From: "Timur Kristóf" <timur.kristof@gmail.com>
To: amd-gfx@lists.freedesktop.org
Cc: Alex Deucher <alexander.deucher@amd.com>,
Alex Deucher <alexander.deucher@amd.com>
Subject: Re: [PATCH 4/4] drm/amdgpu/gfx11: enable gfx pipe1 hardware support
Date: Tue, 07 Jul 2026 10:08:45 +0200 [thread overview]
Message-ID: <1961599.dNmn5OnKVQ@timur-max> (raw)
In-Reply-To: <20260626204101.31172-4-alexander.deucher@amd.com>
On 2026. június 26., péntek 22:41:01 közép-európai nyári idő Alex Deucher
wrote:
> Enable gfx pipe1 hardware support. This is only available
> on gfx11 chips using the F32 microcontroller. Chips using
> the RS64 microcontroller are not able to use the second gfx
> pipe. In practice this means the second pipe is only
> available on APUs. This explains the stability issues
> Pierre-Eric saw previously with this on Navi33.
>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c | 22 +++++++++++++++++-----
> 1 file changed, 17 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
> b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 5e7d0cd85c365..c0e34519b6b06
> 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
> @@ -51,7 +51,7 @@
> #include "mes_userqueue.h"
> #include "amdgpu_userq_fence.h"
>
> -#define GFX11_NUM_GFX_RINGS 1
> +#define GFX11_NUM_GFX_RINGS 2
> #define GFX11_MEC_HPD_SIZE 2048
>
> #define RLCG_UCODE_LOADING_START_ADDRESS 0x00002000L
> @@ -1612,7 +1612,10 @@ static int gfx_v11_0_sw_init(struct amdgpu_ip_block
> *ip_block) case IP_VERSION(11, 5, 4):
> case IP_VERSION(11, 5, 6):
> adev->gfx.me.num_me = 1;
> - adev->gfx.me.num_pipe_per_me = 1;
> + if (adev->gfx.rs64_enable)
> + adev->gfx.me.num_pipe_per_me = 1;
> + else
> + adev->gfx.me.num_pipe_per_me = 2;
> adev->gfx.me.num_queue_per_pipe = 2;
> adev->gfx.mec.num_mec = 1;
> adev->gfx.mec.num_pipe_per_mec = 4;
> @@ -5355,6 +5358,7 @@ static void gfx_v11_0_ring_emit_gds_switch(struct
> amdgpu_ring *ring, static int gfx_v11_0_early_init(struct amdgpu_ip_block
> *ip_block) {
> struct amdgpu_device *adev = ip_block->adev;
> + int r;
>
> switch (amdgpu_user_queue) {
> case -1:
> @@ -5375,6 +5379,11 @@ static int gfx_v11_0_early_init(struct
> amdgpu_ip_block *ip_block)
>
> adev->gfx.funcs = &gfx_v11_0_gfx_funcs;
>
> + gfx_v11_0_set_imu_funcs(adev);
> + r = gfx_v11_0_init_microcode(adev);
> + if (r)
> + return r;
> +
> if (adev->gfx.disable_kq) {
> /* We need one GFX ring temporarily to set up
> * the clear state.
> @@ -5382,7 +5391,11 @@ static int gfx_v11_0_early_init(struct
> amdgpu_ip_block *ip_block) adev->gfx.num_gfx_rings = 1;
> adev->gfx.num_compute_rings = 0;
> } else {
> - adev->gfx.num_gfx_rings = GFX11_NUM_GFX_RINGS;
> + /* rs64 only supports one gfx pipe */
> + if (adev->gfx.rs64_enable)
> + adev->gfx.num_gfx_rings = 1;
> + else
> + adev->gfx.num_gfx_rings =
GFX11_NUM_GFX_RINGS;
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Could you maybe move this code to a small helper function, eg. we could call
it gfx_v11_0_calc_num_pipes_per_me() and call that from both places.
Thanks & best regards,
Timur
> adev->gfx.num_compute_rings =
min(amdgpu_gfx_get_num_kcq(adev),
>
AMDGPU_MAX_COMPUTE_RINGS);
> }
> @@ -5393,13 +5406,12 @@ static int gfx_v11_0_early_init(struct
> amdgpu_ip_block *ip_block) gfx_v11_0_set_gds_init(adev);
> gfx_v11_0_set_rlc_funcs(adev);
> gfx_v11_0_set_mqd_funcs(adev);
> - gfx_v11_0_set_imu_funcs(adev);
>
> gfx_v11_0_init_rlcg_reg_access_ctrl(adev);
>
> amdgpu_init_rlc_reg_funcs(adev);
>
> - return gfx_v11_0_init_microcode(adev);
> + return 0;
> }
>
> static bool gfx_v11_0_is_rlc_enabled(struct amdgpu_device *adev)
next prev parent reply other threads:[~2026-07-07 8:08 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-26 20:40 [PATCH 1/4] drm/gfx10: Program DB_RING_CONTROL Alex Deucher
2026-06-26 20:40 ` [PATCH 2/4] drm/gfx11: " Alex Deucher
2026-06-26 20:41 ` [PATCH 3/4] drm/gfx12: " Alex Deucher
2026-06-26 20:41 ` [PATCH 4/4] drm/amdgpu/gfx11: enable gfx pipe1 hardware support Alex Deucher
2026-07-07 8:08 ` Timur Kristóf [this message]
2026-07-06 15:00 ` [PATCH 1/4] drm/gfx10: Program DB_RING_CONTROL Alex Deucher
2026-07-07 8:06 ` Timur Kristóf
2026-07-07 13:44 ` Alex Deucher
2026-07-07 15:14 ` Timur Kristóf
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=1961599.dNmn5OnKVQ@timur-max \
--to=timur.kristof@gmail.com \
--cc=alexander.deucher@amd.com \
--cc=amd-gfx@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 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.