Nouveau - Oct 2020 - [PATCH] device: return error for unknown chipsets

Previously the code relied on device->pri to be NULL and to fail probing
later. We really should just return an error inside nvkm_device_ctor for
unsupported GPUs.

Fixes: 24d5ff40a732 ("drm/nouveau/device: rework mmio mapping code to get
rid of second map")

Signed-off-by: Karol Herbst <kherbst at redhat.com>
Cc: dann frazier <dann.frazier at canonical.com>
---
 drm/nouveau/nvkm/engine/device/base.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drm/nouveau/nvkm/engine/device/base.c
b/drm/nouveau/nvkm/engine/device/base.c
index 9f4ac2672..68d665414 100644
--- a/drm/nouveau/nvkm/engine/device/base.c
+++ b/drm/nouveau/nvkm/engine/device/base.c
@@ -3149,6 +3149,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
 		case 0x168: device->chip = &nv168_chipset; break;
 		default:
 			nvdev_error(device, "unknown chipset (%08x)\n", boot0);
+			ret = -ENOSYS;
 			goto done;
 		}
 
-- 
2.26.2

On Mon, Oct 05, 2020 at 03:52:30PM +0200, Karol Herbst
wrote:
> Previously the code relied on device->pri to be NULL and to fail probing
> later. We really should just return an error inside nvkm_device_ctor for
> unsupported GPUs.
> 
> Fixes: 24d5ff40a732 ("drm/nouveau/device: rework mmio mapping code to
get rid of second map")
> Signed-off-by: Karol Herbst <kherbst at redhat.com>
> Cc: dann frazier <dann.frazier at canonical.com>
Thanks Karol! Got rid of the Oops for me.

Tested-by: dann frazier <dann.frazier at canonical.com>

Would it be appropriate to tag this for stable v5.8+?

  -dann
> ---
>  drm/nouveau/nvkm/engine/device/base.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drm/nouveau/nvkm/engine/device/base.c
b/drm/nouveau/nvkm/engine/device/base.c
> index 9f4ac2672..68d665414 100644
> --- a/drm/nouveau/nvkm/engine/device/base.c
> +++ b/drm/nouveau/nvkm/engine/device/base.c
> @@ -3149,6 +3149,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func,
>  		case 0x168: device->chip = &nv168_chipset; break;
>  		default:
>  			nvdev_error(device, "unknown chipset (%08x)\n", boot0);
> +			ret = -ENOSYS;
>  			goto done;
>  		}
>

On Tue, Oct 6, 2020 at 6:43 PM dann frazier <dann.frazier at
canonical.com> wrote:
>
> On Mon, Oct 05, 2020 at 03:52:30PM +0200, Karol Herbst wrote:
> > Previously the code relied on device->pri to be NULL and to fail
probing
> > later. We really should just return an error inside nvkm_device_ctor
for
> > unsupported GPUs.
> >
> > Fixes: 24d5ff40a732 ("drm/nouveau/device: rework mmio mapping
code to get rid of second map")
> > Signed-off-by: Karol Herbst <kherbst at redhat.com>
> > Cc: dann frazier <dann.frazier at canonical.com>
>
> Thanks Karol! Got rid of the Oops for me.
>
> Tested-by: dann frazier <dann.frazier at canonical.com>
>
> Would it be appropriate to tag this for stable v5.8+?
>
Due to the Fixes: tags it will get auto selected for stable releases
afaik. But I think an explicit send might speed it up? I think it's
fine either way.
>   -dann
>
> > ---
> >  drm/nouveau/nvkm/engine/device/base.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drm/nouveau/nvkm/engine/device/base.c
b/drm/nouveau/nvkm/engine/device/base.c
> > index 9f4ac2672..68d665414 100644
> > --- a/drm/nouveau/nvkm/engine/device/base.c
> > +++ b/drm/nouveau/nvkm/engine/device/base.c
> > @@ -3149,6 +3149,7 @@ nvkm_device_ctor(const struct nvkm_device_func
*func,
> >               case 0x168: device->chip = &nv168_chipset; break;
> >               default:
> >                       nvdev_error(device, "unknown chipset
(%08x)\n", boot0);
> > +                     ret = -ENOSYS;
> >                       goto done;
> >               }
> >
>