<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<p style="font-family:Arial;font-size:10pt;color:#008000;margin:15pt;" align="Left">
[Public]<br>
</p>
<br>
<div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
Gotcha,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
Can you add prior to the implementation a small description on how the interface and the different scratch registers work? It may be easier to review with a clear idea of the operation. I know the earlier implementation did not include it, but now that we are
modifying it, it would be good to document it.</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">
<br>
</div>
<div id="appendonsend"></div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>From:</b> Skvortsov, Victor <Victor.Skvortsov@amd.com><br>
<b>Sent:</b> Wednesday, December 15, 2021 11:18 AM<br>
<b>To:</b> Nieto, David M <David.Nieto@amd.com>; amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org>; Deng, Emily <Emily.Deng@amd.com>; Liu, Monk <Monk.Liu@amd.com>; Ming, Davis <Davis.Ming@amd.com>; Liu, Shaoyun <Shaoyun.Liu@amd.com>; Zhou, Peng
Ju <PengJu.Zhou@amd.com>; Chen, JingWen <JingWen.Chen2@amd.com>; Chen, Horace <Horace.Chen@amd.com><br>
<b>Subject:</b> RE: [PATCH 5/5] drm/amdgpu: Modify indirect register access for gfx9 sriov</font>
<div> </div>
</div>
<style>
<!--
@font-face
{font-family:"Cambria Math"}
@font-face
{font-family:Calibri}
@font-face
{font-family:"Segoe UI"}
p.x_MsoNormal, li.x_MsoNormal, div.x_MsoNormal
{margin:0in;
font-size:11.0pt;
font-family:"Calibri",sans-serif}
a:link, span.x_MsoHyperlink
{color:#0563C1;
text-decoration:underline}
span.x_EmailStyle18
{font-family:"Calibri",sans-serif;
color:windowtext}
p.x_msipheadera4477989, li.x_msipheadera4477989, div.x_msipheadera4477989
{margin-right:0in;
margin-left:0in;
font-size:11.0pt;
font-family:"Calibri",sans-serif}
.x_MsoChpDefault
{font-size:10.0pt}
@page WordSection1
{margin:1.0in 1.0in 1.0in 1.0in}
div.x_WordSection1
{}
-->
</style>
<div lang="EN-US" link="#0563C1" vlink="#954F72" style="word-wrap:break-word">
<div class="x_WordSection1">
<p class="x_msipheadera4477989" style="margin:0in"><span style="font-size:10.0pt; font-family:"Arial",sans-serif; color:blue">[AMD Official Use Only]</span></p>
<p class="x_MsoNormal"> </p>
<p class="x_MsoNormal">This was a bug in the original definition, but it functionally it makes no difference (in both cases the macros resolve to the same value).</p>
<p class="x_MsoNormal"> </p>
<div>
<div style="border:none; border-top:solid #E1E1E1 1.0pt; padding:3.0pt 0in 0in 0in">
<p class="x_MsoNormal"><b>From:</b> Nieto, David M <David.Nieto@amd.com> <br>
<b>Sent:</b> Wednesday, December 15, 2021 2:16 PM<br>
<b>To:</b> Skvortsov, Victor <Victor.Skvortsov@amd.com>; amd-gfx@lists.freedesktop.org; Deng, Emily <Emily.Deng@amd.com>; Liu, Monk <Monk.Liu@amd.com>; Ming, Davis <Davis.Ming@amd.com>; Liu, Shaoyun <Shaoyun.Liu@amd.com>; Zhou, Peng Ju <PengJu.Zhou@amd.com>;
Chen, JingWen <JingWen.Chen2@amd.com>; Chen, Horace <Horace.Chen@amd.com><br>
<b>Subject:</b> Re: [PATCH 5/5] drm/amdgpu: Modify indirect register access for gfx9 sriov</p>
</div>
</div>
<p class="x_MsoNormal"> </p>
<p style="margin:5.0pt"><span style="font-size:10.0pt; font-family:"Arial",sans-serif; color:blue">[AMD Official Use Only]</span></p>
<p class="x_MsoNormal"> </p>
<div>
<div>
<p class="x_MsoNormal" style="background:white"><span style="font-family:"Segoe UI",sans-serif; color:black; background:white"> scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4;</span><span style="font-size:12.0pt; color:black"><br>
</span><span style="font-family:"Segoe UI",sans-serif; color:black; background:white"> scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4;</span><span style="font-family:"Segoe UI",sans-serif; color:black"><br>
<span style="background:white">- scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2)*4;</span><br>
<span style="background:white">- scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3)*4;</span><br>
<span style="background:white">+ scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG2_BASE_IDX] + mmSCRATCH_REG2)*4;</span><br>
<span style="background:white">+ scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG3_BASE_IDX] + mmSCRATCH_REG3)*4;</span><br>
<span style="background:white"> spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT)*4;</span></span><span style="font-size:12.0pt; color:black"></span></p>
</div>
<div>
<p class="x_MsoNormal" style="background:white"><span style="font-size:12.0pt; color:black"> </span></p>
</div>
<div>
<p class="x_MsoNormal" style="background:white"><span style="font-family:"Segoe UI",sans-serif; color:black; background:white">the definition of scratch_reg2 and 3 has here.... will this be backwards compatible? Was it a bug in the definition?</span><span style="font-size:12.0pt; color:black"></span></p>
</div>
<div class="x_MsoNormal" align="center" style="text-align:center">
<hr size="1" width="98%" align="center">
</div>
<div id="x_divRplyFwdMsg">
<p class="x_MsoNormal"><b><span style="color:black">From:</span></b><span style="color:black"> Skvortsov, Victor <<a href="mailto:Victor.Skvortsov@amd.com">Victor.Skvortsov@amd.com</a>><br>
<b>Sent:</b> Wednesday, December 15, 2021 10:55 AM<br>
<b>To:</b> <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>>; Deng, Emily <<a href="mailto:Emily.Deng@amd.com">Emily.Deng@amd.com</a>>; Liu, Monk
<<a href="mailto:Monk.Liu@amd.com">Monk.Liu@amd.com</a>>; Ming, Davis <<a href="mailto:Davis.Ming@amd.com">Davis.Ming@amd.com</a>>; Liu, Shaoyun <<a href="mailto:Shaoyun.Liu@amd.com">Shaoyun.Liu@amd.com</a>>; Zhou, Peng Ju <<a href="mailto:PengJu.Zhou@amd.com">PengJu.Zhou@amd.com</a>>;
Chen, JingWen <<a href="mailto:JingWen.Chen2@amd.com">JingWen.Chen2@amd.com</a>>; Chen, Horace <<a href="mailto:Horace.Chen@amd.com">Horace.Chen@amd.com</a>>; Nieto, David M <<a href="mailto:David.Nieto@amd.com">David.Nieto@amd.com</a>><br>
<b>Cc:</b> Skvortsov, Victor <<a href="mailto:Victor.Skvortsov@amd.com">Victor.Skvortsov@amd.com</a>><br>
<b>Subject:</b> [PATCH 5/5] drm/amdgpu: Modify indirect register access for gfx9 sriov</span>
</p>
<div>
<p class="x_MsoNormal"> </p>
</div>
</div>
<div>
<div>
<p class="x_MsoNormal" style="margin-bottom:12.0pt">Expand RLCG interface for new GC read & write commands.<br>
New interface will only be used if the PF enables the flag in pf2vf msg.<br>
<br>
Signed-off-by: Victor Skvortsov <<a href="mailto:victor.skvortsov@amd.com">victor.skvortsov@amd.com</a>><br>
---<br>
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 111 +++++++++++++++++++-------<br>
1 file changed, 83 insertions(+), 28 deletions(-)<br>
<br>
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c<br>
index d252b06efa43..bce6ab52cae0 100644<br>
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c<br>
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c<br>
@@ -63,6 +63,13 @@<br>
#define mmGCEA_PROBE_MAP 0x070c<br>
#define mmGCEA_PROBE_MAP_BASE_IDX 0<br>
<br>
+#define GFX9_RLCG_GC_WRITE_OLD (0x8 << 28)<br>
+#define GFX9_RLCG_GC_WRITE (0x0 << 28)<br>
+#define GFX9_RLCG_GC_READ (0x1 << 28)<br>
+#define GFX9_RLCG_VFGATE_DISABLED 0x4000000<br>
+#define GFX9_RLCG_WRONG_OPERATION_TYPE 0x2000000<br>
+#define GFX9_RLCG_NOT_IN_RANGE 0x1000000<br>
+<br>
MODULE_FIRMWARE("amdgpu/vega10_ce.bin");<br>
MODULE_FIRMWARE("amdgpu/vega10_pfp.bin");<br>
MODULE_FIRMWARE("amdgpu/vega10_me.bin");<br>
@@ -739,7 +746,7 @@ static const u32 GFX_RLC_SRM_INDEX_CNTL_DATA_OFFSETS[] =<br>
mmRLC_SRM_INDEX_CNTL_DATA_7 - mmRLC_SRM_INDEX_CNTL_DATA_0,<br>
};<br>
<br>
-static void gfx_v9_0_rlcg_w(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag)<br>
+static u32 gfx_v9_0_rlcg_rw(struct amdgpu_device *adev, u32 offset, u32 v, uint32_t flag)<br>
{<br>
static void *scratch_reg0;<br>
static void *scratch_reg1;<br>
@@ -748,21 +755,20 @@ static void gfx_v9_0_rlcg_w(struct amdgpu_device *adev, u32 offset, u32 v, u32 f<br>
static void *spare_int;<br>
static uint32_t grbm_cntl;<br>
static uint32_t grbm_idx;<br>
+ uint32_t i = 0;<br>
+ uint32_t retries = 50000;<br>
+ u32 ret = 0;<br>
+ u32 tmp;<br>
<br>
scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4;<br>
scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4;<br>
- scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG2)*4;<br>
- scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG3)*4;<br>
+ scratch_reg2 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG2_BASE_IDX] + mmSCRATCH_REG2)*4;<br>
+ scratch_reg3 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG3_BASE_IDX] + mmSCRATCH_REG3)*4;<br>
spare_int = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmRLC_SPARE_INT_BASE_IDX] + mmRLC_SPARE_INT)*4;<br>
<br>
grbm_cntl = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_CNTL_BASE_IDX] + mmGRBM_GFX_CNTL;<br>
grbm_idx = adev->reg_offset[GC_HWIP][0][mmGRBM_GFX_INDEX_BASE_IDX] + mmGRBM_GFX_INDEX;<br>
<br>
- if (amdgpu_sriov_runtime(adev)) {<br>
- pr_err("shouldn't call rlcg write register during runtime\n");<br>
- return;<br>
- }<br>
-<br>
if (offset == grbm_cntl || offset == grbm_idx) {<br>
if (offset == grbm_cntl)<br>
writel(v, scratch_reg2);<br>
@@ -771,41 +777,89 @@ static void gfx_v9_0_rlcg_w(struct amdgpu_device *adev, u32 offset, u32 v, u32 f<br>
<br>
writel(v, ((void __iomem *)adev->rmmio) + (offset * 4));<br>
} else {<br>
- uint32_t i = 0;<br>
- uint32_t retries = 50000;<br>
-<br>
writel(v, scratch_reg0);<br>
- writel(offset | 0x80000000, scratch_reg1);<br>
+ writel(offset | flag, scratch_reg1);<br>
writel(1, spare_int);<br>
- for (i = 0; i < retries; i++) {<br>
- u32 tmp;<br>
<br>
+ for (i = 0; i < retries; i++) {<br>
tmp = readl(scratch_reg1);<br>
- if (!(tmp & 0x80000000))<br>
+ if (!(tmp & flag))<br>
break;<br>
<br>
udelay(10);<br>
}<br>
- if (i >= retries)<br>
- pr_err("timeout: rlcg program reg:0x%05x failed !\n", offset);<br>
+<br>
+ if (i >= retries) {<br>
+ if (amdgpu_sriov_reg_indirect_gc(adev)) {<br>
+ if (tmp & GFX9_RLCG_VFGATE_DISABLED)<br>
+ pr_err("The vfgate is disabled, program reg:0x%05x failed!\n", offset);<br>
+ else if (tmp & GFX9_RLCG_WRONG_OPERATION_TYPE)<br>
+ pr_err("Wrong operation type, program reg:0x%05x failed!\n", offset);<br>
+ else if (tmp & GFX9_RLCG_NOT_IN_RANGE)<br>
+ pr_err("The register is not in range, program reg:0x%05x failed!\n", offset);<br>
+ else<br>
+ pr_err("Unknown error type, program reg:0x%05x failed!\n", offset);<br>
+ } else<br>
+ pr_err("timeout: rlcg program reg:0x%05x failed!\n", offset);<br>
+ }<br>
+ }<br>
+<br>
+ ret = readl(scratch_reg0);<br>
+<br>
+ return ret;<br>
+}<br>
+<br>
+static bool gfx_v9_0_get_rlcg_flag(struct amdgpu_device *adev, u32 acc_flags, u32 hwip,<br>
+ int write, u32 *rlcg_flag)<br>
+{<br>
+<br>
+ switch (hwip) {<br>
+ case GC_HWIP:<br>
+ if (amdgpu_sriov_reg_indirect_gc(adev)) {<br>
+ *rlcg_flag = write ? GFX9_RLCG_GC_WRITE : GFX9_RLCG_GC_READ;<br>
+<br>
+ return true;<br>
+ /* only in new version, AMDGPU_REGS_NO_KIQ and AMDGPU_REGS_RLC enabled simultaneously */<br>
+ } else if ((acc_flags & AMDGPU_REGS_RLC) && !(acc_flags & AMDGPU_REGS_NO_KIQ) && write) {<br>
+ *rlcg_flag = GFX9_RLCG_GC_WRITE_OLD;<br>
+ return true;<br>
+ }<br>
+<br>
+ break;<br>
+ default:<br>
+ return false;<br>
}<br>
<br>
+ return false;<br>
+}<br>
+<br>
+static u32 gfx_v9_0_sriov_rreg(struct amdgpu_device *adev, u32 offset, u32 acc_flags, u32 hwip)<br>
+{<br>
+ u32 rlcg_flag;<br>
+<br>
+ if (!amdgpu_sriov_runtime(adev) && gfx_v9_0_get_rlcg_flag(adev, acc_flags, hwip, 0, &rlcg_flag))<br>
+ return gfx_v9_0_rlcg_rw(adev, offset, 0, rlcg_flag);<br>
+<br>
+ if (acc_flags & AMDGPU_REGS_NO_KIQ)<br>
+ return RREG32_NO_KIQ(offset);<br>
+ else<br>
+ return RREG32(offset);<br>
}<br>
<br>
static void gfx_v9_0_sriov_wreg(struct amdgpu_device *adev, u32 offset,<br>
- u32 v, u32 acc_flags, u32 hwip)<br>
+ u32 value, u32 acc_flags, u32 hwip)<br>
{<br>
- if ((acc_flags & AMDGPU_REGS_RLC) &&<br>
- amdgpu_sriov_fullaccess(adev)) {<br>
- gfx_v9_0_rlcg_w(adev, offset, v, acc_flags);<br>
+ u32 rlcg_flag;<br>
<br>
+ if (!amdgpu_sriov_runtime(adev) && gfx_v9_0_get_rlcg_flag(adev, acc_flags, hwip, 1, &rlcg_flag)) {<br>
+ gfx_v9_0_rlcg_rw(adev, offset, value, rlcg_flag);<br>
return;<br>
}<br>
<br>
if (acc_flags & AMDGPU_REGS_NO_KIQ)<br>
- WREG32_NO_KIQ(offset, v);<br>
+ WREG32_NO_KIQ(offset, value);<br>
else<br>
- WREG32(offset, v);<br>
+ WREG32(offset, value);<br>
}<br>
<br>
#define VEGA10_GB_ADDR_CONFIG_GOLDEN 0x2a114042<br>
@@ -5134,7 +5188,7 @@ static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)<br>
if (amdgpu_sriov_is_pp_one_vf(adev))<br>
data = RREG32_NO_KIQ(reg);<br>
else<br>
- data = RREG32(reg);<br>
+ data = RREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL);<br>
<br>
data &= ~RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK;<br>
data |= (vmid & RLC_SPM_MC_CNTL__RLC_SPM_VMID_MASK) << RLC_SPM_MC_CNTL__RLC_SPM_VMID__SHIFT;<br>
@@ -5190,6 +5244,7 @@ static const struct amdgpu_rlc_funcs gfx_v9_0_rlc_funcs = {<br>
.start = gfx_v9_0_rlc_start,<br>
.update_spm_vmid = gfx_v9_0_update_spm_vmid,<br>
.sriov_wreg = gfx_v9_0_sriov_wreg,<br>
+ .sriov_rreg = gfx_v9_0_sriov_rreg,<br>
.is_rlcg_access_range = gfx_v9_0_is_rlcg_access_range,<br>
};<br>
<br>
@@ -5795,16 +5850,16 @@ static void gfx_v9_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,<br>
<br>
switch (state) {<br>
case AMDGPU_IRQ_STATE_DISABLE:<br>
- mec_int_cntl = RREG32(mec_int_cntl_reg);<br>
+ mec_int_cntl = RREG32_SOC15_IP(GC,mec_int_cntl_reg);<br>
mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,<br>
TIME_STAMP_INT_ENABLE, 0);<br>
- WREG32(mec_int_cntl_reg, mec_int_cntl);<br>
+ WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);<br>
break;<br>
case AMDGPU_IRQ_STATE_ENABLE:<br>
- mec_int_cntl = RREG32(mec_int_cntl_reg);<br>
+ mec_int_cntl = RREG32_SOC15_IP(GC, mec_int_cntl_reg);<br>
mec_int_cntl = REG_SET_FIELD(mec_int_cntl, CP_ME1_PIPE0_INT_CNTL,<br>
TIME_STAMP_INT_ENABLE, 1);<br>
- WREG32(mec_int_cntl_reg, mec_int_cntl);<br>
+ WREG32_SOC15_IP(GC, mec_int_cntl_reg, mec_int_cntl);<br>
break;<br>
default:<br>
break;<br>
-- <br>
2.25.1</p>
</div>
</div>
</div>
</div>
</div>
</div>
</body>
</html>