[PATCH] sh: Fix building j2_defconfig

[PATCH] sh: Fix building j2_defconfig

[Jason Cooper]

When targeting the j2, we need to retain '-m2'.  Previously, the
Makefile blew out -m2 on the next line via :=.

Fix this by s/:=/+=/ when building for the J2.

Fixes: 5a846abad07f6 ("sh: add support for J-Core J2 processor")
Signed-off-by: Jason Cooper <jason@lakedaemon.net>
---
Thanks to Richard Felker for helping me locate the problem.

 arch/sh/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/sh/Makefile b/arch/sh/Makefile
index 00476662ac2c..336f33a419d9 100644
--- a/arch/sh/Makefile
+++ b/arch/sh/Makefile
@@ -31,7 +31,7 @@ isa-y					:= $(isa-y)-up
 endif
 
 cflags-$(CONFIG_CPU_SH2)		:= $(call cc-option,-m2,)
-cflags-$(CONFIG_CPU_J2)			:= $(call cc-option,-mj2,)
+cflags-$(CONFIG_CPU_J2)			+= $(call cc-option,-mj2,)
 cflags-$(CONFIG_CPU_SH2A)		+= $(call cc-option,-m2a,) \
 					   $(call cc-option,-m2a-nofpu,) \
 					   $(call cc-option,-m4-nofpu,)
-- 
2.9.3

Re: [PATCH] sh: Fix building j2_defconfig

[Rob Landley]

On 08/16/2016 10:41 AM, Jason Cooper wrote:
> When targeting the j2, we need to retain '-m2'.  Previously, the
> Makefile blew out -m2 on the next line via :=.
> 
> Fix this by s/:=/+=/ when building for the J2.
>
> Fixes: 5a846abad07f6 ("sh: add support for J-Core J2 processor")
> Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Speaking of j2, any status on the missing pieces of infratsructure that
went in through other trees, without which booting hangs awaiting the
first interrupt?

  http://lists.j-core.org/pipermail/j-core/2016-August/000326.html

It would be nice if the rest of the board support could make it in this
release. Which trees are they going through?

Rob

Re: [PATCH] sh: Fix building j2_defconfig

[Jason Cooper]

Hi Rob,

On Tue, Aug 16, 2016 at 04:15:22PM -0500, Rob Landley wrote:
> On 08/16/2016 10:41 AM, Jason Cooper wrote:
> > When targeting the j2, we need to retain '-m2'.  Previously, the
> > Makefile blew out -m2 on the next line via :=.
> > 
> > Fix this by s/:=/+=/ when building for the J2.
> >
> > Fixes: 5a846abad07f6 ("sh: add support for J-Core J2 processor")
> > Signed-off-by: Jason Cooper <jason@lakedaemon.net>
> 
> Speaking of j2, any status on the missing pieces of infratsructure that
> went in through other trees, without which booting hangs awaiting the
> first interrupt?
> 
>   http://lists.j-core.org/pipermail/j-core/2016-August/000326.html
> 
> It would be nice if the rest of the board support could make it in this
> release. Which trees are they going through?

I'm not aware of the status of other bits, but the irqchip driver can be
found [1] in a stable, based off of v4.8-rc1, branch here:

  git://git.infradead.org/users/jcooper/linux.git irqchip/jcore

hth,

Jason.

[1] Now that I got the damn toolchain fixed and j2_defconfig building...

Re: [PATCH] sh: Fix building j2_defconfig

[Rob Landley]

On 08/16/2016 04:23 PM, Jason Cooper wrote:
> Hi Rob,
> 
> On Tue, Aug 16, 2016 at 04:15:22PM -0500, Rob Landley wrote:
>> On 08/16/2016 10:41 AM, Jason Cooper wrote:
>>> When targeting the j2, we need to retain '-m2'.  Previously, the
>>> Makefile blew out -m2 on the next line via :=.
>>>
>>> Fix this by s/:=/+=/ when building for the J2.
>>>
>>> Fixes: 5a846abad07f6 ("sh: add support for J-Core J2 processor")
>>> Signed-off-by: Jason Cooper <jason@lakedaemon.net>
>>
>> Speaking of j2, any status on the missing pieces of infratsructure that
>> went in through other trees, without which booting hangs awaiting the
>> first interrupt?
>>
>>   http://lists.j-core.org/pipermail/j-core/2016-August/000326.html
>>
>> It would be nice if the rest of the board support could make it in this
>> release. Which trees are they going through?
> 
> I'm not aware of the status of other bits, but the irqchip driver can be
> found [1] in a stable, based off of v4.8-rc1, branch here:
> 
>   git://git.infradead.org/users/jcooper/linux.git irqchip/jcore

That's got the interrupt controller, and presumably Thomas' tree has the
timer.

Is it likely to go upstream this dev cycle? Basic j2 board support did,
and as I said it hangs before userspace without the rest of the
interrupt controller and timer plumbing (which are currently only used
by this board).

The above message to the j-core list had an attached patch that adds the
missing bits to -rc2. I tested that patch and it worked for me:

Tested-by: Rob Landley <rob@landley.net>

I just checked the current git pull (not quite rc3) and vanilla is still
hanging at the same place, and the patch still applies cleanly. I'm
aware we're in bugfix-only mode, but "kernel hangs before launching
init" seems bug-ish to me.

Rob