2021-08-25 18:28:12

by Robin Murphy

[permalink] [raw]
Subject: [PATCH] drm/cma-helper: Set VM_DONTEXPAND for mmap

drm_gem_cma_mmap() cannot assume every implementation of dma_mmap_wc()
will end up calling remap_pfn_range() (which happens to set the relevant
vma flag, among others), so in order to make sure expectations around
VM_DONTEXPAND are met, let it explicitly set the flag like most other
GEM mmap implementations do.

This avoids repeated warnings on a small minority of systems where the
display is behind an IOMMU, and has a simple driver which does not
override drm_gem_cma_default_funcs.

Signed-off-by: Robin Murphy <[email protected]>
---
drivers/gpu/drm/drm_gem_cma_helper.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
index d53388199f34..63e48d98263d 100644
--- a/drivers/gpu/drm/drm_gem_cma_helper.c
+++ b/drivers/gpu/drm/drm_gem_cma_helper.c
@@ -510,6 +510,7 @@ int drm_gem_cma_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
*/
vma->vm_pgoff -= drm_vma_node_start(&obj->vma_node);
vma->vm_flags &= ~VM_PFNMAP;
+ vma->vm_flags |= VM_DONTEXPAND;

cma_obj = to_drm_gem_cma_obj(obj);

--
2.25.1


2021-08-26 01:17:18

by Alyssa Rosenzweig

[permalink] [raw]
Subject: Re: [PATCH] drm/cma-helper: Set VM_DONTEXPAND for mmap

> drm_gem_cma_mmap() cannot assume every implementation of dma_mmap_wc()
> will end up calling remap_pfn_range() (which happens to set the relevant
> vma flag, among others), so in order to make sure expectations around
> VM_DONTEXPAND are met, let it explicitly set the flag like most other
> GEM mmap implementations do.
>
> This avoids repeated warnings on a small minority of systems where the
> display is behind an IOMMU, and has a simple driver which does not
> override drm_gem_cma_default_funcs.

Apple system-on-chips have their display behind an IOMMU. Actually, a
separate IOMMU for each display, and a separate IOMMU for each display
controller -- so there are 4 IOMMUs total for display on the M1.

I've tested this patch against my work-in-progress display driver for
the M1. It indeed fixes the annoying warnings every frame (wayland) and
on mode setting (x11). So this is

Tested-by: Alyssa Rosenzweig <[email protected]>

I've cherry-picked the patch into my M1 staging/downstream tree, so I
guess that's an Acked-by. I don't know anything about the vm_* stuff in
the kernel yet, though, since can't give a reviewed-by. Will leave that
one to the pro's.

I know you were trying to fix an HDLCD issue, but I needed this patch
too, so thank you! ????