<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=us-ascii"> <meta name="Generator" content="Microsoft Word 15 (filtered medium)"> <style></style> </head> <body lang="EN-US" link="#467886" vlink="#96607D" style="word-wrap:break-word"> [Public] <div> <div class="WordSection1"> It seems that there is no reason to write the same register bitfield separately as no HW sequence required that. Moreover, there is a similar program sequence in the KFD which directly programs the RLC_CP_SCHEDULERS register once.<o:p></o:p> <o:p> </o:p> <div> Thanks,<o:p></o:p> Prike<o:p></o:p> </div> <o:p> </o:p> <div style="border:none;border-left:solid blue 1.5pt;padding:0in 0in 0in 4.0pt"> <div> <div style="border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0in 0in 0in"> From: Deucher, Alexander <Alexander.Deucher@amd.com> Sent: Wednesday, November 20, 2024 2:29 AM To: Liang, Prike <Prike.Liang@amd.com>; amd-gfx@lists.freedesktop.org Subject: Re: [PATCH] drm/amdgpu: reduce the mmio writes in kiq setting<o:p></o:p> </div> </div> <o:p> </o:p> [Public]<o:p></o:p> <o:p> </o:p> <div> <div> As long as there is no hw programming sequence requirement to do it as two separate writes this looks fine to me.<o:p></o:p> </div> <div> Acked-by: Alex Deucher <<a href="mailto:alexander.deucher@amd.com">alexander.deucher@amd.com</a>><o:p></o:p> </div> <div class="MsoNormal" align="center" style="text-align:center"> <hr size="2" width="98%" align="center"> </div> <div id="divRplyFwdMsg"> From: Liang, Prike <<a href="mailto:Prike.Liang@amd.com">Prike.Liang@amd.com</a>> Sent: Friday, November 15, 2024 3:39 AM To: <a href="mailto:amd-gfx@lists.freedesktop.org">amd-gfx@lists.freedesktop.org</a> <<a href="mailto:amd-gfx@lists.freedesktop.org">amd-gfx@lists.freedesktop.org</a>> Cc: Deucher, Alexander <<a href="mailto:Alexander.Deucher@amd.com">Alexander.Deucher@amd.com</a>>; Liang, Prike <<a href="mailto:Prike.Liang@amd.com">Prike.Liang@amd.com</a>> Subject: [PATCH] drm/amdgpu: reduce the mmio writes in kiq setting <o:p> </o:p> <div> <o:p></o:p> </div> </div> <div> <div> There's no need to perform the two MMIO writes in the KIQ Setting registers programmed period, and reducing the MMIO writes will save the driver loading time. Signed-off-by: Prike Liang <<a href="mailto:Prike.Liang@amd.com">Prike.Liang@amd.com</a>> --- drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 8 ++------ drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c | 4 +--- drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c | 4 +--- drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 4 +--- drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 4 +--- drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c | 4 +--- drivers/gpu/drm/amd/amdgpu/mes_v11_0.c | 4 +--- drivers/gpu/drm/amd/amdgpu/mes_v12_0.c | 4 +--- 8 files changed, 9 insertions(+), 27 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index 9da95b25e158..ffd3c45b4ccd 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -6593,17 +6593,13 @@ static void gfx_v10_0_kiq_setting(struct amdgpu_ring *ring) tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp); - tmp |= 0x80; - WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp); + WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS_Sienna_Cichlid, tmp | 0x80); break; default: tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp); + WREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS, tmp | 0x80); break; } } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 5aff8f72de9c..a2aedcabae65 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -3890,9 +3890,7 @@ static void gfx_v11_0_kiq_setting(struct amdgpu_ring *ring) tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); + WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80); } static void gfx_v11_0_cp_set_doorbell_range(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c index 9fec28d8a5fc..c61d383c45c4 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c @@ -2826,9 +2826,7 @@ static void gfx_v12_0_kiq_setting(struct amdgpu_ring *ring) tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); + WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80); } static void gfx_v12_0_cp_set_doorbell_range(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 480c41ee947e..d465c3ea6e6c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -4304,9 +4304,7 @@ static void gfx_v8_0_kiq_setting(struct amdgpu_ring *ring) tmp = RREG32(mmRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32(mmRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32(mmRLC_CP_SCHEDULERS, tmp); + WREG32(mmRLC_CP_SCHEDULERS, tmp | 0x80); } static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index e9248a855ba7..a6cb30558f2f 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -3482,9 +3482,7 @@ static void gfx_v9_0_kiq_setting(struct amdgpu_ring *ring) tmp = RREG32_SOC15(GC, 0, mmRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp); + WREG32_SOC15_RLC(GC, 0, mmRLC_CP_SCHEDULERS, tmp | 0x80); } static void gfx_v9_0_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c index 016290f00592..ae3c8645633b 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c @@ -1771,9 +1771,7 @@ static void gfx_v9_4_3_xcc_kiq_setting(struct amdgpu_ring *ring, int xcc_id) tmp = RREG32_SOC15(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp); + WREG32_SOC15_RLC(GC, GET_INST(GC, xcc_id), regRLC_CP_SCHEDULERS, tmp | 0x80); } static void gfx_v9_4_3_mqd_set_priority(struct amdgpu_ring *ring, struct v9_mqd *mqd) diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index 8ca137313961..3b818441c7ba 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -1491,9 +1491,7 @@ static void mes_v11_0_kiq_setting(struct amdgpu_ring *ring) tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); + WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80); } static void mes_v11_0_kiq_clear(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c index 9d0e342a2f81..44b0d0f76944 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c @@ -1453,9 +1453,7 @@ static void mes_v12_0_kiq_setting(struct amdgpu_ring *ring) tmp = RREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS); tmp &= 0xffffff00; tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue); - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); - tmp |= 0x80; - WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp); + WREG32_SOC15(GC, 0, regRLC_CP_SCHEDULERS, tmp | 0x80); } static int mes_v12_0_kiq_hw_init(struct amdgpu_device *adev) -- 2.34.1<o:p></o:p> </div> </div> </div> </div> </div> </div> </body> </html>