linux-next: manual merge of the bkl-llseek tree with the rr tree

linux-next: manual merge of the bkl-llseek tree with the rr tree

[Stephen Rothwell]

Hi Arnd,

Today's linux-next merge of the bkl-llseek tree got a conflict in
drivers/char/virtio_console.c between commit
a0e1acd65fa6c2da3b039a67ac30cc3b73466c37 ("virtio: console: Send SIGIO to
processes that request it for host events") from the rr tree and commit
7f09912dd2169ca9ce56c5e3428c7b4c5e92e727 ("llseek: automatically
add .llseek fop") from the bkl-llseek tree.

Just context changes.  I fixed it up (see below) and can carry the fix as
necessary.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc drivers/char/virtio_console.c
index f7adfd3,524907b..0000000
--- a/drivers/char/virtio_console.c
+++ b/drivers/char/virtio_console.c
@@@ -824,7 -728,7 +824,8 @@@ static const struct file_operations por
  	.write = port_fops_write,
  	.poll  = port_fops_poll,
  	.release = port_fops_release,
 +	.fasync = port_fops_fasync,
+ 	.llseek = noop_llseek,/* read and write both use no f_pos */
  };
  
  /*

Re: linux-next: manual merge of the bkl-llseek tree with the rr tree

[Amit Shah]

On (Wed) Sep 15 2010 [13:26:46], Stephen Rothwell wrote:
> Hi Arnd,
> 
> Today's linux-next merge of the bkl-llseek tree got a conflict in
> drivers/char/virtio_console.c between commit
> a0e1acd65fa6c2da3b039a67ac30cc3b73466c37 ("virtio: console: Send SIGIO to
> processes that request it for host events") from the rr tree and commit
> 7f09912dd2169ca9ce56c5e3428c7b4c5e92e727 ("llseek: automatically
> add .llseek fop") from the bkl-llseek tree.
> 
> Just context changes.  I fixed it up (see below) and can carry the fix as
> necessary.

Thanks, this looks correct.

Arnd, the device is supposed to be non-seekable so I'll add a
nonseekable_open() to the open() call.

So I guess the llseek operation should ne no_llseek instead of
noop_llseek.  Will you change that in your patchset?  Should I do that
in the patch I'll queue up?

> diff --cc drivers/char/virtio_console.c
> index f7adfd3,524907b..0000000
> --- a/drivers/char/virtio_console.c
> +++ b/drivers/char/virtio_console.c
> @@@ -824,7 -728,7 +824,8 @@@ static const struct file_operations por
>   	.write = port_fops_write,
>   	.poll  = port_fops_poll,
>   	.release = port_fops_release,
>  +	.fasync = port_fops_fasync,
> + 	.llseek = noop_llseek,/* read and write both use no f_pos */
>   };
>   
>   /*

		Amit

Re: linux-next: manual merge of the bkl-llseek tree with the rr tree

[Arnd Bergmann]

On Thursday 16 September 2010 08:54:39 Amit Shah wrote:
> Arnd, the device is supposed to be non-seekable so I'll add a
> nonseekable_open() to the open() call.
> 
> So I guess the llseek operation should ne no_llseek instead of
> noop_llseek.  Will you change that in your patchset?  Should I do that
> in the patch I'll queue up?

Yes, I think it's best if you just do both changes in your patch, I'll
drop this file from my series then.

Any driver that we can make use no_llseek instead of noop_llseek
is a step forward.

	Arnd

Re: linux-next: manual merge of the bkl-llseek tree with the rr tree

[Amit Shah]

On (Thu) Sep 16 2010 [10:20:01], Arnd Bergmann wrote:
> On Thursday 16 September 2010 08:54:39 Amit Shah wrote:
> > Arnd, the device is supposed to be non-seekable so I'll add a
> > nonseekable_open() to the open() call.
> > 
> > So I guess the llseek operation should ne no_llseek instead of
> > noop_llseek.  Will you change that in your patchset?  Should I do that
> > in the patch I'll queue up?
> 
> Yes, I think it's best if you just do both changes in your patch, I'll
> drop this file from my series then.

Great, I'll queue that up.

		Amit

Re: linux-next: manual merge of the bkl-llseek tree with the rr tree

[Arnd Bergmann]

On Thursday 16 September 2010 10:56:08 Amit Shah wrote:
> On (Thu) Sep 16 2010 [10:20:01], Arnd Bergmann wrote:
> > On Thursday 16 September 2010 08:54:39 Amit Shah wrote:
> > > Arnd, the device is supposed to be non-seekable so I'll add a
> > > nonseekable_open() to the open() call.
> > > 
> > > So I guess the llseek operation should ne no_llseek instead of
> > > noop_llseek.  Will you change that in your patchset?  Should I do that
> > > in the patch I'll queue up?
> > 
> > Yes, I think it's best if you just do both changes in your patch, I'll
> > drop this file from my series then.
> 
> Great, I'll queue that up.

I don't see this change in linux-next yet. What happened?

	Arnd