Re: [PATCH v2 RESEND 1/2] arm, arm64: change_memory_common with numpages == 0 should be no-op.

[Will Deacon]

Hi Mika,

On Tue, Jan 26, 2016 at 04:59:52PM +0200, mika.penttila@xxxxxxxxxxxx wrote:
> From: Mika Penttilä <mika.penttila@xxxxxxxxxxxx>
> 
> This makes the caller set_memory_xx() consistent with x86.
> 
> arm64 part is rebased on 4.5.0-rc1 with Ard's patch
>  lkml.kernel.org/g/<1453125665-26627-1-git-send-email-ard.biesheuvel@xxxxxxxxxx>
> applied.
> 
> Signed-off-by: Mika Penttilä mika.penttila@xxxxxxxxxxxx
> Reviewed-by: Laura Abbott <labbott@xxxxxxxxxx>
> Acked-by: David Rientjes <rientjes@xxxxxxxxxx>
> 
> ---
>  arch/arm/mm/pageattr.c   | 3 +++
>  arch/arm64/mm/pageattr.c | 3 +++
>  2 files changed, 6 insertions(+)
> 
> diff --git a/arch/arm/mm/pageattr.c b/arch/arm/mm/pageattr.c
> index cf30daf..d19b1ad 100644
> --- a/arch/arm/mm/pageattr.c
> +++ b/arch/arm/mm/pageattr.c
> @@ -49,6 +49,9 @@ static int change_memory_common(unsigned long addr, int numpages,
>  		WARN_ON_ONCE(1);
>  	}
>  
> +	if (!numpages)
> +		return 0;
> +
>  	if (start < MODULES_VADDR || start >= MODULES_END)
>  		return -EINVAL;
>  
> diff --git a/arch/arm64/mm/pageattr.c b/arch/arm64/mm/pageattr.c
> index 1360a02..b582fc2 100644
> --- a/arch/arm64/mm/pageattr.c
> +++ b/arch/arm64/mm/pageattr.c
> @@ -53,6 +53,9 @@ static int change_memory_common(unsigned long addr, int numpages,
>  		WARN_ON_ONCE(1);
>  	}
>  
> +	if (!numpages)
> +		return 0;
> +

Thanks for this. I can reproduce the failure on my Juno board, so I'd
like to queue this for 4.5 since it fixes a real issue. I've taken the
liberty of rebasing the arm64 part to my fixes branch and writing a
commit message. Does the patch below look ok to you?

Will

--->8