Re: [PATCH 05/31] Constify struct dma_map_ops for 2.6.32 v1

[Emese Revfy <re.emese@gmail.com>]

Stephen Rothwell wrote:
> Hi Emese,
> 
> On Sat, 05 Dec 2009 03:00:25 +0100 Emese Revfy <re.emese@gmail.com> wrote:
>> Ok, I made the changes, I also removed the vio_cmo_set_dma_ops function
>> because I think it is now unnecessary.
> 
> It is necessary because vio_dma_mapping_ops is only defined when
> CONFIG_PPC_SMLPAR is set, but that function was called from unconditional
> code (and there was a trivial version for when CONFIG_PPC_SMLPAR was not
> set) i.e. we only want the dma_ops set if we were built with
> CONFIG_PPC_SMLPAR.

I already took care of these, look at these hunks:

@@ -1033,7 +1026,6 @@ int vio_cmo_entitlement_update(size_t new_entitlement) { return 0; }
 void vio_cmo_set_dev_desired(struct vio_dev *viodev, size_t desired) {}
 static int vio_cmo_bus_probe(struct vio_dev *viodev) { return 0; }
 static void vio_cmo_bus_remove(struct vio_dev *viodev) {}
-static void vio_cmo_set_dma_ops(struct vio_dev *viodev) {}
 static void vio_cmo_bus_init(void) {}
 static void vio_cmo_sysfs_init(void) { }
 #endif /* CONFIG_PPC_SMLPAR */
@@ -1233,7 +1225,7 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node)
 	viodev->dev.archdata.of_node = of_node_get(of_node);
 
 	if (firmware_has_feature(FW_FEATURE_CMO))
-		vio_cmo_set_dma_ops(viodev);
+		viodev->dev.archdata.dma_ops = &vio_dma_mapping_ops;
 	else
 		viodev->dev.archdata.dma_ops = &dma_iommu_ops;
 	set_iommu_table_base(&viodev->dev, vio_build_iommu_table(viodev));


If you think it's not enough please let me know.
--
Emese