[PATCH v4] drm/nouveau: use tile_mode and pte_kind for VM_BIND bo allocations

[Danilo Krummrich]

Hi Mohamed,

Thank you for fixing this up!

On 5/9/24 22:43, Mohamed Ahmed wrote:
> Allows PTE kind and tile mode on BO create with VM_BIND,
> and adds a GETPARAM to indicate this change. This is needed to support

It's usually better to use imperative verb form for commit messages. No
need to send a new version though.

> modifiers in NVK and ensure correctness when dealing with the nouveau
> GL driver.
> 
> The userspace modifiers implementation this is for can be found here:
> https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24795
> 
> Fixes: b88baab82871 ("drm/nouveau: implement new VM_BIND uAPI")
> Signed-off-by: Mohamed Ahmed <mohamedahmedegypt2001 at gmail.com>

Applied to drm-misc-next-fixes.

Generally, please make sure to use scripts/get_maintainer.pl before sending
patches.

- Danilo

> ---
>   drivers/gpu/drm/nouveau/nouveau_abi16.c |  3 ++
>   drivers/gpu/drm/nouveau/nouveau_bo.c    | 44 +++++++++++--------------
>   include/uapi/drm/nouveau_drm.h          |  7 ++++
>   3 files changed, 29 insertions(+), 25 deletions(-)
> 
> diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
> index 80f74ee0f..47e53e17b 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
> @@ -272,6 +272,9 @@ nouveau_abi16_ioctl_getparam(ABI16_IOCTL_ARGS)
>   		getparam->value = (u64)ttm_resource_manager_usage(vram_mgr);
>   		break;
>   	}
> +	case NOUVEAU_GETPARAM_HAS_VMA_TILEMODE:
> +		getparam->value = 1;
> +		break;
>   	default:
>   		NV_PRINTK(dbg, cli, "unknown parameter %lld\n", getparam->param);
>   		return -EINVAL;
> diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
> index db8cbf615..186add400 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_bo.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
> @@ -241,28 +241,28 @@ nouveau_bo_alloc(struct nouveau_cli *cli, u64 *size, int *align, u32 domain,
>   	}
>   
>   	nvbo->contig = !(tile_flags & NOUVEAU_GEM_TILE_NONCONTIG);
> -	if (!nouveau_cli_uvmm(cli) || internal) {
> -		/* for BO noVM allocs, don't assign kinds */
> -		if (cli->device.info.family >= NV_DEVICE_INFO_V0_FERMI) {
> -			nvbo->kind = (tile_flags & 0x0000ff00) >> 8;
> -			if (!nvif_mmu_kind_valid(mmu, nvbo->kind)) {
> -				kfree(nvbo);
> -				return ERR_PTR(-EINVAL);
> -			}
>   
> -			nvbo->comp = mmu->kind[nvbo->kind] != nvbo->kind;
> -		} else if (cli->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
> -			nvbo->kind = (tile_flags & 0x00007f00) >> 8;
> -			nvbo->comp = (tile_flags & 0x00030000) >> 16;
> -			if (!nvif_mmu_kind_valid(mmu, nvbo->kind)) {
> -				kfree(nvbo);
> -				return ERR_PTR(-EINVAL);
> -			}
> -		} else {
> -			nvbo->zeta = (tile_flags & 0x00000007);
> +	if (cli->device.info.family >= NV_DEVICE_INFO_V0_FERMI) {
> +		nvbo->kind = (tile_flags & 0x0000ff00) >> 8;
> +		if (!nvif_mmu_kind_valid(mmu, nvbo->kind)) {
> +			kfree(nvbo);
> +			return ERR_PTR(-EINVAL);
> +		}
> +
> +		nvbo->comp = mmu->kind[nvbo->kind] != nvbo->kind;
> +	} else if (cli->device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
> +		nvbo->kind = (tile_flags & 0x00007f00) >> 8;
> +		nvbo->comp = (tile_flags & 0x00030000) >> 16;
> +		if (!nvif_mmu_kind_valid(mmu, nvbo->kind)) {
> +			kfree(nvbo);
> +			return ERR_PTR(-EINVAL);
>   		}
> -		nvbo->mode = tile_mode;
> +	} else {
> +		nvbo->zeta = (tile_flags & 0x00000007);
> +	}
> +	nvbo->mode = tile_mode;
>   
> +	if (!nouveau_cli_uvmm(cli) || internal) {
>   		/* Determine the desirable target GPU page size for the buffer. */
>   		for (i = 0; i < vmm->page_nr; i++) {
>   			/* Because we cannot currently allow VMM maps to fail
> @@ -304,12 +304,6 @@ nouveau_bo_alloc(struct nouveau_cli *cli, u64 *size, int *align, u32 domain,
>   		}
>   		nvbo->page = vmm->page[pi].shift;
>   	} else {
> -		/* reject other tile flags when in VM mode. */
> -		if (tile_mode)
> -			return ERR_PTR(-EINVAL);
> -		if (tile_flags & ~NOUVEAU_GEM_TILE_NONCONTIG)
> -			return ERR_PTR(-EINVAL);
> -
>   		/* Determine the desirable target GPU page size for the buffer. */
>   		for (i = 0; i < vmm->page_nr; i++) {
>   			/* Because we cannot currently allow VMM maps to fail
> diff --git a/include/uapi/drm/nouveau_drm.h b/include/uapi/drm/nouveau_drm.h
> index cd84227f1..5402f77ee 100644
> --- a/include/uapi/drm/nouveau_drm.h
> +++ b/include/uapi/drm/nouveau_drm.h
> @@ -68,6 +68,13 @@ extern "C" {
>    */
>   #define NOUVEAU_GETPARAM_VRAM_USED 19
>   
> +/*
> + * NOUVEAU_GETPARAM_HAS_VMA_TILEMODE
> + *
> + * Query whether tile mode and PTE kind are accepted with VM allocs or not.
> + */
> +#define NOUVEAU_GETPARAM_HAS_VMA_TILEMODE 20
> +
>   struct drm_nouveau_getparam {
>   	__u64 param;
>   	__u64 value;