[PATCH] drm/cma: correctly handle non-zero offset for mmap
Steven Price
Steven.Price at arm.com
Wed Sep 27 14:01:31 UTC 2017
From: Tu Vuong <tu.vuong at arm.com>
When a CMA GEM object is exported via DRM PRIME it should be possible
to mmap the object using an offset. However drm_gem_cma_mmap_obj always
zeroed vm_pgoff.
Fix this by moving the zeroing of vm_pgoff to drm_gem_cma_mmap (which
is only used for non-PRIME mmap) and correct the size parameter in the
call to dma_mmap_wc as the offset may not be non-zero.
Signed-off-by: Tu Vuong <tu.vuong at arm.com>
Signed-off-by: Steven Price <steven.price at arm.com>
Reviewed-by: Brian Starkey <brian.starkey at arm.com>
CC: Brian Starkey <brian.starkey at arm.com>
CC: Liviu Dudau <Liviu.Dudau at arm.com>
---
drivers/gpu/drm/drm_gem_cma_helper.c | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
index 373e33f22be4..25828b33c5be 100644
--- a/drivers/gpu/drm/drm_gem_cma_helper.c
+++ b/drivers/gpu/drm/drm_gem_cma_helper.c
@@ -276,15 +276,12 @@ static int drm_gem_cma_mmap_obj(struct drm_gem_cma_object *cma_obj,
int ret;
/*
- * Clear the VM_PFNMAP flag that was set by drm_gem_mmap(), and set the
- * vm_pgoff (used as a fake buffer offset by DRM) to 0 as we want to map
- * the whole buffer.
+ * Clear the VM_PFNMAP flag that was set by drm_gem_mmap()
*/
vma->vm_flags &= ~VM_PFNMAP;
- vma->vm_pgoff = 0;
ret = dma_mmap_wc(cma_obj->base.dev->dev, vma, cma_obj->vaddr,
- cma_obj->paddr, vma->vm_end - vma->vm_start);
+ cma_obj->paddr, cma_obj->base.size);
if (ret)
drm_gem_vm_close(vma);
@@ -322,6 +319,12 @@ int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
gem_obj = vma->vm_private_data;
cma_obj = to_drm_gem_cma_obj(gem_obj);
+ /*
+ * Set the vm_pgoff (used as a fake buffer offset by DRM) to 0 as we
+ * want to map the whole buffer.
+ */
+ vma->vm_pgoff = 0;
+
return drm_gem_cma_mmap_obj(cma_obj, vma);
}
EXPORT_SYMBOL_GPL(drm_gem_cma_mmap);
--
2.11.0
More information about the dri-devel
mailing list