[PATCH] omap: dsp: fix for ioremap()

[PATCH] omap: dsp: fix for ioremap()

[Felipe Contreras]

Since ioremap() on RAM will be soon disallowed, this is required.

This also requires a patch like:
http://article.gmane.org/gmane.linux.kernel/1047146

Felipe Contreras (1):
  omap: dsp: remove shm from normal memory

 arch/arm/plat-omap/devices.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

-- 
1.7.3.1.2.g7fe2b

[PATCH] omap: dsp: remove shm from normal memory

[Felipe Contreras]

Also, don't be picky about the location.

Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
---
 arch/arm/plat-omap/devices.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/arch/arm/plat-omap/devices.c b/arch/arm/plat-omap/devices.c
index 4c8f9b9..da10433 100644
--- a/arch/arm/plat-omap/devices.c
+++ b/arch/arm/plat-omap/devices.c
@@ -285,12 +285,14 @@ void __init omap_dsp_reserve_sdram_memblock(void)
 	if (!size)
 		return;
 
-	paddr = __memblock_alloc_base(size, SZ_1M, MEMBLOCK_REAL_LIMIT);
+	paddr = memblock_alloc(size, SZ_1M);
 	if (!paddr) {
 		pr_err("%s: failed to reserve %x bytes\n",
 				__func__, size);
 		return;
 	}
+	memblock_free(paddr, size);
+	memblock_remove(paddr, size);
 
 	omap_dsp_phys_mempool_base = paddr;
 }
-- 
1.7.3.1.2.g7fe2b

Re: [PATCH] omap: dsp: remove shm from normal memory

[Tony Lindgren]

* Felipe Contreras <felipe.contreras@gmail.com> [101019 00:28]:
> Also, don't be picky about the location.

As this code is in CONFIG_TIDSPBRIDGE block and AFAIK that is
not working, this should probably wait for the next merge window?

Regards,

Tony

 
> Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
> ---
>  arch/arm/plat-omap/devices.c |    4 +++-
>  1 files changed, 3 insertions(+), 1 deletions(-)
> 
> diff --git a/arch/arm/plat-omap/devices.c b/arch/arm/plat-omap/devices.c
> index 4c8f9b9..da10433 100644
> --- a/arch/arm/plat-omap/devices.c
> +++ b/arch/arm/plat-omap/devices.c
> @@ -285,12 +285,14 @@ void __init omap_dsp_reserve_sdram_memblock(void)
>  	if (!size)
>  		return;
>  
> -	paddr = __memblock_alloc_base(size, SZ_1M, MEMBLOCK_REAL_LIMIT);
> +	paddr = memblock_alloc(size, SZ_1M);
>  	if (!paddr) {
>  		pr_err("%s: failed to reserve %x bytes\n",
>  				__func__, size);
>  		return;
>  	}
> +	memblock_free(paddr, size);
> +	memblock_remove(paddr, size);
>  
>  	omap_dsp_phys_mempool_base = paddr;
>  }
> -- 
> 1.7.3.1.2.g7fe2b
> 

Re: [PATCH] omap: dsp: remove shm from normal memory

[Felipe Contreras]

On Fri, Nov 5, 2010 at 6:34 PM, Tony Lindgren <tony@atomide.com> wrote:
> * Felipe Contreras <felipe.contreras@gmail.com> [101019 00:28]:
>> Also, don't be picky about the location.
>
> As this code is in CONFIG_TIDSPBRIDGE block and AFAIK that is
> not working, this should probably wait for the next merge window?

I plan to send a patch to revert the iommu changes to staging so that
it does work... But this patch would be needed as well.

-- 
Felipe Contreras

Re: [PATCH] omap: dsp: remove shm from normal memory

[Tony Lindgren]

* Felipe Contreras <felipe.contreras@gmail.com> [101105 10:44]:
> On Fri, Nov 5, 2010 at 6:34 PM, Tony Lindgren <tony@atomide.com> wrote:
> > * Felipe Contreras <felipe.contreras@gmail.com> [101019 00:28]:
> >> Also, don't be picky about the location.
> >
> > As this code is in CONFIG_TIDSPBRIDGE block and AFAIK that is
> > not working, this should probably wait for the next merge window?
> 
> I plan to send a patch to revert the iommu changes to staging so that
> it does work... But this patch would be needed as well.

OK. Let's wait and see what your minimal set of fixes to make it
work looks.

Regards,

Tony

Re: [PATCH] omap: dsp: remove shm from normal memory

[Felipe Contreras]

On Fri, Nov 5, 2010 at 8:16 PM, Tony Lindgren <tony@atomide.com> wrote:
> * Felipe Contreras <felipe.contreras@gmail.com> [101105 10:44]:
>> On Fri, Nov 5, 2010 at 6:34 PM, Tony Lindgren <tony@atomide.com> wrote:
>> > * Felipe Contreras <felipe.contreras@gmail.com> [101019 00:28]:
>> >> Also, don't be picky about the location.
>> >
>> > As this code is in CONFIG_TIDSPBRIDGE block and AFAIK that is
>> > not working, this should probably wait for the next merge window?
>>
>> I plan to send a patch to revert the iommu changes to staging so that
>> it does work... But this patch would be needed as well.
>
> OK. Let's wait and see what your minimal set of fixes to make it
> work looks.

Huh? Reverting the iommu changes will affect only tidspbridge (staging
tree) the omap codes is not affected at all. I say this patch needs to
be merged regardless.

Anyway, I'm about to send a pull request.

-- 
Felipe Contreras

Re: [PATCH] omap: dsp: remove shm from normal memory

[Ramirez Luna, Omar]

Hi Tony,

On Fri, Nov 5, 2010 at 12:53 PM, Felipe Contreras
<felipe.contreras@gmail.com> wrote:
> On Fri, Nov 5, 2010 at 6:34 PM, Tony Lindgren <tony@atomide.com> wrote:
>> * Felipe Contreras <felipe.contreras@gmail.com> [101019 00:28]:
>>> Also, don't be picky about the location.
>>
>> As this code is in CONFIG_TIDSPBRIDGE block and AFAIK that is
>> not working, this should probably wait for the next merge window?
>
> I plan to send a patch to revert the iommu changes to staging so that
> it does work... But this patch would be needed as well.

Would it be possible to include this series too (if nothing is
missing)? The only open question was whether to move iva2dsp.h
contents into dsp.h

tidspbridge: SCM layer violation fixes [1]

Patch 4 is to be sent through staging path, once these are pushed.

Regards,

Omar

---
[1] http://marc.info/?l=linux-omap&m=128871356703161&w=2

Re: [PATCH] omap: dsp: fix for ioremap()

[Felipe Balbi]

On Tue, Oct 19, 2010 at 02:37:23AM -0500, Felipe Contreras wrote:
>Since ioremap() on RAM will be soon disallowed, this is required.
>
>This also requires a patch like:
>http://article.gmane.org/gmane.linux.kernel/1047146
>
>Felipe Contreras (1):
>  omap: dsp: remove shm from normal memory
>
> arch/arm/plat-omap/devices.c |    4 +++-
> 1 files changed, 3 insertions(+), 1 deletions(-)

-ENOPATCH ??

-- 
balbi

Re: [PATCH] omap: dsp: fix for ioremap()

[Felipe Balbi]

On Tue, Oct 19, 2010 at 02:38:34AM -0500, Balbi, Felipe wrote:
>On Tue, Oct 19, 2010 at 02:37:23AM -0500, Felipe Contreras wrote:
>>Since ioremap() on RAM will be soon disallowed, this is required.
>>
>>This also requires a patch like:
>>http://article.gmane.org/gmane.linux.kernel/1047146
>>
>>Felipe Contreras (1):
>>  omap: dsp: remove shm from normal memory
>>
>> arch/arm/plat-omap/devices.c |    4 +++-
>> 1 files changed, 3 insertions(+), 1 deletions(-)
>
>-ENOPATCH ??

ok, now I got the patch. Just $SUBJECT is a bit misleading, sorry

-- 
balbi

Re: [PATCH] omap: dsp: fix for ioremap()

[Felipe Contreras]

On Tue, Oct 19, 2010 at 10:39 AM, Felipe Balbi <balbi@ti.com> wrote:
> On Tue, Oct 19, 2010 at 02:38:34AM -0500, Balbi, Felipe wrote:
>>
>> On Tue, Oct 19, 2010 at 02:37:23AM -0500, Felipe Contreras wrote:
>>>
>>> Since ioremap() on RAM will be soon disallowed, this is required.
>>>
>>> This also requires a patch like:
>>> http://article.gmane.org/gmane.linux.kernel/1047146
>>>
>>> Felipe Contreras (1):
>>>  omap: dsp: remove shm from normal memory
>>>
>>> arch/arm/plat-omap/devices.c |    4 +++-
>>> 1 files changed, 3 insertions(+), 1 deletions(-)
>>
>> -ENOPATCH ??
>
> ok, now I got the patch. Just $SUBJECT is a bit misleading, sorry

:)

I thought on just sending the patch, but the last time there wasn't
enough explanation that it depended on another one.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html