[PATCH v2] drm/gem: fix mmap vma size calculations

[Sedat Dilek]

On Fri, Jul 26, 2013 at 10:15 PM, Daniel Vetter <daniel at ffwll.ch> wrote:
> On Fri, Jul 26, 2013 at 12:09:32PM +0200, David Herrmann wrote:
>> The VMA manager is page-size based so drm_vma_node_size() returns the size
>> in pages. However, drm_gem_mmap_obj() requires the size in bytes. Apply
>> PAGE_SHIFT so we no longer get EINVAL during mmaps due to too small
>> buffers.
>>
>> This bug was introduced in commit:
>>   0de23977cfeb5b357ec884ba15417ae118ff9e9b
>>   "drm/gem: convert to new unified vma manager"
>>
>> Fixes i915 gtt mmap failure reported by Sedat Dilek in:
>>   Re: linux-next: Tree for Jul 25 [ call-trace: drm | drm-intel related? ]
>>
>> Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
>> Cc: Chris Wilson <chris at chris-wilson.co.uk>
>> Signed-off-by: David Herrmann <dh.herrmann at gmail.com>
>> Reported-by: Sedat Dilek <sedat.dilek at gmail.com>
>> Tested-by: Sedat Dilek <sedat.dilek at gmail.com>
>
> I remember that I even checked whether v4 was consistent with pages vs.
> bytes ;-) So
>

Hi David, Daniel, and Dave,

Just reading quickly "drm: add unified vma offset manager"... is that
below in the docs?

"The VMA manager is page-size based so drm_vma_node_size() returns the size
in pages. However, drm_gem_mmap_obj() requires the size in bytes. Apply
PAGE_SHIFT so we no longer get EINVAL during mmaps due to too small
buffers."

If not, do you mind adding it?

Thanks in advance!

- Sedat -

> Reviewed-by: Daniel Vetter <daniel.vetter at ffwll.ch>
>
> on this little fixup.
> -Daniel
>
>> ---
>>  drivers/gpu/drm/drm_gem.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
>> index 3613b50..1f76572 100644
>> --- a/drivers/gpu/drm/drm_gem.c
>> +++ b/drivers/gpu/drm/drm_gem.c
>> @@ -666,7 +666,7 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
>>       }
>>
>>       obj = container_of(node, struct drm_gem_object, vma_node);
>> -     ret = drm_gem_mmap_obj(obj, drm_vma_node_size(node), vma);
>> +     ret = drm_gem_mmap_obj(obj, drm_vma_node_size(node) << PAGE_SHIFT, vma);
>>
>>       mutex_unlock(&dev->struct_mutex);
>>
>> --
>> 1.8.3.4
>>
>
> --
> Daniel Vetter
> Software Engineer, Intel Corporation
> +41 (0) 79 365 57 48 - http://blog.ffwll.ch