usb: uhci-platform driver fails after patch changes during merge

DISCUSSION Tony Prisk 2 participants 7 messages Oct 4, 2012

Share & tools

Tony Prisk

author

2012-10-04 06:48 UTC #1 lore raw

Quoted 32 lines · Tony Prisk

Mike,
I see someone made changes to the uhci-platform.c driver I submitted during v3.7 which results in it not working on mach-vt8500.
Could you clarify why the changes were made, and what the suggested resolution would be to solve the problem that it introduced?
Lines indicated by ---> below were removed from the patch, which means that on arch-vt8500 there is no dma_mask set, and its fails to communicate with attached devices.
Regards
Tony P
static int __devinit uhci_hcd_platform_probe(struct platform_device *pdev)
...
	if (usb_disabled())
		return -ENODEV;
--->
	/* Right now device-tree probed devices don't get dma_mask set.
	 * Since shared usb code relies on it, set it here for now.
	 * Once we have dma capability bindings this can go away.
	 */
	if (!pdev->dev.dma_mask)
		pdev->dev.dma_mask = &platform_uhci_dma_mask;
--->
	hcd = usb_create_hcd(&uhci_platform_hc_driver, &pdev->dev,
				pdev->name);

Quoted 2 lines

...

Apologies Mike,

This isn't intended for you.

Tony Prisk

author

2012-10-04 06:55 UTC #2 lore raw

In reply to Tony Prisk

Quoted 32 lines · Tony Prisk

Mike,
I see someone made changes to the uhci-platform.c driver I submitted during v3.7 which results in it not working on mach-vt8500.
Could you clarify why the changes were made, and what the suggested resolution would be to solve the problem that it introduced?
Lines indicated by ---> below were removed from the patch, which means that on arch-vt8500 there is no dma_mask set, and its fails to communicate with attached devices.
Regards
Tony P
static int __devinit uhci_hcd_platform_probe(struct platform_device *pdev)
...
	if (usb_disabled())
		return -ENODEV;
--->
	/* Right now device-tree probed devices don't get dma_mask set.
	 * Since shared usb code relies on it, set it here for now.
	 * Once we have dma capability bindings this can go away.
	 */
	if (!pdev->dev.dma_mask)
		pdev->dev.dma_mask = &platform_uhci_dma_mask;
--->
	hcd = usb_create_hcd(&uhci_platform_hc_driver, &pdev->dev,
				pdev->name);

Quoted 7 lines

...
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Greg,

This message was intended for you rather than Mike. Could you clarify what happened and the expected resolution?

Regards

Tony P

Greg KH

2012-10-04 14:26 UTC #3 lore raw

Quoted 41 lines · Tony Prisk

Quoted 32 lines · Tony Prisk
Mike,
I see someone made changes to the uhci-platform.c driver I submitted during v3.7 which results in it not working on mach-vt8500.
Could you clarify why the changes were made, and what the suggested resolution would be to solve the problem that it introduced?
Lines indicated by ---> below were removed from the patch, which means that on arch-vt8500 there is no dma_mask set, and its fails to communicate with attached devices.
Regards
Tony P
static int __devinit uhci_hcd_platform_probe(struct platform_device *pdev)
...
	if (usb_disabled())
		return -ENODEV;
--->
	/* Right now device-tree probed devices don't get dma_mask set.
	 * Since shared usb code relies on it, set it here for now.
	 * Once we have dma capability bindings this can go away.
	 */
	if (!pdev->dev.dma_mask)
		pdev->dev.dma_mask = &platform_uhci_dma_mask;
--->
	hcd = usb_create_hcd(&uhci_platform_hc_driver, &pdev->dev,
				pdev->name);
Quoted 7 lines
...
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Quoted 4 lines

Greg,
This message was intended for you rather than Mike. Could you clarify what happened and the expected resolution?

I don't know, this should be directed at the person who made the change that is causing the problem, and the linux-usb at vger.kernel.org mailing list.

Who changed the patch? What patch exactly are you referring to? Who signed off on it? Where was it discussed?

thanks,

greg k-h

Tony Prisk

author

2012-10-04 17:42 UTC #4 lore raw

Quoted 54 lines · Greg KH

Quoted 41 lines · Tony Prisk
Quoted 32 lines · Tony Prisk
Mike,
I see someone made changes to the uhci-platform.c driver I submitted during v3.7 which results in it not working on mach-vt8500.
Could you clarify why the changes were made, and what the suggested resolution would be to solve the problem that it introduced?
Lines indicated by ---> below were removed from the patch, which means that on arch-vt8500 there is no dma_mask set, and its fails to communicate with attached devices.
Regards
Tony P
static int __devinit uhci_hcd_platform_probe(struct platform_device *pdev)
...
	if (usb_disabled())
		return -ENODEV;
--->
	/* Right now device-tree probed devices don't get dma_mask set.
	 * Since shared usb code relies on it, set it here for now.
	 * Once we have dma capability bindings this can go away.
	 */
	if (!pdev->dev.dma_mask)
		pdev->dev.dma_mask = &platform_uhci_dma_mask;
--->
	hcd = usb_create_hcd(&uhci_platform_hc_driver, &pdev->dev,
				pdev->name);
Quoted 7 lines
...
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Quoted 4 lines
Greg,
This message was intended for you rather than Mike. Could you clarify what happened and the expected resolution?
I don't know, this should be directed at the person who made the change that is causing the problem, and the linux-usb at vger.kernel.org mailing list.
Who changed the patch? What patch exactly are you referring to? Who signed off on it? Where was it discussed?

Quoted 8 lines

thanks,
greg k-h
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Greg,

Thats the problem I have - I can't track where the changes came from.

The commit details show:

author
Tony Prisk
<[email protected]>

Sat, 21 Jul 2012
10:58:53 +0000 (22:58
+1200)
committer
Greg Kroah-Hartman
<[email protected]>

Thu, 16 Aug 2012
21:00:37 +0000 (14:00
-0700)
commit
100d45970327f78584ff4846deeca14bba511e28

But this file -
drivers/usb/host/uhci-platform.c

isn't the version I supplied in the patchset. There are no other commits against it so I assume it was changed before it was committed.

Confused?!?

Regards

Tony P

Greg KH

2012-10-04 17:52 UTC #5 lore raw

Quoted 89 lines · Tony Prisk

Quoted 54 lines · Greg KH
Quoted 41 lines · Tony Prisk
Quoted 32 lines · Tony Prisk
Mike,
I see someone made changes to the uhci-platform.c driver I submitted during v3.7 which results in it not working on mach-vt8500.
Could you clarify why the changes were made, and what the suggested resolution would be to solve the problem that it introduced?
Lines indicated by ---> below were removed from the patch, which means that on arch-vt8500 there is no dma_mask set, and its fails to communicate with attached devices.
Regards
Tony P
static int __devinit uhci_hcd_platform_probe(struct platform_device *pdev)
...
	if (usb_disabled())
		return -ENODEV;
--->
	/* Right now device-tree probed devices don't get dma_mask set.
	 * Since shared usb code relies on it, set it here for now.
	 * Once we have dma capability bindings this can go away.
	 */
	if (!pdev->dev.dma_mask)
		pdev->dev.dma_mask = &platform_uhci_dma_mask;
--->
	hcd = usb_create_hcd(&uhci_platform_hc_driver, &pdev->dev,
				pdev->name);
Quoted 7 lines
...
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Quoted 4 lines
Greg,
This message was intended for you rather than Mike. Could you clarify what happened and the expected resolution?
I don't know, this should be directed at the person who made the change that is causing the problem, and the linux-usb at vger.kernel.org mailing list.
Who changed the patch? What patch exactly are you referring to? Who signed off on it? Where was it discussed?
Quoted 8 lines
thanks,
greg k-h
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Greg,
Thats the problem I have - I can't track where the changes came from.
The commit details show:
author
Tony Prisk
<[email protected]>
Sat, 21 Jul 2012
10:58:53 +0000 (22:58
+1200)
committer
Greg Kroah-Hartman
<[email protected]>
Thu, 16 Aug 2012
21:00:37 +0000 (14:00
-0700)
commit
100d45970327f78584ff4846deeca14bba511e28

Quoted 5 lines

But this file -
drivers/usb/host/uhci-platform.c
isn't the version I supplied in the patchset. There are no other commits against it so I assume it was changed before it was committed.

No, this commit matches up exactly with the email I got from you with
Message-Id: <[email protected]>
which looks to be the patch I applied here. The diffstat matches up exactly with that email.

Do you see that as well?

You should have gotten an automated email response from me when the patch was applied, with the patch in it. My email system doesn't keep those around, but you should have it somewhere, so you could compare it to the commit in Linus's tree now, right? Does it match up?

If I accidentally applied the incorrect patch (note, this looks like it was version 4 of the patch, was there a later version that I missed?), please feel free to send an add-on patch that fixes up any errors.

thanks,

greg k-h

linux at prisktech.co.nz

author Applied

2012-10-04 19:01 UTC #6 lore raw

Quoted 7 lines · Tony Prisk

Quoted 6 lines · Greg KH
Quoted 5 lines · Tony Prisk
Quoted 4 lines · Tony Prisk
Quoted 3 lines
Mike,
I see someone made changes to the uhci-platform.c driver I submitted

during v3.7 which results in it not working on mach-vt8500.

Quoted 2 lines

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Could you clarify why the changes were made, and what the suggested

resolution would be to solve the problem that it introduced?

Quoted 2 lines

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Lines indicated by ---> below were removed from the patch, which means

that on arch-vt8500 there is no dma_mask set, and its fails to communicate with attached devices.

Quoted 6 lines

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Regards
Tony P

Quoted 7 lines

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
static int __devinit uhci_hcd_platform_probe(struct platform_device *pdev)
...
	if (usb_disabled())
		return -ENODEV;
--->
	/* Right now device-tree probed devices don't get dma_mask set.
	 * Since shared usb code relies on it, set it here for now. * Once we

have dma capability bindings this can go away. */

Quoted 8 lines

	if (!pdev->dev.dma_mask)
		pdev->dev.dma_mask = &platform_uhci_dma_mask;
--->
	hcd = usb_create_hcd(&uhci_platform_hc_driver, &pdev->dev,
				pdev->name);

...

Quoted 8 lines

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Greg,
This message was intended for you rather than Mike. Could you clarify

what happened and the expected resolution?

Quoted 2 lines

Quoted 2 lines
Quoted 2 lines
I don't know, this should be directed at the person who made the change

that is causing the problem, and the linux-usb at vger.kernel.org mailing list.

Quoted 2 lines

Quoted 2 lines
Quoted 2 lines
Who changed the patch? What patch exactly are you referring to? Who

signed off on it? Where was it discussed?

Quoted 11 lines

Quoted 11 lines
Quoted 9 lines
thanks,
greg k-h
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Greg,
Thats the problem I have - I can't track where the changes came from. The

commit details show:

Quoted 9 lines

Quoted 9 lines
author
Tony Prisk
<[email protected]>
Sat, 21 Jul 2012
10:58:53 +0000 (22:58
+1200)
committer
Greg Kroah-Hartman
<[email protected]>

Quoted 8 lines

Quoted 8 lines
Thu, 16 Aug 2012
21:00:37 +0000 (14:00
-0700)
commit
100d45970327f78584ff4846deeca14bba511e28
But this file -
drivers/usb/host/uhci-platform.c
isn't the version I supplied in the patchset. There are no other commits

against it so I assume it was changed before it was committed.

Quoted 4 lines

No, this commit matches up exactly with the email I got from you with
Message-Id: <[email protected]>
which looks to be the patch I applied here. The diffstat matches up exactly

with that email.

Quoted 4 lines

Do you see that as well?
You should have gotten an automated email response from me when the patch was

applied, with the patch in it. My email system doesn't keep those around, but you should have it somewhere, so you could compare it to the commit in Linus's tree now, right? Does it match up?

Quoted 2 lines

If I accidentally applied the incorrect patch (note, this looks like it was

version 4 of the patch, was there a later version that I

Quoted 6 lines

missed?), please feel free to send an add-on patch that fixes up any errors.
thanks,
greg k-h

*facepalm*

I see what has happened now. When you said v4 I got even more confused as the missing code was in every version from v1->v5 of the current patchset (August/September). The patch is question is from July 21st and was standalone. I assumed that patch had been lost as I never recieved any feedback outside of Arnd and the internal vt8500 mailing list. A new patchset was done later which included the UHCI driver with fixes. This also explains the merge-conflict between arm-soc and usb trees just recently.

Apologies for the messup. I will send out a fix for 3.7-rc

Regards

Tony P