[patch for-3.8-rc8] usb, chipidea: fix link error when USB_EHCI_HCD is a module

[patch for-3.8-rc8] usb, chipidea: fix link error when USB_EHCI_HCD is a module

[David Rientjes]

Fixes link error when USB_EHCI_HCD=m and USB_CHIPIDEA_HOST=y:

	drivers/built-in.o: In function `ci_hdrc_host_init':
	drivers/usb/chipidea/host.c:104: undefined reference to `ehci_init_driver'

as a result of commit 09f6ffde2ece ("USB: EHCI: fix build error by making
ChipIdea host a normal EHCI driver").

Cc: stable@vger.kernel.org [v3.7+]
Signed-off-by: David Rientjes <rientjes@google.com>
---
 drivers/usb/chipidea/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/usb/chipidea/Kconfig b/drivers/usb/chipidea/Kconfig
--- a/drivers/usb/chipidea/Kconfig
+++ b/drivers/usb/chipidea/Kconfig
@@ -20,7 +20,7 @@ config USB_CHIPIDEA_UDC
 config USB_CHIPIDEA_HOST
 	bool "ChipIdea host controller"
 	depends on USB=y || USB=USB_CHIPIDEA
-	depends on USB_EHCI_HCD
+	depends on USB_EHCI_HCD=y
 	select USB_EHCI_ROOT_HUB_TT
 	help
 	  Say Y here to enable host controller functionality of the

Re: [patch for-3.8-rc8] usb, chipidea: fix link error when USB_EHCI_HCD is a module

[Greg Kroah-Hartman]

I think you mean "3.9" in your subject: right?

On Fri, Apr 26, 2013 at 01:25:01PM -0700, David Rientjes wrote:
> Fixes link error when USB_EHCI_HCD=m and USB_CHIPIDEA_HOST=y:
> 
> 	drivers/built-in.o: In function `ci_hdrc_host_init':
> 	drivers/usb/chipidea/host.c:104: undefined reference to `ehci_init_driver'
> 
> as a result of commit 09f6ffde2ece ("USB: EHCI: fix build error by making
> ChipIdea host a normal EHCI driver").
> 
> Cc: stable@vger.kernel.org [v3.7+]
> Signed-off-by: David Rientjes <rientjes@google.com>
> ---
>  drivers/usb/chipidea/Kconfig | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

There have been a ton of chipidea changes recently in linux-next.  This
really just looks like you got the configuration wrong, nothing that is
needed at this point in the release cycle, right?

thanks,

greg k-h

Re: [patch for-3.8-rc8] usb, chipidea: fix link error when USB_EHCI_HCD is a module

[David Rientjes]

On Fri, 26 Apr 2013, Greg Kroah-Hartman wrote:

> I think you mean "3.9" in your subject: right?
> 

It's your subsystem, if you want to release a kernel with a known-broken 
config, that's your call.

> On Fri, Apr 26, 2013 at 01:25:01PM -0700, David Rientjes wrote:
> > Fixes link error when USB_EHCI_HCD=m and USB_CHIPIDEA_HOST=y:
> > 
> > 	drivers/built-in.o: In function `ci_hdrc_host_init':
> > 	drivers/usb/chipidea/host.c:104: undefined reference to `ehci_init_driver'
> > 
> > as a result of commit 09f6ffde2ece ("USB: EHCI: fix build error by making
> > ChipIdea host a normal EHCI driver").
> > 
> > Cc: stable@vger.kernel.org [v3.7+]
> > Signed-off-by: David Rientjes <rientjes@google.com>
> > ---
> >  drivers/usb/chipidea/Kconfig | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> There have been a ton of chipidea changes recently in linux-next.  This
> really just looks like you got the configuration wrong, nothing that is
> needed at this point in the release cycle, right?
> 

This happens in both linux-next as well as Linus's tree, and has since 
3.7.

Re: [patch for-3.8-rc8] usb, chipidea: fix link error when USB_EHCI_HCD is a module

[Greg Kroah-Hartman]

On Fri, Apr 26, 2013 at 02:50:56PM -0700, David Rientjes wrote:
> On Fri, 26 Apr 2013, Greg Kroah-Hartman wrote:
> 
> > I think you mean "3.9" in your subject: right?
> > 
> 
> It's your subsystem, if you want to release a kernel with a known-broken 
> config, that's your call.

As you point out, it's been "known broken" since 3.7, no rush now to fix
it, especially so late in the merge window. :)

> > On Fri, Apr 26, 2013 at 01:25:01PM -0700, David Rientjes wrote:
> > > Fixes link error when USB_EHCI_HCD=m and USB_CHIPIDEA_HOST=y:
> > > 
> > > 	drivers/built-in.o: In function `ci_hdrc_host_init':
> > > 	drivers/usb/chipidea/host.c:104: undefined reference to `ehci_init_driver'
> > > 
> > > as a result of commit 09f6ffde2ece ("USB: EHCI: fix build error by making
> > > ChipIdea host a normal EHCI driver").
> > > 
> > > Cc: stable@vger.kernel.org [v3.7+]
> > > Signed-off-by: David Rientjes <rientjes@google.com>
> > > ---
> > >  drivers/usb/chipidea/Kconfig | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > There have been a ton of chipidea changes recently in linux-next.  This
> > really just looks like you got the configuration wrong, nothing that is
> > needed at this point in the release cycle, right?
> > 
> 
> This happens in both linux-next as well as Linus's tree, and has since 
> 3.7.

The chipidea maintainer has a big update all queued up in my trees for
3.10-rc1, if you could verify that linux-next fixes your problem, I
would appreciate it.

If not, please let us know.

thanks,

greg k-h

Re: [patch for-3.8-rc8] usb, chipidea: fix link error when USB_EHCI_HCD is a module

[David Rientjes]

On Fri, 26 Apr 2013, Greg Kroah-Hartman wrote:

> > It's your subsystem, if you want to release a kernel with a known-broken 
> > config, that's your call.
> 
> As you point out, it's been "known broken" since 3.7, no rush now to fix
> it, especially so late in the merge window. :)
> 

It's "known broken" as of today because I've made you aware of it and 
provided a patch to fix it.  I annotated the changelog with 
stable@vger.kernel.org since v3.7 since the problem goes back to that time 
because of the commit quoted in the changelog.

The Kconfig allows for the broken config in 3.8-rc8, so I see no risk 
whatosever in fixing it to prevent specifically that broken combination.  
But again, it's your subsystem so if you'd like to drop this plainly 
obvious fix, that's completely your call.

> The chipidea maintainer has a big update all queued up in my trees for
> 3.10-rc1, if you could verify that linux-next fixes your problem, I
> would appreciate it.
> 

As stated, linux-next as of today has this problem.  If you have a public 
git tree with patches that touch this area that are not pushed to 
linux-next, please send me the location and I'll try the combination 
stated in the changelog.  (You could do that, as well, if you want.)