On Fri, May 25, 2007 at 12:48:52AM +0900, tak...@op... wrote:
> BTW, Which tree should we check hard during rc time?
> 1. sh-2.6
> 2. sh-2.6.x
> 3. archived -rc tree
> 4. linus's tree
>
> This time I checked 1,2 and 3.
> Presicely, I found in 1, then check 2 if it is solved, and confirmd in 3.
> Are there any extra work to do(perhaps 4)? (or easyer way?)
>
The general scheme is that sh-2.6.git is only a temporary staging tree
where Linus pulls for the next -rc. I usually aim for about one pull for
every -rc, but occasionally one gets missed. After -rc2, it's simply bug
fixes. It's good to test on as many configurations as possible simply to
make sure that the -final ends up working, too.
Once -rc2 rolls around, I usually open up the next sh-2.6.x.git and start
queueing up the development work for the next merge window there. It is
helpful to have this tested and to have patches go in, but it's in a much
greater state of flux, so it really depends on what sort of testing and
bug fixing you're comfortable with. Once we're later in the -rc stages,
sh-2.6.x.git should settle down in preparation for the merge window. You
may want to do testing before or after that point, it really depends what
sort of patches you have that you want to aim for the merge window.
You'll have to use your judgment on that.
In general, I would suggest just using Linus's tree. If there's a problem
there, it may have already been fixed in sh-2.6.git, in which case, it
will trickle down naturally within a few days (and you can simply pull
that down if it's already been resolved, too). In general, it's good to
know what sort of regressions we have to handle for whatever the next -rc
is.
> Add lost in_nmi definition to solve pcrel too far.
>
> Signed-off-by: Takashi YOSHII <tak...@op...>
>
> diff --git a/arch/sh/kernel/cpu/sh3/entry.S b/arch/sh/kernel/cpu/sh3/entry.S
> index 832c0b4..659cc08 100644
> --- a/arch/sh/kernel/cpu/sh3/entry.S
> +++ b/arch/sh/kernel/cpu/sh3/entry.S
> @@ -320,6 +320,7 @@ skip_restore:
>
> .align 2
> 5: .long 0x00001000 ! DSP
> +6: .long in_nmi
> 7: .long 0x30000000
>
> ! common exception handler
>
Thanks, I'll queue it up.
|
