From 47659738fbd2f06730635a487605002ea9b11f3d Mon Sep 17 00:00:00 2001
From: Le Ma <le.ma@amd.com>
Date: Fri, 19 Nov 2021 15:35:30 +0800
Subject: drm/amdgpu: allocate doorbell index for multi-die case
Allocate different doorbell index for kiq/kcq rings
on each die
Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
index 8fd11497faba..ffb75d23d2fc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
@@ -83,6 +83,8 @@ struct amdgpu_doorbell_index {
};
uint32_t first_non_cp;
uint32_t last_non_cp;
+ uint32_t xcc1_kiq_start;
+ uint32_t xcc1_mec_ring0_start;
uint32_t max_assignment;
/* Per engine SDMA doorbell size in dword */
uint32_t sdma_doorbell_range;
@@ -164,7 +166,12 @@ typedef enum _AMDGPU_VEGA20_DOORBELL_ASSIGNMENT
AMDGPU_VEGA20_DOORBELL64_FIRST_NON_CP = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE0,
AMDGPU_VEGA20_DOORBELL64_LAST_NON_CP = AMDGPU_VEGA20_DOORBELL64_VCE_RING6_7,
- AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT = 0x18F,
+ /* kiq/kcq from second XCD. Max 8 XCDs */
+ AMDGPU_VEGA20_DOORBELL_XCC1_KIQ_START = 0x190,
+ /* 8 compute rings per GC. Max to 0x1CE */
+ AMDGPU_VEGA20_DOORBELL_XCC1_MEC_RING0_START = 0x197,
+
+ AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT = 0x1CE,
AMDGPU_VEGA20_DOORBELL_INVALID = 0xFFFF
} AMDGPU_VEGA20_DOORBELL_ASSIGNMENT;
--
cgit v1.2.3
From 2a47a2d90e5cd96c24503061c8920a1e6ee248a0 Mon Sep 17 00:00:00 2001
From: Le Ma <le.ma@amd.com>
Date: Tue, 1 Mar 2022 19:42:29 +0800
Subject: drm/amdgpu: assign the doorbell index for sdma on non-AID0
Allocate new sdma doorbell index for the instances only on AID1 for now.
Todo: there's limitation that SDMA doorbell index on SDMA 4.4.2 needs to be
less than 0x1FF, so the tail part in _AMDGPU_VEGA20_DOORBELL_ASSIGNMENT is not
enough to store sdma doorbell range on maximum 4 AIDs if doorbell_range is 20.
So it looks better to create a new doorbell index assignment table for 4.4.2.
v2: change "(x << 1) + 2" to "(x + 1) << 1" for readability.
Signed-off-by: Le Ma <le.ma@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h | 6 +++++-
drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c | 18 ++++++++++++++++--
2 files changed, 21 insertions(+), 3 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
index ffb75d23d2fc..b036d2f01930 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
@@ -85,6 +85,7 @@ struct amdgpu_doorbell_index {
uint32_t last_non_cp;
uint32_t xcc1_kiq_start;
uint32_t xcc1_mec_ring0_start;
+ uint32_t aid1_sdma_start;
uint32_t max_assignment;
/* Per engine SDMA doorbell size in dword */
uint32_t sdma_doorbell_range;
@@ -171,7 +172,10 @@ typedef enum _AMDGPU_VEGA20_DOORBELL_ASSIGNMENT
/* 8 compute rings per GC. Max to 0x1CE */
AMDGPU_VEGA20_DOORBELL_XCC1_MEC_RING0_START = 0x197,
- AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT = 0x1CE,
+ /* AID1 SDMA: 0x1D0 ~ 0x1F7 */
+ AMDGPU_VEGA20_DOORBELL_AID1_sDMA_START = 0x1D0,
+
+ AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT = 0x1F7,
AMDGPU_VEGA20_DOORBELL_INVALID = 0xFFFF
} AMDGPU_VEGA20_DOORBELL_ASSIGNMENT;
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
index 7deadea03caa..6935a24d1e89 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
@@ -1310,7 +1310,14 @@ static int sdma_v4_4_2_sw_init(void *handle)
ring->use_doorbell?"true":"false");
/* doorbell size is 2 dwords, get DWORD offset */
- ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1;
+ if (aid_id > 0)
+ ring->doorbell_index =
+ (adev->doorbell_index.aid1_sdma_start << 1)
+ + adev->doorbell_index.sdma_doorbell_range
+ * (i - adev->sdma.num_inst_per_aid);
+ else
+ ring->doorbell_index =
+ adev->doorbell_index.sdma_engine[i] << 1;
ring->vm_hub = AMDGPU_MMHUB0(aid_id);
sprintf(ring->name, "sdma%d.%d", aid_id,
@@ -1329,7 +1336,14 @@ static int sdma_v4_4_2_sw_init(void *handle)
/* doorbell index of page queue is assigned right after
* gfx queue on the same instance
*/
- ring->doorbell_index = (adev->doorbell_index.sdma_engine[i] + 1) << 1;
+ if (aid_id > 0)
+ ring->doorbell_index =
+ ((adev->doorbell_index.aid1_sdma_start + 1) << 1)
+ + adev->doorbell_index.sdma_doorbell_range
+ * (i - adev->sdma.num_inst_per_aid);
+ else
+ ring->doorbell_index =
+ (adev->doorbell_index.sdma_engine[i] + 1) << 1;
ring->vm_hub = AMDGPU_MMHUB0(aid_id);
sprintf(ring->name, "page%d.%d", aid_id,
--
cgit v1.2.3
From 20bedf1379b1d4b060e3f43661f35e5447d0cfed Mon Sep 17 00:00:00 2001
From: Le Ma <le.ma@amd.com>
Date: Mon, 25 Apr 2022 22:19:58 +0800
Subject: drm/amdgpu: introduce new doorbell assignment table for GC 9.4.3
Four basic reasons as below to do the change:
1. number of ring expand a lot on GC 9.4.3, and adjustment on old
assignment cannot make each ring in a continuous doorbell space.
2. the SDMA doorbell index should not exceed 0x1FF on SDMA 4.2.2 due to
regDOORBELLx_CTRL_ENTRY.BIF_DOORBELLx_RANGE_OFFSET_ENTRY field width.
3. re-design the doorbell assignment and unify the calculation as
"start + ring/inst id" will make the code much concise.
4. only defining the START/END makes the table look simple
v2: (Lijo)
1. replace name
2. use num_inst_per_aid/sdma_doorbell_range instead of hardcoding
Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h | 32 ++++++++++++++++++++++++----
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 6 +-----
drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c | 8 +------
drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c | 19 +++--------------
4 files changed, 33 insertions(+), 32 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
index b036d2f01930..a29a018ec84e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
@@ -59,7 +59,7 @@ struct amdgpu_doorbell_index {
uint32_t gfx_ring1;
uint32_t gfx_userqueue_start;
uint32_t gfx_userqueue_end;
- uint32_t sdma_engine[8];
+ uint32_t sdma_engine[16];
uint32_t mes_ring0;
uint32_t mes_ring1;
uint32_t ih;
@@ -83,9 +83,6 @@ struct amdgpu_doorbell_index {
};
uint32_t first_non_cp;
uint32_t last_non_cp;
- uint32_t xcc1_kiq_start;
- uint32_t xcc1_mec_ring0_start;
- uint32_t aid1_sdma_start;
uint32_t max_assignment;
/* Per engine SDMA doorbell size in dword */
uint32_t sdma_doorbell_range;
@@ -312,6 +309,33 @@ typedef enum _AMDGPU_DOORBELL64_ASSIGNMENT
AMDGPU_DOORBELL64_INVALID = 0xFFFF
} AMDGPU_DOORBELL64_ASSIGNMENT;
+typedef enum _AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1
+{
+ /* KIQ: 0~7 for maximum 8 XCD */
+ AMDGPU_DOORBELL_LAYOUT1_KIQ_START = 0x000,
+ AMDGPU_DOORBELL_LAYOUT1_HIQ = 0x008,
+ AMDGPU_DOORBELL_LAYOUT1_DIQ = 0x009,
+ /* Compute: 0x0A ~ 0x49 */
+ AMDGPU_DOORBELL_LAYOUT1_MEC_RING_START = 0x00A,
+ AMDGPU_DOORBELL_LAYOUT1_MEC_RING_END = 0x049,
+ AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START = 0x04A,
+ AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END = 0x0C9,
+ /* SDMA: 0x100 ~ 0x19F */
+ AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START = 0x100,
+ AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_END = 0x19F,
+ /* IH: 0x1A0 ~ 0x1AF */
+ AMDGPU_DOORBELL_LAYOUT1_IH = 0x1A0,
+ /* VCN: 0x1B0 ~ 0x1C2 */
+ AMDGPU_DOORBELL_LAYOUT1_VCN_START = 0x1B0,
+ AMDGPU_DOORBELL_LAYOUT1_VCN_END = 0x1C2,
+
+ AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START,
+ AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_VCN_END,
+
+ AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT = 0x1C2,
+ AMDGPU_DOORBELL_LAYOUT1_INVALID = 0xFFFF
+} AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1;
+
u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index);
void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v);
u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 49b18cf987aa..76438f197de1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -315,11 +315,7 @@ int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
ring->use_doorbell = true;
ring->xcc_id = xcc_id;
ring->vm_hub = AMDGPU_GFXHUB(xcc_id);
- if (xcc_id >= 1)
- ring->doorbell_index = (adev->doorbell_index.xcc1_kiq_start +
- xcc_id - 1) << 1;
- else
- ring->doorbell_index = adev->doorbell_index.kiq << 1;
+ ring->doorbell_index = (adev->doorbell_index.kiq + xcc_id) << 1;
r = amdgpu_gfx_kiq_acquire(adev, ring, xcc_id);
if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
index a9fab8de29e8..52185b1d5d31 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -817,13 +817,7 @@ static int gfx_v9_4_3_compute_ring_init(struct amdgpu_device *adev, int ring_id,
ring->ring_obj = NULL;
ring->use_doorbell = true;
- if (xcc_id >= 1)
- ring->doorbell_index =
- (adev->doorbell_index.xcc1_mec_ring0_start +
- ring_id - adev->gfx.num_compute_rings) << 1;
- else
- ring->doorbell_index =
- (adev->doorbell_index.mec_ring0 + ring_id) << 1;
+ ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
+ (ring_id * GFX9_MEC_HPD_SIZE);
ring->vm_hub = AMDGPU_GFXHUB(xcc_id);
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
index d3c7f9a43ef1..9b53174925f8 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
@@ -1310,14 +1310,7 @@ static int sdma_v4_4_2_sw_init(void *handle)
ring->use_doorbell?"true":"false");
/* doorbell size is 2 dwords, get DWORD offset */
- if (aid_id > 0)
- ring->doorbell_index =
- (adev->doorbell_index.aid1_sdma_start << 1)
- + adev->doorbell_index.sdma_doorbell_range
- * (i - adev->sdma.num_inst_per_aid);
- else
- ring->doorbell_index =
- adev->doorbell_index.sdma_engine[i] << 1;
+ ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1;
ring->vm_hub = AMDGPU_MMHUB0(aid_id);
sprintf(ring->name, "sdma%d.%d", aid_id,
@@ -1336,14 +1329,8 @@ static int sdma_v4_4_2_sw_init(void *handle)
/* doorbell index of page queue is assigned right after
* gfx queue on the same instance
*/
- if (aid_id > 0)
- ring->doorbell_index =
- ((adev->doorbell_index.aid1_sdma_start + 1) << 1)
- + adev->doorbell_index.sdma_doorbell_range
- * (i - adev->sdma.num_inst_per_aid);
- else
- ring->doorbell_index =
- (adev->doorbell_index.sdma_engine[i] + 1) << 1;
+ ring->doorbell_index =
+ (adev->doorbell_index.sdma_engine[i] + 1) << 1;
ring->vm_hub = AMDGPU_MMHUB0(aid_id);
sprintf(ring->name, "page%d.%d", aid_id,
--
cgit v1.2.3
From 45ed97ad36b744dfc2754c47cfd4423aab6322a2 Mon Sep 17 00:00:00 2001
From: James Zhu <James.Zhu@amd.com>
Date: Mon, 12 Dec 2022 12:29:04 -0500
Subject: drm/amdgpu: increase MAX setting to hold more jpeg instances
vcn_v4_0_3 increased jpeg instances,
need increasing MAX resources setting accordlingly.
Signed-off-by: James Zhu <James.Zhu@amd.com>
Acked-by Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h | 6 +++---
drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 2 +-
3 files changed, 5 insertions(+), 5 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
index a29a018ec84e..3c3ae2b4dbc8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
@@ -325,14 +325,14 @@ typedef enum _AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1
AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_END = 0x19F,
/* IH: 0x1A0 ~ 0x1AF */
AMDGPU_DOORBELL_LAYOUT1_IH = 0x1A0,
- /* VCN: 0x1B0 ~ 0x1C2 */
+ /* VCN: 0x1B0 ~ 0x1D4 */
AMDGPU_DOORBELL_LAYOUT1_VCN_START = 0x1B0,
- AMDGPU_DOORBELL_LAYOUT1_VCN_END = 0x1C2,
+ AMDGPU_DOORBELL_LAYOUT1_VCN_END = 0x1D4,
AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START,
AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_VCN_END,
- AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT = 0x1C2,
+ AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT = 0x1D4,
AMDGPU_DOORBELL_LAYOUT1_INVALID = 0xFFFF
} AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
index 5c200a508fa3..bb700a2b97c2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.h
@@ -26,7 +26,7 @@
#include "amdgpu_ras.h"
-#define AMDGPU_MAX_JPEG_INSTANCES 2
+#define AMDGPU_MAX_JPEG_INSTANCES 4
#define AMDGPU_MAX_JPEG_RINGS 8
#define AMDGPU_JPEG_HARVEST_JPEG0 (1 << 0)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
index 8f4b416a92e6..bed118d7add0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
@@ -37,7 +37,7 @@ struct amdgpu_job;
struct amdgpu_vm;
/* max number of rings */
-#define AMDGPU_MAX_RINGS 96
+#define AMDGPU_MAX_RINGS 102
#define AMDGPU_MAX_HWIP_RINGS 32
#define AMDGPU_MAX_GFX_RINGS 2
#define AMDGPU_MAX_SW_GFX_RINGS 2
--
cgit v1.2.3
From 233bb3733bd43966696f4a5e95129476e86bf4e3 Mon Sep 17 00:00:00 2001
From: Lijo Lazar <lijo.lazar@amd.com>
Date: Thu, 19 Jan 2023 14:47:22 +0530
Subject: drm/amdgpu: Use unique doorbell range per xcc
Program different ranges in each XCC with MEC_DOORBELL_RANGE_LOWER/HIGHER.
Keeping the same range causes CPF in other XCCs also to be busy when an IB
packet is submitted to KCQ. Only the XCC which processes the packet
comes back to idle afterwards and this causes other CPs not be idle.
This in turn affects clockgating behavior as RLC doesn't get idle
interrupt.
LOWER/HIGHER covers only KIQ/KCQs which are per XCC queues. Assigning
different ranges doesn't seem to have any side effect as user queue ranges
are outside of this range. User queue tests - PM4 through KFD and AQL
through rocr - have the same results after this change.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h | 35 ++++++++++++----------
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 5 +++-
.../gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c | 1 +
drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c | 31 +++++++++++++------
4 files changed, 47 insertions(+), 25 deletions(-)
(limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h')
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
index 3c3ae2b4dbc8..f637574644c0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h
@@ -86,6 +86,8 @@ struct amdgpu_doorbell_index {
uint32_t max_assignment;
/* Per engine SDMA doorbell size in dword */
uint32_t sdma_doorbell_range;
+ /* Per xcc doorbell size for KIQ/KCQ */
+ uint32_t xcc_doorbell_range;
};
typedef enum _AMDGPU_DOORBELL_ASSIGNMENT
@@ -309,28 +311,31 @@ typedef enum _AMDGPU_DOORBELL64_ASSIGNMENT
AMDGPU_DOORBELL64_INVALID = 0xFFFF
} AMDGPU_DOORBELL64_ASSIGNMENT;
-typedef enum _AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1
-{
- /* KIQ: 0~7 for maximum 8 XCD */
- AMDGPU_DOORBELL_LAYOUT1_KIQ_START = 0x000,
- AMDGPU_DOORBELL_LAYOUT1_HIQ = 0x008,
- AMDGPU_DOORBELL_LAYOUT1_DIQ = 0x009,
- /* Compute: 0x0A ~ 0x49 */
- AMDGPU_DOORBELL_LAYOUT1_MEC_RING_START = 0x00A,
- AMDGPU_DOORBELL_LAYOUT1_MEC_RING_END = 0x049,
- AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START = 0x04A,
- AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END = 0x0C9,
+typedef enum _AMDGPU_DOORBELL_ASSIGNMENT_LAYOUT1 {
+ /* XCC0: 0x00 ~20, XCC1: 20 ~ 2F ... */
+
+ /* KIQ/HIQ/DIQ */
+ AMDGPU_DOORBELL_LAYOUT1_KIQ_START = 0x000,
+ AMDGPU_DOORBELL_LAYOUT1_HIQ = 0x001,
+ AMDGPU_DOORBELL_LAYOUT1_DIQ = 0x002,
+ /* Compute: 0x08 ~ 0x20 */
+ AMDGPU_DOORBELL_LAYOUT1_MEC_RING_START = 0x008,
+ AMDGPU_DOORBELL_LAYOUT1_MEC_RING_END = 0x00F,
+ AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START = 0x010,
+ AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END = 0x01F,
+ AMDGPU_DOORBELL_LAYOUT1_XCC_RANGE = 0x020,
+
/* SDMA: 0x100 ~ 0x19F */
- AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START = 0x100,
- AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_END = 0x19F,
+ AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START = 0x100,
+ AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_END = 0x19F,
/* IH: 0x1A0 ~ 0x1AF */
AMDGPU_DOORBELL_LAYOUT1_IH = 0x1A0,
/* VCN: 0x1B0 ~ 0x1D4 */
AMDGPU_DOORBELL_LAYOUT1_VCN_START = 0x1B0,
AMDGPU_DOORBELL_LAYOUT1_VCN_END = 0x1D4,
- AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START,
- AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_VCN_END,
+ AMDGPU_DOORBELL_LAYOUT1_FIRST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_sDMA_ENGINE_START,
+ AMDGPU_DOORBELL_LAYOUT1_LAST_NON_CP = AMDGPU_DOORBELL_LAYOUT1_VCN_END,
AMDGPU_DOORBELL_LAYOUT1_MAX_ASSIGNMENT = 0x1D4,
AMDGPU_DOORBELL_LAYOUT1_INVALID = 0xFFFF
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index f895a4b8ca0d..70c6099353b8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -316,7 +316,10 @@ int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
ring->use_doorbell = true;
ring->xcc_id = xcc_id;
ring->vm_hub = AMDGPU_GFXHUB(xcc_id);
- ring->doorbell_index = (adev->doorbell_index.kiq + xcc_id) << 1;
+ ring->doorbell_index =
+ (adev->doorbell_index.kiq +
+ xcc_id * adev->doorbell_index.xcc_doorbell_range)
+ << 1;
r = amdgpu_gfx_kiq_acquire(adev, ring, xcc_id);
if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c
index 6591d39c6518..55a6ebb940ba 100644
--- a/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c
+++ b/drivers/gpu/drm/amd/amdgpu/aqua_vanjaram_reg_init.c
@@ -44,6 +44,7 @@ void aqua_vanjaram_doorbell_index_init(struct amdgpu_device *adev)
adev->doorbell_index.userqueue_start = AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_START;
adev->doorbell_index.userqueue_end = AMDGPU_DOORBELL_LAYOUT1_USERQUEUE_END;
+ adev->doorbell_index.xcc_doorbell_range = AMDGPU_DOORBELL_LAYOUT1_XCC_RANGE;
adev->doorbell_index.sdma_doorbell_range = 20;
for (i = 0; i < adev->sdma.num_instances; i++)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
index 55d99c4ea48c..557a2458ef5e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -729,8 +729,10 @@ static int gfx_v9_4_3_compute_ring_init(struct amdgpu_device *adev, int ring_id,
unsigned irq_type;
struct amdgpu_ring *ring = &adev->gfx.compute_ring[ring_id];
unsigned int hw_prio;
+ uint32_t xcc_doorbell_start;
- ring = &adev->gfx.compute_ring[ring_id];
+ ring = &adev->gfx.compute_ring[xcc_id * adev->gfx.num_compute_rings +
+ ring_id];
/* mec0 is me1 */
ring->xcc_id = xcc_id;
@@ -740,9 +742,12 @@ static int gfx_v9_4_3_compute_ring_init(struct amdgpu_device *adev, int ring_id,
ring->ring_obj = NULL;
ring->use_doorbell = true;
- ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
- ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
- + (ring_id * GFX9_MEC_HPD_SIZE);
+ xcc_doorbell_start = adev->doorbell_index.mec_ring0 +
+ xcc_id * adev->doorbell_index.xcc_doorbell_range;
+ ring->doorbell_index = (xcc_doorbell_start + ring_id) << 1;
+ ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr +
+ (ring_id + xcc_id * adev->gfx.num_compute_rings) *
+ GFX9_MEC_HPD_SIZE;
ring->vm_hub = AMDGPU_GFXHUB(xcc_id);
sprintf(ring->name, "comp_%d.%d.%d.%d",
ring->xcc_id, ring->me, ring->pipe, ring->queue);
@@ -801,8 +806,8 @@ static int gfx_v9_4_3_sw_init(void *handle)
}
/* set up the compute queues - allocate horizontally across pipes */
- ring_id = 0;
for (xcc_id = 0; xcc_id < num_xcc; xcc_id++) {
+ ring_id = 0;
for (i = 0; i < adev->gfx.mec.num_mec; ++i) {
for (j = 0; j < adev->gfx.mec.num_queue_per_pipe; j++) {
for (k = 0; k < adev->gfx.mec.num_pipe_per_mec;
@@ -1654,10 +1659,18 @@ static int gfx_v9_4_3_xcc_kiq_init_register(struct amdgpu_ring *ring,
/* enable the doorbell if requested */
if (ring->use_doorbell) {
- WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEC_DOORBELL_RANGE_LOWER,
- (adev->doorbell_index.kiq * 2) << 2);
- WREG32_SOC15(GC, GET_INST(GC, xcc_id), regCP_MEC_DOORBELL_RANGE_UPPER,
- (adev->doorbell_index.userqueue_end * 2) << 2);
+ WREG32_SOC15(
+ GC, GET_INST(GC, xcc_id),
+ regCP_MEC_DOORBELL_RANGE_LOWER,
+ ((adev->doorbell_index.kiq +
+ xcc_id * adev->doorbell_index.xcc_doorbell_range) *
+ 2) << 2);
+ WREG32_SOC15(
+ GC, GET_INST(GC, xcc_id),
+ regCP_MEC_DOORBELL_RANGE_UPPER,
+ ((adev->doorbell_index.userqueue_end +
+ xcc_id * adev->doorbell_index.xcc_doorbell_range) *
+ 2) << 2);
}
WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regCP_HQD_PQ_DOORBELL_CONTROL,
--
cgit v1.2.3