Re: [PATCH 1/4] AMD IOMMU: use status bit instead of memory write-back for completion wait

[Ingo Oeser]

Hi Joerg,

On Thursday 14 August 2008, Joerg Roedel wrote:
> diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c
> index 22d7d05..028e945 100644
> --- a/arch/x86/kernel/amd_iommu.c
> +++ b/arch/x86/kernel/amd_iommu.c
> @@ -122,9 +119,15 @@ static int iommu_completion_wait(struct amd_iommu *iommu)
>  
>  	while (!ready && (i < EXIT_LOOP_COUNT)) {
>  		++i;
> -		cpu_relax();

Could you elaborate, why you had to remove this?

Busy waiting loops should always do cpu_relay()
to reduce heat procution and power consumption while idle.

> +		/* wait for the bit to become one */
> +		status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
> +		ready = status & MMIO_STATUS_COM_WAIT_INT_MASK;
>  	}
>  

Best Regards

Ingo Oeser
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/