Re: linux-next: usb/pxa trees build failure

Re: linux-next: usb/pxa trees build failure

[Greg KH]

On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
> Hi Eric, Greg,
> 
> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
> failed like this:
> 
> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
> 
> Caused by the interaction of commit
> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
> power settings on resume") from the usb tree with commit
> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
> 
> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
> 
> [this also broke the following configs: magician_defconfig,
> em_x270_defconfig and cm_x2xx_defconfig.]

Ick, not good.  Aric, should I drop this patch from my tree?

thanks,

greg k-h

Re: linux-next: usb/pxa trees build failure

[Eric Miao]

Greg KH wrote:
> On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
>> Hi Eric, Greg,
>>
>> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
>> failed like this:
>>
>> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
>>
>> Caused by the interaction of commit
>> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
>> power settings on resume") from the usb tree with commit
>> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
>> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
>>
>> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
>> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
>>
>> [this also broke the following configs: magician_defconfig,
>> em_x270_defconfig and cm_x2xx_defconfig.]
> 
> Ick, not good.  Aric, should I drop this patch from my tree?
> 

I think so, might be better to go through the pxa tree so that I can give
another review. (the fiber across the pacific seems to be broken again, so
I have to look into this tomorrow, still fetching -next, sigh).

Re: linux-next: usb/pxa trees build failure

[Greg KH]

On Tue, Aug 18, 2009 at 10:05:25PM +0800, Eric Miao wrote:
> Greg KH wrote:
> > On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
> >> Hi Eric, Greg,
> >>
> >> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
> >> failed like this:
> >>
> >> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
> >>
> >> Caused by the interaction of commit
> >> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
> >> power settings on resume") from the usb tree with commit
> >> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
> >> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
> >>
> >> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
> >> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
> >>
> >> [this also broke the following configs: magician_defconfig,
> >> em_x270_defconfig and cm_x2xx_defconfig.]
> > 
> > Ick, not good.  Aric, should I drop this patch from my tree?
> > 
> 
> I think so, might be better to go through the pxa tree so that I can give
> another review. (the fiber across the pacific seems to be broken again, so
> I have to look into this tomorrow, still fetching -next, sigh).

Ok, let me know if you want me to drop these patches from my tree.

thanks,

greg k-h

Re: linux-next: usb/pxa trees build failure

[Aric D. Blumer]

Greg KH wrote:
> On Tue, Aug 18, 2009 at 10:05:25PM +0800, Eric Miao wrote:
>   
>> Greg KH wrote:
>>     
>>> On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
>>>       
>>>> Hi Eric, Greg,
>>>>
>>>> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
>>>> failed like this:
>>>>
>>>> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
>>>>
>>>> Caused by the interaction of commit
>>>> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
>>>> power settings on resume") from the usb tree with commit
>>>> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
>>>> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
>>>>
>>>> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
>>>> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
>>>>
>>>> [this also broke the following configs: magician_defconfig,
>>>> em_x270_defconfig and cm_x2xx_defconfig.]
>>>>         
>>> Ick, not good.  Aric, should I drop this patch from my tree?
>>>       

Looks like it, but I'll let Eric make the call.

>> I think so, might be better to go through the pxa tree so that I can give
>> another review. (the fiber across the pacific seems to be broken again, so
>> I have to look into this tomorrow, still fetching -next, sigh).
>>     
>
> Ok, let me know if you want me to drop these patches from my tree.
>
> thanks,
>
> greg k-h
>   

Re: linux-next: usb/pxa trees build failure

[Eric Miao]

Greg KH wrote:
> On Tue, Aug 18, 2009 at 10:05:25PM +0800, Eric Miao wrote:
>> Greg KH wrote:
>>> On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
>>>> Hi Eric, Greg,
>>>>
>>>> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
>>>> failed like this:
>>>>
>>>> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
>>>>
>>>> Caused by the interaction of commit
>>>> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
>>>> power settings on resume") from the usb tree with commit
>>>> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
>>>> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
>>>>
>>>> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
>>>> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
>>>>
>>>> [this also broke the following configs: magician_defconfig,
>>>> em_x270_defconfig and cm_x2xx_defconfig.]
>>> Ick, not good.  Aric, should I drop this patch from my tree?
>>>
>> I think so, might be better to go through the pxa tree so that I can give
>> another review. (the fiber across the pacific seems to be broken again, so
>> I have to look into this tomorrow, still fetching -next, sigh).
> 
> Ok, let me know if you want me to drop these patches from my tree.
> 

OK, so the problem is 'pdev' is no longer part of the dev_pm_ops, ('dev' is used
instead), and the 'inf' pointer can just be referenced from 'dev->platform_data',
so the fix is actually trivial, but it does have dependency issue.

Greg,

It looks to me would be better if this goes the pxa tree, and I'll send out the
updated patch of 6458baec later.

Re: linux-next: usb/pxa trees build failure

[Eric Miao]

Aric D. Blumer wrote:
> Greg KH wrote:
>> On Tue, Aug 18, 2009 at 10:05:25PM +0800, Eric Miao wrote:
>>   
>>> Greg KH wrote:
>>>     
>>>> On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
>>>>       
>>>>> Hi Eric, Greg,
>>>>>
>>>>> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
>>>>> failed like this:
>>>>>
>>>>> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
>>>>>
>>>>> Caused by the interaction of commit
>>>>> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
>>>>> power settings on resume") from the usb tree with commit
>>>>> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
>>>>> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
>>>>>
>>>>> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
>>>>> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
>>>>>
>>>>> [this also broke the following configs: magician_defconfig,
>>>>> em_x270_defconfig and cm_x2xx_defconfig.]
>>>>>         
>>>> Ick, not good.  Aric, should I drop this patch from my tree?
>>>>       
> 
> Looks like it, but I'll let Eric make the call.
> 

Thanks Aric,

Looks like the additional 'if' statement of 'inf' is unnecessary (otherwise
the _probe has already been failed), patch modified a little bit into the
following, please help confirm before I commit:

commit cc6446f6d3e81939dcc1945aab9ad734b175ce66
Author: Aric D. Blumer <aric@sdgsystems.com>
Date:   Thu Jul 30 13:42:30 2009 -0400

    USB: ohci-pxa27x: Reconfigure power settings on resume
    
    On resume, the power-related bits in UHCRHDA were not being set, so
    they would default to the reset state.  For PXA3xx devices, OCPM must
    be cleared, but it was remaining set from resume reset.
    
    Signed-off-by: Aric D. Blumer <aric@sdgsystems.com>
    Cc: David Brownell <dbrownell@users.sourceforge.net>
    Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

diff --git a/drivers/usb/host/ohci-pxa27x.c b/drivers/usb/host/ohci-pxa27x.c
index 6180074..1229cc0 100644
--- a/drivers/usb/host/ohci-pxa27x.c
+++ b/drivers/usb/host/ohci-pxa27x.c
@@ -497,6 +497,7 @@ static int ohci_hcd_pxa27x_drv_resume(struct device *dev)
 {
        struct usb_hcd *hcd = dev_get_drvdata(dev);
        struct pxa27x_ohci *ohci = to_pxa27x_ohci(hcd);
+       struct pxaohci_platform_data *inf = dev->platform_data;
        int status;
 
        if (time_before(jiffies, ohci->ohci.next_statechange))
@@ -506,6 +507,9 @@ static int ohci_hcd_pxa27x_drv_resume(struct device *dev)
        if ((status = pxa27x_start_hc(ohci, dev)) < 0)
                return status;
 
+       /* Select Power Management Mode */
+       pxa27x_ohci_select_pmm(ohci, inf->port_mode);
+
        ohci_finish_controller_resume(hcd);
        return 0;
 }

Re: linux-next: usb/pxa trees build failure

[Greg KH]

On Wed, Aug 19, 2009 at 10:35:10AM +0800, Eric Miao wrote:
> Greg KH wrote:
> > On Tue, Aug 18, 2009 at 10:05:25PM +0800, Eric Miao wrote:
> >> Greg KH wrote:
> >>> On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
> >>>> Hi Eric, Greg,
> >>>>
> >>>> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
> >>>> failed like this:
> >>>>
> >>>> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
> >>>>
> >>>> Caused by the interaction of commit
> >>>> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
> >>>> power settings on resume") from the usb tree with commit
> >>>> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
> >>>> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
> >>>>
> >>>> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
> >>>> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
> >>>>
> >>>> [this also broke the following configs: magician_defconfig,
> >>>> em_x270_defconfig and cm_x2xx_defconfig.]
> >>> Ick, not good.  Aric, should I drop this patch from my tree?
> >>>
> >> I think so, might be better to go through the pxa tree so that I can give
> >> another review. (the fiber across the pacific seems to be broken again, so
> >> I have to look into this tomorrow, still fetching -next, sigh).
> > 
> > Ok, let me know if you want me to drop these patches from my tree.
> > 
> 
> OK, so the problem is 'pdev' is no longer part of the dev_pm_ops, ('dev' is used
> instead), and the 'inf' pointer can just be referenced from 'dev->platform_data',
> so the fix is actually trivial, but it does have dependency issue.
> 
> Greg,
> 
> It looks to me would be better if this goes the pxa tree, and I'll send out the
> updated patch of 6458baec later.

Ok, should I drop just this one patch from my tree, or any others as
well?

thanks,

greg k-h

Re: linux-next: usb/pxa trees build failure

[Aric D. Blumer]

Greg KH wrote:
> On Wed, Aug 19, 2009 at 10:35:10AM +0800, Eric Miao wrote:
>   
>> Greg KH wrote:
>>     
>>> On Tue, Aug 18, 2009 at 10:05:25PM +0800, Eric Miao wrote:
>>>       
>>>> Greg KH wrote:
>>>>         
>>>>> On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
>>>>>           
>>>>>> Hi Eric, Greg,
>>>>>>
>>>>>> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
>>>>>> failed like this:
>>>>>>
>>>>>> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
>>>>>>
>>>>>> Caused by the interaction of commit
>>>>>> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
>>>>>> power settings on resume") from the usb tree with commit
>>>>>> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
>>>>>> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
>>>>>>
>>>>>> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
>>>>>> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
>>>>>>
>>>>>> [this also broke the following configs: magician_defconfig,
>>>>>> em_x270_defconfig and cm_x2xx_defconfig.]
>>>>>>             
>>>>> Ick, not good.  Aric, should I drop this patch from my tree?
>>>>>
>>>>>           
>>>> I think so, might be better to go through the pxa tree so that I can give
>>>> another review. (the fiber across the pacific seems to be broken again, so
>>>> I have to look into this tomorrow, still fetching -next, sigh).
>>>>         
>>> Ok, let me know if you want me to drop these patches from my tree.
>>>
>>>       
>> OK, so the problem is 'pdev' is no longer part of the dev_pm_ops, ('dev' is used
>> instead), and the 'inf' pointer can just be referenced from 'dev->platform_data',
>> so the fix is actually trivial, but it does have dependency issue.
>>
>> Greg,
>>
>> It looks to me would be better if this goes the pxa tree, and I'll send out the
>> updated patch of 6458baec later.
>>     
>
> Ok, should I drop just this one patch from my tree, or any others as
> well

It is an isolated patch, so dropping it alone should be fine.

Re: linux-next: usb/pxa trees build failure

[Greg KH]

On Wed, Aug 19, 2009 at 02:24:08PM -0400, Aric D. Blumer wrote:
> Greg KH wrote:
> > On Wed, Aug 19, 2009 at 10:35:10AM +0800, Eric Miao wrote:
> >   
> >> Greg KH wrote:
> >>     
> >>> On Tue, Aug 18, 2009 at 10:05:25PM +0800, Eric Miao wrote:
> >>>       
> >>>> Greg KH wrote:
> >>>>         
> >>>>> On Tue, Aug 18, 2009 at 01:58:52PM +1000, Stephen Rothwell wrote:
> >>>>>           
> >>>>>> Hi Eric, Greg,
> >>>>>>
> >>>>>> The next-20090817 (and a few before) build (arm trizeps4_defconfig)
> >>>>>> failed like this:
> >>>>>>
> >>>>>> drivers/usb/host/ohci-pxa27x.c:507: error: 'pdev' undeclared (first use in this function)
> >>>>>>
> >>>>>> Caused by the interaction of commit
> >>>>>> 6458baec17e19d38b9d4eb1f96c213f91609c0e3 ("USB: ohci-pxa27x: Reconfigure
> >>>>>> power settings on resume") from the usb tree with commit
> >>>>>> b8fb7d4317bcea1b48c9560d23dff81121870ece ("[ARM] pxa: update
> >>>>>> ohci-pxa27x.c to use 'struct dev_pm_ops'"). from the pxa tree.
> >>>>>>
> >>>>>> Greg/Aric, maybe that usb tree patch needs to go through the pxa tree?
> >>>>>> (and maybe "USB: ohci-pxa27x: Allow NOCP and OCPM to be cleared" as well?)
> >>>>>>
> >>>>>> [this also broke the following configs: magician_defconfig,
> >>>>>> em_x270_defconfig and cm_x2xx_defconfig.]
> >>>>>>             
> >>>>> Ick, not good.  Aric, should I drop this patch from my tree?
> >>>>>
> >>>>>           
> >>>> I think so, might be better to go through the pxa tree so that I can give
> >>>> another review. (the fiber across the pacific seems to be broken again, so
> >>>> I have to look into this tomorrow, still fetching -next, sigh).
> >>>>         
> >>> Ok, let me know if you want me to drop these patches from my tree.
> >>>
> >>>       
> >> OK, so the problem is 'pdev' is no longer part of the dev_pm_ops, ('dev' is used
> >> instead), and the 'inf' pointer can just be referenced from 'dev->platform_data',
> >> so the fix is actually trivial, but it does have dependency issue.
> >>
> >> Greg,
> >>
> >> It looks to me would be better if this goes the pxa tree, and I'll send out the
> >> updated patch of 6458baec later.
> >>     
> >
> > Ok, should I drop just this one patch from my tree, or any others as
> > well
> 
> It is an isolated patch, so dropping it alone should be fine.

Ok, it's now dropped from my tree, linux-next should now be happy :)

thanks,

greg k-h