[PATCHv3,2/2] stubs: Provide parallel_mm_init stub version

From: Miroslav Rezanina <mrezanin@redhat.com>

mips build fail with link error in case PARALLEL_CONFIG is disabled as
hw/mips/mips_jazz.c calls parallel_mm_init. Due to dependecies to content
of parallel.c we can't simply move it to hw/isa/isa-devices.c.

This patch adds stubs/parallel.c file that contains stub version of
parallel_mm_init. This ensure successful build with PARALLEL_CONFIG disabled.

Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
---
 stubs/Makefile.objs | 1 +
 stubs/parallel.c    | 8 ++++++++
 2 files changed, 9 insertions(+)
 create mode 100644 stubs/parallel.c

On 12/05/2015 08:22, mrezanin@redhat.com wrote:
> From: Miroslav Rezanina <mrezanin@redhat.com>
> 
> mips build fail with link error in case PARALLEL_CONFIG is disabled as
> hw/mips/mips_jazz.c calls parallel_mm_init. Due to dependecies to content
> of parallel.c we can't simply move it to hw/isa/isa-devices.c.
> 
> This patch adds stubs/parallel.c file that contains stub version of
> parallel_mm_init. This ensure successful build with PARALLEL_CONFIG disabled.
> 
> Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
> ---
>  stubs/Makefile.objs | 1 +
>  stubs/parallel.c    | 8 ++++++++
>  2 files changed, 9 insertions(+)
>  create mode 100644 stubs/parallel.c
> 
> diff --git a/stubs/Makefile.objs b/stubs/Makefile.objs
> index 8beff4c..ad4e110 100644
> --- a/stubs/Makefile.objs
> +++ b/stubs/Makefile.objs
> @@ -24,6 +24,7 @@ stub-obj-y += mon-printf.o
>  stub-obj-y += mon-set-error.o
>  stub-obj-y += monitor-init.o
>  stub-obj-y += notify-event.o
> +stub-obj-y += parallel.o
>  stub-obj-$(CONFIG_SPICE) += qemu-chr-open-spice.o
>  stub-obj-y += qtest.o
>  stub-obj-y += reset.o
> diff --git a/stubs/parallel.c b/stubs/parallel.c
> new file mode 100644
> index 0000000..8293d52
> --- /dev/null
> +++ b/stubs/parallel.c
> @@ -0,0 +1,8 @@
> +#include "hw/i386/pc.h"
> +
> +bool parallel_mm_init(MemoryRegion *address_space,
> +                      hwaddr base, int it_shift, qemu_irq irq,
> +                      CharDriverState *chr)
> +{
> +    return false;
> +}
> 

I think removing CONFIG_PARALLEL from a board that hardcodes its
presence makes little sense, so I would just drop this patch.

Paolo

Paolo Bonzini <pbonzini@redhat.com> writes:

> On 12/05/2015 08:22, mrezanin@redhat.com wrote:
>> From: Miroslav Rezanina <mrezanin@redhat.com>
>> 
>> mips build fail with link error in case PARALLEL_CONFIG is disabled as
>> hw/mips/mips_jazz.c calls parallel_mm_init. Due to dependecies to content
>> of parallel.c we can't simply move it to hw/isa/isa-devices.c.
>> 
>> This patch adds stubs/parallel.c file that contains stub version of
>> parallel_mm_init. This ensure successful build with PARALLEL_CONFIG disabled.
>> 
>> Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
>> ---
>>  stubs/Makefile.objs | 1 +
>>  stubs/parallel.c    | 8 ++++++++
>>  2 files changed, 9 insertions(+)
>>  create mode 100644 stubs/parallel.c
>> 
>> diff --git a/stubs/Makefile.objs b/stubs/Makefile.objs
>> index 8beff4c..ad4e110 100644
>> --- a/stubs/Makefile.objs
>> +++ b/stubs/Makefile.objs
>> @@ -24,6 +24,7 @@ stub-obj-y += mon-printf.o
>>  stub-obj-y += mon-set-error.o
>>  stub-obj-y += monitor-init.o
>>  stub-obj-y += notify-event.o
>> +stub-obj-y += parallel.o
>>  stub-obj-$(CONFIG_SPICE) += qemu-chr-open-spice.o
>>  stub-obj-y += qtest.o
>>  stub-obj-y += reset.o
>> diff --git a/stubs/parallel.c b/stubs/parallel.c
>> new file mode 100644
>> index 0000000..8293d52
>> --- /dev/null
>> +++ b/stubs/parallel.c
>> @@ -0,0 +1,8 @@
>> +#include "hw/i386/pc.h"
>> +
>> +bool parallel_mm_init(MemoryRegion *address_space,
>> +                      hwaddr base, int it_shift, qemu_irq irq,
>> +                      CharDriverState *chr)
>> +{
>> +    return false;
>> +}
>> 
>
> I think removing CONFIG_PARALLEL from a board that hardcodes its
> presence makes little sense, so I would just drop this patch.

I pointed Mirek to parallel_mm_init().  Second thoughts: since we don't
know omitting the device breaks guests, and aren't really interested in
finding out, let's leave things as they are, i.e. drop this patch.

On Wed, May 13, 2015 at 10:04:23AM +0200, Markus Armbruster wrote:
> Paolo Bonzini <pbonzini@redhat.com> writes:
> 
> > On 12/05/2015 08:22, mrezanin@redhat.com wrote:
> >> From: Miroslav Rezanina <mrezanin@redhat.com>
> >> 
> >> mips build fail with link error in case PARALLEL_CONFIG is disabled as
> >> hw/mips/mips_jazz.c calls parallel_mm_init. Due to dependecies to content
> >> of parallel.c we can't simply move it to hw/isa/isa-devices.c.
> >> 
> >> This patch adds stubs/parallel.c file that contains stub version of
> >> parallel_mm_init. This ensure successful build with PARALLEL_CONFIG disabled.
> >> 
> >> Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
> >> ---
> >>  stubs/Makefile.objs | 1 +
> >>  stubs/parallel.c    | 8 ++++++++
> >>  2 files changed, 9 insertions(+)
> >>  create mode 100644 stubs/parallel.c
> >> 
> >> diff --git a/stubs/Makefile.objs b/stubs/Makefile.objs
> >> index 8beff4c..ad4e110 100644
> >> --- a/stubs/Makefile.objs
> >> +++ b/stubs/Makefile.objs
> >> @@ -24,6 +24,7 @@ stub-obj-y += mon-printf.o
> >>  stub-obj-y += mon-set-error.o
> >>  stub-obj-y += monitor-init.o
> >>  stub-obj-y += notify-event.o
> >> +stub-obj-y += parallel.o
> >>  stub-obj-$(CONFIG_SPICE) += qemu-chr-open-spice.o
> >>  stub-obj-y += qtest.o
> >>  stub-obj-y += reset.o
> >> diff --git a/stubs/parallel.c b/stubs/parallel.c
> >> new file mode 100644
> >> index 0000000..8293d52
> >> --- /dev/null
> >> +++ b/stubs/parallel.c
> >> @@ -0,0 +1,8 @@
> >> +#include "hw/i386/pc.h"
> >> +
> >> +bool parallel_mm_init(MemoryRegion *address_space,
> >> +                      hwaddr base, int it_shift, qemu_irq irq,
> >> +                      CharDriverState *chr)
> >> +{
> >> +    return false;
> >> +}
> >> 
> >
> > I think removing CONFIG_PARALLEL from a board that hardcodes its
> > presence makes little sense, so I would just drop this patch.
> 
> I pointed Mirek to parallel_mm_init().  Second thoughts: since we don't
> know omitting the device breaks guests, and aren't really interested in
> finding out, let's leave things as they are, i.e. drop this patch.

Ok, v4 will be patch 1 only.

Mirek