dri-devel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Christian König" <christian.koenig@amd.com>
To: Dave Airlie <airlied@gmail.com>, dri-devel@lists.freedesktop.org
Cc: bskeggs@redhat.com
Subject: Re: [PATCH 10/10] drm/ttm: reverse move calling pattern.
Date: Wed, 23 Sep 2020 16:45:03 +0200	[thread overview]
Message-ID: <a66afa9c-2f3c-3e0a-8637-2cea7e7918c1@amd.com> (raw)
In-Reply-To: <20200923030454.362731-11-airlied@gmail.com>

Am 23.09.20 um 05:04 schrieb Dave Airlie:
> From: Dave Airlie <airlied@redhat.com>
>
> Call the driver move function if it exists, otherwise use the
> fallback ttm/memcpy paths.

I would rather like to see the move callback made mandatory instead.

Christian.

>
> Signed-off-by: Dave Airlie <airlied@redhat.com>
> ---
>   drivers/gpu/drm/ttm/ttm_bo.c | 28 +++++++++++++++++++---------
>   1 file changed, 19 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
> index 993a87443c37..3d9c62cdf38d 100644
> --- a/drivers/gpu/drm/ttm/ttm_bo.c
> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
> @@ -229,6 +229,23 @@ void ttm_bo_bulk_move_lru_tail(struct ttm_lru_bulk_move *bulk)
>   }
>   EXPORT_SYMBOL(ttm_bo_bulk_move_lru_tail);
>   
> +static int ttm_bo_move_fallback(struct ttm_buffer_object *bo,
> +				struct ttm_operation_ctx *ctx,
> +				struct ttm_resource *mem)
> +{
> +	struct ttm_resource_manager *old_man = ttm_manager_type(bo->bdev, bo->mem.mem_type);
> +	struct ttm_resource_manager *new_man = ttm_manager_type(bo->bdev, mem->mem_type);
> +
> +	if (old_man->use_tt && new_man->use_tt) {
> +		if (bo->mem.mem_type == TTM_PL_SYSTEM) {
> +			ttm_bo_assign_mem(bo, mem);
> +			return 0;
> +		} else
> +			return ttm_bo_move_ttm(bo, ctx, mem);
> +	} else
> +		return ttm_bo_move_memcpy(bo, ctx, mem);
> +}
> +
>   static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
>   				  struct ttm_resource *mem, bool evict,
>   				  struct ttm_operation_ctx *ctx)
> @@ -270,17 +287,10 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
>   	if (bdev->driver->move_notify)
>   		bdev->driver->move_notify(bo, evict, mem);
>   
> -	if (old_man->use_tt && new_man->use_tt) {
> -		if (bo->mem.mem_type == TTM_PL_SYSTEM) {
> -			ttm_bo_assign_mem(bo, mem);
> -			ret = 0;
> -		} else
> -			ret = ttm_bo_move_ttm(bo, ctx, mem);
> -	}
> -	else if (bdev->driver->move)
> +	if (bdev->driver->move)
>   		ret = bdev->driver->move(bo, evict, ctx, mem);
>   	else
> -		ret = ttm_bo_move_memcpy(bo, ctx, mem);
> +		ret = ttm_bo_move_fallback(bo, ctx, mem);
>   
>   	if (ret) {
>   		if (bdev->driver->move_notify) {

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

  reply	other threads:[~2020-09-23 14:45 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-23  3:04 [PATCH 00/10] ttm driver cleanups and invert move Dave Airlie
2020-09-23  3:04 ` [PATCH 01/10] drm/radeon: kill radeon_bo_wait Dave Airlie
2020-09-23 14:34   ` Christian König
2020-09-23  3:04 ` [PATCH 02/10] drm/qxl: kill unused bo wait wrapper Dave Airlie
2020-09-23 14:35   ` Christian König
2020-09-23  3:04 ` [PATCH 03/10] drm/radeon: cleanup ttm operation ctx usage Dave Airlie
2020-09-23 14:38   ` Christian König
2020-09-23  3:04 ` [PATCH 04/10] drm/nouveau/ttm: plumb ctx through move functions Dave Airlie
2020-09-23 14:39   ` Christian König
2020-09-23  3:04 ` [PATCH 05/10] drm/ttm: add bo wait that takes a ctx wrapper Dave Airlie
2020-09-23 14:40   ` Christian König
2020-09-23  3:04 ` [PATCH 06/10] drm/ttm: handle the SYSTEM->TT path in same place as others Dave Airlie
2020-09-23  3:04 ` [PATCH 07/10] drm/amdgpu/ttm: handle tt moves properly Dave Airlie
2020-09-23 14:45   ` Christian König
2020-09-24  0:46     ` Dave Airlie
2020-09-24  2:35       ` Dave Airlie
2020-09-24  9:48       ` Christian König
2020-09-23  3:04 ` [PATCH 08/10] drm/radeon/ttm: handle ttm " Dave Airlie
2020-09-23  3:04 ` [PATCH 09/10] drm/nouveau/ttm: " Dave Airlie
2020-09-23  3:04 ` [PATCH 10/10] drm/ttm: reverse move calling pattern Dave Airlie
2020-09-23 14:45   ` Christian König [this message]
2020-09-23 20:43     ` Dave Airlie

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=a66afa9c-2f3c-3e0a-8637-2cea7e7918c1@amd.com \
    --to=christian.koenig@amd.com \
    --cc=airlied@gmail.com \
    --cc=bskeggs@redhat.com \
    --cc=dri-devel@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox