linux-next: firmware merge conflicts

linux-next: firmware merge conflicts

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1358 bytes --]

Hi David,

Today's linux-next merge of the firmware tree got a few conflicts:

Commit 33f1dd02ed25db0790fec8a62b008c71831c49d9 ("ti_usb: kick firmware
into user space") from the usb tree against commit
e009cdbc2e96e72fe84df7cf014abcd7a7eb7dc7 ("ti_usb_3410_5052: use
request_firmware()") from the firmware tree.  These seem to have similar
goals but are using different methods.  This affected the files
drivers/usb/serial/Kconfig,
drivers/usb/serial/{ti_fw_3410.h,ti_fw_5052.h,ti_usb_3410_5052.c}.  I
have used the firmware tree version for now.

Some simple conflicts in sound/Kconfig due to commit
bd7d34e759d7537b42cffb7205b7b50dda6535bb ("sound: Convert to menuconfig")
from the sound tree.

Simple conflicts in sound/pci/maestro3.c and
sound/pci/ymfpci/ymfpci_main.c because "maestro3: treat firmware data as
const" and "ymfpci: treat firmware data as const" appear in both the
driver-core tree and the firmware tree.  The conflicts are because the
firmware tree further modifies these files.

So, you and Alan need to come to an agreement about ti_usb (and inform
Greg :-)).  And you should probably ask Greg to drop the to "const"
patches (or maybe the other way around).  And I can carry the
sound/Kconfig fixup.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: linux-next: firmware merge conflicts

[Takashi Iwai]

At Wed, 4 Jun 2008 16:07:42 +1000,
Stephen Rothwell wrote:
> 
> Hi David,
> 
> Today's linux-next merge of the firmware tree got a few conflicts:
> 
> Commit 33f1dd02ed25db0790fec8a62b008c71831c49d9 ("ti_usb: kick firmware
> into user space") from the usb tree against commit
> e009cdbc2e96e72fe84df7cf014abcd7a7eb7dc7 ("ti_usb_3410_5052: use
> request_firmware()") from the firmware tree.  These seem to have similar
> goals but are using different methods.  This affected the files
> drivers/usb/serial/Kconfig,
> drivers/usb/serial/{ti_fw_3410.h,ti_fw_5052.h,ti_usb_3410_5052.c}.  I
> have used the firmware tree version for now.
> 
> Some simple conflicts in sound/Kconfig due to commit
> bd7d34e759d7537b42cffb7205b7b50dda6535bb ("sound: Convert to menuconfig")
> from the sound tree.
> 
> Simple conflicts in sound/pci/maestro3.c and
> sound/pci/ymfpci/ymfpci_main.c because "maestro3: treat firmware data as
> const" and "ymfpci: treat firmware data as const" appear in both the
> driver-core tree and the firmware tree.  The conflicts are because the
> firmware tree further modifies these files.
> 
> So, you and Alan need to come to an agreement about ti_usb (and inform
> Greg :-)).  And you should probably ask Greg to drop the to "const"
> patches (or maybe the other way around).

Or just let me know if you guys want to put the changes into sound
tree.  I'm willing to apply them...

>  And I can carry the
> sound/Kconfig fixup.

... together with this :)


thanks,

Takashi

Re: linux-next: firmware merge conflicts

[David Woodhouse]

On Wed, 2008-06-04 at 16:07 +1000, Stephen Rothwell wrote:
> Hi David,
> 
> Today's linux-next merge of the firmware tree got a few conflicts:
> 
> Commit 33f1dd02ed25db0790fec8a62b008c71831c49d9 ("ti_usb: kick firmware
> into user space") from the usb tree against commit
> e009cdbc2e96e72fe84df7cf014abcd7a7eb7dc7 ("ti_usb_3410_5052: use
> request_firmware()") from the firmware tree.  These seem to have similar
> goals but are using different methods.  This affected the files
> drivers/usb/serial/Kconfig,
> drivers/usb/serial/{ti_fw_3410.h,ti_fw_5052.h,ti_usb_3410_5052.c}.  I
> have used the firmware tree version for now.

I think that's the right choice; the main difference is the fact that my
version rips out the built-in firmware capability completely, and uses
the firmware loader unconditionally -- which is what we want. Alan?

http://git.infradead.org/users/dwmw2/firmware-2.6.git?a=commitdiff;h=7eb1c508575d0e1c98e8560f8215d35b10e6cd7a#patch4

> Some simple conflicts in sound/Kconfig due to commit
> bd7d34e759d7537b42cffb7205b7b50dda6535bb ("sound: Convert to menuconfig")
> from the sound tree.
> 
> Simple conflicts in sound/pci/maestro3.c and
> sound/pci/ymfpci/ymfpci_main.c because "maestro3: treat firmware data as
> const" and "ymfpci: treat firmware data as const" appear in both the
> driver-core tree and the firmware tree.  The conflicts are because the
> firmware tree further modifies these files.
>
> So, you and Alan need to come to an agreement about ti_usb (and inform
> Greg :-)).  And you should probably ask Greg to drop the to "const"
> patches (or maybe the other way around). 

I don't mind much. I was kind of hoping we might sneak the 'const' thing
into 2.6.26 since it's fairly straightforward and should help to stop
people adding more drivers which don't treat fw->data as const.
But then again, that's what linux-next is for too, isn't it?

Greg, want to either push the const changes to Linus or drop them? I
think I've picked up your extra ones, including the iSight one now that
that driver is upstream.

-- 
dwmw2

Re: linux-next: firmware merge conflicts

[David Woodhouse]

On Wed, 2008-06-04 at 11:27 +0200, Takashi Iwai wrote:
> Or just let me know if you guys want to put the changes into sound
> tree.  I'm willing to apply them...

Probably best to handle all the "const" changes together.

> >  And I can carry the sound/Kconfig fixup.

That conflict is just selecting FW_LOADER unconditionally, and changing
the default of SND_*_FIRMWARE_IN_KERNEL to 'n'. The latter you could do
in the sound tree quite happily. The former too, perhaps -- but then
there's Adrian's patch to make FW_LOADER unconditional if !EMBEDDED
anyway, which I think is probably a good idea.

-- 
dwmw2

Re: linux-next: firmware merge conflicts

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1085 bytes --]

Hi Takashi,

On Wed, 04 Jun 2008 11:27:52 +0200 Takashi Iwai <tiwai@suse.de> wrote:
>
> At Wed, 4 Jun 2008 16:07:42 +1000,
> Stephen Rothwell wrote:
> > 
> > Simple conflicts in sound/pci/maestro3.c and
> > sound/pci/ymfpci/ymfpci_main.c because "maestro3: treat firmware data as
> > const" and "ymfpci: treat firmware data as const" appear in both the
> > driver-core tree and the firmware tree.  The conflicts are because the
> > firmware tree further modifies these files.
> > 
> > So, you and Alan need to come to an agreement about ti_usb (and inform
> > Greg :-)).  And you should probably ask Greg to drop the to "const"
> > patches (or maybe the other way around).
> 
> Or just let me know if you guys want to put the changes into sound
> tree.  I'm willing to apply them...

That would also work.

> >  And I can carry the
> > sound/Kconfig fixup.
> 
> ... together with this :)

I guess that conflict would go away if they did the above.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: linux-next: firmware merge conflicts

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 834 bytes --]

On Wed, 04 Jun 2008 10:56:29 +0100 David Woodhouse <dwmw2@infradead.org> wrote:
>
> > >  And I can carry the sound/Kconfig fixup.
> 
> That conflict is just selecting FW_LOADER unconditionally, and changing
> the default of SND_*_FIRMWARE_IN_KERNEL to 'n'. The latter you could do
> in the sound tree quite happily. The former too, perhaps -- but then
> there's Adrian's patch to make FW_LOADER unconditional if !EMBEDDED
> anyway, which I think is probably a good idea.

The conflict is actually caused because all the sections lose a "depends
on SND" (realtive to the firmware tree) which can only go away when the
two trees get merged i.e. in my tree.  I have no problem carrying these
trivial merge fixups.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]