Re: [PATCH 1/4] drm/radeon: handle PCIe root ports with addressing limitations

["Koenig, Christian" <Christian.Koenig@amd.com>]

Am 15.08.19 um 09:27 schrieb Christoph Hellwig:
> radeon uses a need_dma32 flag to indicate to the drm core that some
> allocations need to be done using GFP_DMA32, but it only checks the
> device addressing capabilities to make that decision.  Unfortunately
> PCIe root ports that have limited addressing exist as well.  Use the
> dma_addressing_limited instead to also take those into account.
>
> Reported-by: Atish Patra <Atish.Patra@wdc.com>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Looks sane to me. Reviewed-by: Christian König <christian.koenig@amd.com>.

Should we merge this through our normal amdgpu/radeon branches or do you 
want to send this upstream somehow else?

Thanks,
Christian.

> ---
>   drivers/gpu/drm/radeon/radeon.h        |  1 -
>   drivers/gpu/drm/radeon/radeon_device.c | 12 +++++-------
>   drivers/gpu/drm/radeon/radeon_ttm.c    |  2 +-
>   3 files changed, 6 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
> index 32808e50be12..1a0b22526a75 100644
> --- a/drivers/gpu/drm/radeon/radeon.h
> +++ b/drivers/gpu/drm/radeon/radeon.h
> @@ -2387,7 +2387,6 @@ struct radeon_device {
>   	struct radeon_wb		wb;
>   	struct radeon_dummy_page	dummy_page;
>   	bool				shutdown;
> -	bool				need_dma32;
>   	bool				need_swiotlb;
>   	bool				accel_working;
>   	bool				fastfb_working; /* IGP feature*/
> diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
> index dceb554e5674..b8cc05826667 100644
> --- a/drivers/gpu/drm/radeon/radeon_device.c
> +++ b/drivers/gpu/drm/radeon/radeon_device.c
> @@ -1365,27 +1365,25 @@ int radeon_device_init(struct radeon_device *rdev,
>   	else
>   		rdev->mc.mc_mask = 0xffffffffULL; /* 32 bit MC */
>   
> -	/* set DMA mask + need_dma32 flags.
> +	/* set DMA mask.
>   	 * PCIE - can handle 40-bits.
>   	 * IGP - can handle 40-bits
>   	 * AGP - generally dma32 is safest
>   	 * PCI - dma32 for legacy pci gart, 40 bits on newer asics
>   	 */
> -	rdev->need_dma32 = false;
> +	dma_bits = 40;
>   	if (rdev->flags & RADEON_IS_AGP)
> -		rdev->need_dma32 = true;
> +		dma_bits = 32;
>   	if ((rdev->flags & RADEON_IS_PCI) &&
>   	    (rdev->family <= CHIP_RS740))
> -		rdev->need_dma32 = true;
> +		dma_bits = 32;
>   #ifdef CONFIG_PPC64
>   	if (rdev->family == CHIP_CEDAR)
> -		rdev->need_dma32 = true;
> +		dma_bits = 32;
>   #endif
>   
> -	dma_bits = rdev->need_dma32 ? 32 : 40;
>   	r = pci_set_dma_mask(rdev->pdev, DMA_BIT_MASK(dma_bits));
>   	if (r) {
> -		rdev->need_dma32 = true;
>   		dma_bits = 32;
>   		pr_warn("radeon: No suitable DMA available\n");
>   	}
> diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
> index fb3696bc616d..116a27b25dc4 100644
> --- a/drivers/gpu/drm/radeon/radeon_ttm.c
> +++ b/drivers/gpu/drm/radeon/radeon_ttm.c
> @@ -794,7 +794,7 @@ int radeon_ttm_init(struct radeon_device *rdev)
>   	r = ttm_bo_device_init(&rdev->mman.bdev,
>   			       &radeon_bo_driver,
>   			       rdev->ddev->anon_inode->i_mapping,
> -			       rdev->need_dma32);
> +			       dma_addressing_limited(&rdev->pdev->dev));
>   	if (r) {
>   		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
>   		return r;