[PATCH] enable/disable dom0_11_mapping from command-line and default to enabled

[PATCH] enable/disable dom0_11_mapping from command-line and default to enabled

[Karim Raslan]

Signed-off-by: Karim Raslan <karim.allah.ahmed@gmail.com>
---
 xen/arch/arm/domain_build.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index faff88e..2831ad0 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -23,6 +23,7 @@ static unsigned int __initdata opt_dom0_max_vcpus;
 integer_param("dom0_max_vcpus", opt_dom0_max_vcpus);
 
 static int dom0_11_mapping = 1;
+integer_param("dom0_11_mapping", dom0_11_mapping);
 
 #define DOM0_MEM_DEFAULT 0x8000000 /* 128 MiB */
 static u64 __initdata dom0_mem = DOM0_MEM_DEFAULT;
-- 
1.7.9.5

Re: [PATCH] enable/disable dom0_11_mapping from command-line and default to enabled

[Stefano Stabellini]

On Sun, 5 Jan 2014, Karim Raslan wrote:
> Signed-off-by: Karim Raslan <karim.allah.ahmed@gmail.com>

Thanks!
Given that dom0_11_mapping is an int, the patch is OK.
However it might be better to turn dom0_11_mapping into a bool_t and use
boolean_param instead.


>  xen/arch/arm/domain_build.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
> index faff88e..2831ad0 100644
> --- a/xen/arch/arm/domain_build.c
> +++ b/xen/arch/arm/domain_build.c
> @@ -23,6 +23,7 @@ static unsigned int __initdata opt_dom0_max_vcpus;
>  integer_param("dom0_max_vcpus", opt_dom0_max_vcpus);
>  
>  static int dom0_11_mapping = 1;
> +integer_param("dom0_11_mapping", dom0_11_mapping);
>  
>  #define DOM0_MEM_DEFAULT 0x8000000 /* 128 MiB */
>  static u64 __initdata dom0_mem = DOM0_MEM_DEFAULT;
> -- 
> 1.7.9.5
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
> 

Re: [PATCH] enable/disable dom0_11_mapping from command-line and default to enabled

[Julien Grall]

On 01/05/2014 07:24 PM, Stefano Stabellini wrote:
> On Sun, 5 Jan 2014, Karim Raslan wrote:
>> Signed-off-by: Karim Raslan <karim.allah.ahmed@gmail.com>
>
> Thanks!
> Given that dom0_11_mapping is an int, the patch is OK.
> However it might be better to turn dom0_11_mapping into a bool_t and use
> boolean_param instead.

After reading the thread "Master not working on AllWinner A20", I still 
don't see why this patch is useful.

Unless if the platform has an IOMMU, the 1:1 mapping for the memory 
SHOULD be enabled. Otherwise DMA-capable won't work.

See my answer on your mail (thread "Master not working on AllWinner for 
A20") for a proper fix.

-- 
Julien Grall

Re: [PATCH] enable/disable dom0_11_mapping from command-line and default to enabled

[Stefano Stabellini]

On Sun, 5 Jan 2014, Julien Grall wrote:
> On 01/05/2014 07:24 PM, Stefano Stabellini wrote:
> > On Sun, 5 Jan 2014, Karim Raslan wrote:
> > > Signed-off-by: Karim Raslan <karim.allah.ahmed@gmail.com>
> > 
> > Thanks!
> > Given that dom0_11_mapping is an int, the patch is OK.
> > However it might be better to turn dom0_11_mapping into a bool_t and use
> > boolean_param instead.
> 
> After reading the thread "Master not working on AllWinner A20", I still don't
> see why this patch is useful.
> 
> Unless if the platform has an IOMMU, the 1:1 mapping for the memory SHOULD be
> enabled. Otherwise DMA-capable won't work.
> 
> See my answer on your mail (thread "Master not working on AllWinner for A20")
> for a proper fix.

It is only useful for debugging and development. It should not be used
in production.

Re: [PATCH] enable/disable dom0_11_mapping from command-line and default to enabled

[Ian Campbell]

On Mon, 2014-01-06 at 11:12 +0000, Stefano Stabellini wrote:
> On Sun, 5 Jan 2014, Julien Grall wrote:
> > On 01/05/2014 07:24 PM, Stefano Stabellini wrote:
> > > On Sun, 5 Jan 2014, Karim Raslan wrote:
> > > > Signed-off-by: Karim Raslan <karim.allah.ahmed@gmail.com>
> > > 
> > > Thanks!
> > > Given that dom0_11_mapping is an int, the patch is OK.
> > > However it might be better to turn dom0_11_mapping into a bool_t and use
> > > boolean_param instead.
> > 
> > After reading the thread "Master not working on AllWinner A20", I still don't
> > see why this patch is useful.
> > 
> > Unless if the platform has an IOMMU, the 1:1 mapping for the memory SHOULD be
> > enabled. Otherwise DMA-capable won't work.
> > 
> > See my answer on your mail (thread "Master not working on AllWinner for A20")
> > for a proper fix.
> 
> It is only useful for debugging and development. It should not be used
> in production.

Assuming we are talking about the dom0_11_mapping command line parameter
then for debugging and development the constant in the code can simply
be changed, then there is no danger of an end user thinking it can be
useful.

That's essentially why I dropped the option from my initial series which
added the dom0_11_mapping variable.

Ian.