Re: 2.6.18-rc strange hotplug/udev/uevent problem

[Andrew Morton]

On Wed, 9 Aug 2006 01:41:04 +0100
Andrew Clayton <andrew@xxxxxxxxxxxxxxxxxx> wrote:

> On Mon, 7 Aug 2006 23:02:11 -0700, Greg KH wrote:
>  
> > Can you use 'git bisect' to try to narrow it down which change caused
> > the problem?
> 
> I did 
> 
> git bisect start drivers/scsi drivers/usb
> git bad v2.6.18-rc1
> git good v2.6.17
> 
> Heres what git bisect came up with
> 
> c32ba30f76eb18b3d4449072fe9c345a9574796b is first bad commit
> commit c32ba30f76eb18b3d4449072fe9c345a9574796b
> Author: Paul Serice <paul@xxxxxxxxxx>
> Date:   Wed Jun 7 10:23:38 2006 -0700
> 
>     [PATCH] USB: EHCI works again on NVidia controllers with >2GB RAM
> 
>     From: Paul Serice <paul@xxxxxxxxxx>
> 
>     The workaround in commit f7201c3dcd7799f2aa3d6ec427b194225360ecee
>     broke.  The work around requires memory for DMA transfers for some
>     NVidia EHCI controllers to be below 2GB, but recent changes have
>     caused some DMA memory to be allocated before the DMA mask is set.
> 
>     Signed-off-by: Paul Serice <paul@xxxxxxxxxx>
>     Signed-off-by: David Brownell <dbrownell@xxxxxxxxxxxxxxxxxxxxx>
>     Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>
> 
> :040000 040000 754a9f8cccedc71e8e3689f2b0dee811d94fdc11 5559c0eafe042377430052272d027f0458805126 M      drivers
> 
> 
> What would be the way to revert that patch from 2.6.18-rc4 to confirm it's the culprit?
>  

Here's a backout patch, against 2.6.18-rc4:

From: Andrew Morton <akpm@xxxxxxxx>

Revert c32ba30f76eb18b3d4449072fe9c345a9574796b: it broke Andrew Clayton's
machine.

Cc: Andrew Clayton <andrew@xxxxxxxxxxxxxxxxxx>
Cc: Paul Serice <paul@xxxxxxxxxx>
Cc: David Brownell <dbrownell@xxxxxxxxxxxxxxxxxxxxx>
Cc: Greg Kroah-Hartman <gregkh@xxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxx>
---

 drivers/usb/host/ehci-pci.c |   39 +++++++++++++---------------------
 1 file changed, 15 insertions(+), 24 deletions(-)

diff -puN drivers/usb/host/ehci-pci.c~revert-usb-ehci-works-again-on-nvidia-controllers-with-2gb-ram drivers/usb/host/ehci-pci.c
--- a/drivers/usb/host/ehci-pci.c~revert-usb-ehci-works-again-on-nvidia-controllers-with-2gb-ram
+++ a/drivers/usb/host/ehci-pci.c
@@ -76,30 +76,6 @@ static int ehci_pci_setup(struct usb_hcd
 	dbg_hcs_params(ehci, "reset");
 	dbg_hcc_params(ehci, "reset");
 
-        /* ehci_init() causes memory for DMA transfers to be
-         * allocated.  Thus, any vendor-specific workarounds based on
-         * limiting the type of memory used for DMA transfers must
-         * happen before ehci_init() is called. */
-	switch (pdev->vendor) {
-	case PCI_VENDOR_ID_NVIDIA:
-		/* NVidia reports that certain chips don't handle
-		 * QH, ITD, or SITD addresses above 2GB.  (But TD,
-		 * data buffer, and periodic schedule are normal.)
-		 */
-		switch (pdev->device) {
-		case 0x003c:	/* MCP04 */
-		case 0x005b:	/* CK804 */
-		case 0x00d8:	/* CK8 */
-		case 0x00e8:	/* CK8S */
-			if (pci_set_consistent_dma_mask(pdev,
-						DMA_31BIT_MASK) < 0)
-				ehci_warn(ehci, "can't enable NVidia "
-					"workaround for >2GB RAM\n");
-			break;
-		}
-		break;
-	}
-
 	/* cache this readonly data; minimize chip reads */
 	ehci->hcs_params = readl(&ehci->caps->hcs_params);
 
@@ -112,6 +88,8 @@ static int ehci_pci_setup(struct usb_hcd
 	if (retval)
 		return retval;
 
+	/* NOTE:  only the parts below this line are PCI-specific */
+
 	switch (pdev->vendor) {
 	case PCI_VENDOR_ID_TDI:
 		if (pdev->device == PCI_DEVICE_ID_TDI_EHCI) {
@@ -129,6 +107,19 @@ static int ehci_pci_setup(struct usb_hcd
 		break;
 	case PCI_VENDOR_ID_NVIDIA:
 		switch (pdev->device) {
+		/* NVidia reports that certain chips don't handle
+		 * QH, ITD, or SITD addresses above 2GB.  (But TD,
+		 * data buffer, and periodic schedule are normal.)
+		 */
+		case 0x003c:	/* MCP04 */
+		case 0x005b:	/* CK804 */
+		case 0x00d8:	/* CK8 */
+		case 0x00e8:	/* CK8S */
+			if (pci_set_consistent_dma_mask(pdev,
+						DMA_31BIT_MASK) < 0)
+				ehci_warn(ehci, "can't enable NVidia "
+					"workaround for >2GB RAM\n");
+			break;
 		/* Some NForce2 chips have problems with selective suspend;
 		 * fixed in newer silicon.
 		 */
_

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/