[Nouveau] [PATCH v2] instmem/gk20a: set DMA mask early

[Alexandre Courbot]

DMA mask is typically set in nouveau_ttm_init(), but this function is
called late during initialization and GK20A's instmem will have called
DMA functions before this happens.

Having a wrongly set DMA mask can result in the use of unneeded bounce
buffers. Set it early to avoid this.

Signed-off-by: Alexandre Courbot <acourbot at nvidia.com>
---
Changes since v1:
- Set mask in tegra.c instead of instmem, as suggested by Ben.

 drm/nouveau/nvkm/engine/device/tegra.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drm/nouveau/nvkm/engine/device/tegra.c b/drm/nouveau/nvkm/engine/device/tegra.c
index 6d89416f0bc1..e1cd665aee2c 100644
--- a/drm/nouveau/nvkm/engine/device/tegra.c
+++ b/drm/nouveau/nvkm/engine/device/tegra.c
@@ -272,6 +272,15 @@ nvkm_device_tegra_new(const struct nvkm_device_tegra_func *func,
 	if (IS_ERR(tdev->clk_pwr))
 		return PTR_ERR(tdev->clk_pwr);
 
+	/**
+	 * The IOMMU bit defines the upper limit of the GPU-addressable space.
+	 * This will be refined in nouveau_ttm_init but we need to do it early
+	 * for instmem to behave properly
+	 */
+	ret = dma_set_mask(&pdev->dev, DMA_BIT_MASK(tdev->func->iommu_bit));
+	if (ret)
+		return ret;
+
 	nvkm_device_tegra_probe_iommu(tdev);
 
 	ret = nvkm_device_tegra_power_up(tdev);
-- 
2.7.1