lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <a71af2ef-d35b-4ef2-baaf-2e68cd3d22de@ursulin.net>
Date: Fri, 31 Oct 2025 11:10:38 +0000
From: Tvrtko Ursulin <tursulin@...ulin.net>
To: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@....com>,
 Alex Deucher <alexander.deucher@....com>,
 Christian König <christian.koenig@....com>,
 David Airlie <airlied@...il.com>, Simona Vetter <simona@...ll.ch>
Cc: amd-gfx@...ts.freedesktop.org, dri-devel@...ts.freedesktop.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/3] drm/amdgpu: add engine_retains_context to
 amdgpu_ring_funcs


On 30/10/2025 09:11, Pierre-Eric Pelloux-Prayer wrote:
> If true, the hw engine retains context among dependent jobs, which means
> load balancing between schedulers cannot be used at the job level.
> 
> amdgpu_ctx_init_entity uses this information to disable load balancing,
> but it's best to store it as a property rather than deduce it based on
> hw_ip.

On a tangent, this patch reminded me on a series I had to remove some 
holes from structs. For the ring specifically:

commit e89d67894fe4824d0163c8831b9bcdaabb4ae3a2
Author: Tvrtko Ursulin <tvrtko.ursulin@...lia.com>
Date:   Fri Feb 21 13:37:25 2025 +0000

     struct amdgpu_ring {

             /* size: 792, cachelines: 13, members: 62 */
             /* sum members: 746, holes: 13, sum holes: 46 */
             /* last cacheline: 24 bytes */

             /* size: 744, cachelines: 12, members: 62 */
             /* sum members: 738, holes: 2, sum holes: 6 */
             /* last cacheline: 40 bytes */

I tried to keep and/or improve logical grouping of fields while 
re-arranging to avoid holes. Only one cacheline saved is not much so not 
sure. Not even sure if I ever sent it. Any interest to clean it up and send?

Regards,

Tvrtko
> Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@....com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 1 +
>   drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c    | 1 +
>   drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c    | 1 +
>   drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c    | 1 +
>   drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c    | 3 +++
>   drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c    | 2 ++
>   drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c    | 2 ++
>   drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c    | 2 ++
>   drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c    | 2 ++
>   drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c    | 3 +++
>   drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c    | 1 +
>   drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c  | 1 +
>   drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c  | 1 +
>   drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c  | 1 +
>   drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c  | 1 +
>   15 files changed, 23 insertions(+)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
> index 4b46e3c26ff3..a10efac2fc54 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
> @@ -211,6 +211,7 @@ struct amdgpu_ring_funcs {
>   	bool			support_64bit_ptrs;
>   	bool			no_user_fence;
>   	bool			secure_submission_supported;
> +	bool			engine_retains_context;
>   
>   	/**
>   	 * @extra_bytes:
> diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
> index 2e79a3afc774..4a85b5465bb2 100644
> --- a/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
> +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v3_1.c
> @@ -181,6 +181,7 @@ static const struct amdgpu_ring_funcs uvd_v3_1_ring_funcs = {
>   	.align_mask = 0xf,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v3_1_ring_get_rptr,
>   	.get_wptr = uvd_v3_1_ring_get_wptr,
>   	.set_wptr = uvd_v3_1_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
> index 4b96fd583772..e7c1d12f0596 100644
> --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
> +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
> @@ -775,6 +775,7 @@ static const struct amdgpu_ring_funcs uvd_v4_2_ring_funcs = {
>   	.align_mask = 0xf,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v4_2_ring_get_rptr,
>   	.get_wptr = uvd_v4_2_ring_get_wptr,
>   	.set_wptr = uvd_v4_2_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
> index 71409ad8b7ed..a62788e4af96 100644
> --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
> @@ -882,6 +882,7 @@ static const struct amdgpu_ring_funcs uvd_v5_0_ring_funcs = {
>   	.align_mask = 0xf,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v5_0_ring_get_rptr,
>   	.get_wptr = uvd_v5_0_ring_get_wptr,
>   	.set_wptr = uvd_v5_0_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
> index ceb94bbb03a4..0435577b9b3b 100644
> --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
> @@ -1552,6 +1552,7 @@ static const struct amdgpu_ring_funcs uvd_v6_0_ring_phys_funcs = {
>   	.align_mask = 0xf,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v6_0_ring_get_rptr,
>   	.get_wptr = uvd_v6_0_ring_get_wptr,
>   	.set_wptr = uvd_v6_0_ring_set_wptr,
> @@ -1578,6 +1579,7 @@ static const struct amdgpu_ring_funcs uvd_v6_0_ring_vm_funcs = {
>   	.align_mask = 0xf,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v6_0_ring_get_rptr,
>   	.get_wptr = uvd_v6_0_ring_get_wptr,
>   	.set_wptr = uvd_v6_0_ring_set_wptr,
> @@ -1607,6 +1609,7 @@ static const struct amdgpu_ring_funcs uvd_v6_0_enc_ring_vm_funcs = {
>   	.nop = HEVC_ENC_CMD_NO_OP,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v6_0_enc_ring_get_rptr,
>   	.get_wptr = uvd_v6_0_enc_ring_get_wptr,
>   	.set_wptr = uvd_v6_0_enc_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
> index 1f8866f3f63c..3720d72f2c3e 100644
> --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
> @@ -1539,6 +1539,7 @@ static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
>   	.align_mask = 0xf,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v7_0_ring_get_rptr,
>   	.get_wptr = uvd_v7_0_ring_get_wptr,
>   	.set_wptr = uvd_v7_0_ring_set_wptr,
> @@ -1571,6 +1572,7 @@ static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
>   	.nop = HEVC_ENC_CMD_NO_OP,
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
> +	.engine_retains_context = true,
>   	.get_rptr = uvd_v7_0_enc_ring_get_rptr,
>   	.get_wptr = uvd_v7_0_enc_ring_get_wptr,
>   	.set_wptr = uvd_v7_0_enc_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
> index a316797875a8..1691d0f955a9 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
> @@ -2117,6 +2117,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
>   	.support_64bit_ptrs = false,
>   	.no_user_fence = true,
>   	.secure_submission_supported = true,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v1_0_dec_ring_get_rptr,
>   	.get_wptr = vcn_v1_0_dec_ring_get_wptr,
>   	.set_wptr = vcn_v1_0_dec_ring_set_wptr,
> @@ -2150,6 +2151,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
>   	.support_64bit_ptrs = false,
> +	.engine_retains_context = true,
>   	.no_user_fence = true,
>   	.get_rptr = vcn_v1_0_enc_ring_get_rptr,
>   	.get_wptr = vcn_v1_0_enc_ring_get_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
> index 8897dcc9c1a0..046dd6b216e9 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
> @@ -2113,6 +2113,7 @@ static const struct amdgpu_ring_funcs vcn_v2_0_dec_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_DEC,
>   	.align_mask = 0xf,
>   	.secure_submission_supported = true,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v2_0_dec_ring_get_rptr,
>   	.get_wptr = vcn_v2_0_dec_ring_get_wptr,
>   	.set_wptr = vcn_v2_0_dec_ring_set_wptr,
> @@ -2144,6 +2145,7 @@ static const struct amdgpu_ring_funcs vcn_v2_0_enc_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v2_0_enc_ring_get_rptr,
>   	.get_wptr = vcn_v2_0_enc_ring_get_wptr,
>   	.set_wptr = vcn_v2_0_enc_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
> index cebee453871c..063f88da120b 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c
> @@ -1777,6 +1777,7 @@ static const struct amdgpu_ring_funcs vcn_v2_5_dec_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_DEC,
>   	.align_mask = 0xf,
>   	.secure_submission_supported = true,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v2_5_dec_ring_get_rptr,
>   	.get_wptr = vcn_v2_5_dec_ring_get_wptr,
>   	.set_wptr = vcn_v2_5_dec_ring_set_wptr,
> @@ -1877,6 +1878,7 @@ static const struct amdgpu_ring_funcs vcn_v2_5_enc_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v2_5_enc_ring_get_rptr,
>   	.get_wptr = vcn_v2_5_enc_ring_get_wptr,
>   	.set_wptr = vcn_v2_5_enc_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
> index d9cf8f0feeb3..8dcc07b3f631 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c
> @@ -1857,6 +1857,7 @@ static const struct amdgpu_ring_funcs vcn_v3_0_dec_sw_ring_vm_funcs = {
>   	.align_mask = 0x3f,
>   	.nop = VCN_DEC_SW_CMD_NO_OP,
>   	.secure_submission_supported = true,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v3_0_dec_ring_get_rptr,
>   	.get_wptr = vcn_v3_0_dec_ring_get_wptr,
>   	.set_wptr = vcn_v3_0_dec_ring_set_wptr,
> @@ -2021,6 +2022,7 @@ static const struct amdgpu_ring_funcs vcn_v3_0_dec_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_DEC,
>   	.align_mask = 0xf,
>   	.secure_submission_supported = true,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v3_0_dec_ring_get_rptr,
>   	.get_wptr = vcn_v3_0_dec_ring_get_wptr,
>   	.set_wptr = vcn_v3_0_dec_ring_set_wptr,
> @@ -2122,6 +2124,7 @@ static const struct amdgpu_ring_funcs vcn_v3_0_enc_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v3_0_enc_ring_get_rptr,
>   	.get_wptr = vcn_v3_0_enc_ring_get_wptr,
>   	.set_wptr = vcn_v3_0_enc_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
> index 3ae666522d57..f1306316dc3c 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0.c
> @@ -1977,6 +1977,7 @@ static struct amdgpu_ring_funcs vcn_v4_0_unified_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.extra_bytes = sizeof(struct amdgpu_vcn_rb_metadata),
>   	.get_rptr = vcn_v4_0_unified_ring_get_rptr,
>   	.get_wptr = vcn_v4_0_unified_ring_get_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
> index eacf4e93ba2f..5a935c07352a 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
> @@ -1628,6 +1628,7 @@ static const struct amdgpu_ring_funcs vcn_v4_0_3_unified_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v4_0_3_unified_ring_get_rptr,
>   	.get_wptr = vcn_v4_0_3_unified_ring_get_wptr,
>   	.set_wptr = vcn_v4_0_3_unified_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
> index b107ee80e472..1a485f5825dd 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c
> @@ -1481,6 +1481,7 @@ static struct amdgpu_ring_funcs vcn_v4_0_5_unified_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v4_0_5_unified_ring_get_rptr,
>   	.get_wptr = vcn_v4_0_5_unified_ring_get_wptr,
>   	.set_wptr = vcn_v4_0_5_unified_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
> index 0202df5db1e1..2d8214f591f1 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_0.c
> @@ -1203,6 +1203,7 @@ static const struct amdgpu_ring_funcs vcn_v5_0_0_unified_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v5_0_0_unified_ring_get_rptr,
>   	.get_wptr = vcn_v5_0_0_unified_ring_get_wptr,
>   	.set_wptr = vcn_v5_0_0_unified_ring_set_wptr,
> diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
> index 714350cabf2f..bd3a04f1414d 100644
> --- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
> +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
> @@ -1328,6 +1328,7 @@ static const struct amdgpu_ring_funcs vcn_v5_0_1_unified_ring_vm_funcs = {
>   	.type = AMDGPU_RING_TYPE_VCN_ENC,
>   	.align_mask = 0x3f,
>   	.nop = VCN_ENC_CMD_NO_OP,
> +	.engine_retains_context = true,
>   	.get_rptr = vcn_v5_0_1_unified_ring_get_rptr,
>   	.get_wptr = vcn_v5_0_1_unified_ring_get_wptr,
>   	.set_wptr = vcn_v5_0_1_unified_ring_set_wptr,


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ