dri-devel.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] drm/vmwgfx: Make sure we unpin no longer needed buffers
@ 2021-01-19 19:31 Zack Rusin
  2021-01-19 19:35 ` Daniel Vetter
  0 siblings, 1 reply; 3+ messages in thread
From: Zack Rusin @ 2021-01-19 19:31 UTC (permalink / raw)
  To: dri-devel; +Cc: Martin Krastev, Roland Scheidegger

We were not correctly unpinning no longer needed buffers. In particular
vmw_buffer_object, which is internally often pinned on creation wasn't
unpinned on destruction and none of the internal MOB buffers were
unpinned before being put back. Technically this existed for a
long time but 57fcd550eb15bce14a7154736379dfd4ed60ae81 introduced
a WARN_ON which was filling up the kernel logs rather quickly.

Quite frankly internal usage of vmw_buffer_object and in general
pinning needs to be refactored in vmwgfx but for now this makes
it work.

Signed-off-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Martin Krastev <krastevm@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Fixes: 57fcd550eb15bce14a7154736379dfd4ed60ae81
---
 drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 2 ++
 drivers/gpu/drm/vmwgfx/vmwgfx_mob.c | 4 ++++
 2 files changed, 6 insertions(+)

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index b45becbb00f8..73225ab691e6 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -1554,6 +1554,8 @@ static inline void vmw_bo_unreference(struct vmw_buffer_object **buf)
 
 	*buf = NULL;
 	if (tmp_buf != NULL) {
+		if (tmp_buf->base.pin_count > 0)
+			ttm_bo_unpin(&tmp_buf->base);
 		ttm_bo_put(&tmp_buf->base);
 	}
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
index 7f95ed6aa224..3c6e69f36767 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
@@ -277,6 +277,7 @@ static int vmw_otable_batch_setup(struct vmw_private *dev_priv,
 						 &batch->otables[i]);
 	}
 
+	ttm_bo_unpin(batch->otable_bo);
 	ttm_bo_put(batch->otable_bo);
 	batch->otable_bo = NULL;
 	return ret;
@@ -342,6 +343,7 @@ static void vmw_otable_batch_takedown(struct vmw_private *dev_priv,
 	vmw_bo_fence_single(bo, NULL);
 	ttm_bo_unreserve(bo);
 
+	ttm_bo_unpin(batch->otable_bo);
 	ttm_bo_put(batch->otable_bo);
 	batch->otable_bo = NULL;
 }
@@ -528,6 +530,7 @@ static void vmw_mob_pt_setup(struct vmw_mob *mob,
 void vmw_mob_destroy(struct vmw_mob *mob)
 {
 	if (mob->pt_bo) {
+		ttm_bo_unpin(mob->pt_bo);
 		ttm_bo_put(mob->pt_bo);
 		mob->pt_bo = NULL;
 	}
@@ -643,6 +646,7 @@ int vmw_mob_bind(struct vmw_private *dev_priv,
 out_no_cmd_space:
 	vmw_fifo_resource_dec(dev_priv);
 	if (pt_set_up) {
+		ttm_bo_unpin(mob->pt_bo);
 		ttm_bo_put(mob->pt_bo);
 		mob->pt_bo = NULL;
 	}
-- 
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/vmwgfx: Make sure we unpin no longer needed buffers
  2021-01-19 19:31 [PATCH] drm/vmwgfx: Make sure we unpin no longer needed buffers Zack Rusin
@ 2021-01-19 19:35 ` Daniel Vetter
  2021-01-19 19:36   ` Daniel Vetter
  0 siblings, 1 reply; 3+ messages in thread
From: Daniel Vetter @ 2021-01-19 19:35 UTC (permalink / raw)
  To: Zack Rusin; +Cc: Martin Krastev, Roland Scheidegger, dri-devel

On Tue, Jan 19, 2021 at 8:31 PM Zack Rusin <zackr@vmware.com> wrote:
>
> We were not correctly unpinning no longer needed buffers. In particular
> vmw_buffer_object, which is internally often pinned on creation wasn't
> unpinned on destruction and none of the internal MOB buffers were
> unpinned before being put back. Technically this existed for a
> long time but 57fcd550eb15bce14a7154736379dfd4ed60ae81 introduced
> a WARN_ON which was filling up the kernel logs rather quickly.
>
> Quite frankly internal usage of vmw_buffer_object and in general
> pinning needs to be refactored in vmwgfx but for now this makes
> it work.
>
> Signed-off-by: Zack Rusin <zackr@vmware.com>
> Reviewed-by: Martin Krastev <krastevm@vmware.com>
> Reviewed-by: Roland Scheidegger <sroland@vmware.com>
> Fixes: 57fcd550eb15bce14a7154736379dfd4ed60ae81

dim will balk on this (or should at least)

$ dim fixes <sha1>

should give you the recommend thing.
-Daniel

> ---
>  drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 2 ++
>  drivers/gpu/drm/vmwgfx/vmwgfx_mob.c | 4 ++++
>  2 files changed, 6 insertions(+)
>
> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> index b45becbb00f8..73225ab691e6 100644
> --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> @@ -1554,6 +1554,8 @@ static inline void vmw_bo_unreference(struct vmw_buffer_object **buf)
>
>         *buf = NULL;
>         if (tmp_buf != NULL) {
> +               if (tmp_buf->base.pin_count > 0)
> +                       ttm_bo_unpin(&tmp_buf->base);
>                 ttm_bo_put(&tmp_buf->base);
>         }
>  }
> diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
> index 7f95ed6aa224..3c6e69f36767 100644
> --- a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
> +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
> @@ -277,6 +277,7 @@ static int vmw_otable_batch_setup(struct vmw_private *dev_priv,
>                                                  &batch->otables[i]);
>         }
>
> +       ttm_bo_unpin(batch->otable_bo);
>         ttm_bo_put(batch->otable_bo);
>         batch->otable_bo = NULL;
>         return ret;
> @@ -342,6 +343,7 @@ static void vmw_otable_batch_takedown(struct vmw_private *dev_priv,
>         vmw_bo_fence_single(bo, NULL);
>         ttm_bo_unreserve(bo);
>
> +       ttm_bo_unpin(batch->otable_bo);
>         ttm_bo_put(batch->otable_bo);
>         batch->otable_bo = NULL;
>  }
> @@ -528,6 +530,7 @@ static void vmw_mob_pt_setup(struct vmw_mob *mob,
>  void vmw_mob_destroy(struct vmw_mob *mob)
>  {
>         if (mob->pt_bo) {
> +               ttm_bo_unpin(mob->pt_bo);
>                 ttm_bo_put(mob->pt_bo);
>                 mob->pt_bo = NULL;
>         }
> @@ -643,6 +646,7 @@ int vmw_mob_bind(struct vmw_private *dev_priv,
>  out_no_cmd_space:
>         vmw_fifo_resource_dec(dev_priv);
>         if (pt_set_up) {
> +               ttm_bo_unpin(mob->pt_bo);
>                 ttm_bo_put(mob->pt_bo);
>                 mob->pt_bo = NULL;
>         }
> --
> 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/vmwgfx: Make sure we unpin no longer needed buffers
  2021-01-19 19:35 ` Daniel Vetter
@ 2021-01-19 19:36   ` Daniel Vetter
  0 siblings, 0 replies; 3+ messages in thread
From: Daniel Vetter @ 2021-01-19 19:36 UTC (permalink / raw)
  To: Zack Rusin; +Cc: Martin Krastev, Roland Scheidegger, dri-devel

On Tue, Jan 19, 2021 at 8:35 PM Daniel Vetter <daniel@ffwll.ch> wrote:
>
> On Tue, Jan 19, 2021 at 8:31 PM Zack Rusin <zackr@vmware.com> wrote:
> >
> > We were not correctly unpinning no longer needed buffers. In particular
> > vmw_buffer_object, which is internally often pinned on creation wasn't
> > unpinned on destruction and none of the internal MOB buffers were
> > unpinned before being put back. Technically this existed for a
> > long time but 57fcd550eb15bce14a7154736379dfd4ed60ae81 introduced

Also this one should be replaced by the output of dim cite <sha1>, but
I think dim doesn't check for these.
-Daniel

> > a WARN_ON which was filling up the kernel logs rather quickly.
> >
> > Quite frankly internal usage of vmw_buffer_object and in general
> > pinning needs to be refactored in vmwgfx but for now this makes
> > it work.
> >
> > Signed-off-by: Zack Rusin <zackr@vmware.com>
> > Reviewed-by: Martin Krastev <krastevm@vmware.com>
> > Reviewed-by: Roland Scheidegger <sroland@vmware.com>
> > Fixes: 57fcd550eb15bce14a7154736379dfd4ed60ae81
>
> dim will balk on this (or should at least)
>
> $ dim fixes <sha1>
>
> should give you the recommend thing.
> -Daniel
>
> > ---
> >  drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 2 ++
> >  drivers/gpu/drm/vmwgfx/vmwgfx_mob.c | 4 ++++
> >  2 files changed, 6 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> > index b45becbb00f8..73225ab691e6 100644
> > --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> > +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
> > @@ -1554,6 +1554,8 @@ static inline void vmw_bo_unreference(struct vmw_buffer_object **buf)
> >
> >         *buf = NULL;
> >         if (tmp_buf != NULL) {
> > +               if (tmp_buf->base.pin_count > 0)
> > +                       ttm_bo_unpin(&tmp_buf->base);
> >                 ttm_bo_put(&tmp_buf->base);
> >         }
> >  }
> > diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
> > index 7f95ed6aa224..3c6e69f36767 100644
> > --- a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
> > +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c
> > @@ -277,6 +277,7 @@ static int vmw_otable_batch_setup(struct vmw_private *dev_priv,
> >                                                  &batch->otables[i]);
> >         }
> >
> > +       ttm_bo_unpin(batch->otable_bo);
> >         ttm_bo_put(batch->otable_bo);
> >         batch->otable_bo = NULL;
> >         return ret;
> > @@ -342,6 +343,7 @@ static void vmw_otable_batch_takedown(struct vmw_private *dev_priv,
> >         vmw_bo_fence_single(bo, NULL);
> >         ttm_bo_unreserve(bo);
> >
> > +       ttm_bo_unpin(batch->otable_bo);
> >         ttm_bo_put(batch->otable_bo);
> >         batch->otable_bo = NULL;
> >  }
> > @@ -528,6 +530,7 @@ static void vmw_mob_pt_setup(struct vmw_mob *mob,
> >  void vmw_mob_destroy(struct vmw_mob *mob)
> >  {
> >         if (mob->pt_bo) {
> > +               ttm_bo_unpin(mob->pt_bo);
> >                 ttm_bo_put(mob->pt_bo);
> >                 mob->pt_bo = NULL;
> >         }
> > @@ -643,6 +646,7 @@ int vmw_mob_bind(struct vmw_private *dev_priv,
> >  out_no_cmd_space:
> >         vmw_fifo_resource_dec(dev_priv);
> >         if (pt_set_up) {
> > +               ttm_bo_unpin(mob->pt_bo);
> >                 ttm_bo_put(mob->pt_bo);
> >                 mob->pt_bo = NULL;
> >         }
> > --
> > 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:[~2021-01-19 19:36 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-01-19 19:31 [PATCH] drm/vmwgfx: Make sure we unpin no longer needed buffers Zack Rusin
2021-01-19 19:35 ` Daniel Vetter
2021-01-19 19:36   ` Daniel Vetter

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).