaining)
Restarting tasks...<6> Strange, khubd not stopped
done

- The e100 driver seems to be broken
The NIC is detected correctly and ifconfig shows eth0 as usually. But for some reason not a single Byte seems to go over the NIC. I can ping the machine itself but I'm not able to send anything to any outside machine. This again works perfectly with test3 but shows the same problems with test5 and test6 Please note: I compiled test6 with the same .config as test3 and I didn't change any init scripts before bootuing test6. I've no idea but might this problem be related to the broken ACPI?

full dmesg: http://www.wnk.at/tmp/test6/dmesg.txt
.config used building the kernel: http://www.wnk.at/tmp/test6/config.txt

Since I'm not subscribed to LKM please CC me on replys. If you need more information please drop me a line.

Thanks,

Signature · 2 lines
Tom Winkler
e-mail: [email protected]
#1 lore raw

Thomas Winkler <[email protected]> wrote on 2003-09-29 15:51:00

Quoted 4 lines

drivers/usb/host/uhci-hcd.c: USB Universal Host Controller Interface driver v2.1
uhci-hcd 0000:00:1f.2: UHCI Host Controller
irq 9: nobody cared!

Quoted 9 lines
Call Trace:
 [<c010ae0a>] __report_bad_irq+0x2a/0x90
 [...]
 [<c01072a9>] kernel_thread_helper+0x5/0xc

handlers:
[<c0204f3f>] (acpi_irq+0x0/0x16)
Disabling IRQ #9

Try reverting (patch -p1 -R) the following patch. It'll give one rejects, but afterwards it works for me.
Greg, Davids patch didn't work out for this particular problem.

Thanks,

Jan

drivers/usb/core/hcd-pci.c +5 -2
diff -urN linux-2.6.0-test4-bk4/drivers/usb/core/hcd-pci.c linux-2.6.0-test4-bk5/drivers/usb/core/hcd-pci.c
@@ -139,6 +139,7 @@
 			return retval;
 		}
 	}
+	// hcd zeroed everything
 	hcd->regs = base;
 	hcd->region = region;
 
@@ -165,6 +166,7 @@
 		dev_err (hcd->controller, "can't reset\n");
 		goto clean_3;
 	}
+	hcd->state = USB_STATE_HALT;
 
 	pci_set_master (dev);
 #ifndef __sparc__
@@ -230,7 +232,8 @@
 		BUG ();
 
 	hub = hcd->self.root_hub;
-	hcd->state = USB_STATE_QUIESCING;
+	if (HCD_IS_RUNNING (hcd->state))
+		hcd->state = USB_STATE_QUIESCING;
 
 	dev_dbg (hcd->controller, "roothub graceful disconnect\n");
 	usb_disconnect (&hub);
@@ -287,8 +290,8 @@
 		pci_save_state (dev, hcd->pci_state);
 
 		/* driver may want to disable DMA etc */
+		hcd->state = USB_STATE_QUIESCING;
 		retval = hcd->driver->suspend (hcd, state);
-		hcd->state = USB_STATE_SUSPENDED;
 	}
 
  	pci_set_power_state (dev, state);
drivers/usb/core/hcd.c +26 -2
diff -urN linux-2.6.0-test4-bk4/drivers/usb/core/hcd.c linux-2.6.0-test4-bk5/drivers/usb/core/hcd.c
@@ -483,7 +483,7 @@
 {
 	struct urb	*urb;
 	struct usb_hcd	*hcd;
-	int		length;
+	int		length = 0;
 	unsigned long	flags;
 
 	urb = (struct urb *) ptr;
@@ -499,7 +499,9 @@
 		return;
 	}
 
-	length = hcd->driver->hub_status_data (hcd, urb->transfer_buffer);
+	if (!HCD_IS_SUSPENDED (hcd->state))
+		length = hcd->driver->hub_status_data (
+					hcd, urb->transfer_buffer);
 
 	/* complete the status urb, or retrigger the timer */
 	spin_lock (&hcd_data_lock);
@@ -1097,6 +1099,8 @@
 static int hcd_get_frame_number (struct usb_device *udev)
 {
 	struct usb_hcd	*hcd = (struct usb_hcd *)udev->bus->hcpriv;
+	if (!HCD_IS_RUNNING (hcd->state))
+		return -ESHUTDOWN;
 	return hcd->driver->get_frame_number (hcd);
 }
 
@@ -1193,6 +1197,12 @@
 		goto done;
 	}
 
+	/* running ~= hc unlink handshake works (irq, timer, etc)
+	 * halted ~= no unlink handshake is needed
+	 * suspended, resuming == should never happen
+	 */
+	WARN_ON (!HCD_IS_RUNNING (hcd->state) && hcd->state != USB_STATE_HALT);
+
 	if (!urb->hcpriv) {
 		retval = -EINVAL;
 		goto done;
@@ -1208,6 +1218,17 @@
 		goto done;
 	}
 
+	/* PCI IRQ setup can easily be broken so that USB controllers
+	 * never get completion IRQs ... maybe even the ones we need to
+	 * finish unlinking the initial failed usb_set_address().
+	 */
+	if (!hcd->saw_irq) {
+		dev_warn (hcd->controller, "Unlink after no-IRQ?  "
+			"Different ACPI or APIC settings may help."
+			"\n");
+		hcd->saw_irq = 1;
+	}
+
 	/* maybe set up to block until the urb's completion fires.  the
 	 * lower level hcd code is always async, locking on urb->status
 	 * updates; an intercepted completion unblocks us.
@@ -1287,6 +1308,8 @@
 	dev = udev->hcpriv;
 	hcd = udev->bus->hcpriv;
 
+	WARN_ON (!HCD_IS_RUNNING (hcd->state) && hcd->state != USB_STATE_HALT);
+
 	local_irq_disable ();
 
 rescan:
@@ -1483,6 +1506,7 @@
 	if (unlikely (hcd->state == USB_STATE_HALT))	/* irq sharing? */
 		return IRQ_NONE;
 
+	hcd->saw_irq = 1;
 	hcd->driver->irq (hcd, r);
 	if (hcd->state != start && hcd->state == USB_STATE_HALT)
 		usb_hc_died (hcd);
drivers/usb/core/hcd.h +1 -2
diff -urN linux-2.6.0-test4-bk4/drivers/usb/core/hcd.h linux-2.6.0-test4-bk5/drivers/usb/core/hcd.h
@@ -73,6 +73,7 @@
 	 * hardware info/state
 	 */
 	struct hc_driver	*driver;	/* hw-specific hooks */
+	unsigned		saw_irq : 1;
 	int			irq;		/* irq allocated */
 	void			*regs;		/* device memory/io */
 	struct device		*controller;	/* handle to hardware */
@@ -89,13 +90,11 @@
 
 	int			state;
 #	define	__ACTIVE		0x01
-#	define	__SLEEPY		0x02
 #	define	__SUSPEND		0x04
 #	define	__TRANSIENT		0x80
 
 #	define	USB_STATE_HALT		0
 #	define	USB_STATE_RUNNING	(__ACTIVE)
-#	define	USB_STATE_READY		(__ACTIVE|__SLEEPY)
 #	define	USB_STATE_QUIESCING	(__SUSPEND|__TRANSIENT|__ACTIVE)
 #	define	USB_STATE_RESUMING	(__SUSPEND|__TRANSIENT)
 #	define	USB_STATE_SUSPENDED	(__SUSPEND)
drivers/usb/core/message.c +17 -13
diff -urN linux-2.6.0-test4-bk4/drivers/usb/core/message.c linux-2.6.0-test4-bk5/drivers/usb/core/message.c
@@ -246,21 +246,22 @@
 		io->status = urb->status;
 
 		/* the previous urbs, and this one, completed already.
-		 * unlink the later ones so they won't rx/tx bad data,
-		 *
-		 * FIXME don't bother unlinking urbs that haven't yet been
-		 * submitted; those non-error cases shouldn't be syslogged
+		 * unlink pending urbs so they won't rx/tx bad data.
 		 */
 		for (i = 0, found = 0; i < io->entries; i++) {
+			if (!io->urbs [i])
+				continue;
 			if (found) {
 				status = usb_unlink_urb (io->urbs [i]);
-				if (status && status != -EINPROGRESS)
-					err ("sg_complete, unlink --> %d",
-							status);
+				if (status != -EINPROGRESS && status != -EBUSY)
+					dev_err (&io->dev->dev,
+						"%s, unlink --> %d\n",
+						__FUNCTION__, status);
 			} else if (urb == io->urbs [i])
 				found = 1;
 		}
 	}
+	urb->dev = 0;
 
 	/* on the last completion, signal usb_sg_wait() */
 	io->bytes += urb->actual_length;
@@ -356,7 +357,7 @@
 			goto nomem;
 		}
 
-		io->urbs [i]->dev = dev;
+		io->urbs [i]->dev = 0;
 		io->urbs [i]->pipe = pipe;
 		io->urbs [i]->interval = period;
 		io->urbs [i]->transfer_flags = urb_flags;
@@ -448,6 +449,7 @@
 	for (i = 0; i < io->entries && !io->status; i++) {
 		int	retval;
 
+		io->urbs [i]->dev = io->dev;
 		retval = usb_submit_urb (io->urbs [i], SLAB_ATOMIC);
 
 		/* after we submit, let completions or cancelations fire;
@@ -459,9 +461,9 @@
 		case -ENXIO:	// hc didn't queue this one
 		case -EAGAIN:
 		case -ENOMEM:
+			io->urbs [i]->dev = 0;
 			retval = 0;
 			i--;
-			// FIXME:  should it usb_sg_cancel() on INTERRUPT?
 			yield ();
 			break;
 
@@ -477,8 +479,10 @@
 
 			/* fail any uncompleted urbs */
 		default:
+			io->urbs [i]->dev = 0;
 			io->urbs [i]->status = retval;
-			dbg ("usb_sg_msg, submit --> %d", retval);
+			dev_dbg (&io->dev->dev, "%s, submit --> %d\n",
+				__FUNCTION__, retval);
 			usb_sg_cancel (io);
 		}
 		spin_lock_irqsave (&io->lock, flags);
@@ -521,9 +525,9 @@
 			if (!io->urbs [i]->dev)
 				continue;
 			retval = usb_unlink_urb (io->urbs [i]);
-			if (retval && retval != -EINPROGRESS)
-				warn ("usb_sg_cancel, unlink --> %d", retval);
-			// FIXME don't warn on "not yet submitted" error
+			if (retval != -EINPROGRESS && retval != -EBUSY)
+				dev_warn (&io->dev->dev, "%s, unlink --> %d\n",
+					__FUNCTION__, retval);
 		}
 	}
 	spin_unlock_irqrestore (&io->lock, flags);
drivers/usb/core/usb.c +37 -3
diff -urN linux-2.6.0-test4-bk4/drivers/usb/core/usb.c linux-2.6.0-test4-bk5/drivers/usb/core/usb.c
@@ -991,8 +991,8 @@
  *
  * This call is synchronous, and may not be used in an interrupt context.
  *
- * Only hub drivers (including virtual root hub drivers for host
- * controllers) should ever call this.
+ * Only the hub driver should ever call this; root hub registration
+ * uses it only indirectly.
  */
 #define NEW_DEVICE_RETRYS	2
 #define SET_ADDRESS_RETRYS	2
@@ -1417,11 +1417,46 @@
 			usb_pipein (pipe) ? DMA_FROM_DEVICE : DMA_TO_DEVICE);
 }
 
+static int usb_device_suspend(struct device *dev, u32 state)
+{
+	struct usb_interface *intf;
+	struct usb_driver *driver;
+
+	if ((dev->driver == &usb_generic_driver) || 
+	    (dev->driver_data == &usb_generic_driver_data))
+		return 0;
+
+	intf = to_usb_interface(dev);
+	driver = to_usb_driver(dev->driver);
+
+	if (driver && driver->suspend)
+		return driver->suspend(intf, state);
+	return 0;
+}
+
+static int usb_device_resume(struct device *dev)
+{
+	struct usb_interface *intf;
+	struct usb_driver *driver;
+
+	if ((dev->driver == &usb_generic_driver) || 
+	    (dev->driver_data == &usb_generic_driver_data))
+		return 0;
+
+	intf = to_usb_interface(dev);
+	driver = to_usb_driver(dev->driver);
+
+	if (driver && driver->resume)
+		return driver->resume(intf);
+	return 0;
+}
 
 struct bus_type usb_bus_type = {
 	.name =		"usb",
 	.match =	usb_device_match,
 	.hotplug =	usb_hotplug,
+	.suspend =	usb_device_suspend,
+	.resume =	usb_device_resume,
 };
 
 #ifndef MODULE
@@ -1509,7 +1544,6 @@
 EXPORT_SYMBOL(usb_find_interface);
 EXPORT_SYMBOL(usb_ifnum_to_if);
 
-EXPORT_SYMBOL(usb_new_device);
 EXPORT_SYMBOL(usb_reset_device);
 EXPORT_SYMBOL(usb_disconnect);
 
drivers/usb/host/uhci-hcd.c +19 -21
diff -urN linux-2.6.0-test4-bk4/drivers/usb/host/uhci-hcd.c linux-2.6.0-test4-bk5/drivers/usb/host/uhci-hcd.c
@@ -2099,7 +2099,7 @@
 	uhci->state_end = jiffies + HZ;
 	outw(USBCMD_RS | USBCMD_CF | USBCMD_MAXP, io_addr + USBCMD);
 
-        uhci->hcd.state = USB_STATE_READY;
+        uhci->hcd.state = USB_STATE_RUNNING;
 }
 
 /*
@@ -2143,6 +2143,20 @@
 #endif
 }
 
+static int uhci_reset(struct usb_hcd *hcd)
+{
+	struct uhci_hcd *uhci = hcd_to_uhci(hcd);
+
+	uhci->io_addr = (unsigned long) hcd->regs;
+
+	/* Maybe kick BIOS off this hardware.  Then reset, so we won't get
+	 * interrupts from any previous setup.
+	 */
+	pci_write_config_word(hcd->pdev, USBLEGSUP, USBLEGSUP_DEFAULT);
+	reset_hc(uhci);
+	return 0;
+}
+
 /*
  * Allocate a frame list, and then setup the skeleton
  *
@@ -2159,7 +2173,7 @@
  *  - The fourth queue is the bandwidth reclamation queue, which loops back
  *    to the high speed control queue.
  */
-static int __devinit uhci_start(struct usb_hcd *hcd)
+static int uhci_start(struct usb_hcd *hcd)
 {
 	struct uhci_hcd *uhci = hcd_to_uhci(hcd);
 	int retval = -EBUSY;
@@ -2171,7 +2185,6 @@
 	struct proc_dir_entry *ent;
 #endif
 
-	uhci->io_addr = (unsigned long) hcd->regs;
 	io_size = pci_resource_len(hcd->pdev, hcd->region);
 
 #ifdef CONFIG_PROC_FS
@@ -2188,10 +2201,6 @@
 	uhci->proc_entry = ent;
 #endif
 
-	/* Reset here so we don't get any interrupts from an old setup */
-	/*  or broken setup */
-	reset_hc(uhci);
-
 	uhci->fsbr = 0;
 	uhci->fsbrtimeout = 0;
 
@@ -2343,9 +2352,6 @@
 
 	init_stall_timer(hcd);
 
-	/* disable legacy emulation */
-	pci_write_config_word(hcd->pdev, USBLEGSUP, USBLEGSUP_DEFAULT);
-
 	udev->speed = USB_SPEED_FULL;
 
 	if (usb_register_root_hub(udev, &hcd->pdev->dev) != 0) {
@@ -2446,7 +2452,7 @@
 		reset_hc(uhci);
 		start_hc(uhci);
 	}
-	uhci->hcd.state = USB_STATE_READY;
+	uhci->hcd.state = USB_STATE_RUNNING;
 	return 0;
 }
 #endif
@@ -2484,6 +2490,7 @@
 	.flags =		HCD_USB11,
 
 	/* Basic lifecycle operations */
+	.reset =		uhci_reset,
 	.start =		uhci_start,
 #ifdef CONFIG_PM
 	.suspend =		uhci_suspend,
@@ -2504,18 +2511,9 @@
 };
 
 static const struct pci_device_id uhci_pci_ids[] = { {
-
 	/* handle any USB UHCI controller */
-	.class = 		((PCI_CLASS_SERIAL_USB << 8) | 0x00),
-	.class_mask = 	~0,
+	PCI_DEVICE_CLASS(((PCI_CLASS_SERIAL_USB << 8) | 0x00), ~0),
 	.driver_data =	(unsigned long) &uhci_driver,
-
-	/* no matter who makes it */
-	.vendor =	PCI_ANY_ID,
-	.device =	PCI_ANY_ID,
-	.subvendor =	PCI_ANY_ID,
-	.subdevice =	PCI_ANY_ID,
-
 	}, { /* end: all zeroes */ }
 };
 
#2 lore raw

Hello,

I just tested to uhci-hsd one-line patch sent by Wim Van Sebroeck. USB works fine again now.

In addition to that the strange e100 problems are also gone now. I double checked this by removing the patch once again. Without the patch e100 is dead, with the patch everything works fine. Interesting what an USB patch can do to a NIC (a side effect of irq problems?).

This only leaves the SWSUP problem open. An
echo 4 > /proc/acpi/sleep
still shows no effect at all (see also the original mail). Is SWSUP supposed to work in test6?

bye,

Signature · 2 lines
Tom Winkler
e-mail: [email protected]
#3 lore raw
In reply to Thomas Winkler

Hi!

Quoted 5 lines

- SWSUSP
In contrast du test5 there now is a /proc/acpi/sleep file again. But an

echo 4 > /proc/acpi/sleep shows no effect at all. SWSUSP is enabled in the kernel (full .config at the end of the mail).

Look for [pm] patches I sent to the list; you need to add __initcall(software_resume).

Quoted 6 lines

echo 3 > /proc/acpi/sleep produces the following output and then the prompt returns again:
Stopping tasks: ==================
stopping tasks failed (1 tasks remaining)
Restarting tasks...<6> Strange, khubd not stopped
done

Turn off config_usb.

Signature · 2 lines
				Pavel
Written on sharp zaurus, because my Velo1 broke. If you have Velo you don't need...