Skip to contentSkip to conversation

Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

RFC In review Andres Rodriguez 5 participants 14 messages Apr 3, 2017 – Apr 5, 2017 13 messages not in the archive

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

RFC 1/1 Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the truncated version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that is the case.

Sending this mostly as an FYI to get the patch removed from the 4.12-wip

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +5 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev)
 		return r;
 	}
 	/* set resources */
-	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 2));
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
 	amdgpu_ring_write(kiq_ring, 0);	/* vmid_mask:0 queue_type:0 (KIQ) */
 	amdgpu_ring_write(kiq_ring, 0x000000FF);	/* queue mask lo */
 	amdgpu_ring_write(kiq_ring, 0);	/* queue mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask lo */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
+	amdgpu_ring_write(kiq_ring, 0);	/* gds heap base:0, gds heap size:0 */
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
 		struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
 		uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
Signature · 6 lines
2.9.3

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 20170403202534.3311-1-andresx7-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

#2 lore raw
Quoted 26 lines · Andres Rodriguez

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the truncated version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that is the case.

Sending this mostly as an FYI to get the patch removed from the 4.12-wip

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev)
 		return r;
 	}
 	/* set resources */
-	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 2));
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev) · line 4660

Quoted 10 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
 	amdgpu_ring_write(kiq_ring, 0);	/* vmid_mask:0 queue_type:0 (KIQ) */
 	amdgpu_ring_write(kiq_ring, 0x000000FF);	/* queue mask lo */
 	amdgpu_ring_write(kiq_ring, 0);	/* queue mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask lo */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
+	amdgpu_ring_write(kiq_ring, 0);	/* gds heap base:0, gds heap size:0 */
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
 		struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
 		uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU clock staying at 1200MHz the entire time.

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 3ae98951-f12f-3860-e10a-00e0be2b9e84-5C7GfCeVMHo@public.gmane.org

#4 lore raw
Quoted 37 lines · Tom St Denis
Quoted 36 lines · Andres Rodriguez

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the truncated version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that is the case.

Sending this mostly as an FYI to get the patch removed from the 4.12-wip

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
         return r;
     }
     /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 2));
+    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
     amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0 queue_type:0

(KIQ) */
amdgpu_ring_write(kiq_ring, 0x000000FF); /* queue mask lo */
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds heap

size:0 */

Quoted 9 lines
Quoted 4 lines
     for (i = 0; i < adev->gfx.num_compute_rings; i++) {
         struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
         uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU clock staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Quoted 5 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): d4fbb81c-be3a-c54a-6d56-17dc4ae27895-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

#6 lore raw
Quoted 53 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]] On Behalf Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

Quoted 16 lines · Tom St Denis
Quoted 15 lines · Andres Rodriguez

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the truncated version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that is the case.

Sending this mostly as an FYI to get the patch removed from the 4.12-wip

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 12 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 12 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
         return r;
     }
     /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 2));
+    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

6));

Quoted 14 lines
Quoted 13 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
     amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0 queue_type:0

(KIQ) */
amdgpu_ring_write(kiq_ring, 0x000000FF); /* queue mask lo */
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds heap

size:0 */

     for (i = 0; i < adev->gfx.num_compute_rings; i++) {
         struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
         uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
Quoted 8 lines
Quoted 4 lines

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU clock staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 11 lines
Quoted 7 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): BN6PR12MB1652D90B12B8A60F1F725AC0F70B0-/b2+HYfkarQqUD6E6FAiowdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org

#8 lore raw

I'm using the latest from all open (our internal FC24 repo). I'll fetch the exact later with "umr -c".

Tom

Quoted 76 lines · Alexander
Quoted 9 lines

-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org] On Behalf Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

Quoted 4 lines
Quoted 3 lines · Tom St Denis
Quoted 2 lines · Andres Rodriguez

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

truncated

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that is the case.

Sending this mostly as an FYI to get the patch removed from the

4.12-wip

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
 drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 29 lines
Quoted 4 lines
Quoted 4 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
index 2d8d7bb..4c3bf51 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
Quoted 8 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 8 lines

+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -1
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
         return r;
     }
     /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 2));
+    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

6));

Quoted 16 lines
Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
     amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0 queue_type:0

(KIQ) */
amdgpu_ring_write(kiq_ring, 0x000000FF); /* queue mask lo */
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds heap

size:0 */

Quoted 4 lines
     for (i = 0; i < adev->gfx.num_compute_rings; i++) {
         struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
         uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU

clock

Quoted 6 lines
Quoted 2 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 7 lines
Quoted 7 lines

Tom
_______________________________________________
amd-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Quoted 9 lines
Quoted 4 lines

_______________________________________________
amd-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): CAAzXoR+TJHk1Eo_h0HJGg-KGfON-uLWE7e27Q-95PXyBY3zcOw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#10 lore raw

This is my info in case that is useful:

    umr.version == ea8e49bb15ed

    fw.VCE == .feature==0 .firmware==0x34040300
    fw.UVD == .feature==0 .firmware==0x014f1000
    fw.MC == .feature==0 .firmware==0x00000000
    fw.ME == .feature==37 .firmware==0x00000094
    fw.PFP == .feature==37 .firmware==0x000000dc
    fw.CE == .feature==37 .firmware==0x00000080
    fw.RLC == .feature==1 .firmware==0x0000010e
    fw.MEC == .feature==37 .firmware==0x000002a2
    fw.MEC2 == .feature==37 .firmware==0x000002a2
    fw.SOS == .feature==0 .firmware==0x00000000
    fw.ASD == .feature==0 .firmware==0x00000000
    fw.SMC == .feature==0 .firmware==0x00170f00
    fw.SDMA0 == .feature==31 .firmware==0x00000036
    fw.SDMA1 == .feature==0 .firmware==0x00000036

    asic.instance == 0


    gfx.max_shader_engines == 4
    gfx.max_tile_pipes == 8
    gfx.max_cu_per_sh == 9
    gfx.max_sh_per_se == 1
    gfx.max_backends_per_se == 2
    gfx.max_texture_channel_caches == 8
    gfx.max_gprs == 256
    gfx.max_gs_threads == 32
    gfx.max_hw_contexts == 8
    gfx.sc_prim_fifo_size_frontend == 32
    gfx.sc_prim_fifo_size_backend == 256
    gfx.sc_hiz_tile_fifo_size == 48
    gfx.sc_earlyz_tile_fifo_size == 304
    gfx.num_tile_pipes == 8
    gfx.backend_enable_mask == 255
    gfx.mem_max_burst_length_bytes == 256
    gfx.mem_row_size_in_kb == 4
    gfx.shader_engine_tile_size == 32
    gfx.num_gpus == 1
    gfx.multi_gpu_tile_size == 64
    gfx.mc_arb_ramcfg == 24738
    gfx.gb_addr_config == 570494979
    gfx.num_rbs == 8
    gfx.family = 130, Volcanic Islands
    gfx.rev_id == 00000001
    gfx.external_rev_id == 00000051
    gfx.cg_flags == 003fffcd
        AMD_CG_SUPPORT_GFX_MGCG
        AMD_CG_SUPPORT_GFX_CGCG
        AMD_CG_SUPPORT_GFX_CGLS
        AMD_CG_SUPPORT_GFX_CP_LS
        AMD_CG_SUPPORT_GFX_RLC_LS
        AMD_CG_SUPPORT_MC_LS
        AMD_CG_SUPPORT_MC_MGCG
        AMD_CG_SUPPORT_SDMA_LS
        AMD_CG_SUPPORT_SDMA_MGCG
        AMD_CG_SUPPORT_BIF_LS
        AMD_CG_SUPPORT_UVD_MGCG
        AMD_CG_SUPPORT_VCE_MGCG
        AMD_CG_SUPPORT_HDP_LS
        AMD_CG_SUPPORT_HDP_MGCG
        AMD_CG_SUPPORT_ROM_MGCG
    gfx.pg_flags == 00000000

Regards,
Andres

Quoted 88 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch the exact later with "umr -c".

Tom

Quoted 80 lines · Alexander
Quoted 9 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]] On Behalf Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

Quoted 4 lines
Quoted 3 lines · Tom St Denis
Quoted 2 lines · Andres Rodriguez

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

truncated

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that is the case.

Sending this mostly as an FYI to get the patch removed from the

4.12-wip

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 28 lines
Quoted 4 lines
Quoted 4 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
index 2d8d7bb..4c3bf51 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
Quoted 8 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 8 lines

+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -1
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 2));
+    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

6));

Quoted 15 lines
Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0 queue_type:0

(KIQ) */
amdgpu_ring_write(kiq_ring, 0x000000FF); /* queue mask lo */
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds heap

size:0 */

Quoted 4 lines
      for (i = 0; i < adev->gfx.num_compute_rings; i++) {
          struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
          uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU

clock

Quoted 5 lines
Quoted 2 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 10 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 32c1f7e8-54a4-8db7-c764-8bae55bf3f94-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

#12 lore raw
Quoted 64 lines

-----Original Message-----
From: Andres Rodriguez [mailto:[email protected]]
Sent: Tuesday, April 04, 2017 4:01 PM
To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

This is my info in case that is useful:

     umr.version == ea8e49bb15ed

     fw.VCE == .feature==0 .firmware==0x34040300
     fw.UVD == .feature==0 .firmware==0x014f1000
     fw.MC == .feature==0 .firmware==0x00000000
     fw.ME == .feature==37 .firmware==0x00000094
     fw.PFP == .feature==37 .firmware==0x000000dc
     fw.CE == .feature==37 .firmware==0x00000080
     fw.RLC == .feature==1 .firmware==0x0000010e
     fw.MEC == .feature==37 .firmware==0x000002a2
     fw.MEC2 == .feature==37 .firmware==0x000002a2
     fw.SOS == .feature==0 .firmware==0x00000000
     fw.ASD == .feature==0 .firmware==0x00000000
     fw.SMC == .feature==0 .firmware==0x00170f00
     fw.SDMA0 == .feature==31 .firmware==0x00000036
     fw.SDMA1 == .feature==0 .firmware==0x00000036

     asic.instance == 0


     gfx.max_shader_engines == 4
     gfx.max_tile_pipes == 8
     gfx.max_cu_per_sh == 9
     gfx.max_sh_per_se == 1
     gfx.max_backends_per_se == 2
     gfx.max_texture_channel_caches == 8
     gfx.max_gprs == 256
     gfx.max_gs_threads == 32
     gfx.max_hw_contexts == 8
     gfx.sc_prim_fifo_size_frontend == 32
     gfx.sc_prim_fifo_size_backend == 256
     gfx.sc_hiz_tile_fifo_size == 48
     gfx.sc_earlyz_tile_fifo_size == 304
     gfx.num_tile_pipes == 8
     gfx.backend_enable_mask == 255
     gfx.mem_max_burst_length_bytes == 256
     gfx.mem_row_size_in_kb == 4
     gfx.shader_engine_tile_size == 32
     gfx.num_gpus == 1
     gfx.multi_gpu_tile_size == 64
     gfx.mc_arb_ramcfg == 24738
     gfx.gb_addr_config == 570494979
     gfx.num_rbs == 8
     gfx.family = 130, Volcanic Islands
     gfx.rev_id == 00000001
     gfx.external_rev_id == 00000051
     gfx.cg_flags == 003fffcd
         AMD_CG_SUPPORT_GFX_MGCG
         AMD_CG_SUPPORT_GFX_CGCG
         AMD_CG_SUPPORT_GFX_CGLS
         AMD_CG_SUPPORT_GFX_CP_LS
         AMD_CG_SUPPORT_GFX_RLC_LS
         AMD_CG_SUPPORT_MC_LS
         AMD_CG_SUPPORT_MC_MGCG
         AMD_CG_SUPPORT_SDMA_LS
Quoted 8 lines
         AMD_CG_SUPPORT_SDMA_MGCG
         AMD_CG_SUPPORT_BIF_LS
         AMD_CG_SUPPORT_UVD_MGCG
         AMD_CG_SUPPORT_VCE_MGCG
         AMD_CG_SUPPORT_HDP_LS
         AMD_CG_SUPPORT_HDP_MGCG
         AMD_CG_SUPPORT_ROM_MGCG
     gfx.pg_flags == 00000000

Is this the stock firmware from upstream or the stuff I pointed you to yesterday on my fdo site?

Alex

Quoted 56 lines

Regards,
Andres

Quoted 6 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch the exact later with "umr -c".

Tom

On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander

<[email protected]>

Quoted 4 lines
Quoted 2 lines · wrote:
Quoted 2 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]] On

Behalf

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · Tom St Denis
Quoted 8 lines
Quoted 8 lines
Quoted 3 lines
Quoted 3 lines · Andres Rodriguez
Quoted 2 lines

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

truncated

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that

is

Quoted 9 lines
Quoted 9 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

the case.

Sending this mostly as an FYI to get the patch removed from the

4.12-wip

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 11 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

2));

Quoted 43 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 1 line
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -0
+    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 41 lines
Quoted 41 lines
Quoted 16 lines

6));

Quoted 15 lines
Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0 queue_type:0

(KIQ) */
amdgpu_ring_write(kiq_ring, 0x000000FF); /* queue mask lo */
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds heap

size:0 */

Quoted 4 lines
      for (i = 0; i < adev->gfx.num_compute_rings; i++) {
          struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
          uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU

clock

Quoted 5 lines
Quoted 2 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 10 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): BN6PR12MB1652C9E0D15D794E02FAB2ADF70B0-/b2+HYfkarQqUD6E6FAiowdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org

#14 lore raw

This should be the the linux-firmware repository @ commit 6d3bc8886517d171068fd1263176b8b5c51df204

I reverted back to that firmware since I didn't want to deal with possible bugs when testing my patches.

Regards,
Andres

Quoted 73 lines · Alexander
Quoted 64 lines

-----Original Message-----
From: Andres Rodriguez [mailto:[email protected]]
Sent: Tuesday, April 04, 2017 4:01 PM
To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

This is my info in case that is useful:

     umr.version == ea8e49bb15ed

     fw.VCE == .feature==0 .firmware==0x34040300
     fw.UVD == .feature==0 .firmware==0x014f1000
     fw.MC == .feature==0 .firmware==0x00000000
     fw.ME == .feature==37 .firmware==0x00000094
     fw.PFP == .feature==37 .firmware==0x000000dc
     fw.CE == .feature==37 .firmware==0x00000080
     fw.RLC == .feature==1 .firmware==0x0000010e
     fw.MEC == .feature==37 .firmware==0x000002a2
     fw.MEC2 == .feature==37 .firmware==0x000002a2
     fw.SOS == .feature==0 .firmware==0x00000000
     fw.ASD == .feature==0 .firmware==0x00000000
     fw.SMC == .feature==0 .firmware==0x00170f00
     fw.SDMA0 == .feature==31 .firmware==0x00000036
     fw.SDMA1 == .feature==0 .firmware==0x00000036

     asic.instance == 0


     gfx.max_shader_engines == 4
     gfx.max_tile_pipes == 8
     gfx.max_cu_per_sh == 9
     gfx.max_sh_per_se == 1
     gfx.max_backends_per_se == 2
     gfx.max_texture_channel_caches == 8
     gfx.max_gprs == 256
     gfx.max_gs_threads == 32
     gfx.max_hw_contexts == 8
     gfx.sc_prim_fifo_size_frontend == 32
     gfx.sc_prim_fifo_size_backend == 256
     gfx.sc_hiz_tile_fifo_size == 48
     gfx.sc_earlyz_tile_fifo_size == 304
     gfx.num_tile_pipes == 8
     gfx.backend_enable_mask == 255
     gfx.mem_max_burst_length_bytes == 256
     gfx.mem_row_size_in_kb == 4
     gfx.shader_engine_tile_size == 32
     gfx.num_gpus == 1
     gfx.multi_gpu_tile_size == 64
     gfx.mc_arb_ramcfg == 24738
     gfx.gb_addr_config == 570494979
     gfx.num_rbs == 8
     gfx.family = 130, Volcanic Islands
     gfx.rev_id == 00000001
     gfx.external_rev_id == 00000051
     gfx.cg_flags == 003fffcd
         AMD_CG_SUPPORT_GFX_MGCG
         AMD_CG_SUPPORT_GFX_CGCG
         AMD_CG_SUPPORT_GFX_CGLS
         AMD_CG_SUPPORT_GFX_CP_LS
         AMD_CG_SUPPORT_GFX_RLC_LS
         AMD_CG_SUPPORT_MC_LS
         AMD_CG_SUPPORT_MC_MGCG
         AMD_CG_SUPPORT_SDMA_LS
Quoted 8 lines
         AMD_CG_SUPPORT_SDMA_MGCG
         AMD_CG_SUPPORT_BIF_LS
         AMD_CG_SUPPORT_UVD_MGCG
         AMD_CG_SUPPORT_VCE_MGCG
         AMD_CG_SUPPORT_HDP_LS
         AMD_CG_SUPPORT_HDP_MGCG
         AMD_CG_SUPPORT_ROM_MGCG
     gfx.pg_flags == 00000000
Quoted 104 lines

Is this the stock firmware from upstream or the stuff I pointed you to yesterday on my fdo site?

Alex

Quoted 99 lines

Regards,
Andres

Quoted 6 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch the exact later with "umr -c".

Tom

On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander

<[email protected]>

Quoted 4 lines
Quoted 2 lines · wrote:
Quoted 2 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]] On

Behalf

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · Tom St Denis
Quoted 8 lines
Quoted 8 lines
Quoted 3 lines
Quoted 3 lines · Andres Rodriguez
Quoted 2 lines

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

truncated

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that

is

Quoted 9 lines
Quoted 9 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

the case.

Sending this mostly as an FYI to get the patch removed from the

4.12-wip

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 11 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

2));

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 1 line
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -0
+    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 41 lines
Quoted 41 lines
Quoted 16 lines

6));

Quoted 15 lines
Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0 queue_type:0

(KIQ) */
amdgpu_ring_write(kiq_ring, 0x000000FF); /* queue mask lo */
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds heap

size:0 */

Quoted 4 lines
      for (i = 0; i < adev->gfx.num_compute_rings; i++) {
          struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
          uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU

clock

Quoted 5 lines
Quoted 2 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 10 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): CAFQ_0eH8wU1NOfj5TXR+SKWAbFcRqqS4mK-xEFWuSG82V-zozg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#16 lore raw

My firmware is

        fw.VCE == .feature==0 .firmware==0x34040300
        fw.UVD == .feature==0 .firmware==0x015b0b00
        fw.MC == .feature==0 .firmware==0x00000000
        fw.ME == .feature==46 .firmware==0x000000a1
        fw.PFP == .feature==46 .firmware==0x000000eb
        fw.CE == .feature==46 .firmware==0x00000086
        fw.RLC == .feature==1 .firmware==0x0000009c
        fw.MEC == .feature==46 .firmware==0x000002c1
        fw.MEC2 == .feature==46 .firmware==0x000002c1
        fw.SOS == .feature==0 .firmware==0x00000000
        fw.ASD == .feature==0 .firmware==0x00000000
        fw.SMC == .feature==0 .firmware==0x00000000
        fw.SDMA0 == .feature==0 .firmware==0x00000022
        fw.SDMA1 == .feature==0 .firmware==0x00000022

It seems like our all-open is behind what he has for SDMA, SMC, RLC but ahead for others? Bizarre.

Tom

Quoted 188 lines · Andres Rodriguez

This should be the the linux-firmware repository @ commit 6d3bc8886517d171068fd1263176b8b5c51df204

I reverted back to that firmware since I didn't want to deal with possible bugs when testing my patches.

Regards,
Andres

Quoted 177 lines · Alexander
Quoted 64 lines

-----Original Message-----
From: Andres Rodriguez [mailto:[email protected]]
Sent: Tuesday, April 04, 2017 4:01 PM
To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

This is my info in case that is useful:

     umr.version == ea8e49bb15ed

     fw.VCE == .feature==0 .firmware==0x34040300
     fw.UVD == .feature==0 .firmware==0x014f1000
     fw.MC == .feature==0 .firmware==0x00000000
     fw.ME == .feature==37 .firmware==0x00000094
     fw.PFP == .feature==37 .firmware==0x000000dc
     fw.CE == .feature==37 .firmware==0x00000080
     fw.RLC == .feature==1 .firmware==0x0000010e
     fw.MEC == .feature==37 .firmware==0x000002a2
     fw.MEC2 == .feature==37 .firmware==0x000002a2
     fw.SOS == .feature==0 .firmware==0x00000000
     fw.ASD == .feature==0 .firmware==0x00000000
     fw.SMC == .feature==0 .firmware==0x00170f00
     fw.SDMA0 == .feature==31 .firmware==0x00000036
     fw.SDMA1 == .feature==0 .firmware==0x00000036

     asic.instance == 0


     gfx.max_shader_engines == 4
     gfx.max_tile_pipes == 8
     gfx.max_cu_per_sh == 9
     gfx.max_sh_per_se == 1
     gfx.max_backends_per_se == 2
     gfx.max_texture_channel_caches == 8
     gfx.max_gprs == 256
     gfx.max_gs_threads == 32
     gfx.max_hw_contexts == 8
     gfx.sc_prim_fifo_size_frontend == 32
     gfx.sc_prim_fifo_size_backend == 256
     gfx.sc_hiz_tile_fifo_size == 48
     gfx.sc_earlyz_tile_fifo_size == 304
     gfx.num_tile_pipes == 8
     gfx.backend_enable_mask == 255
     gfx.mem_max_burst_length_bytes == 256
     gfx.mem_row_size_in_kb == 4
     gfx.shader_engine_tile_size == 32
     gfx.num_gpus == 1
     gfx.multi_gpu_tile_size == 64
     gfx.mc_arb_ramcfg == 24738
     gfx.gb_addr_config == 570494979
     gfx.num_rbs == 8
     gfx.family = 130, Volcanic Islands
     gfx.rev_id == 00000001
     gfx.external_rev_id == 00000051
     gfx.cg_flags == 003fffcd
         AMD_CG_SUPPORT_GFX_MGCG
         AMD_CG_SUPPORT_GFX_CGCG
         AMD_CG_SUPPORT_GFX_CGLS
         AMD_CG_SUPPORT_GFX_CP_LS
         AMD_CG_SUPPORT_GFX_RLC_LS
         AMD_CG_SUPPORT_MC_LS
         AMD_CG_SUPPORT_MC_MGCG
         AMD_CG_SUPPORT_SDMA_LS
Quoted 8 lines
         AMD_CG_SUPPORT_SDMA_MGCG
         AMD_CG_SUPPORT_BIF_LS
         AMD_CG_SUPPORT_UVD_MGCG
         AMD_CG_SUPPORT_VCE_MGCG
         AMD_CG_SUPPORT_HDP_LS
         AMD_CG_SUPPORT_HDP_MGCG
         AMD_CG_SUPPORT_ROM_MGCG
     gfx.pg_flags == 00000000

Is this the stock firmware from upstream or the stuff I pointed you to yesterday on my fdo site?

Alex

Quoted 99 lines

Regards,
Andres

Quoted 6 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch the exact later with "umr -c".

Tom

On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander

<[email protected]>

Quoted 4 lines
Quoted 2 lines · wrote:
Quoted 2 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]] On

Behalf

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · Tom St Denis
Quoted 8 lines
Quoted 8 lines
Quoted 3 lines
Quoted 3 lines · Andres Rodriguez
Quoted 2 lines

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

truncated

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that

is

Quoted 9 lines
Quoted 9 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

the case.

Sending this mostly as an FYI to get the patch removed from the

4.12-wip

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 11 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

2));

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 1 line
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -0
+    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 41 lines
Quoted 41 lines
Quoted 16 lines

6));

Quoted 15 lines
Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0 queue_type:0

(KIQ) */
amdgpu_ring_write(kiq_ring, 0x000000FF); /* queue mask lo */
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds heap

size:0 */

Quoted 4 lines
      for (i = 0; i < adev->gfx.num_compute_rings; i++) {
          struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
          uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock problem (CP is pegged at 100% as read by umr) resulting in the GPU

clock

Quoted 5 lines
Quoted 2 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 10 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 08b6ee6a-f201-272c-d812-7f7404a51863-5C7GfCeVMHo@public.gmane.org

#18 lore raw
Quoted 17 lines · Tom St Denis

My firmware is

        fw.VCE == .feature==0 .firmware==0x34040300
        fw.UVD == .feature==0 .firmware==0x015b0b00
        fw.MC == .feature==0 .firmware==0x00000000
        fw.ME == .feature==46 .firmware==0x000000a1
        fw.PFP == .feature==46 .firmware==0x000000eb
        fw.CE == .feature==46 .firmware==0x00000086
        fw.RLC == .feature==1 .firmware==0x0000009c
        fw.MEC == .feature==46 .firmware==0x000002c1
        fw.MEC2 == .feature==46 .firmware==0x000002c1
        fw.SOS == .feature==0 .firmware==0x00000000
        fw.ASD == .feature==0 .firmware==0x00000000
        fw.SMC == .feature==0 .firmware==0x00000000
        fw.SDMA0 == .feature==0 .firmware==0x00000022
        fw.SDMA1 == .feature==0 .firmware==0x00000022
Quoted 2 lines

It seems like our all-open is behind what he has for SDMA, SMC, RLC but ahead for others? Bizarre.

Can you try the upstream ucode? That is what Andres is using.
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git

Alex

Quoted 228 lines

Tom

Quoted 4 lines · Andres Rodriguez

This should be the the linux-firmware repository @ commit 6d3bc8886517d171068fd1263176b8b5c51df204

Quoted 218 lines

I reverted back to that firmware since I didn't want to deal with possible bugs when testing my patches.

Regards,
Andres

Quoted 210 lines · Alexander
Quoted 65 lines

-----Original Message-----
From: Andres Rodriguez [mailto:[email protected]]
Sent: Tuesday, April 04, 2017 4:01 PM
To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

This is my info in case that is useful:

     umr.version == ea8e49bb15ed

     fw.VCE == .feature==0 .firmware==0x34040300
     fw.UVD == .feature==0 .firmware==0x014f1000
     fw.MC == .feature==0 .firmware==0x00000000
     fw.ME == .feature==37 .firmware==0x00000094
     fw.PFP == .feature==37 .firmware==0x000000dc
     fw.CE == .feature==37 .firmware==0x00000080
     fw.RLC == .feature==1 .firmware==0x0000010e
     fw.MEC == .feature==37 .firmware==0x000002a2
     fw.MEC2 == .feature==37 .firmware==0x000002a2
     fw.SOS == .feature==0 .firmware==0x00000000
     fw.ASD == .feature==0 .firmware==0x00000000
     fw.SMC == .feature==0 .firmware==0x00170f00
     fw.SDMA0 == .feature==31 .firmware==0x00000036
     fw.SDMA1 == .feature==0 .firmware==0x00000036

     asic.instance == 0


     gfx.max_shader_engines == 4
     gfx.max_tile_pipes == 8
     gfx.max_cu_per_sh == 9
     gfx.max_sh_per_se == 1
     gfx.max_backends_per_se == 2
     gfx.max_texture_channel_caches == 8
     gfx.max_gprs == 256
     gfx.max_gs_threads == 32
     gfx.max_hw_contexts == 8
     gfx.sc_prim_fifo_size_frontend == 32
     gfx.sc_prim_fifo_size_backend == 256
     gfx.sc_hiz_tile_fifo_size == 48
     gfx.sc_earlyz_tile_fifo_size == 304
     gfx.num_tile_pipes == 8
     gfx.backend_enable_mask == 255
     gfx.mem_max_burst_length_bytes == 256
     gfx.mem_row_size_in_kb == 4
     gfx.shader_engine_tile_size == 32
     gfx.num_gpus == 1
     gfx.multi_gpu_tile_size == 64
     gfx.mc_arb_ramcfg == 24738
     gfx.gb_addr_config == 570494979
     gfx.num_rbs == 8
     gfx.family = 130, Volcanic Islands
     gfx.rev_id == 00000001
     gfx.external_rev_id == 00000051
     gfx.cg_flags == 003fffcd
         AMD_CG_SUPPORT_GFX_MGCG
         AMD_CG_SUPPORT_GFX_CGCG
         AMD_CG_SUPPORT_GFX_CGLS
         AMD_CG_SUPPORT_GFX_CP_LS
         AMD_CG_SUPPORT_GFX_RLC_LS
         AMD_CG_SUPPORT_MC_LS
         AMD_CG_SUPPORT_MC_MGCG
         AMD_CG_SUPPORT_SDMA_LS
Quoted 8 lines
         AMD_CG_SUPPORT_SDMA_MGCG
         AMD_CG_SUPPORT_BIF_LS
         AMD_CG_SUPPORT_UVD_MGCG
         AMD_CG_SUPPORT_VCE_MGCG
         AMD_CG_SUPPORT_HDP_LS
         AMD_CG_SUPPORT_HDP_MGCG
         AMD_CG_SUPPORT_ROM_MGCG
     gfx.pg_flags == 00000000

Is this the stock firmware from upstream or the stuff I pointed you to yesterday on my fdo site?

Alex

Quoted 129 lines

Regards,
Andres

Quoted 8 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch
the exact later with "umr -c".

Tom

On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander

<[email protected]>

Quoted 5 lines
Quoted 2 lines · wrote:
Quoted 2 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]] On

Behalf

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 11 lines
Quoted 11 lines
Quoted 10 lines · Tom St Denis
Quoted 5 lines · Andres Rodriguez
Quoted 3 lines

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

Quoted 6 lines
Quoted 6 lines

truncated

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that

is

Quoted 12 lines
Quoted 12 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

the case.

Sending this mostly as an FYI to get the patch removed from the

4.12-wip

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 11 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

2));

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 2 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -0
+    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 52 lines
Quoted 52 lines
Quoted 22 lines

6));

Quoted 20 lines
Quoted 13 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0

queue_type:0
(KIQ) */

      amdgpu_ring_write(kiq_ring, 0x000000FF);    /* queue mask lo

*/
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds

heap
size:0 */

Quoted 4 lines
      for (i = 0; i < adev->gfx.num_compute_rings; i++) {
          struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
          uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock
problem (CP is pegged at 100% as read by umr) resulting in the GPU

clock

Quoted 6 lines
Quoted 3 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 11 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Quoted 4 lines

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): CADnq5_O7KZy08QQ2P0mqGMkOdFtFYRh5jrJx_Gfm5Ahxvse8Vw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#20 lore raw
Quoted 18 lines · Alex Deucher
Quoted 17 lines · Tom St Denis

My firmware is

        fw.VCE == .feature==0 .firmware==0x34040300
        fw.UVD == .feature==0 .firmware==0x015b0b00
        fw.MC == .feature==0 .firmware==0x00000000
        fw.ME == .feature==46 .firmware==0x000000a1
        fw.PFP == .feature==46 .firmware==0x000000eb
        fw.CE == .feature==46 .firmware==0x00000086
        fw.RLC == .feature==1 .firmware==0x0000009c
        fw.MEC == .feature==46 .firmware==0x000002c1
        fw.MEC2 == .feature==46 .firmware==0x000002c1
        fw.SOS == .feature==0 .firmware==0x00000000
        fw.ASD == .feature==0 .firmware==0x00000000
        fw.SMC == .feature==0 .firmware==0x00000000
        fw.SDMA0 == .feature==0 .firmware==0x00000022
        fw.SDMA1 == .feature==0 .firmware==0x00000022
Quoted 5 lines
Quoted 2 lines

It seems like our all-open is behind what he has for SDMA, SMC, RLC but ahead for others? Bizarre.

Can you try the upstream ucode? That is what Andres is using.
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git

Copying the tip of master from that to my /lib/firmware/ results in the same firmware versions being loaded.

Quoted 235 lines

Alex

Quoted 232 lines

Tom

Quoted 4 lines · Andres Rodriguez

This should be the the linux-firmware repository @ commit 6d3bc8886517d171068fd1263176b8b5c51df204

Quoted 218 lines

I reverted back to that firmware since I didn't want to deal with possible bugs when testing my patches.

Regards,
Andres

Quoted 210 lines · Alexander
Quoted 65 lines

-----Original Message-----
From: Andres Rodriguez [mailto:[email protected]]
Sent: Tuesday, April 04, 2017 4:01 PM
To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

This is my info in case that is useful:

     umr.version == ea8e49bb15ed

     fw.VCE == .feature==0 .firmware==0x34040300
     fw.UVD == .feature==0 .firmware==0x014f1000
     fw.MC == .feature==0 .firmware==0x00000000
     fw.ME == .feature==37 .firmware==0x00000094
     fw.PFP == .feature==37 .firmware==0x000000dc
     fw.CE == .feature==37 .firmware==0x00000080
     fw.RLC == .feature==1 .firmware==0x0000010e
     fw.MEC == .feature==37 .firmware==0x000002a2
     fw.MEC2 == .feature==37 .firmware==0x000002a2
     fw.SOS == .feature==0 .firmware==0x00000000
     fw.ASD == .feature==0 .firmware==0x00000000
     fw.SMC == .feature==0 .firmware==0x00170f00
     fw.SDMA0 == .feature==31 .firmware==0x00000036
     fw.SDMA1 == .feature==0 .firmware==0x00000036

     asic.instance == 0


     gfx.max_shader_engines == 4
     gfx.max_tile_pipes == 8
     gfx.max_cu_per_sh == 9
     gfx.max_sh_per_se == 1
     gfx.max_backends_per_se == 2
     gfx.max_texture_channel_caches == 8
     gfx.max_gprs == 256
     gfx.max_gs_threads == 32
     gfx.max_hw_contexts == 8
     gfx.sc_prim_fifo_size_frontend == 32
     gfx.sc_prim_fifo_size_backend == 256
     gfx.sc_hiz_tile_fifo_size == 48
     gfx.sc_earlyz_tile_fifo_size == 304
     gfx.num_tile_pipes == 8
     gfx.backend_enable_mask == 255
     gfx.mem_max_burst_length_bytes == 256
     gfx.mem_row_size_in_kb == 4
     gfx.shader_engine_tile_size == 32
     gfx.num_gpus == 1
     gfx.multi_gpu_tile_size == 64
     gfx.mc_arb_ramcfg == 24738
     gfx.gb_addr_config == 570494979
     gfx.num_rbs == 8
     gfx.family = 130, Volcanic Islands
     gfx.rev_id == 00000001
     gfx.external_rev_id == 00000051
     gfx.cg_flags == 003fffcd
         AMD_CG_SUPPORT_GFX_MGCG
         AMD_CG_SUPPORT_GFX_CGCG
         AMD_CG_SUPPORT_GFX_CGLS
         AMD_CG_SUPPORT_GFX_CP_LS
         AMD_CG_SUPPORT_GFX_RLC_LS
         AMD_CG_SUPPORT_MC_LS
         AMD_CG_SUPPORT_MC_MGCG
         AMD_CG_SUPPORT_SDMA_LS
Quoted 8 lines
         AMD_CG_SUPPORT_SDMA_MGCG
         AMD_CG_SUPPORT_BIF_LS
         AMD_CG_SUPPORT_UVD_MGCG
         AMD_CG_SUPPORT_VCE_MGCG
         AMD_CG_SUPPORT_HDP_LS
         AMD_CG_SUPPORT_HDP_MGCG
         AMD_CG_SUPPORT_ROM_MGCG
     gfx.pg_flags == 00000000

Is this the stock firmware from upstream or the stuff I pointed you to yesterday on my fdo site?

Alex

Quoted 129 lines

Regards,
Andres

Quoted 8 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch
the exact later with "umr -c".

Tom

On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander

<[email protected]>

Quoted 5 lines
Quoted 2 lines · wrote:
Quoted 2 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]] On

Behalf

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 11 lines
Quoted 11 lines
Quoted 10 lines · Tom St Denis
Quoted 5 lines · Andres Rodriguez
Quoted 3 lines

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

Quoted 6 lines
Quoted 6 lines

truncated

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the KCQ suspend/resume errors. I *think* it may no longer be needed if that

is

Quoted 12 lines
Quoted 12 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

the case.

Sending this mostly as an FYI to get the patch removed from the

4.12-wip

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

This reverts commit e579f564c333a7eaddf7e12b1b8414410c36e80f.

---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 11 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int gfx_v8_0_kiq_kcq_enable(struct
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES,

2));

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int gfx_v8_0_kiq_kcq_enable(struct · line 4651

Quoted 2 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -0
+    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 52 lines
Quoted 52 lines
Quoted 22 lines

6));

Quoted 20 lines
Quoted 13 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0

queue_type:0
(KIQ) */

      amdgpu_ring_write(kiq_ring, 0x000000FF);    /* queue mask lo

*/
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds

heap
size:0 */

Quoted 4 lines
      for (i = 0; i < adev->gfx.num_compute_rings; i++) {
          struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
          uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);

This fixes init issues with my polaris board but there's still a clock
problem (CP is pegged at 100% as read by umr) resulting in the GPU

clock

Quoted 6 lines
Quoted 3 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as you mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 11 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 682c9bb0-19d6-f64a-84d1-73dd77824aa4-5C7GfCeVMHo@public.gmane.org

#22 lore raw
Quoted 31 lines

-----Original Message-----
From: StDenis, Tom
Sent: Wednesday, April 05, 2017 9:14 AM
To: Alex Deucher
Cc: Andres Rodriguez; Deucher, Alexander; Tom St Denis; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

Quoted 1 line · Alex Deucher

On Wed, Apr 5, 2017 at 7:30 AM, Tom St Denis <[email protected]>

Quoted 20 lines · wrote:
Quoted 19 lines

My firmware is

        fw.VCE == .feature==0 .firmware==0x34040300
        fw.UVD == .feature==0 .firmware==0x015b0b00
        fw.MC == .feature==0 .firmware==0x00000000
        fw.ME == .feature==46 .firmware==0x000000a1
        fw.PFP == .feature==46 .firmware==0x000000eb
        fw.CE == .feature==46 .firmware==0x00000086
        fw.RLC == .feature==1 .firmware==0x0000009c
        fw.MEC == .feature==46 .firmware==0x000002c1
        fw.MEC2 == .feature==46 .firmware==0x000002c1
        fw.SOS == .feature==0 .firmware==0x00000000
        fw.ASD == .feature==0 .firmware==0x00000000
        fw.SMC == .feature==0 .firmware==0x00000000
        fw.SDMA0 == .feature==0 .firmware==0x00000022
        fw.SDMA1 == .feature==0 .firmware==0x00000022

It seems like our all-open is behind what he has for SDMA, SMC, RLC but ahead for others? Bizarre.

Quoted 7 lines
Quoted 2 lines

Can you try the upstream ucode? That is what Andres is using.
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-

firmware.git

Copying the tip of master from that to my /lib/firmware/ results in the same firmware versions being loaded.

Same as you or same as Andres?

Alex

Quoted 211 lines
Quoted 27 lines

Alex

Quoted 24 lines

Tom

Quoted 19 lines · Andres Rodriguez

This should be the the linux-firmware repository @ commit 6d3bc8886517d171068fd1263176b8b5c51df204

I reverted back to that firmware since I didn't want to deal with possible bugs when testing my patches.

Regards,
Andres

Quoted 7 lines · Alexander
Quoted 7 lines

-----Original Message-----
From: Andres Rodriguez [mailto:[email protected]]
Sent: Tuesday, April 04, 2017 4:01 PM
To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 85 lines
Quoted 85 lines
Quoted 85 lines
Quoted 85 lines
Quoted 58 lines

This is my info in case that is useful:

     umr.version == ea8e49bb15ed

     fw.VCE == .feature==0 .firmware==0x34040300
     fw.UVD == .feature==0 .firmware==0x014f1000
     fw.MC == .feature==0 .firmware==0x00000000
     fw.ME == .feature==37 .firmware==0x00000094
     fw.PFP == .feature==37 .firmware==0x000000dc
     fw.CE == .feature==37 .firmware==0x00000080
     fw.RLC == .feature==1 .firmware==0x0000010e
     fw.MEC == .feature==37 .firmware==0x000002a2
     fw.MEC2 == .feature==37 .firmware==0x000002a2
     fw.SOS == .feature==0 .firmware==0x00000000
     fw.ASD == .feature==0 .firmware==0x00000000
     fw.SMC == .feature==0 .firmware==0x00170f00
     fw.SDMA0 == .feature==31 .firmware==0x00000036
     fw.SDMA1 == .feature==0 .firmware==0x00000036

     asic.instance == 0


     gfx.max_shader_engines == 4
     gfx.max_tile_pipes == 8
     gfx.max_cu_per_sh == 9
     gfx.max_sh_per_se == 1
     gfx.max_backends_per_se == 2
     gfx.max_texture_channel_caches == 8
     gfx.max_gprs == 256
     gfx.max_gs_threads == 32
     gfx.max_hw_contexts == 8
     gfx.sc_prim_fifo_size_frontend == 32
     gfx.sc_prim_fifo_size_backend == 256
     gfx.sc_hiz_tile_fifo_size == 48
     gfx.sc_earlyz_tile_fifo_size == 304
     gfx.num_tile_pipes == 8
     gfx.backend_enable_mask == 255
     gfx.mem_max_burst_length_bytes == 256
     gfx.mem_row_size_in_kb == 4
     gfx.shader_engine_tile_size == 32
     gfx.num_gpus == 1
     gfx.multi_gpu_tile_size == 64
     gfx.mc_arb_ramcfg == 24738
     gfx.gb_addr_config == 570494979
     gfx.num_rbs == 8
     gfx.family = 130, Volcanic Islands
     gfx.rev_id == 00000001
     gfx.external_rev_id == 00000051
     gfx.cg_flags == 003fffcd
         AMD_CG_SUPPORT_GFX_MGCG
         AMD_CG_SUPPORT_GFX_CGCG
         AMD_CG_SUPPORT_GFX_CGLS
         AMD_CG_SUPPORT_GFX_CP_LS
         AMD_CG_SUPPORT_GFX_RLC_LS
         AMD_CG_SUPPORT_MC_LS
         AMD_CG_SUPPORT_MC_MGCG
         AMD_CG_SUPPORT_SDMA_LS
Quoted 8 lines
         AMD_CG_SUPPORT_SDMA_MGCG
         AMD_CG_SUPPORT_BIF_LS
         AMD_CG_SUPPORT_UVD_MGCG
         AMD_CG_SUPPORT_VCE_MGCG
         AMD_CG_SUPPORT_HDP_LS
         AMD_CG_SUPPORT_HDP_MGCG
         AMD_CG_SUPPORT_ROM_MGCG
     gfx.pg_flags == 00000000

Is this the stock firmware from upstream or the stuff I pointed you to yesterday on my fdo site?

Alex

Quoted 12 lines

Regards,
Andres

Quoted 7 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch
the exact later with "umr -c".

Tom

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 1 line

On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander

<[email protected]>

Quoted 5 lines
Quoted 2 lines · wrote:
Quoted 2 lines

-----Original Message-----
From: amd-gfx [mailto:[email protected]]

On

Quoted 17 lines
Quoted 17 lines
Quoted 17 lines
Quoted 17 lines
Quoted 17 lines

Behalf

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines · Tom St Denis
Quoted 3 lines · Andres Rodriguez
Quoted 1 line
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Commit e579f56 results in a KCQ initialization error for polaris10 cards. Providing the full SET_RESOURCES packet instead of the

truncated

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing the

KCQ

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

suspend/resume errors. I *think* it may no longer be needed

if that

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

is

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

the case.

Sending this mostly as an FYI to get the patch removed from

the

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

4.12-wip

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

This reverts commit

e579f564c333a7eaddf7e12b1b8414410c36e80f.

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int

Quoted 10 lines
---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int

gfx_v8_0_kiq_kcq_enable(struct

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int · line 4651

Quoted 5 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

2));

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -0
+    amdgpu_ring_write(kiq_ring,
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

PACKET3(PACKET3_SET_RESOURCES,

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

6));

Quoted 5 lines
Quoted 5 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0

queue_type:0
(KIQ) */

      amdgpu_ring_write(kiq_ring, 0x000000FF);    /* queue mask

lo

Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

*/
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0, gds

heap
size:0 */

      for (i = 0; i < adev->gfx.num_compute_rings; i++) {
          struct amdgpu_ring *ring = &adev->gfx.compute_ring[i];
          uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring-
Quoted 5 lines

mqd_obj);

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 1 line

This fixes init issues with my polaris board but there's still a clock
problem (CP is pegged at 100% as read by umr) resulting in the

GPU

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 36 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

clock

Quoted 5 lines
Quoted 3 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as

you

Quoted 29 lines
Quoted 29 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 1 line

mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 11 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): BN6PR12MB1652D439580DE76681B0EFA7F70A0-/b2+HYfkarQqUD6E6FAiowdYzm3356FpvxpqHgZTriW3zl9H0oFU5g@public.gmane.org

#24 lore raw

Hey Tom,

If it's the same as your old firmware you might be missing an extra step. I usually just run the kernel 'make install' as that usually take care of the distro specific quirks.

Regards,
Andres

Quoted 252 lines · Alexander
Quoted 8 lines

-----Original Message-----
From: StDenis, Tom
Sent: Wednesday, April 05, 2017 9:14 AM
To: Alex Deucher
Cc: Andres Rodriguez; Deucher, Alexander; Tom St Denis; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

Quoted 21 lines
Quoted 1 line · Alex Deucher

On Wed, Apr 5, 2017 at 7:30 AM, Tom St Denis <[email protected]>

Quoted 18 lines · wrote:
Quoted 18 lines

My firmware is

        fw.VCE == .feature==0 .firmware==0x34040300
        fw.UVD == .feature==0 .firmware==0x015b0b00
        fw.MC == .feature==0 .firmware==0x00000000
        fw.ME == .feature==46 .firmware==0x000000a1
        fw.PFP == .feature==46 .firmware==0x000000eb
        fw.CE == .feature==46 .firmware==0x00000086
        fw.RLC == .feature==1 .firmware==0x0000009c
        fw.MEC == .feature==46 .firmware==0x000002c1
        fw.MEC2 == .feature==46 .firmware==0x000002c1
        fw.SOS == .feature==0 .firmware==0x00000000
        fw.ASD == .feature==0 .firmware==0x00000000
        fw.SMC == .feature==0 .firmware==0x00000000
        fw.SDMA0 == .feature==0 .firmware==0x00000022
        fw.SDMA1 == .feature==0 .firmware==0x00000022

It seems like our all-open is behind what he has for SDMA, SMC, RLC

but

Quoted 9 lines
Quoted 4 lines
Quoted 1 line

ahead for others? Bizarre.

Can you try the upstream ucode? That is what Andres is using.
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-

firmware.git

Copying the tip of master from that to my /lib/firmware/ results in the same firmware versions being loaded.

Same as you or same as Andres?

Alex

Quoted 21 lines
Quoted 21 lines

Alex

Quoted 18 lines

Tom

Quoted 13 lines · Andres Rodriguez

This should be the the linux-firmware repository @ commit 6d3bc8886517d171068fd1263176b8b5c51df204

I reverted back to that firmware since I didn't want to deal with possible bugs when testing my patches.

Regards,
Andres

Quoted 1 line · Alexander
Quoted 1 line
Quoted 76 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

-----Original Message-----
From: Andres Rodriguez [mailto:[email protected]]
Sent: Tuesday, April 04, 2017 4:01 PM
To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd- [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 69 lines
Quoted 69 lines
Quoted 69 lines
Quoted 69 lines
Quoted 58 lines

This is my info in case that is useful:

     umr.version == ea8e49bb15ed

     fw.VCE == .feature==0 .firmware==0x34040300
     fw.UVD == .feature==0 .firmware==0x014f1000
     fw.MC == .feature==0 .firmware==0x00000000
     fw.ME == .feature==37 .firmware==0x00000094
     fw.PFP == .feature==37 .firmware==0x000000dc
     fw.CE == .feature==37 .firmware==0x00000080
     fw.RLC == .feature==1 .firmware==0x0000010e
     fw.MEC == .feature==37 .firmware==0x000002a2
     fw.MEC2 == .feature==37 .firmware==0x000002a2
     fw.SOS == .feature==0 .firmware==0x00000000
     fw.ASD == .feature==0 .firmware==0x00000000
     fw.SMC == .feature==0 .firmware==0x00170f00
     fw.SDMA0 == .feature==31 .firmware==0x00000036
     fw.SDMA1 == .feature==0 .firmware==0x00000036

     asic.instance == 0


     gfx.max_shader_engines == 4
     gfx.max_tile_pipes == 8
     gfx.max_cu_per_sh == 9
     gfx.max_sh_per_se == 1
     gfx.max_backends_per_se == 2
     gfx.max_texture_channel_caches == 8
     gfx.max_gprs == 256
     gfx.max_gs_threads == 32
     gfx.max_hw_contexts == 8
     gfx.sc_prim_fifo_size_frontend == 32
     gfx.sc_prim_fifo_size_backend == 256
     gfx.sc_hiz_tile_fifo_size == 48
     gfx.sc_earlyz_tile_fifo_size == 304
     gfx.num_tile_pipes == 8
     gfx.backend_enable_mask == 255
     gfx.mem_max_burst_length_bytes == 256
     gfx.mem_row_size_in_kb == 4
     gfx.shader_engine_tile_size == 32
     gfx.num_gpus == 1
     gfx.multi_gpu_tile_size == 64
     gfx.mc_arb_ramcfg == 24738
     gfx.gb_addr_config == 570494979
     gfx.num_rbs == 8
     gfx.family = 130, Volcanic Islands
     gfx.rev_id == 00000001
     gfx.external_rev_id == 00000051
     gfx.cg_flags == 003fffcd
         AMD_CG_SUPPORT_GFX_MGCG
         AMD_CG_SUPPORT_GFX_CGCG
         AMD_CG_SUPPORT_GFX_CGLS
         AMD_CG_SUPPORT_GFX_CP_LS
         AMD_CG_SUPPORT_GFX_RLC_LS
         AMD_CG_SUPPORT_MC_LS
         AMD_CG_SUPPORT_MC_MGCG
         AMD_CG_SUPPORT_SDMA_LS
Quoted 8 lines
         AMD_CG_SUPPORT_SDMA_MGCG
         AMD_CG_SUPPORT_BIF_LS
         AMD_CG_SUPPORT_UVD_MGCG
         AMD_CG_SUPPORT_VCE_MGCG
         AMD_CG_SUPPORT_HDP_LS
         AMD_CG_SUPPORT_HDP_MGCG
         AMD_CG_SUPPORT_ROM_MGCG
     gfx.pg_flags == 00000000

Is this the stock firmware from upstream or the stuff I pointed you

to

Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines

yesterday on my fdo site?

Alex

Quoted 14 lines

Regards,
Andres

Quoted 8 lines · Tom St Denis

I'm using the latest from all open (our internal FC24 repo). I'll fetch
the exact later with "umr -c".

Tom

On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander

Quoted 25 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

<[email protected]>

Quoted 5 lines
Quoted 2 lines · wrote:
Quoted 2 lines

-----Original Message-----
From: amd-gfx [mailto:amd-gfx-bounces-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org]

On

Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines

Behalf

Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

Of Andres Rodriguez
Sent: Tuesday, April 04, 2017 3:39 PM
To: StDenis, Tom; [email protected]
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES

packet"

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines · Tom St Denis
Quoted 4 lines · Andres Rodriguez
Quoted 2 lines

Commit e579f56 results in a KCQ initialization error for

polaris10

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

cards. Providing the full SET_RESOURCES packet instead of the

truncated

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

version fixes the problem.

I think this patch may have been one of the stabs at fixing

the

Quoted 26 lines

KCQ

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

suspend/resume errors. I *think* it may no longer be needed

if that

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

is

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

the case.

Sending this mostly as an FYI to get the patch removed from

the

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

4.12-wip

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

This reverts commit

e579f564c333a7eaddf7e12b1b8414410c36e80f.

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int

Quoted 10 lines
---
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
index 2d8d7bb..4c3bf51 100644
@@ -4647,10 +4647,14 @@ static int

gfx_v8_0_kiq_kcq_enable(struct

Quoted 19 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int · line 4651

Quoted 5 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -1
 amdgpu_device *adev)
          return r;
      }
      /* set resources */
-    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 13 lines
Quoted 13 lines
Quoted 13 lines
Quoted 13 lines
Quoted 13 lines

2));

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c static int · line 4651

Quoted 2 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1 -0
+    amdgpu_ring_write(kiq_ring,

PACKET3(PACKET3_SET_RESOURCES,

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

6));

Quoted 5 lines
Quoted 5 lines
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0 -0
      amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0

queue_type:0
(KIQ) */

      amdgpu_ring_write(kiq_ring, 0x000000FF);    /* queue

mask

Quoted 7 lines

lo

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

*/
amdgpu_ring_write(kiq_ring, 0); /* queue mask hi */

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +4 -0
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
+    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
+    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
+    amdgpu_ring_write(kiq_ring, 0);    /* gds heap base:0,

gds

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines

heap
size:0 */
for (i = 0; i < adev->gfx.num_compute_rings; i++) {
struct amdgpu_ring *ring =

Quoted 48 lines

&adev->gfx.compute_ring[i];

Quoted 4 lines
Quoted 4 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring-

mqd_obj);

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 1 line

This fixes init issues with my polaris board but there's still

a

Quoted 40 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

clock
problem (CP is pegged at 100% as read by umr) resulting in the

GPU

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

clock

Quoted 5 lines
Quoted 3 lines

staying at 1200MHz the entire time.

Yeah I sent this out because I was having KCQ init issues on my polaris10 card. However, I'm not seeing the CP pegged at 100% as

you

Quoted 29 lines
Quoted 29 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 18 lines
Quoted 1 line

mentioned. Everything sitting quietly at 0% here.

Which firmware are you using?

Alex

Quoted 11 lines
Quoted 6 lines

Tom
_______________________________________________
amd-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

_______________________________________________
amd-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): CAFQ_0eHiU=HNCpPA=2+Op0944anFsb1w-=Q3923LXPY02Pia3Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#26 lore raw
Quoted 5 lines · Andres Rodriguez

Hey Tom,

If it's the same as your old firmware you might be missing an extra step. I usually just run the kernel 'make install' as that usually take care of the distro specific quirks.

I manually load the module but for sanity I did a make install and it didn't change the rlc/smc versions.

Tom

Quoted 315 lines
Quoted 10 lines · On Apr 5, 2017 9:56 AM, "Deucher, Alexander" <[email protected]…

-----Original Message-----
From: StDenis, Tom
Sent: Wednesday, April 05, 2017 9:14 AM
To: Alex Deucher
Cc: Andres Rodriguez; Deucher, Alexander; Tom St Denis; amd- [email protected] <mailto:[email protected]>
Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES packet"

Quoted 1 line · Alex Deucher

On Wed, Apr 5, 2017 at 7:30 AM, Tom St Denis

    <[email protected] <mailto:[email protected]>>
    > wrote:
    > >> My firmware is
    > >>
    > >>         fw.VCE == .feature==0 .firmware==0x34040300
    > >>         fw.UVD == .feature==0 .firmware==0x015b0b00
    > >>         fw.MC == .feature==0 .firmware==0x00000000
    > >>         fw.ME == .feature==46 .firmware==0x000000a1
    > >>         fw.PFP == .feature==46 .firmware==0x000000eb
    > >>         fw.CE == .feature==46 .firmware==0x00000086
    > >>         fw.RLC == .feature==1 .firmware==0x0000009c
    > >>         fw.MEC == .feature==46 .firmware==0x000002c1
    > >>         fw.MEC2 == .feature==46 .firmware==0x000002c1
    > >>         fw.SOS == .feature==0 .firmware==0x00000000
    > >>         fw.ASD == .feature==0 .firmware==0x00000000
    > >>         fw.SMC == .feature==0 .firmware==0x00000000
    > >>         fw.SDMA0 == .feature==0 .firmware==0x00000022
    > >>         fw.SDMA1 == .feature==0 .firmware==0x00000022
    > >>
    > >> It seems like our all-open is behind what he has for SDMA, SMC,
    RLC but
    > >> ahead for others?  Bizarre.
    > >
    > > Can you try the upstream ucode?  That is what Andres is using.
    > > https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-
    <https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux->
    > firmware.git
    >
    > Copying the tip of master from that to my /lib/firmware/ results
    in the
    > same firmware versions being loaded.
    >

    Same as you or same as Andres?

    Alex

    > >
    > > Alex
    > >
    > >>
    > >> Tom
    > >>
    > >>
    > >> On 04/04/17 04:17 PM, Andres Rodriguez wrote:
    > >>>
    > >>> This should be the the linux-firmware repository @ commit
    > >>> 6d3bc8886517d171068fd1263176b8b5c51df204
    > >>>
    > >>> I reverted back to that firmware since I didn't want to deal with
    > >>> possible bugs when testing my patches.
    > >>>
    > >>> Regards,
    > >>> Andres
    > >>>
    > >>> On Tue, Apr 4, 2017 at 4:03 PM, Deucher, Alexander
    > >>> <[email protected] <mailto:[email protected]>>
    wrote:
    > >>>>>
    > >>>>> -----Original Message-----
    > >>>>> From: Andres Rodriguez [mailto:[email protected]
    <mailto:[email protected]>]
    > >>>>> Sent: Tuesday, April 04, 2017 4:01 PM
    > >>>>> To: Tom St Denis; Deucher, Alexander; StDenis, Tom; amd-
    > >>>>> [email protected] <mailto:[email protected]>
    > >>>>> Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES
    > packet"
    > >>>>>
    > >>>>> This is my info in case that is useful:
    > >>>>>
    > >>>>>      umr.version == ea8e49bb15ed
    > >>>>>
    > >>>>>      fw.VCE == .feature==0 .firmware==0x34040300
    > >>>>>      fw.UVD == .feature==0 .firmware==0x014f1000
    > >>>>>      fw.MC == .feature==0 .firmware==0x00000000
    > >>>>>      fw.ME == .feature==37 .firmware==0x00000094
    > >>>>>      fw.PFP == .feature==37 .firmware==0x000000dc
    > >>>>>      fw.CE == .feature==37 .firmware==0x00000080
    > >>>>>      fw.RLC == .feature==1 .firmware==0x0000010e
    > >>>>>      fw.MEC == .feature==37 .firmware==0x000002a2
    > >>>>>      fw.MEC2 == .feature==37 .firmware==0x000002a2
    > >>>>>      fw.SOS == .feature==0 .firmware==0x00000000
    > >>>>>      fw.ASD == .feature==0 .firmware==0x00000000
    > >>>>>      fw.SMC == .feature==0 .firmware==0x00170f00
    > >>>>>      fw.SDMA0 == .feature==31 .firmware==0x00000036
    > >>>>>      fw.SDMA1 == .feature==0 .firmware==0x00000036
    > >>>>>
    > >>>>>      asic.instance == 0
    > >>>>>
    > >>>>>
    > >>>>>      gfx.max_shader_engines == 4
    > >>>>>      gfx.max_tile_pipes == 8
    > >>>>>      gfx.max_cu_per_sh == 9
    > >>>>>      gfx.max_sh_per_se == 1
    > >>>>>      gfx.max_backends_per_se == 2
    > >>>>>      gfx.max_texture_channel_caches == 8
    > >>>>>      gfx.max_gprs == 256
    > >>>>>      gfx.max_gs_threads == 32
    > >>>>>      gfx.max_hw_contexts == 8
    > >>>>>      gfx.sc_prim_fifo_size_frontend == 32
    > >>>>>      gfx.sc_prim_fifo_size_backend == 256
    > >>>>>      gfx.sc_hiz_tile_fifo_size == 48
    > >>>>>      gfx.sc_earlyz_tile_fifo_size == 304
    > >>>>>      gfx.num_tile_pipes == 8
    > >>>>>      gfx.backend_enable_mask == 255
    > >>>>>      gfx.mem_max_burst_length_bytes == 256
    > >>>>>      gfx.mem_row_size_in_kb == 4
    > >>>>>      gfx.shader_engine_tile_size == 32
    > >>>>>      gfx.num_gpus == 1
    > >>>>>      gfx.multi_gpu_tile_size == 64
    > >>>>>      gfx.mc_arb_ramcfg == 24738
    > >>>>>      gfx.gb_addr_config == 570494979
    > >>>>>      gfx.num_rbs == 8
    > >>>>>      gfx.family = 130, Volcanic Islands
    > >>>>>      gfx.rev_id == 00000001
    > >>>>>      gfx.external_rev_id == 00000051
    > >>>>>      gfx.cg_flags == 003fffcd
    > >>>>>          AMD_CG_SUPPORT_GFX_MGCG
    > >>>>>          AMD_CG_SUPPORT_GFX_CGCG
    > >>>>>          AMD_CG_SUPPORT_GFX_CGLS
    > >>>>>          AMD_CG_SUPPORT_GFX_CP_LS
    > >>>>>          AMD_CG_SUPPORT_GFX_RLC_LS
    > >>>>>          AMD_CG_SUPPORT_MC_LS
    > >>>>>          AMD_CG_SUPPORT_MC_MGCG
    > >>>>>          AMD_CG_SUPPORT_SDMA_LS
    > >>>>>          AMD_CG_SUPPORT_SDMA_MGCG
    > >>>>>          AMD_CG_SUPPORT_BIF_LS
    > >>>>>          AMD_CG_SUPPORT_UVD_MGCG
    > >>>>>          AMD_CG_SUPPORT_VCE_MGCG
    > >>>>>          AMD_CG_SUPPORT_HDP_LS
    > >>>>>          AMD_CG_SUPPORT_HDP_MGCG
    > >>>>>          AMD_CG_SUPPORT_ROM_MGCG
    > >>>>>      gfx.pg_flags == 00000000
    > >>>>
    > >>>>
    > >>>> Is this the stock firmware from upstream or the stuff I
    pointed you to
    > >>>> yesterday on my fdo site?
    > >>>>
    > >>>> Alex
    > >>>>
    > >>>>>
    > >>>>> Regards,
    > >>>>> Andres
    > >>>>>
    > >>>>> On 2017-04-04 03:59 PM, Tom St Denis wrote:
    > >>>>>>
    > >>>>>> I'm using the latest from all open (our internal FC24
    repo).  I'll
    > >>>>>> fetch
    > >>>>>> the exact later with "umr -c".
    > >>>>>>
    > >>>>>> Tom
    > >>>>>>
    > >>>>>> On Tue, Apr 4, 2017 at 3:48 PM Deucher, Alexander
    > >>>>>
    > >>>>> <[email protected] <mailto:[email protected]>>
    > >>>>>>
    > >>>>>> wrote:
    > >>>>>>
    > >>>>>>>> -----Original Message-----
    > >>>>>>>> From: amd-gfx
    [mailto:[email protected]
    <mailto:[email protected]>]
    > On
    > >>>>>
    > >>>>> Behalf
    > >>>>>>>>
    > >>>>>>>> Of Andres Rodriguez
    > >>>>>>>> Sent: Tuesday, April 04, 2017 3:39 PM
    > >>>>>>>> To: StDenis, Tom; [email protected]
    <mailto:[email protected]>
    > >>>>>>>> Subject: Re: [RFC] Revert "drm/amdgpu/gfx8: Fix SET_RESOURCES
    > >>>>>
    > >>>>> packet"
    > >>>>>>>>
    > >>>>>>>>
    > >>>>>>>>
    > >>>>>>>>
    > >>>>>>>> On 2017-04-04 08:27 AM, Tom St Denis wrote:
    > >>>>>>>>>
    > >>>>>>>>> On 03/04/17 04:25 PM, Andres Rodriguez wrote:
    > >>>>>>>>>>
    > >>>>>>>>>> Commit e579f56 results in a KCQ initialization error
    for polaris10
    > >>>>>>>>>> cards. Providing the full SET_RESOURCES packet instead
    of the
    > >>>>>>>
    > >>>>>>> truncated
    > >>>>>>>>>>
    > >>>>>>>>>> version fixes the problem.
    > >>>>>>>>>>
    > >>>>>>>>>> I think this patch may have been one of the stabs at
    fixing the
    > KCQ
    > >>>>>>>>>> suspend/resume errors. I *think* it may no longer be needed
    > if that
    > >>>>>
    > >>>>> is
    > >>>>>>>>>>
    > >>>>>>>>>> the case.
    > >>>>>>>>>>
    > >>>>>>>>>> Sending this mostly as an FYI to get the patch removed from
    > the
    > >>>>>>>
    > >>>>>>> 4.12-wip
    > >>>>>>>>>>
    > >>>>>>>>>> This reverts commit
    > e579f564c333a7eaddf7e12b1b8414410c36e80f.
    > >>>>>>>>>> ---
    > >>>>>>>>>>   drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 6 +++++-
    > >>>>>>>>>>   1 file changed, 5 insertions(+), 1 deletion(-)
    > >>>>>>>>>>
    > >>>>>>>>>> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
    > >>>>>>>>>> b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
    > >>>>>>>>>> index 2d8d7bb..4c3bf51 100644
    > >>>>>>>>>> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
    > >>>>>>>>>> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
    > >>>>>>>>>> @@ -4647,10 +4647,14 @@ static int
    > gfx_v8_0_kiq_kcq_enable(struct
    > >>>>>>>>>> amdgpu_device *adev)
    > >>>>>>>>>>           return r;
    > >>>>>>>>>>       }
    > >>>>>>>>>>       /* set resources */
    > >>>>>>>>>> -    amdgpu_ring_write(kiq_ring,
    > PACKET3(PACKET3_SET_RESOURCES,
    > >>>>>
    > >>>>> 2));
    > >>>>>>>>>>
    > >>>>>>>>>> +    amdgpu_ring_write(kiq_ring,
    > >>>>>
    > >>>>> PACKET3(PACKET3_SET_RESOURCES,
    > >>>>>>>>
    > >>>>>>>> 6));
    > >>>>>>>>>>
    > >>>>>>>>>>       amdgpu_ring_write(kiq_ring, 0);    /* vmid_mask:0
    > >>>>>>>>>> queue_type:0
    > >>>>>>>>>> (KIQ) */
    > >>>>>>>>>>       amdgpu_ring_write(kiq_ring, 0x000000FF);    /*
    queue mask
    > lo
    > >>>>>>>>>> */
    > >>>>>>>>>>       amdgpu_ring_write(kiq_ring, 0);    /* queue mask
    hi */
    > >>>>>>>>>> +    amdgpu_ring_write(kiq_ring, 0);    /* gws mask lo */
    > >>>>>>>>>> +    amdgpu_ring_write(kiq_ring, 0);    /* gws mask hi */
    > >>>>>>>>>> +    amdgpu_ring_write(kiq_ring, 0);    /* oac mask */
    > >>>>>>>>>> +    amdgpu_ring_write(kiq_ring, 0);    /* gds heap
    base:0, gds
    > >>>>>>>>>> heap
    > >>>>>>>>>> size:0 */
    > >>>>>>>>>>       for (i = 0; i < adev->gfx.num_compute_rings; i++) {
    > >>>>>>>>>>           struct amdgpu_ring *ring =
    &adev->gfx.compute_ring[i];
    > >>>>>>>>>>           uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring-
    > >mqd_obj);
    > >>>>>>>>>>
    > >>>>>>>>> This fixes init issues with my polaris board but there's
    still a
    > >>>>>>>>> clock
    > >>>>>>>>> problem (CP is pegged at 100% as read by umr) resulting
    in the
    > GPU
    > >>>>>>>
    > >>>>>>> clock
    > >>>>>>>>>
    > >>>>>>>>> staying at 1200MHz the entire time.
    > >>>>>>>>>
    > >>>>>>>> Yeah I sent this out because I was having KCQ init issues
    on my
    > >>>>>>>> polaris10 card. However, I'm not seeing the CP pegged at
    100% as
    > you
    > >>>>>>>> mentioned. Everything sitting quietly at 0% here.
    > >>>>>>>
    > >>>>>>> Which firmware are you using?
    > >>>>>>>
    > >>>>>>> Alex
    > >>>>>>>
    > >>>>>>>>
    > >>>>>>>>> Tom
    > >>>>>>>>> _______________________________________________
    > >>>>>>>>> amd-gfx mailing list
    > >>>>>>>>> [email protected]
    <mailto:[email protected]>
    > >>>>>>>>> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
    <https://lists.freedesktop.org/mailman/listinfo/amd-gfx>
    > >>>>>>>>
    > >>>>>>>> _______________________________________________
    > >>>>>>>> amd-gfx mailing list
    > >>>>>>>> [email protected]
    <mailto:[email protected]>
    > >>>>>>>> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
    <https://lists.freedesktop.org/mailman/listinfo/amd-gfx>
    > >>>>>>>
    > >>>>>>> _______________________________________________
    > >>>>>>> amd-gfx mailing list
    > >>>>>>> [email protected]
    <mailto:[email protected]>
    > >>>>>>> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
    <https://lists.freedesktop.org/mailman/listinfo/amd-gfx>
    > >>>>>>>
Quoted 9 lines
    > >>>>
    > >>
    > >> _______________________________________________
    > >> amd-gfx mailing list
    > >> [email protected]
    <mailto:[email protected]>
    > >> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
    <https://lists.freedesktop.org/mailman/listinfo/amd-gfx>

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx