* [PATCH] drm/nouveau: fix multihop when move doesn't work.
@ 2020-12-17 20:09 Dave Airlie
2020-12-17 20:29 ` Daniel Vetter
0 siblings, 1 reply; 3+ messages in thread
From: Dave Airlie @ 2020-12-17 20:09 UTC (permalink / raw)
To: dri-devel
From: Dave Airlie <airlied@redhat.com>
As per the radeon/amdgpu fix don't use multihop is hw moves
aren't enabled.
Reported-by: Mike Galbraith <efault@gmx.de>
Tested-by: Mike Galbraith <efault@gmx.de>
Fixes: 0c8c0659d74 ("drm/nouveau/ttm: use multihop")
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
drivers/gpu/drm/nouveau/nouveau_bo.c | 31 ++++++++++++++--------------
1 file changed, 16 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 1386b0fc1640..c85b1af06b7b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -942,16 +942,6 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
struct nouveau_drm_tile *new_tile = NULL;
int ret = 0;
- if ((old_reg->mem_type == TTM_PL_SYSTEM &&
- new_reg->mem_type == TTM_PL_VRAM) ||
- (old_reg->mem_type == TTM_PL_VRAM &&
- new_reg->mem_type == TTM_PL_SYSTEM)) {
- hop->fpfn = 0;
- hop->lpfn = 0;
- hop->mem_type = TTM_PL_TT;
- hop->flags = 0;
- return -EMULTIHOP;
- }
if (new_reg->mem_type == TTM_PL_TT) {
ret = nouveau_ttm_tt_bind(bo->bdev, bo->ttm, new_reg);
@@ -995,14 +985,25 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
/* Hardware assisted copy. */
if (drm->ttm.move) {
+ if ((old_reg->mem_type == TTM_PL_SYSTEM &&
+ new_reg->mem_type == TTM_PL_VRAM) ||
+ (old_reg->mem_type == TTM_PL_VRAM &&
+ new_reg->mem_type == TTM_PL_SYSTEM)) {
+ hop->fpfn = 0;
+ hop->lpfn = 0;
+ hop->mem_type = TTM_PL_TT;
+ hop->flags = 0;
+ return -EMULTIHOP;
+ }
ret = nouveau_bo_move_m2mf(bo, evict, ctx,
new_reg);
- if (!ret)
- goto out;
- }
+ } else
+ ret = -ENODEV;
- /* Fallback to software copy. */
- ret = ttm_bo_move_memcpy(bo, ctx, new_reg);
+ if (ret) {
+ /* Fallback to software copy. */
+ ret = ttm_bo_move_memcpy(bo, ctx, new_reg);
+ }
out:
if (drm->client.device.info.family < NV_DEVICE_INFO_V0_TESLA) {
--
2.27.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH] drm/nouveau: fix multihop when move doesn't work.
2020-12-17 20:09 [PATCH] drm/nouveau: fix multihop when move doesn't work Dave Airlie
@ 2020-12-17 20:29 ` Daniel Vetter
2020-12-18 14:15 ` Daniel Vetter
0 siblings, 1 reply; 3+ messages in thread
From: Daniel Vetter @ 2020-12-17 20:29 UTC (permalink / raw)
To: Dave Airlie; +Cc: dri-devel
On Fri, Dec 18, 2020 at 06:09:43AM +1000, Dave Airlie wrote:
> From: Dave Airlie <airlied@redhat.com>
>
> As per the radeon/amdgpu fix don't use multihop is hw moves
> aren't enabled.
>
> Reported-by: Mike Galbraith <efault@gmx.de>
> Tested-by: Mike Galbraith <efault@gmx.de>
> Fixes: 0c8c0659d74 ("drm/nouveau/ttm: use multihop")
> Signed-off-by: Dave Airlie <airlied@redhat.com>
Pulled to drm-next directly, will go out to Linus tomorrow.
-Daniel
> ---
> drivers/gpu/drm/nouveau/nouveau_bo.c | 31 ++++++++++++++--------------
> 1 file changed, 16 insertions(+), 15 deletions(-)
>
> diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
> index 1386b0fc1640..c85b1af06b7b 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_bo.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
> @@ -942,16 +942,6 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
> struct nouveau_drm_tile *new_tile = NULL;
> int ret = 0;
>
> - if ((old_reg->mem_type == TTM_PL_SYSTEM &&
> - new_reg->mem_type == TTM_PL_VRAM) ||
> - (old_reg->mem_type == TTM_PL_VRAM &&
> - new_reg->mem_type == TTM_PL_SYSTEM)) {
> - hop->fpfn = 0;
> - hop->lpfn = 0;
> - hop->mem_type = TTM_PL_TT;
> - hop->flags = 0;
> - return -EMULTIHOP;
> - }
>
> if (new_reg->mem_type == TTM_PL_TT) {
> ret = nouveau_ttm_tt_bind(bo->bdev, bo->ttm, new_reg);
> @@ -995,14 +985,25 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
>
> /* Hardware assisted copy. */
> if (drm->ttm.move) {
> + if ((old_reg->mem_type == TTM_PL_SYSTEM &&
> + new_reg->mem_type == TTM_PL_VRAM) ||
> + (old_reg->mem_type == TTM_PL_VRAM &&
> + new_reg->mem_type == TTM_PL_SYSTEM)) {
> + hop->fpfn = 0;
> + hop->lpfn = 0;
> + hop->mem_type = TTM_PL_TT;
> + hop->flags = 0;
> + return -EMULTIHOP;
> + }
> ret = nouveau_bo_move_m2mf(bo, evict, ctx,
> new_reg);
> - if (!ret)
> - goto out;
> - }
> + } else
> + ret = -ENODEV;
>
> - /* Fallback to software copy. */
> - ret = ttm_bo_move_memcpy(bo, ctx, new_reg);
> + if (ret) {
> + /* Fallback to software copy. */
> + ret = ttm_bo_move_memcpy(bo, ctx, new_reg);
> + }
>
> out:
> if (drm->client.device.info.family < NV_DEVICE_INFO_V0_TESLA) {
> --
> 2.27.0
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH] drm/nouveau: fix multihop when move doesn't work.
2020-12-17 20:29 ` Daniel Vetter
@ 2020-12-18 14:15 ` Daniel Vetter
0 siblings, 0 replies; 3+ messages in thread
From: Daniel Vetter @ 2020-12-18 14:15 UTC (permalink / raw)
To: Dave Airlie; +Cc: dri-devel
On Thu, Dec 17, 2020 at 09:29:03PM +0100, Daniel Vetter wrote:
> On Fri, Dec 18, 2020 at 06:09:43AM +1000, Dave Airlie wrote:
> > From: Dave Airlie <airlied@redhat.com>
> >
> > As per the radeon/amdgpu fix don't use multihop is hw moves
> > aren't enabled.
> >
> > Reported-by: Mike Galbraith <efault@gmx.de>
> > Tested-by: Mike Galbraith <efault@gmx.de>
> > Fixes: 0c8c0659d74 ("drm/nouveau/ttm: use multihop")
Also fixed the Fixes: line here, it was pointing at something else (old
rebased version from your own tree maybe?).
-Daniel
> > Signed-off-by: Dave Airlie <airlied@redhat.com>
>
> Pulled to drm-next directly, will go out to Linus tomorrow.
> -Daniel
>
> > ---
> > drivers/gpu/drm/nouveau/nouveau_bo.c | 31 ++++++++++++++--------------
> > 1 file changed, 16 insertions(+), 15 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
> > index 1386b0fc1640..c85b1af06b7b 100644
> > --- a/drivers/gpu/drm/nouveau/nouveau_bo.c
> > +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
> > @@ -942,16 +942,6 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
> > struct nouveau_drm_tile *new_tile = NULL;
> > int ret = 0;
> >
> > - if ((old_reg->mem_type == TTM_PL_SYSTEM &&
> > - new_reg->mem_type == TTM_PL_VRAM) ||
> > - (old_reg->mem_type == TTM_PL_VRAM &&
> > - new_reg->mem_type == TTM_PL_SYSTEM)) {
> > - hop->fpfn = 0;
> > - hop->lpfn = 0;
> > - hop->mem_type = TTM_PL_TT;
> > - hop->flags = 0;
> > - return -EMULTIHOP;
> > - }
> >
> > if (new_reg->mem_type == TTM_PL_TT) {
> > ret = nouveau_ttm_tt_bind(bo->bdev, bo->ttm, new_reg);
> > @@ -995,14 +985,25 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict,
> >
> > /* Hardware assisted copy. */
> > if (drm->ttm.move) {
> > + if ((old_reg->mem_type == TTM_PL_SYSTEM &&
> > + new_reg->mem_type == TTM_PL_VRAM) ||
> > + (old_reg->mem_type == TTM_PL_VRAM &&
> > + new_reg->mem_type == TTM_PL_SYSTEM)) {
> > + hop->fpfn = 0;
> > + hop->lpfn = 0;
> > + hop->mem_type = TTM_PL_TT;
> > + hop->flags = 0;
> > + return -EMULTIHOP;
> > + }
> > ret = nouveau_bo_move_m2mf(bo, evict, ctx,
> > new_reg);
> > - if (!ret)
> > - goto out;
> > - }
> > + } else
> > + ret = -ENODEV;
> >
> > - /* Fallback to software copy. */
> > - ret = ttm_bo_move_memcpy(bo, ctx, new_reg);
> > + if (ret) {
> > + /* Fallback to software copy. */
> > + ret = ttm_bo_move_memcpy(bo, ctx, new_reg);
> > + }
> >
> > out:
> > if (drm->client.device.info.family < NV_DEVICE_INFO_V0_TESLA) {
> > --
> > 2.27.0
> >
> > _______________________________________________
> > dri-devel mailing list
> > dri-devel@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/dri-devel
>
> --
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-12-18 14:15 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-12-17 20:09 [PATCH] drm/nouveau: fix multihop when move doesn't work Dave Airlie
2020-12-17 20:29 ` Daniel Vetter
2020-12-18 14:15 ` Daniel Vetter
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.