Blackfin STAMP537 compiling error in 2.6.25-rc8

Blackfin STAMP537 compiling error in 2.6.25-rc8

[Bryan Wu]

We found the latest kernel compile failed with Blackfin STAMP537
configuration on our daily build machine.
---
  CC      drivers/net/smc91x.o
drivers/net/smc91x.c: In function 'smc_rcv':
drivers/net/smc91x.c:513: warning: passing argument 1 of 'insw' makes
integer from pointer without a cast
drivers/net/smc91x.c: In function 'smc_hardware_send_pkt':
drivers/net/smc91x.c:620: warning: passing argument 1 of 'outsw' makes
integer from pointer without a cast
drivers/net/smc91x.c:1859:32: error: macro "SMC_GET_MAC_ADDR" requires
2 arguments, but only 1 given
drivers/net/smc91x.c: In function 'smc_probe':
drivers/net/smc91x.c:1859: error: 'SMC_GET_MAC_ADDR' undeclared (first
use in this function)
drivers/net/smc91x.c:1859: error: (Each undeclared identifier is
reported only once
drivers/net/smc91x.c:1859: error: for each function it appears in.)
make[2]: *** [drivers/net/smc91x.o] Error 1
make[1]: *** [drivers/net] Error 2
make: *** [drivers] Error 2
---

That is because my -mm tree fixing bug patch:
(http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9e6db60825ef7e7999abc610ce256ba768e58162)
was merged into mainline before the smc91x API change patch in -mm:
http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25-rc5/2.6.25-rc5-mm1/broken-out/smc91x-pass-along-private-data-v2.patch

Did I miss something here? or this patch is on the way before 2.6.25
final release?

Thanks
-Bryan Wu

Re: Blackfin STAMP537 compiling error in 2.6.25-rc8

[Jeff Garzik]

Bryan Wu wrote:
> We found the latest kernel compile failed with Blackfin STAMP537
> configuration on our daily build machine.
> ---
>   CC      drivers/net/smc91x.o
> drivers/net/smc91x.c: In function 'smc_rcv':
> drivers/net/smc91x.c:513: warning: passing argument 1 of 'insw' makes
> integer from pointer without a cast
> drivers/net/smc91x.c: In function 'smc_hardware_send_pkt':
> drivers/net/smc91x.c:620: warning: passing argument 1 of 'outsw' makes
> integer from pointer without a cast
> drivers/net/smc91x.c:1859:32: error: macro "SMC_GET_MAC_ADDR" requires
> 2 arguments, but only 1 given
> drivers/net/smc91x.c: In function 'smc_probe':
> drivers/net/smc91x.c:1859: error: 'SMC_GET_MAC_ADDR' undeclared (first
> use in this function)
> drivers/net/smc91x.c:1859: error: (Each undeclared identifier is
> reported only once
> drivers/net/smc91x.c:1859: error: for each function it appears in.)
> make[2]: *** [drivers/net/smc91x.o] Error 1
> make[1]: *** [drivers/net] Error 2
> make: *** [drivers] Error 2
> ---
> 
> That is because my -mm tree fixing bug patch:
> (http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9e6db60825ef7e7999abc610ce256ba768e58162)
> was merged into mainline before the smc91x API change patch in -mm:
> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25-rc5/2.6.25-rc5-mm1/broken-out/smc91x-pass-along-private-data-v2.patch
> 
> Did I miss something here? or this patch is on the way before 2.6.25
> final release?

It looks like I missed something...  I thought akpm was going to send 
the patch for -mm in?

If not, let's revert this one...

	Jeff

Re: Blackfin STAMP537 compiling error in 2.6.25-rc8

[Bryan Wu]

On Mon, Apr 7, 2008 at 1:13 AM, Jeff Garzik <jeff@garzik.org> wrote:
>
> Bryan Wu wrote:
>
> > We found the latest kernel compile failed with Blackfin STAMP537
> > configuration on our daily build machine.
> > ---
> >  CC      drivers/net/smc91x.o
> > drivers/net/smc91x.c: In function 'smc_rcv':
> > drivers/net/smc91x.c:513: warning: passing argument 1 of 'insw' makes
> > integer from pointer without a cast
> > drivers/net/smc91x.c: In function 'smc_hardware_send_pkt':
> > drivers/net/smc91x.c:620: warning: passing argument 1 of 'outsw' makes
> > integer from pointer without a cast
> > drivers/net/smc91x.c:1859:32: error: macro "SMC_GET_MAC_ADDR" requires
> > 2 arguments, but only 1 given
> > drivers/net/smc91x.c: In function 'smc_probe':
> > drivers/net/smc91x.c:1859: error: 'SMC_GET_MAC_ADDR' undeclared (first
> > use in this function)
> > drivers/net/smc91x.c:1859: error: (Each undeclared identifier is
> > reported only once
> > drivers/net/smc91x.c:1859: error: for each function it appears in.)
> > make[2]: *** [drivers/net/smc91x.o] Error 1
> > make[1]: *** [drivers/net] Error 2
> > make: *** [drivers] Error 2
> > ---
> >
> > That is because my -mm tree fixing bug patch:
> >
> (http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9e6db60825ef7e7999abc610ce256ba768e58162)
> > was merged into mainline before the smc91x API change patch in -mm:
> >
> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25-rc5/2.6.25-rc5-mm1/broken-out/smc91x-pass-along-private-data-v2.patch
> >
> > Did I miss something here? or this patch is on the way before 2.6.25
> > final release?
> >
>
>  It looks like I missed something...  I thought akpm was going to send the
> patch for -mm in?
>
>  If not, let's revert this one...
>

ok, no problem, it is ok for me.

Thanks
-Bryan

Re: Blackfin STAMP537 compiling error in 2.6.25-rc8

[Jeff Garzik]

Bryan Wu wrote:
> On Mon, Apr 7, 2008 at 1:13 AM, Jeff Garzik <jeff@garzik.org> wrote:
>> Bryan Wu wrote:
>>
>>> We found the latest kernel compile failed with Blackfin STAMP537
>>> configuration on our daily build machine.
>>> ---
>>>  CC      drivers/net/smc91x.o
>>> drivers/net/smc91x.c: In function 'smc_rcv':
>>> drivers/net/smc91x.c:513: warning: passing argument 1 of 'insw' makes
>>> integer from pointer without a cast
>>> drivers/net/smc91x.c: In function 'smc_hardware_send_pkt':
>>> drivers/net/smc91x.c:620: warning: passing argument 1 of 'outsw' makes
>>> integer from pointer without a cast
>>> drivers/net/smc91x.c:1859:32: error: macro "SMC_GET_MAC_ADDR" requires
>>> 2 arguments, but only 1 given
>>> drivers/net/smc91x.c: In function 'smc_probe':
>>> drivers/net/smc91x.c:1859: error: 'SMC_GET_MAC_ADDR' undeclared (first
>>> use in this function)
>>> drivers/net/smc91x.c:1859: error: (Each undeclared identifier is
>>> reported only once
>>> drivers/net/smc91x.c:1859: error: for each function it appears in.)
>>> make[2]: *** [drivers/net/smc91x.o] Error 1
>>> make[1]: *** [drivers/net] Error 2
>>> make: *** [drivers] Error 2
>>> ---
>>>
>>> That is because my -mm tree fixing bug patch:
>>>
>> (http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9e6db60825ef7e7999abc610ce256ba768e58162)
>>> was merged into mainline before the smc91x API change patch in -mm:
>>>
>> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25-rc5/2.6.25-rc5-mm1/broken-out/smc91x-pass-along-private-data-v2.patch
>>> Did I miss something here? or this patch is on the way before 2.6.25
>>> final release?
>>>
>>  It looks like I missed something...  I thought akpm was going to send the
>> patch for -mm in?
>>
>>  If not, let's revert this one...
>>
> 
> ok, no problem, it is ok for me.

Cool.

David or Linus, please revert 9e6db60825ef7e7999abc610ce256ba768e58162

It was merged without the API it needed, causing build breakage.

Acked-by: Jeff Garzik <jgarzik@redhat.com>



Andrew,

Since it was an arch patch to a driver I cannot build, I made the 
[mistaken] assumption that your sending it to me was an indication that 
it should actually be merged :)  Especially with a subject like "fix 
build breakage", for a platform I cannot easily build...

Are there any process tweaks that could be made here?  Maybe -mm stuff 
sent to maintainers could be tagged, noting dependencies on other -mm 
patches that are not yet upstream?

Thanks,

	Jeff

Re: Blackfin STAMP537 compiling error in 2.6.25-rc8

[Andrew Morton]

On Mon, 07 Apr 2008 16:17:01 -0400
Jeff Garzik <jeff@garzik.org> wrote:

> Since it was an arch patch to a driver I cannot build, I made the 
> [mistaken] assumption that your sending it to me was an indication that 
> it should actually be merged :)  Especially with a subject like "fix 
> build breakage", for a platform I cannot easily build...
> 
> Are there any process tweaks that could be made here?  Maybe -mm stuff 
> sent to maintainers could be tagged, noting dependencies on other -mm 
> patches that are not yet upstream?

Well.  It really helps if submitters tell us that a patch fixes another
pending patch, and which one that is.  Usually Ihave to ask if I can't work
it out.


But if a) we weren't told that and b) I have no reason to think it's not a
mainline problem and c) the patch applies to mainline and d) the patch
affects an architecture which I'm not cross-compiling for, it's going to
sneak through.

Fortunately, a && b && c && d doesn't happen at all often.

Re: Blackfin STAMP537 compiling error in 2.6.25-rc8

[Bryan Wu]

On Tue, Apr 8, 2008 at 4:17 AM, Jeff Garzik <jeff@garzik.org> wrote:
>
> Bryan Wu wrote:
>
> > On Mon, Apr 7, 2008 at 1:13 AM, Jeff Garzik <jeff@garzik.org> wrote:
> >
> > > Bryan Wu wrote:
> > >
> > >
> > > > We found the latest kernel compile failed with Blackfin STAMP537
> > > > configuration on our daily build machine.
> > > > ---
> > > >  CC      drivers/net/smc91x.o
> > > > drivers/net/smc91x.c: In function 'smc_rcv':
> > > > drivers/net/smc91x.c:513: warning: passing argument 1 of 'insw' makes
> > > > integer from pointer without a cast
> > > > drivers/net/smc91x.c: In function 'smc_hardware_send_pkt':
> > > > drivers/net/smc91x.c:620: warning: passing argument 1 of 'outsw' makes
> > > > integer from pointer without a cast
> > > > drivers/net/smc91x.c:1859:32: error: macro "SMC_GET_MAC_ADDR" requires
> > > > 2 arguments, but only 1 given
> > > > drivers/net/smc91x.c: In function 'smc_probe':
> > > > drivers/net/smc91x.c:1859: error: 'SMC_GET_MAC_ADDR' undeclared (first
> > > > use in this function)
> > > > drivers/net/smc91x.c:1859: error: (Each undeclared identifier is
> > > > reported only once
> > > > drivers/net/smc91x.c:1859: error: for each function it appears in.)
> > > > make[2]: *** [drivers/net/smc91x.o] Error 1
> > > > make[1]: *** [drivers/net] Error 2
> > > > make: *** [drivers] Error 2
> > > > ---
> > > >
> > > > That is because my -mm tree fixing bug patch:
> > > >
> > > >
> > >
> (http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9e6db60825ef7e7999abc610ce256ba768e58162)
> > >
> > > > was merged into mainline before the smc91x API change patch in -mm:
> > > >
> > > >
> > >
> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25-rc5/2.6.25-rc5-mm1/broken-out/smc91x-pass-along-private-data-v2.patch
> > >
> > > > Did I miss something here? or this patch is on the way before 2.6.25
> > > > final release?
> > > >
> > > >
> > >  It looks like I missed something...  I thought akpm was going to send
> the
> > > patch for -mm in?
> > >
> > >  If not, let's revert this one...
> > >
> > >
> >
> > ok, no problem, it is ok for me.
> >
>
>  Cool.
>
>  David or Linus, please revert 9e6db60825ef7e7999abc610ce256ba768e58162
>
>  It was merged without the API it needed, causing build breakage.
>
>  Acked-by: Jeff Garzik <jgarzik@redhat.com>
>
>

Thanks a lot.

>
>  Andrew,
>
>  Since it was an arch patch to a driver I cannot build, I made the
> [mistaken] assumption that your sending it to me was an indication that it
> should actually be merged :)  Especially with a subject like "fix build
> breakage", for a platform I cannot easily build...
>
>  Are there any process tweaks that could be made here?  Maybe -mm stuff sent
> to maintainers could be tagged, noting dependencies on other -mm patches
> that are not yet upstream?
>

Oh, IMO, this confusing is initially introduced by me. I should choose
more meaningful name of this patch.
such as smc91x-pass-along-private-data-v2-fix-build-fail-on-blackfin.patch.
Then things become more clear for everyone.

-Bryan