[PATCH 1/2] drm/amd/display: Reduce stack size for dml31_ModeSupportAndSystemConfigurationFull
From: Michel Dänzer mdaenzer@redhat.com
Move code using the Pipe struct to a new helper function.
Works around[0] this warning (resulting in failure to build a RHEL debug kernel with Werror enabled):
../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c: In function ‘dml31_ModeSupportAndSystemConfigurationFull’: ../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c:5740:1: warning: the frame size of 2144 bytes is larger than 2048 bytes [-Wframe-larger-than=]
The culprit seems to be the Pipe struct, so pull the relevant block out into its own sub-function. (This is porting a62427ef9b55 "drm/amd/display: Reduce stack size for dml21_ModeSupportAndSystemConfigurationFull" from dml31 to dml21)
[0] AFAICT this doesn't actually reduce the total amount of stack which can be used, just moves some of it from dml31_ModeSupportAndSystemConfigurationFull to the new helper function, so the former happens to no longer exceed the limit for a single function.
Signed-off-by: Michel Dänzer mdaenzer@redhat.com --- .../dc/dml/dcn31/display_mode_vba_31.c | 185 ++++++++++-------- 1 file changed, 99 insertions(+), 86 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index 7e937bdcea00..8965f9af9d0a 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -3949,6 +3949,102 @@ static double TruncToValidBPP( return BPP_INVALID; }
+static noinline void CalculatePrefetchSchedulePerPlane( + struct display_mode_lib *mode_lib, + double HostVMInefficiencyFactor, + int i, + unsigned j, + unsigned k) +{ + struct vba_vars_st *v = &mode_lib->vba; + Pipe myPipe; + + myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; + myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; + myPipe.PixelClock = v->PixelClock[k]; + myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; + myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; + myPipe.ScalerEnabled = v->ScalerEnabled[k]; + myPipe.SourceScan = v->SourceScan[k]; + myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; + myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; + myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; + myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; + myPipe.InterlaceEnable = v->Interlace[k]; + myPipe.NumberOfCursors = v->NumberOfCursors[k]; + myPipe.VBlank = v->VTotal[k] - v->VActive[k]; + myPipe.HTotal = v->HTotal[k]; + myPipe.DCCEnable = v->DCCEnable[k]; + myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 + || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; + myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; + myPipe.BytePerPixelY = v->BytePerPixelY[k]; + myPipe.BytePerPixelC = v->BytePerPixelC[k]; + myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; + v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( + mode_lib, + HostVMInefficiencyFactor, + &myPipe, + v->DSCDelayPerState[i][k], + v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, + v->DPPCLKDelaySCL, + v->DPPCLKDelaySCLLBOnly, + v->DPPCLKDelayCNVCCursor, + v->DISPCLKDelaySubtotal, + v->SwathWidthYThisState[k] / v->HRatio[k], + v->OutputFormat[k], + v->MaxInterDCNTileRepeaters, + dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), + v->MaximumVStartup[i][j][k], + v->GPUVMMaxPageTableLevels, + v->GPUVMEnable, + v->HostVMEnable, + v->HostVMMaxNonCachedPageTableLevels, + v->HostVMMinPageSize, + v->DynamicMetadataEnable[k], + v->DynamicMetadataVMEnabled, + v->DynamicMetadataLinesBeforeActiveRequired[k], + v->DynamicMetadataTransmittedBytes[k], + v->UrgLatency[i], + v->ExtraLatency, + v->TimeCalc, + v->PDEAndMetaPTEBytesPerFrame[i][j][k], + v->MetaRowBytes[i][j][k], + v->DPTEBytesPerRow[i][j][k], + v->PrefetchLinesY[i][j][k], + v->SwathWidthYThisState[k], + v->PrefillY[k], + v->MaxNumSwY[k], + v->PrefetchLinesC[i][j][k], + v->SwathWidthCThisState[k], + v->PrefillC[k], + v->MaxNumSwC[k], + v->swath_width_luma_ub_this_state[k], + v->swath_width_chroma_ub_this_state[k], + v->SwathHeightYThisState[k], + v->SwathHeightCThisState[k], + v->TWait, + &v->DSTXAfterScaler[k], + &v->DSTYAfterScaler[k], + &v->LineTimesForPrefetch[k], + &v->PrefetchBW[k], + &v->LinesForMetaPTE[k], + &v->LinesForMetaAndDPTERow[k], + &v->VRatioPreY[i][j][k], + &v->VRatioPreC[i][j][k], + &v->RequiredPrefetchPixelDataBWLuma[i][j][k], + &v->RequiredPrefetchPixelDataBWChroma[i][j][k], + &v->NoTimeForDynamicMetadata[i][j][k], + &v->Tno_bw[k], + &v->prefetch_vmrow_bw[k], + &v->dummy7[k], + &v->dummy8[k], + &v->dummy13[k], + &v->VUpdateOffsetPix[k], + &v->VUpdateWidthPix[k], + &v->VReadyOffsetPix[k]); +} + void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) { struct vba_vars_st *v = &mode_lib->vba; @@ -5276,92 +5372,9 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->SREnterPlusExitTime);
for (k = 0; k < v->NumberOfActivePlanes; k++) { - Pipe myPipe; - - myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; - myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; - myPipe.PixelClock = v->PixelClock[k]; - myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; - myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; - myPipe.ScalerEnabled = v->ScalerEnabled[k]; - myPipe.SourceScan = v->SourceScan[k]; - myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; - myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; - myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; - myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; - myPipe.InterlaceEnable = v->Interlace[k]; - myPipe.NumberOfCursors = v->NumberOfCursors[k]; - myPipe.VBlank = v->VTotal[k] - v->VActive[k]; - myPipe.HTotal = v->HTotal[k]; - myPipe.DCCEnable = v->DCCEnable[k]; - myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 - || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; - myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; - myPipe.BytePerPixelY = v->BytePerPixelY[k]; - myPipe.BytePerPixelC = v->BytePerPixelC[k]; - myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; - v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( - mode_lib, - HostVMInefficiencyFactor, - &myPipe, - v->DSCDelayPerState[i][k], - v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, - v->DPPCLKDelaySCL, - v->DPPCLKDelaySCLLBOnly, - v->DPPCLKDelayCNVCCursor, - v->DISPCLKDelaySubtotal, - v->SwathWidthYThisState[k] / v->HRatio[k], - v->OutputFormat[k], - v->MaxInterDCNTileRepeaters, - dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), - v->MaximumVStartup[i][j][k], - v->GPUVMMaxPageTableLevels, - v->GPUVMEnable, - v->HostVMEnable, - v->HostVMMaxNonCachedPageTableLevels, - v->HostVMMinPageSize, - v->DynamicMetadataEnable[k], - v->DynamicMetadataVMEnabled, - v->DynamicMetadataLinesBeforeActiveRequired[k], - v->DynamicMetadataTransmittedBytes[k], - v->UrgLatency[i], - v->ExtraLatency, - v->TimeCalc, - v->PDEAndMetaPTEBytesPerFrame[i][j][k], - v->MetaRowBytes[i][j][k], - v->DPTEBytesPerRow[i][j][k], - v->PrefetchLinesY[i][j][k], - v->SwathWidthYThisState[k], - v->PrefillY[k], - v->MaxNumSwY[k], - v->PrefetchLinesC[i][j][k], - v->SwathWidthCThisState[k], - v->PrefillC[k], - v->MaxNumSwC[k], - v->swath_width_luma_ub_this_state[k], - v->swath_width_chroma_ub_this_state[k], - v->SwathHeightYThisState[k], - v->SwathHeightCThisState[k], - v->TWait, - &v->DSTXAfterScaler[k], - &v->DSTYAfterScaler[k], - &v->LineTimesForPrefetch[k], - &v->PrefetchBW[k], - &v->LinesForMetaPTE[k], - &v->LinesForMetaAndDPTERow[k], - &v->VRatioPreY[i][j][k], - &v->VRatioPreC[i][j][k], - &v->RequiredPrefetchPixelDataBWLuma[i][j][k], - &v->RequiredPrefetchPixelDataBWChroma[i][j][k], - &v->NoTimeForDynamicMetadata[i][j][k], - &v->Tno_bw[k], - &v->prefetch_vmrow_bw[k], - &v->dummy7[k], - &v->dummy8[k], - &v->dummy13[k], - &v->VUpdateOffsetPix[k], - &v->VUpdateWidthPix[k], - &v->VReadyOffsetPix[k]); + CalculatePrefetchSchedulePerPlane(mode_lib, + HostVMInefficiencyFactor, + i, j, k); }
for (k = 0; k < v->NumberOfActivePlanes; k++) {
From: Michel Dänzer mdaenzer@redhat.com
Use the struct display_mode_lib pointer instead of passing lots of large arrays as parameters by value.
Addresses this warning (resulting in failure to build a RHEL debug kernel with Werror enabled):
../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c: In function ‘UseMinimumDCFCLK’: ../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c:7478:1: warning: the frame size of 2128 bytes is larger than 2048 bytes [-Wframe-larger-than=]
NOTE: AFAICT this function previously had no observable effect, since it only modified parameters passed by value and doesn't return anything. Now it may modify some values in struct display_mode_lib passed in by reference.
Signed-off-by: Michel Dänzer mdaenzer@redhat.com --- .../dc/dml/dcn31/display_mode_vba_31.c | 304 ++++-------------- 1 file changed, 69 insertions(+), 235 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index 8965f9af9d0a..6feb23432f8d 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -422,62 +422,8 @@ static void CalculateUrgentBurstFactor(
static void UseMinimumDCFCLK( struct display_mode_lib *mode_lib, - int MaxInterDCNTileRepeaters, int MaxPrefetchMode, - double FinalDRAMClockChangeLatency, - double SREnterPlusExitTime, - int ReturnBusWidth, - int RoundTripPingLatencyCycles, - int ReorderingBytes, - int PixelChunkSizeInKByte, - int MetaChunkSize, - bool GPUVMEnable, - int GPUVMMaxPageTableLevels, - bool HostVMEnable, - int NumberOfActivePlanes, - double HostVMMinPageSize, - int HostVMMaxNonCachedPageTableLevels, - bool DynamicMetadataVMEnabled, - enum immediate_flip_requirement ImmediateFlipRequirement, - bool ProgressiveToInterlaceUnitInOPP, - double MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation, - double PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency, - int VTotal[], - int VActive[], - int DynamicMetadataTransmittedBytes[], - int DynamicMetadataLinesBeforeActiveRequired[], - bool Interlace[], - double RequiredDPPCLK[][2][DC__NUM_DPP__MAX], - double RequiredDISPCLK[][2], - double UrgLatency[], - unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX], - double ProjectedDCFCLKDeepSleep[][2], - double MaximumVStartup[][2][DC__NUM_DPP__MAX], - double TotalVActivePixelBandwidth[][2], - double TotalVActiveCursorBandwidth[][2], - double TotalMetaRowBandwidth[][2], - double TotalDPTERowBandwidth[][2], - unsigned int TotalNumberOfActiveDPP[][2], - unsigned int TotalNumberOfDCCActiveDPP[][2], - int dpte_group_bytes[], - double PrefetchLinesY[][2][DC__NUM_DPP__MAX], - double PrefetchLinesC[][2][DC__NUM_DPP__MAX], - int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX], - int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX], - int BytePerPixelY[], - int BytePerPixelC[], - int HTotal[], - double PixelClock[], - double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX], - double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX], - double MetaRowBytes[][2][DC__NUM_DPP__MAX], - bool DynamicMetadataEnable[], - double VActivePixelBandwidth[][2][DC__NUM_DPP__MAX], - double VActiveCursorBandwidth[][2][DC__NUM_DPP__MAX], - double ReadBandwidthLuma[], - double ReadBandwidthChroma[], - double DCFCLKPerState[], - double DCFCLKState[][2]); + int ReorderingBytes);
static void CalculatePixelDeliveryTimes( unsigned int NumberOfActivePlanes, @@ -5175,66 +5121,8 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l } }
- if (v->UseMinimumRequiredDCFCLK == true) { - UseMinimumDCFCLK( - mode_lib, - v->MaxInterDCNTileRepeaters, - MaxPrefetchMode, - v->DRAMClockChangeLatency, - v->SREnterPlusExitTime, - v->ReturnBusWidth, - v->RoundTripPingLatencyCycles, - ReorderingBytes, - v->PixelChunkSizeInKByte, - v->MetaChunkSize, - v->GPUVMEnable, - v->GPUVMMaxPageTableLevels, - v->HostVMEnable, - v->NumberOfActivePlanes, - v->HostVMMinPageSize, - v->HostVMMaxNonCachedPageTableLevels, - v->DynamicMetadataVMEnabled, - v->ImmediateFlipRequirement[0], - v->ProgressiveToInterlaceUnitInOPP, - v->MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation, - v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency, - v->VTotal, - v->VActive, - v->DynamicMetadataTransmittedBytes, - v->DynamicMetadataLinesBeforeActiveRequired, - v->Interlace, - v->RequiredDPPCLK, - v->RequiredDISPCLK, - v->UrgLatency, - v->NoOfDPP, - v->ProjectedDCFCLKDeepSleep, - v->MaximumVStartup, - v->TotalVActivePixelBandwidth, - v->TotalVActiveCursorBandwidth, - v->TotalMetaRowBandwidth, - v->TotalDPTERowBandwidth, - v->TotalNumberOfActiveDPP, - v->TotalNumberOfDCCActiveDPP, - v->dpte_group_bytes, - v->PrefetchLinesY, - v->PrefetchLinesC, - v->swath_width_luma_ub_all_states, - v->swath_width_chroma_ub_all_states, - v->BytePerPixelY, - v->BytePerPixelC, - v->HTotal, - v->PixelClock, - v->PDEAndMetaPTEBytesPerFrame, - v->DPTEBytesPerRow, - v->MetaRowBytes, - v->DynamicMetadataEnable, - v->VActivePixelBandwidth, - v->VActiveCursorBandwidth, - v->ReadBandwidthLuma, - v->ReadBandwidthChroma, - v->DCFCLKPerState, - v->DCFCLKState); - } + if (v->UseMinimumRequiredDCFCLK == true) + UseMinimumDCFCLK(mode_lib, MaxPrefetchMode, ReorderingBytes);
for (i = 0; i < v->soc.num_states; ++i) { for (j = 0; j <= 1; ++j) { @@ -7262,69 +7150,15 @@ static double CalculateUrgentLatency(
static void UseMinimumDCFCLK( struct display_mode_lib *mode_lib, - int MaxInterDCNTileRepeaters, int MaxPrefetchMode, - double FinalDRAMClockChangeLatency, - double SREnterPlusExitTime, - int ReturnBusWidth, - int RoundTripPingLatencyCycles, - int ReorderingBytes, - int PixelChunkSizeInKByte, - int MetaChunkSize, - bool GPUVMEnable, - int GPUVMMaxPageTableLevels, - bool HostVMEnable, - int NumberOfActivePlanes, - double HostVMMinPageSize, - int HostVMMaxNonCachedPageTableLevels, - bool DynamicMetadataVMEnabled, - enum immediate_flip_requirement ImmediateFlipRequirement, - bool ProgressiveToInterlaceUnitInOPP, - double MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation, - double PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency, - int VTotal[], - int VActive[], - int DynamicMetadataTransmittedBytes[], - int DynamicMetadataLinesBeforeActiveRequired[], - bool Interlace[], - double RequiredDPPCLK[][2][DC__NUM_DPP__MAX], - double RequiredDISPCLK[][2], - double UrgLatency[], - unsigned int NoOfDPP[][2][DC__NUM_DPP__MAX], - double ProjectedDCFCLKDeepSleep[][2], - double MaximumVStartup[][2][DC__NUM_DPP__MAX], - double TotalVActivePixelBandwidth[][2], - double TotalVActiveCursorBandwidth[][2], - double TotalMetaRowBandwidth[][2], - double TotalDPTERowBandwidth[][2], - unsigned int TotalNumberOfActiveDPP[][2], - unsigned int TotalNumberOfDCCActiveDPP[][2], - int dpte_group_bytes[], - double PrefetchLinesY[][2][DC__NUM_DPP__MAX], - double PrefetchLinesC[][2][DC__NUM_DPP__MAX], - int swath_width_luma_ub_all_states[][2][DC__NUM_DPP__MAX], - int swath_width_chroma_ub_all_states[][2][DC__NUM_DPP__MAX], - int BytePerPixelY[], - int BytePerPixelC[], - int HTotal[], - double PixelClock[], - double PDEAndMetaPTEBytesPerFrame[][2][DC__NUM_DPP__MAX], - double DPTEBytesPerRow[][2][DC__NUM_DPP__MAX], - double MetaRowBytes[][2][DC__NUM_DPP__MAX], - bool DynamicMetadataEnable[], - double VActivePixelBandwidth[][2][DC__NUM_DPP__MAX], - double VActiveCursorBandwidth[][2][DC__NUM_DPP__MAX], - double ReadBandwidthLuma[], - double ReadBandwidthChroma[], - double DCFCLKPerState[], - double DCFCLKState[][2]) + int ReorderingBytes) { struct vba_vars_st *v = &mode_lib->vba; int dummy1, i, j, k; double NormalEfficiency, dummy2, dummy3; double TotalMaxPrefetchFlipDPTERowBandwidth[DC__VOLTAGE_STATES][2];
- NormalEfficiency = PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0; + NormalEfficiency = v->PercentOfIdealFabricAndSDPPortBWReceivedAfterUrgLatency / 100.0; for (i = 0; i < v->soc.num_states; ++i) { for (j = 0; j <= 1; ++j) { double PixelDCFCLKCyclesRequiredInPrefetch[DC__NUM_DPP__MAX]; @@ -7342,61 +7176,61 @@ static void UseMinimumDCFCLK( double MinimumTvmPlus2Tr0;
TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = 0; - for (k = 0; k < NumberOfActivePlanes; ++k) { + for (k = 0; k < v->NumberOfActivePlanes; ++k) { TotalMaxPrefetchFlipDPTERowBandwidth[i][j] = TotalMaxPrefetchFlipDPTERowBandwidth[i][j] - + NoOfDPP[i][j][k] * DPTEBytesPerRow[i][j][k] / (15.75 * HTotal[k] / PixelClock[k]); + + v->NoOfDPP[i][j][k] * v->DPTEBytesPerRow[i][j][k] / (15.75 * v->HTotal[k] / v->PixelClock[k]); }
- for (k = 0; k <= NumberOfActivePlanes - 1; ++k) { - NoOfDPPState[k] = NoOfDPP[i][j][k]; + for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k) { + NoOfDPPState[k] = v->NoOfDPP[i][j][k]; }
- MinimumTWait = CalculateTWait(MaxPrefetchMode, FinalDRAMClockChangeLatency, UrgLatency[i], SREnterPlusExitTime); - NonDPTEBandwidth = TotalVActivePixelBandwidth[i][j] + TotalVActiveCursorBandwidth[i][j] + TotalMetaRowBandwidth[i][j]; - DPTEBandwidth = (HostVMEnable == true || ImmediateFlipRequirement == dm_immediate_flip_required) ? - TotalMaxPrefetchFlipDPTERowBandwidth[i][j] : TotalDPTERowBandwidth[i][j]; + MinimumTWait = CalculateTWait(MaxPrefetchMode, v->FinalDRAMClockChangeLatency, v->UrgLatency[i], v->SREnterPlusExitTime); + NonDPTEBandwidth = v->TotalVActivePixelBandwidth[i][j] + v->TotalVActiveCursorBandwidth[i][j] + v->TotalMetaRowBandwidth[i][j]; + DPTEBandwidth = (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) ? + TotalMaxPrefetchFlipDPTERowBandwidth[i][j] : v->TotalDPTERowBandwidth[i][j]; DCFCLKRequiredForAverageBandwidth = dml_max3( - ProjectedDCFCLKDeepSleep[i][j], - (NonDPTEBandwidth + TotalDPTERowBandwidth[i][j]) / ReturnBusWidth - / (MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation / 100), - (NonDPTEBandwidth + DPTEBandwidth / NormalEfficiency) / NormalEfficiency / ReturnBusWidth); + v->ProjectedDCFCLKDeepSleep[i][j], + (NonDPTEBandwidth + v->TotalDPTERowBandwidth[i][j]) / v->ReturnBusWidth + / (v->MaxAveragePercentOfIdealFabricAndSDPPortBWDisplayCanUseInNormalSystemOperation / 100), + (NonDPTEBandwidth + DPTEBandwidth / NormalEfficiency) / NormalEfficiency / v->ReturnBusWidth);
ExtraLatencyBytes = CalculateExtraLatencyBytes( ReorderingBytes, - TotalNumberOfActiveDPP[i][j], - PixelChunkSizeInKByte, - TotalNumberOfDCCActiveDPP[i][j], - MetaChunkSize, - GPUVMEnable, - HostVMEnable, - NumberOfActivePlanes, + v->TotalNumberOfActiveDPP[i][j], + v->PixelChunkSizeInKByte, + v->TotalNumberOfDCCActiveDPP[i][j], + v->MetaChunkSize, + v->GPUVMEnable, + v->HostVMEnable, + v->NumberOfActivePlanes, NoOfDPPState, - dpte_group_bytes, + v->dpte_group_bytes, 1, - HostVMMinPageSize, - HostVMMaxNonCachedPageTableLevels); - ExtraLatencyCycles = RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__ + ExtraLatencyBytes / NormalEfficiency / ReturnBusWidth; - for (k = 0; k < NumberOfActivePlanes; ++k) { + v->HostVMMinPageSize, + v->HostVMMaxNonCachedPageTableLevels); + ExtraLatencyCycles = v->RoundTripPingLatencyCycles + __DML_ARB_TO_RET_DELAY__ + ExtraLatencyBytes / NormalEfficiency / v->ReturnBusWidth; + for (k = 0; k < v->NumberOfActivePlanes; ++k) { double DCFCLKCyclesRequiredInPrefetch; double ExpectedPrefetchBWAcceleration; double PrefetchTime;
- PixelDCFCLKCyclesRequiredInPrefetch[k] = (PrefetchLinesY[i][j][k] * swath_width_luma_ub_all_states[i][j][k] * BytePerPixelY[k] - + PrefetchLinesC[i][j][k] * swath_width_chroma_ub_all_states[i][j][k] * BytePerPixelC[k]) / NormalEfficiency / ReturnBusWidth; + PixelDCFCLKCyclesRequiredInPrefetch[k] = (v->PrefetchLinesY[i][j][k] * v->swath_width_luma_ub_all_states[i][j][k] * v->BytePerPixelY[k] + + v->PrefetchLinesC[i][j][k] * v->swath_width_chroma_ub_all_states[i][j][k] * v->BytePerPixelC[k]) / NormalEfficiency / v->ReturnBusWidth; DCFCLKCyclesRequiredInPrefetch = 2 * ExtraLatencyCycles / NoOfDPPState[k] - + PDEAndMetaPTEBytesPerFrame[i][j][k] / NormalEfficiency / NormalEfficiency / ReturnBusWidth * (GPUVMMaxPageTableLevels > 2 ? 1 : 0) - + 2 * DPTEBytesPerRow[i][j][k] / NormalEfficiency / NormalEfficiency / ReturnBusWidth - + 2 * MetaRowBytes[i][j][k] / NormalEfficiency / ReturnBusWidth + PixelDCFCLKCyclesRequiredInPrefetch[k]; - PrefetchPixelLinesTime[k] = dml_max(PrefetchLinesY[i][j][k], PrefetchLinesC[i][j][k]) * HTotal[k] / PixelClock[k]; - ExpectedPrefetchBWAcceleration = (VActivePixelBandwidth[i][j][k] + VActiveCursorBandwidth[i][j][k]) - / (ReadBandwidthLuma[k] + ReadBandwidthChroma[k]); + + v->PDEAndMetaPTEBytesPerFrame[i][j][k] / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth * (v->GPUVMMaxPageTableLevels > 2 ? 1 : 0) + + 2 * v->DPTEBytesPerRow[i][j][k] / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth + + 2 * v->MetaRowBytes[i][j][k] / NormalEfficiency / v->ReturnBusWidth + PixelDCFCLKCyclesRequiredInPrefetch[k]; + PrefetchPixelLinesTime[k] = dml_max(v->PrefetchLinesY[i][j][k], v->PrefetchLinesC[i][j][k]) * v->HTotal[k] / v->PixelClock[k]; + ExpectedPrefetchBWAcceleration = (v->VActivePixelBandwidth[i][j][k] + v->VActiveCursorBandwidth[i][j][k]) + / (v->ReadBandwidthLuma[k] + v->ReadBandwidthChroma[k]); DynamicMetadataVMExtraLatency[k] = - (GPUVMEnable == true && DynamicMetadataEnable[k] == true && DynamicMetadataVMEnabled == true) ? - UrgLatency[i] * GPUVMMaxPageTableLevels * (HostVMEnable == true ? HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0; - PrefetchTime = (MaximumVStartup[i][j][k] - 1) * HTotal[k] / PixelClock[k] - MinimumTWait - - UrgLatency[i] - * ((GPUVMMaxPageTableLevels <= 2 ? GPUVMMaxPageTableLevels : GPUVMMaxPageTableLevels - 2) - * (HostVMEnable == true ? HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1) + (v->GPUVMEnable == true && v->DynamicMetadataEnable[k] == true && v->DynamicMetadataVMEnabled == true) ? + v->UrgLatency[i] * v->GPUVMMaxPageTableLevels * (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) : 0; + PrefetchTime = (v->MaximumVStartup[i][j][k] - 1) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait + - v->UrgLatency[i] + * ((v->GPUVMMaxPageTableLevels <= 2 ? v->GPUVMMaxPageTableLevels : v->GPUVMMaxPageTableLevels - 2) + * (v->HostVMEnable == true ? v->HostVMMaxNonCachedPageTableLevels + 1 : 1) - 1) - DynamicMetadataVMExtraLatency[k];
if (PrefetchTime > 0) { @@ -7405,14 +7239,14 @@ static void UseMinimumDCFCLK( / (PrefetchTime * PixelDCFCLKCyclesRequiredInPrefetch[k] / DCFCLKCyclesRequiredInPrefetch); DCFCLKRequiredForPeakBandwidthPerPlane[k] = NoOfDPPState[k] * PixelDCFCLKCyclesRequiredInPrefetch[k] / PrefetchPixelLinesTime[k] * dml_max(1.0, ExpectedVRatioPrefetch) * dml_max(1.0, ExpectedVRatioPrefetch / 4) * ExpectedPrefetchBWAcceleration; - if (HostVMEnable == true || ImmediateFlipRequirement == dm_immediate_flip_required) { + if (v->HostVMEnable == true || v->ImmediateFlipRequirement[0] == dm_immediate_flip_required) { DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKRequiredForPeakBandwidthPerPlane[k] - + NoOfDPPState[k] * DPTEBandwidth / NormalEfficiency / NormalEfficiency / ReturnBusWidth; + + NoOfDPPState[k] * DPTEBandwidth / NormalEfficiency / NormalEfficiency / v->ReturnBusWidth; } } else { - DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKPerState[i]; + DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i]; } - if (DynamicMetadataEnable[k] == true) { + if (v->DynamicMetadataEnable[k] == true) { double TSetupPipe; double TdmbfPipe; double TdmsksPipe; @@ -7420,17 +7254,17 @@ static void UseMinimumDCFCLK( double AllowedTimeForUrgentExtraLatency;
CalculateVupdateAndDynamicMetadataParameters( - MaxInterDCNTileRepeaters, - RequiredDPPCLK[i][j][k], - RequiredDISPCLK[i][j], - ProjectedDCFCLKDeepSleep[i][j], - PixelClock[k], - HTotal[k], - VTotal[k] - VActive[k], - DynamicMetadataTransmittedBytes[k], - DynamicMetadataLinesBeforeActiveRequired[k], - Interlace[k], - ProgressiveToInterlaceUnitInOPP, + v->MaxInterDCNTileRepeaters, + v->RequiredDPPCLK[i][j][k], + v->RequiredDISPCLK[i][j], + v->ProjectedDCFCLKDeepSleep[i][j], + v->PixelClock[k], + v->HTotal[k], + v->VTotal[k] - v->VActive[k], + v->DynamicMetadataTransmittedBytes[k], + v->DynamicMetadataLinesBeforeActiveRequired[k], + v->Interlace[k], + v->ProgressiveToInterlaceUnitInOPP, &TSetupPipe, &TdmbfPipe, &TdmecPipe, @@ -7438,31 +7272,31 @@ static void UseMinimumDCFCLK( &dummy1, &dummy2, &dummy3); - AllowedTimeForUrgentExtraLatency = MaximumVStartup[i][j][k] * HTotal[k] / PixelClock[k] - MinimumTWait - TSetupPipe - TdmbfPipe - TdmecPipe + AllowedTimeForUrgentExtraLatency = v->MaximumVStartup[i][j][k] * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - TSetupPipe - TdmbfPipe - TdmecPipe - TdmsksPipe - DynamicMetadataVMExtraLatency[k]; if (AllowedTimeForUrgentExtraLatency > 0) { DCFCLKRequiredForPeakBandwidthPerPlane[k] = dml_max( DCFCLKRequiredForPeakBandwidthPerPlane[k], ExtraLatencyCycles / AllowedTimeForUrgentExtraLatency); } else { - DCFCLKRequiredForPeakBandwidthPerPlane[k] = DCFCLKPerState[i]; + DCFCLKRequiredForPeakBandwidthPerPlane[k] = v->DCFCLKPerState[i]; } } } DCFCLKRequiredForPeakBandwidth = 0; - for (k = 0; k <= NumberOfActivePlanes - 1; ++k) { + for (k = 0; k <= v->NumberOfActivePlanes - 1; ++k) { DCFCLKRequiredForPeakBandwidth = DCFCLKRequiredForPeakBandwidth + DCFCLKRequiredForPeakBandwidthPerPlane[k]; } - MinimumTvmPlus2Tr0 = UrgLatency[i] - * (GPUVMEnable == true ? - (HostVMEnable == true ? - (GPUVMMaxPageTableLevels + 2) * (HostVMMaxNonCachedPageTableLevels + 1) - 1 : GPUVMMaxPageTableLevels + 1) : + MinimumTvmPlus2Tr0 = v->UrgLatency[i] + * (v->GPUVMEnable == true ? + (v->HostVMEnable == true ? + (v->GPUVMMaxPageTableLevels + 2) * (v->HostVMMaxNonCachedPageTableLevels + 1) - 1 : v->GPUVMMaxPageTableLevels + 1) : 0); - for (k = 0; k < NumberOfActivePlanes; ++k) { + for (k = 0; k < v->NumberOfActivePlanes; ++k) { double MaximumTvmPlus2Tr0PlusTsw; - MaximumTvmPlus2Tr0PlusTsw = (MaximumVStartup[i][j][k] - 2) * HTotal[k] / PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k]; + MaximumTvmPlus2Tr0PlusTsw = (v->MaximumVStartup[i][j][k] - 2) * v->HTotal[k] / v->PixelClock[k] - MinimumTWait - DynamicMetadataVMExtraLatency[k]; if (MaximumTvmPlus2Tr0PlusTsw <= MinimumTvmPlus2Tr0 + PrefetchPixelLinesTime[k] / 4) { - DCFCLKRequiredForPeakBandwidth = DCFCLKPerState[i]; + DCFCLKRequiredForPeakBandwidth = v->DCFCLKPerState[i]; } else { DCFCLKRequiredForPeakBandwidth = dml_max3( DCFCLKRequiredForPeakBandwidth, @@ -7470,7 +7304,7 @@ static void UseMinimumDCFCLK( (2 * ExtraLatencyCycles + PixelDCFCLKCyclesRequiredInPrefetch[k]) / (MaximumTvmPlus2Tr0PlusTsw - MinimumTvmPlus2Tr0)); } } - DCFCLKState[i][j] = dml_min(DCFCLKPerState[i], 1.05 * dml_max(DCFCLKRequiredForAverageBandwidth, DCFCLKRequiredForPeakBandwidth)); + v->DCFCLKState[i][j] = dml_min(v->DCFCLKPerState[i], 1.05 * dml_max(DCFCLKRequiredForAverageBandwidth, DCFCLKRequiredForPeakBandwidth)); } } }
On 2021-12-09 11:46 a.m., Michel Dänzer wrote:
From: Michel Dänzer mdaenzer@redhat.com
Move code using the Pipe struct to a new helper function.
Works around[0] this warning (resulting in failure to build a RHEL debug kernel with Werror enabled):
../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c: In function ‘dml31_ModeSupportAndSystemConfigurationFull’: ../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c:5740:1: warning: the frame size of 2144 bytes is larger than 2048 bytes [-Wframe-larger-than=]
The culprit seems to be the Pipe struct, so pull the relevant block out into its own sub-function. (This is porting a62427ef9b55 "drm/amd/display: Reduce stack size for dml21_ModeSupportAndSystemConfigurationFull" from dml31 to dml21)
[0] AFAICT this doesn't actually reduce the total amount of stack which can be used, just moves some of it from dml31_ModeSupportAndSystemConfigurationFull to the new helper function, so the former happens to no longer exceed the limit for a single function.
Signed-off-by: Michel Dänzer mdaenzer@redhat.com
.../dc/dml/dcn31/display_mode_vba_31.c | 185 ++++++++++-------- 1 file changed, 99 insertions(+), 86 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index 7e937bdcea00..8965f9af9d0a 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -3949,6 +3949,102 @@ static double TruncToValidBPP( return BPP_INVALID; }
+static noinline void CalculatePrefetchSchedulePerPlane(
struct display_mode_lib *mode_lib,double HostVMInefficiencyFactor,int i,unsigned j,unsigned k)+{
- struct vba_vars_st *v = &mode_lib->vba;
- Pipe myPipe;
- myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k];
- myPipe.DISPCLK = v->RequiredDISPCLK[i][j];
- myPipe.PixelClock = v->PixelClock[k];
- myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j];
- myPipe.DPPPerPlane = v->NoOfDPP[i][j][k];
- myPipe.ScalerEnabled = v->ScalerEnabled[k];
- myPipe.SourceScan = v->SourceScan[k];
- myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k];
- myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k];
- myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k];
- myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k];
- myPipe.InterlaceEnable = v->Interlace[k];
- myPipe.NumberOfCursors = v->NumberOfCursors[k];
- myPipe.VBlank = v->VTotal[k] - v->VActive[k];
- myPipe.HTotal = v->HTotal[k];
- myPipe.DCCEnable = v->DCCEnable[k];
- myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1
|| v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1;- myPipe.SourcePixelFormat = v->SourcePixelFormat[k];
- myPipe.BytePerPixelY = v->BytePerPixelY[k];
- myPipe.BytePerPixelC = v->BytePerPixelC[k];
- myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP;
- v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule(
mode_lib,HostVMInefficiencyFactor,&myPipe,v->DSCDelayPerState[i][k],v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater,v->DPPCLKDelaySCL,v->DPPCLKDelaySCLLBOnly,v->DPPCLKDelayCNVCCursor,v->DISPCLKDelaySubtotal,v->SwathWidthYThisState[k] / v->HRatio[k],v->OutputFormat[k],v->MaxInterDCNTileRepeaters,dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]),v->MaximumVStartup[i][j][k],v->GPUVMMaxPageTableLevels,v->GPUVMEnable,v->HostVMEnable,v->HostVMMaxNonCachedPageTableLevels,v->HostVMMinPageSize,v->DynamicMetadataEnable[k],v->DynamicMetadataVMEnabled,v->DynamicMetadataLinesBeforeActiveRequired[k],v->DynamicMetadataTransmittedBytes[k],v->UrgLatency[i],v->ExtraLatency,v->TimeCalc,v->PDEAndMetaPTEBytesPerFrame[i][j][k],v->MetaRowBytes[i][j][k],v->DPTEBytesPerRow[i][j][k],v->PrefetchLinesY[i][j][k],v->SwathWidthYThisState[k],v->PrefillY[k],v->MaxNumSwY[k],v->PrefetchLinesC[i][j][k],v->SwathWidthCThisState[k],v->PrefillC[k],v->MaxNumSwC[k],v->swath_width_luma_ub_this_state[k],v->swath_width_chroma_ub_this_state[k],v->SwathHeightYThisState[k],v->SwathHeightCThisState[k],v->TWait,&v->DSTXAfterScaler[k],&v->DSTYAfterScaler[k],&v->LineTimesForPrefetch[k],&v->PrefetchBW[k],&v->LinesForMetaPTE[k],&v->LinesForMetaAndDPTERow[k],&v->VRatioPreY[i][j][k],&v->VRatioPreC[i][j][k],&v->RequiredPrefetchPixelDataBWLuma[i][j][k],&v->RequiredPrefetchPixelDataBWChroma[i][j][k],&v->NoTimeForDynamicMetadata[i][j][k],&v->Tno_bw[k],&v->prefetch_vmrow_bw[k],&v->dummy7[k],&v->dummy8[k],&v->dummy13[k],&v->VUpdateOffsetPix[k],&v->VUpdateWidthPix[k],&v->VReadyOffsetPix[k]);+}
- void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) { struct vba_vars_st *v = &mode_lib->vba;
@@ -5276,92 +5372,9 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->SREnterPlusExitTime);
for (k = 0; k < v->NumberOfActivePlanes; k++) {
Pipe myPipe;myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k];myPipe.DISPCLK = v->RequiredDISPCLK[i][j];myPipe.PixelClock = v->PixelClock[k];myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j];myPipe.DPPPerPlane = v->NoOfDPP[i][j][k];myPipe.ScalerEnabled = v->ScalerEnabled[k];myPipe.SourceScan = v->SourceScan[k];myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k];myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k];myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k];myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k];myPipe.InterlaceEnable = v->Interlace[k];myPipe.NumberOfCursors = v->NumberOfCursors[k];myPipe.VBlank = v->VTotal[k] - v->VActive[k];myPipe.HTotal = v->HTotal[k];myPipe.DCCEnable = v->DCCEnable[k];myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1|| v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1;myPipe.SourcePixelFormat = v->SourcePixelFormat[k];myPipe.BytePerPixelY = v->BytePerPixelY[k];myPipe.BytePerPixelC = v->BytePerPixelC[k];myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP;v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule(mode_lib,HostVMInefficiencyFactor,&myPipe,v->DSCDelayPerState[i][k],v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater,v->DPPCLKDelaySCL,v->DPPCLKDelaySCLLBOnly,v->DPPCLKDelayCNVCCursor,v->DISPCLKDelaySubtotal,v->SwathWidthYThisState[k] / v->HRatio[k],v->OutputFormat[k],v->MaxInterDCNTileRepeaters,dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]),v->MaximumVStartup[i][j][k],v->GPUVMMaxPageTableLevels,v->GPUVMEnable,v->HostVMEnable,v->HostVMMaxNonCachedPageTableLevels,v->HostVMMinPageSize,v->DynamicMetadataEnable[k],v->DynamicMetadataVMEnabled,v->DynamicMetadataLinesBeforeActiveRequired[k],v->DynamicMetadataTransmittedBytes[k],v->UrgLatency[i],v->ExtraLatency,v->TimeCalc,v->PDEAndMetaPTEBytesPerFrame[i][j][k],v->MetaRowBytes[i][j][k],v->DPTEBytesPerRow[i][j][k],v->PrefetchLinesY[i][j][k],v->SwathWidthYThisState[k],v->PrefillY[k],v->MaxNumSwY[k],v->PrefetchLinesC[i][j][k],v->SwathWidthCThisState[k],v->PrefillC[k],v->MaxNumSwC[k],v->swath_width_luma_ub_this_state[k],v->swath_width_chroma_ub_this_state[k],v->SwathHeightYThisState[k],v->SwathHeightCThisState[k],v->TWait,&v->DSTXAfterScaler[k],&v->DSTYAfterScaler[k],&v->LineTimesForPrefetch[k],&v->PrefetchBW[k],&v->LinesForMetaPTE[k],&v->LinesForMetaAndDPTERow[k],&v->VRatioPreY[i][j][k],&v->VRatioPreC[i][j][k],&v->RequiredPrefetchPixelDataBWLuma[i][j][k],&v->RequiredPrefetchPixelDataBWChroma[i][j][k],&v->NoTimeForDynamicMetadata[i][j][k],&v->Tno_bw[k],&v->prefetch_vmrow_bw[k],&v->dummy7[k],&v->dummy8[k],&v->dummy13[k],&v->VUpdateOffsetPix[k],&v->VUpdateWidthPix[k],&v->VReadyOffsetPix[k]);
CalculatePrefetchSchedulePerPlane(mode_lib,HostVMInefficiencyFactor,i, j, k); } for (k = 0; k < v->NumberOfActivePlanes; k++) {
Hi Michel,
Overwall I think this series is good. I also run it in our internal CI and everything looks fine.
Reviewed-by: Rodrigo Siqueira Rodrigo.Siqueira@amd.com
And applied to amd-staging-drm-next.
Thanks
On 2021-12-11 13:20, Rodrigo Siqueira Jordao wrote:
On 2021-12-09 11:46 a.m., Michel Dänzer wrote:
From: Michel Dänzer mdaenzer@redhat.com
Move code using the Pipe struct to a new helper function.
Works around[0] this warning (resulting in failure to build a RHEL debug kernel with Werror enabled):
../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c: In function ‘dml31_ModeSupportAndSystemConfigurationFull’: ../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c:5740:1: warning: the frame size of 2144 bytes is larger than 2048 bytes [-Wframe-larger-than=]
The culprit seems to be the Pipe struct, so pull the relevant block out into its own sub-function. (This is porting a62427ef9b55 "drm/amd/display: Reduce stack size for dml21_ModeSupportAndSystemConfigurationFull" from dml31 to dml21)
[0] AFAICT this doesn't actually reduce the total amount of stack which can be used, just moves some of it from dml31_ModeSupportAndSystemConfigurationFull to the new helper function, so the former happens to no longer exceed the limit for a single function.
Signed-off-by: Michel Dänzer mdaenzer@redhat.com
.../dc/dml/dcn31/display_mode_vba_31.c | 185 ++++++++++-------- 1 file changed, 99 insertions(+), 86 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index 7e937bdcea00..8965f9af9d0a 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -3949,6 +3949,102 @@ static double TruncToValidBPP( return BPP_INVALID; } +static noinline void CalculatePrefetchSchedulePerPlane( + struct display_mode_lib *mode_lib, + double HostVMInefficiencyFactor, + int i, + unsigned j, + unsigned k) +{ + struct vba_vars_st *v = &mode_lib->vba; + Pipe myPipe;
+ myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; + myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; + myPipe.PixelClock = v->PixelClock[k]; + myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; + myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; + myPipe.ScalerEnabled = v->ScalerEnabled[k]; + myPipe.SourceScan = v->SourceScan[k]; + myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; + myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; + myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; + myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; + myPipe.InterlaceEnable = v->Interlace[k]; + myPipe.NumberOfCursors = v->NumberOfCursors[k]; + myPipe.VBlank = v->VTotal[k] - v->VActive[k]; + myPipe.HTotal = v->HTotal[k]; + myPipe.DCCEnable = v->DCCEnable[k]; + myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 + || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; + myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; + myPipe.BytePerPixelY = v->BytePerPixelY[k]; + myPipe.BytePerPixelC = v->BytePerPixelC[k]; + myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; + v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( + mode_lib, + HostVMInefficiencyFactor, + &myPipe, + v->DSCDelayPerState[i][k], + v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, + v->DPPCLKDelaySCL, + v->DPPCLKDelaySCLLBOnly, + v->DPPCLKDelayCNVCCursor, + v->DISPCLKDelaySubtotal, + v->SwathWidthYThisState[k] / v->HRatio[k], + v->OutputFormat[k], + v->MaxInterDCNTileRepeaters, + dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), + v->MaximumVStartup[i][j][k], + v->GPUVMMaxPageTableLevels, + v->GPUVMEnable, + v->HostVMEnable, + v->HostVMMaxNonCachedPageTableLevels, + v->HostVMMinPageSize, + v->DynamicMetadataEnable[k], + v->DynamicMetadataVMEnabled, + v->DynamicMetadataLinesBeforeActiveRequired[k], + v->DynamicMetadataTransmittedBytes[k], + v->UrgLatency[i], + v->ExtraLatency, + v->TimeCalc, + v->PDEAndMetaPTEBytesPerFrame[i][j][k], + v->MetaRowBytes[i][j][k], + v->DPTEBytesPerRow[i][j][k], + v->PrefetchLinesY[i][j][k], + v->SwathWidthYThisState[k], + v->PrefillY[k], + v->MaxNumSwY[k], + v->PrefetchLinesC[i][j][k], + v->SwathWidthCThisState[k], + v->PrefillC[k], + v->MaxNumSwC[k], + v->swath_width_luma_ub_this_state[k], + v->swath_width_chroma_ub_this_state[k], + v->SwathHeightYThisState[k], + v->SwathHeightCThisState[k], + v->TWait, + &v->DSTXAfterScaler[k], + &v->DSTYAfterScaler[k], + &v->LineTimesForPrefetch[k], + &v->PrefetchBW[k], + &v->LinesForMetaPTE[k], + &v->LinesForMetaAndDPTERow[k], + &v->VRatioPreY[i][j][k], + &v->VRatioPreC[i][j][k], + &v->RequiredPrefetchPixelDataBWLuma[i][j][k], + &v->RequiredPrefetchPixelDataBWChroma[i][j][k], + &v->NoTimeForDynamicMetadata[i][j][k], + &v->Tno_bw[k], + &v->prefetch_vmrow_bw[k], + &v->dummy7[k], + &v->dummy8[k], + &v->dummy13[k], + &v->VUpdateOffsetPix[k], + &v->VUpdateWidthPix[k], + &v->VReadyOffsetPix[k]); +}
void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) { struct vba_vars_st *v = &mode_lib->vba; @@ -5276,92 +5372,9 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->SREnterPlusExitTime); for (k = 0; k < v->NumberOfActivePlanes; k++) { - Pipe myPipe;
- myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; - myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; - myPipe.PixelClock = v->PixelClock[k]; - myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; - myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; - myPipe.ScalerEnabled = v->ScalerEnabled[k]; - myPipe.SourceScan = v->SourceScan[k]; - myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; - myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; - myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; - myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; - myPipe.InterlaceEnable = v->Interlace[k]; - myPipe.NumberOfCursors = v->NumberOfCursors[k]; - myPipe.VBlank = v->VTotal[k] - v->VActive[k]; - myPipe.HTotal = v->HTotal[k]; - myPipe.DCCEnable = v->DCCEnable[k]; - myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 - || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; - myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; - myPipe.BytePerPixelY = v->BytePerPixelY[k]; - myPipe.BytePerPixelC = v->BytePerPixelC[k]; - myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; - v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( - mode_lib, - HostVMInefficiencyFactor, - &myPipe, - v->DSCDelayPerState[i][k], - v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, - v->DPPCLKDelaySCL, - v->DPPCLKDelaySCLLBOnly, - v->DPPCLKDelayCNVCCursor, - v->DISPCLKDelaySubtotal, - v->SwathWidthYThisState[k] / v->HRatio[k], - v->OutputFormat[k], - v->MaxInterDCNTileRepeaters, - dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), - v->MaximumVStartup[i][j][k], - v->GPUVMMaxPageTableLevels, - v->GPUVMEnable, - v->HostVMEnable, - v->HostVMMaxNonCachedPageTableLevels, - v->HostVMMinPageSize, - v->DynamicMetadataEnable[k], - v->DynamicMetadataVMEnabled, - v->DynamicMetadataLinesBeforeActiveRequired[k], - v->DynamicMetadataTransmittedBytes[k], - v->UrgLatency[i], - v->ExtraLatency, - v->TimeCalc, - v->PDEAndMetaPTEBytesPerFrame[i][j][k], - v->MetaRowBytes[i][j][k], - v->DPTEBytesPerRow[i][j][k], - v->PrefetchLinesY[i][j][k], - v->SwathWidthYThisState[k], - v->PrefillY[k], - v->MaxNumSwY[k], - v->PrefetchLinesC[i][j][k], - v->SwathWidthCThisState[k], - v->PrefillC[k], - v->MaxNumSwC[k], - v->swath_width_luma_ub_this_state[k], - v->swath_width_chroma_ub_this_state[k], - v->SwathHeightYThisState[k], - v->SwathHeightCThisState[k], - v->TWait, - &v->DSTXAfterScaler[k], - &v->DSTYAfterScaler[k], - &v->LineTimesForPrefetch[k], - &v->PrefetchBW[k], - &v->LinesForMetaPTE[k], - &v->LinesForMetaAndDPTERow[k], - &v->VRatioPreY[i][j][k], - &v->VRatioPreC[i][j][k], - &v->RequiredPrefetchPixelDataBWLuma[i][j][k], - &v->RequiredPrefetchPixelDataBWChroma[i][j][k], - &v->NoTimeForDynamicMetadata[i][j][k], - &v->Tno_bw[k], - &v->prefetch_vmrow_bw[k], - &v->dummy7[k], - &v->dummy8[k], - &v->dummy13[k], - &v->VUpdateOffsetPix[k], - &v->VUpdateWidthPix[k], - &v->VReadyOffsetPix[k]); + CalculatePrefetchSchedulePerPlane(mode_lib, + HostVMInefficiencyFactor, + i, j, k); } for (k = 0; k < v->NumberOfActivePlanes; k++) {
Hi Michel,
Overwall I think this series is good. I also run it in our internal CI and everything looks fine.
Reviewed-by: Rodrigo Siqueira Rodrigo.Siqueira@amd.com
And applied to amd-staging-drm-next.
Thanks!
Are there any plans for actually reducing the combined amount of stack used by ModeSupportAndSystemConfigurationFull + CalculatePrefetchSchedulePerPlane?
Also, did you check that UseMinimumDCFCLK now modifying mode_lib->vba.DCFCLKState[i][j] and possibly other values in mode_lib->vba makes sense?
On 2021-12-13 4:46 a.m., Michel Dänzer wrote:
On 2021-12-11 13:20, Rodrigo Siqueira Jordao wrote:
On 2021-12-09 11:46 a.m., Michel Dänzer wrote:
From: Michel Dänzer mdaenzer@redhat.com
Move code using the Pipe struct to a new helper function.
Works around[0] this warning (resulting in failure to build a RHEL debug kernel with Werror enabled):
../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c: In function ‘dml31_ModeSupportAndSystemConfigurationFull’: ../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c:5740:1: warning: the frame size of 2144 bytes is larger than 2048 bytes [-Wframe-larger-than=]
The culprit seems to be the Pipe struct, so pull the relevant block out into its own sub-function. (This is porting a62427ef9b55 "drm/amd/display: Reduce stack size for dml21_ModeSupportAndSystemConfigurationFull" from dml31 to dml21)
[0] AFAICT this doesn't actually reduce the total amount of stack which can be used, just moves some of it from dml31_ModeSupportAndSystemConfigurationFull to the new helper function, so the former happens to no longer exceed the limit for a single function.
Signed-off-by: Michel Dänzer mdaenzer@redhat.com
.../dc/dml/dcn31/display_mode_vba_31.c | 185 ++++++++++-------- 1 file changed, 99 insertions(+), 86 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index 7e937bdcea00..8965f9af9d0a 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -3949,6 +3949,102 @@ static double TruncToValidBPP( return BPP_INVALID; } +static noinline void CalculatePrefetchSchedulePerPlane( + struct display_mode_lib *mode_lib, + double HostVMInefficiencyFactor, + int i, + unsigned j, + unsigned k) +{ + struct vba_vars_st *v = &mode_lib->vba; + Pipe myPipe;
+ myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; + myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; + myPipe.PixelClock = v->PixelClock[k]; + myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; + myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; + myPipe.ScalerEnabled = v->ScalerEnabled[k]; + myPipe.SourceScan = v->SourceScan[k]; + myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; + myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; + myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; + myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; + myPipe.InterlaceEnable = v->Interlace[k]; + myPipe.NumberOfCursors = v->NumberOfCursors[k]; + myPipe.VBlank = v->VTotal[k] - v->VActive[k]; + myPipe.HTotal = v->HTotal[k]; + myPipe.DCCEnable = v->DCCEnable[k]; + myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 + || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; + myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; + myPipe.BytePerPixelY = v->BytePerPixelY[k]; + myPipe.BytePerPixelC = v->BytePerPixelC[k]; + myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; + v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( + mode_lib, + HostVMInefficiencyFactor, + &myPipe, + v->DSCDelayPerState[i][k], + v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, + v->DPPCLKDelaySCL, + v->DPPCLKDelaySCLLBOnly, + v->DPPCLKDelayCNVCCursor, + v->DISPCLKDelaySubtotal, + v->SwathWidthYThisState[k] / v->HRatio[k], + v->OutputFormat[k], + v->MaxInterDCNTileRepeaters, + dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), + v->MaximumVStartup[i][j][k], + v->GPUVMMaxPageTableLevels, + v->GPUVMEnable, + v->HostVMEnable, + v->HostVMMaxNonCachedPageTableLevels, + v->HostVMMinPageSize, + v->DynamicMetadataEnable[k], + v->DynamicMetadataVMEnabled, + v->DynamicMetadataLinesBeforeActiveRequired[k], + v->DynamicMetadataTransmittedBytes[k], + v->UrgLatency[i], + v->ExtraLatency, + v->TimeCalc, + v->PDEAndMetaPTEBytesPerFrame[i][j][k], + v->MetaRowBytes[i][j][k], + v->DPTEBytesPerRow[i][j][k], + v->PrefetchLinesY[i][j][k], + v->SwathWidthYThisState[k], + v->PrefillY[k], + v->MaxNumSwY[k], + v->PrefetchLinesC[i][j][k], + v->SwathWidthCThisState[k], + v->PrefillC[k], + v->MaxNumSwC[k], + v->swath_width_luma_ub_this_state[k], + v->swath_width_chroma_ub_this_state[k], + v->SwathHeightYThisState[k], + v->SwathHeightCThisState[k], + v->TWait, + &v->DSTXAfterScaler[k], + &v->DSTYAfterScaler[k], + &v->LineTimesForPrefetch[k], + &v->PrefetchBW[k], + &v->LinesForMetaPTE[k], + &v->LinesForMetaAndDPTERow[k], + &v->VRatioPreY[i][j][k], + &v->VRatioPreC[i][j][k], + &v->RequiredPrefetchPixelDataBWLuma[i][j][k], + &v->RequiredPrefetchPixelDataBWChroma[i][j][k], + &v->NoTimeForDynamicMetadata[i][j][k], + &v->Tno_bw[k], + &v->prefetch_vmrow_bw[k], + &v->dummy7[k], + &v->dummy8[k], + &v->dummy13[k], + &v->VUpdateOffsetPix[k], + &v->VUpdateWidthPix[k], + &v->VReadyOffsetPix[k]); +}
void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) { struct vba_vars_st *v = &mode_lib->vba; @@ -5276,92 +5372,9 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->SREnterPlusExitTime); for (k = 0; k < v->NumberOfActivePlanes; k++) { - Pipe myPipe;
- myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; - myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; - myPipe.PixelClock = v->PixelClock[k]; - myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; - myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; - myPipe.ScalerEnabled = v->ScalerEnabled[k]; - myPipe.SourceScan = v->SourceScan[k]; - myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; - myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; - myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; - myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; - myPipe.InterlaceEnable = v->Interlace[k]; - myPipe.NumberOfCursors = v->NumberOfCursors[k]; - myPipe.VBlank = v->VTotal[k] - v->VActive[k]; - myPipe.HTotal = v->HTotal[k]; - myPipe.DCCEnable = v->DCCEnable[k]; - myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 - || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; - myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; - myPipe.BytePerPixelY = v->BytePerPixelY[k]; - myPipe.BytePerPixelC = v->BytePerPixelC[k]; - myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; - v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( - mode_lib, - HostVMInefficiencyFactor, - &myPipe, - v->DSCDelayPerState[i][k], - v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, - v->DPPCLKDelaySCL, - v->DPPCLKDelaySCLLBOnly, - v->DPPCLKDelayCNVCCursor, - v->DISPCLKDelaySubtotal, - v->SwathWidthYThisState[k] / v->HRatio[k], - v->OutputFormat[k], - v->MaxInterDCNTileRepeaters, - dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), - v->MaximumVStartup[i][j][k], - v->GPUVMMaxPageTableLevels, - v->GPUVMEnable, - v->HostVMEnable, - v->HostVMMaxNonCachedPageTableLevels, - v->HostVMMinPageSize, - v->DynamicMetadataEnable[k], - v->DynamicMetadataVMEnabled, - v->DynamicMetadataLinesBeforeActiveRequired[k], - v->DynamicMetadataTransmittedBytes[k], - v->UrgLatency[i], - v->ExtraLatency, - v->TimeCalc, - v->PDEAndMetaPTEBytesPerFrame[i][j][k], - v->MetaRowBytes[i][j][k], - v->DPTEBytesPerRow[i][j][k], - v->PrefetchLinesY[i][j][k], - v->SwathWidthYThisState[k], - v->PrefillY[k], - v->MaxNumSwY[k], - v->PrefetchLinesC[i][j][k], - v->SwathWidthCThisState[k], - v->PrefillC[k], - v->MaxNumSwC[k], - v->swath_width_luma_ub_this_state[k], - v->swath_width_chroma_ub_this_state[k], - v->SwathHeightYThisState[k], - v->SwathHeightCThisState[k], - v->TWait, - &v->DSTXAfterScaler[k], - &v->DSTYAfterScaler[k], - &v->LineTimesForPrefetch[k], - &v->PrefetchBW[k], - &v->LinesForMetaPTE[k], - &v->LinesForMetaAndDPTERow[k], - &v->VRatioPreY[i][j][k], - &v->VRatioPreC[i][j][k], - &v->RequiredPrefetchPixelDataBWLuma[i][j][k], - &v->RequiredPrefetchPixelDataBWChroma[i][j][k], - &v->NoTimeForDynamicMetadata[i][j][k], - &v->Tno_bw[k], - &v->prefetch_vmrow_bw[k], - &v->dummy7[k], - &v->dummy8[k], - &v->dummy13[k], - &v->VUpdateOffsetPix[k], - &v->VUpdateWidthPix[k], - &v->VReadyOffsetPix[k]); + CalculatePrefetchSchedulePerPlane(mode_lib, + HostVMInefficiencyFactor, + i, j, k); } for (k = 0; k < v->NumberOfActivePlanes; k++) {
Hi Michel,
Overwall I think this series is good. I also run it in our internal CI and everything looks fine.
Reviewed-by: Rodrigo Siqueira Rodrigo.Siqueira@amd.com
And applied to amd-staging-drm-next.
Thanks!
Are there any plans for actually reducing the combined amount of stack used by ModeSupportAndSystemConfigurationFull + CalculatePrefetchSchedulePerPlane?
Hi Michel,
Tbh, I'm not fully aware of the problem with the stack size used by "ModeSupportAndSystemConfigurationFull + CalculatePrefetchSchedulePerPlane". Could you help me to understand it better? Could you provide some background? Also, could you help me better understand the impact of this stack size issue in the DML code? Any information will be helpful.
Also, did you check that UseMinimumDCFCLK now modifying mode_lib->vba.DCFCLKState[i][j] and possibly other values in mode_lib->vba makes sense?
To check this patch, I submitted it to our Internal CI, where we ran a couple of IGT tests in multiple ASICs, and I conducted a simple smoke test using 5600XT and a Raven system. Everything was fine.
Finally, I checked Dmytro's opinion about this change, and he agreed that your patch is fine.
Thanks Siqueira
On 2021-12-13 15:22, Rodrigo Siqueira Jordao wrote:
On 2021-12-13 4:46 a.m., Michel Dänzer wrote:
On 2021-12-11 13:20, Rodrigo Siqueira Jordao wrote:
On 2021-12-09 11:46 a.m., Michel Dänzer wrote:
From: Michel Dänzer mdaenzer@redhat.com
Move code using the Pipe struct to a new helper function.
Works around[0] this warning (resulting in failure to build a RHEL debug kernel with Werror enabled):
../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c: In function ‘dml31_ModeSupportAndSystemConfigurationFull’: ../drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_mode_vba_31.c:5740:1: warning: the frame size of 2144 bytes is larger than 2048 bytes [-Wframe-larger-than=]
The culprit seems to be the Pipe struct, so pull the relevant block out into its own sub-function. (This is porting a62427ef9b55 "drm/amd/display: Reduce stack size for dml21_ModeSupportAndSystemConfigurationFull" from dml31 to dml21)
[0] AFAICT this doesn't actually reduce the total amount of stack which can be used, just moves some of it from dml31_ModeSupportAndSystemConfigurationFull to the new helper function, so the former happens to no longer exceed the limit for a single function.
Signed-off-by: Michel Dänzer mdaenzer@redhat.com
.../dc/dml/dcn31/display_mode_vba_31.c | 185 ++++++++++-------- 1 file changed, 99 insertions(+), 86 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index 7e937bdcea00..8965f9af9d0a 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -3949,6 +3949,102 @@ static double TruncToValidBPP( return BPP_INVALID; } +static noinline void CalculatePrefetchSchedulePerPlane( + struct display_mode_lib *mode_lib, + double HostVMInefficiencyFactor, + int i, + unsigned j, + unsigned k) +{ + struct vba_vars_st *v = &mode_lib->vba; + Pipe myPipe;
+ myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; + myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; + myPipe.PixelClock = v->PixelClock[k]; + myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; + myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; + myPipe.ScalerEnabled = v->ScalerEnabled[k]; + myPipe.SourceScan = v->SourceScan[k]; + myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; + myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; + myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; + myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; + myPipe.InterlaceEnable = v->Interlace[k]; + myPipe.NumberOfCursors = v->NumberOfCursors[k]; + myPipe.VBlank = v->VTotal[k] - v->VActive[k]; + myPipe.HTotal = v->HTotal[k]; + myPipe.DCCEnable = v->DCCEnable[k]; + myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 + || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; + myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; + myPipe.BytePerPixelY = v->BytePerPixelY[k]; + myPipe.BytePerPixelC = v->BytePerPixelC[k]; + myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; + v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( + mode_lib, + HostVMInefficiencyFactor, + &myPipe, + v->DSCDelayPerState[i][k], + v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, + v->DPPCLKDelaySCL, + v->DPPCLKDelaySCLLBOnly, + v->DPPCLKDelayCNVCCursor, + v->DISPCLKDelaySubtotal, + v->SwathWidthYThisState[k] / v->HRatio[k], + v->OutputFormat[k], + v->MaxInterDCNTileRepeaters, + dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), + v->MaximumVStartup[i][j][k], + v->GPUVMMaxPageTableLevels, + v->GPUVMEnable, + v->HostVMEnable, + v->HostVMMaxNonCachedPageTableLevels, + v->HostVMMinPageSize, + v->DynamicMetadataEnable[k], + v->DynamicMetadataVMEnabled, + v->DynamicMetadataLinesBeforeActiveRequired[k], + v->DynamicMetadataTransmittedBytes[k], + v->UrgLatency[i], + v->ExtraLatency, + v->TimeCalc, + v->PDEAndMetaPTEBytesPerFrame[i][j][k], + v->MetaRowBytes[i][j][k], + v->DPTEBytesPerRow[i][j][k], + v->PrefetchLinesY[i][j][k], + v->SwathWidthYThisState[k], + v->PrefillY[k], + v->MaxNumSwY[k], + v->PrefetchLinesC[i][j][k], + v->SwathWidthCThisState[k], + v->PrefillC[k], + v->MaxNumSwC[k], + v->swath_width_luma_ub_this_state[k], + v->swath_width_chroma_ub_this_state[k], + v->SwathHeightYThisState[k], + v->SwathHeightCThisState[k], + v->TWait, + &v->DSTXAfterScaler[k], + &v->DSTYAfterScaler[k], + &v->LineTimesForPrefetch[k], + &v->PrefetchBW[k], + &v->LinesForMetaPTE[k], + &v->LinesForMetaAndDPTERow[k], + &v->VRatioPreY[i][j][k], + &v->VRatioPreC[i][j][k], + &v->RequiredPrefetchPixelDataBWLuma[i][j][k], + &v->RequiredPrefetchPixelDataBWChroma[i][j][k], + &v->NoTimeForDynamicMetadata[i][j][k], + &v->Tno_bw[k], + &v->prefetch_vmrow_bw[k], + &v->dummy7[k], + &v->dummy8[k], + &v->dummy13[k], + &v->VUpdateOffsetPix[k], + &v->VUpdateWidthPix[k], + &v->VReadyOffsetPix[k]); +}
void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) { struct vba_vars_st *v = &mode_lib->vba; @@ -5276,92 +5372,9 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->SREnterPlusExitTime); for (k = 0; k < v->NumberOfActivePlanes; k++) { - Pipe myPipe;
- myPipe.DPPCLK = v->RequiredDPPCLK[i][j][k]; - myPipe.DISPCLK = v->RequiredDISPCLK[i][j]; - myPipe.PixelClock = v->PixelClock[k]; - myPipe.DCFCLKDeepSleep = v->ProjectedDCFCLKDeepSleep[i][j]; - myPipe.DPPPerPlane = v->NoOfDPP[i][j][k]; - myPipe.ScalerEnabled = v->ScalerEnabled[k]; - myPipe.SourceScan = v->SourceScan[k]; - myPipe.BlockWidth256BytesY = v->Read256BlockWidthY[k]; - myPipe.BlockHeight256BytesY = v->Read256BlockHeightY[k]; - myPipe.BlockWidth256BytesC = v->Read256BlockWidthC[k]; - myPipe.BlockHeight256BytesC = v->Read256BlockHeightC[k]; - myPipe.InterlaceEnable = v->Interlace[k]; - myPipe.NumberOfCursors = v->NumberOfCursors[k]; - myPipe.VBlank = v->VTotal[k] - v->VActive[k]; - myPipe.HTotal = v->HTotal[k]; - myPipe.DCCEnable = v->DCCEnable[k]; - myPipe.ODMCombineIsEnabled = v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_4to1 - || v->ODMCombineEnablePerState[i][k] == dm_odm_combine_mode_2to1; - myPipe.SourcePixelFormat = v->SourcePixelFormat[k]; - myPipe.BytePerPixelY = v->BytePerPixelY[k]; - myPipe.BytePerPixelC = v->BytePerPixelC[k]; - myPipe.ProgressiveToInterlaceUnitInOPP = v->ProgressiveToInterlaceUnitInOPP; - v->NoTimeForPrefetch[i][j][k] = CalculatePrefetchSchedule( - mode_lib, - HostVMInefficiencyFactor, - &myPipe, - v->DSCDelayPerState[i][k], - v->DPPCLKDelaySubtotal + v->DPPCLKDelayCNVCFormater, - v->DPPCLKDelaySCL, - v->DPPCLKDelaySCLLBOnly, - v->DPPCLKDelayCNVCCursor, - v->DISPCLKDelaySubtotal, - v->SwathWidthYThisState[k] / v->HRatio[k], - v->OutputFormat[k], - v->MaxInterDCNTileRepeaters, - dml_min(v->MaxVStartup, v->MaximumVStartup[i][j][k]), - v->MaximumVStartup[i][j][k], - v->GPUVMMaxPageTableLevels, - v->GPUVMEnable, - v->HostVMEnable, - v->HostVMMaxNonCachedPageTableLevels, - v->HostVMMinPageSize, - v->DynamicMetadataEnable[k], - v->DynamicMetadataVMEnabled, - v->DynamicMetadataLinesBeforeActiveRequired[k], - v->DynamicMetadataTransmittedBytes[k], - v->UrgLatency[i], - v->ExtraLatency, - v->TimeCalc, - v->PDEAndMetaPTEBytesPerFrame[i][j][k], - v->MetaRowBytes[i][j][k], - v->DPTEBytesPerRow[i][j][k], - v->PrefetchLinesY[i][j][k], - v->SwathWidthYThisState[k], - v->PrefillY[k], - v->MaxNumSwY[k], - v->PrefetchLinesC[i][j][k], - v->SwathWidthCThisState[k], - v->PrefillC[k], - v->MaxNumSwC[k], - v->swath_width_luma_ub_this_state[k], - v->swath_width_chroma_ub_this_state[k], - v->SwathHeightYThisState[k], - v->SwathHeightCThisState[k], - v->TWait, - &v->DSTXAfterScaler[k], - &v->DSTYAfterScaler[k], - &v->LineTimesForPrefetch[k], - &v->PrefetchBW[k], - &v->LinesForMetaPTE[k], - &v->LinesForMetaAndDPTERow[k], - &v->VRatioPreY[i][j][k], - &v->VRatioPreC[i][j][k], - &v->RequiredPrefetchPixelDataBWLuma[i][j][k], - &v->RequiredPrefetchPixelDataBWChroma[i][j][k], - &v->NoTimeForDynamicMetadata[i][j][k], - &v->Tno_bw[k], - &v->prefetch_vmrow_bw[k], - &v->dummy7[k], - &v->dummy8[k], - &v->dummy13[k], - &v->VUpdateOffsetPix[k], - &v->VUpdateWidthPix[k], - &v->VReadyOffsetPix[k]); + CalculatePrefetchSchedulePerPlane(mode_lib, + HostVMInefficiencyFactor, + i, j, k); } for (k = 0; k < v->NumberOfActivePlanes; k++) {
Hi Michel,
Overwall I think this series is good. I also run it in our internal CI and everything looks fine.
Reviewed-by: Rodrigo Siqueira Rodrigo.Siqueira@amd.com
And applied to amd-staging-drm-next.
Thanks!
Are there any plans for actually reducing the combined amount of stack used by ModeSupportAndSystemConfigurationFull + CalculatePrefetchSchedulePerPlane?
Hi Michel,
Tbh, I'm not fully aware of the problem with the stack size used by "ModeSupportAndSystemConfigurationFull + CalculatePrefetchSchedulePerPlane". Could you help me to understand it better?
The warning which inspired this patch (and the corresponding change to dml21_ModeSupportAndSystemConfigurationFull) is about ModeSupportAndSystemConfigurationFull exceeding a threshold of stack usage. The patch moves some of that stack usage to the new CalculatePrefetchSchedulePerPlane helper function. However, since the former calls the latter, together they still use as much stack (possibly even slightly more, due to the function calling convention) as before. So while we've silenced the warning, we haven't actually improved the situation the warning is about.
Could you provide some background? Also, could you help me better understand the impact of this stack size issue in the DML code? Any information will be helpful.
I don't know the exact reasons for the warning offhand. Presumably the fact that this warning is enabled indicates that stack usage should be minimized though, or at least stack shouldn't be lightly wasted.
Also, did you check that UseMinimumDCFCLK now modifying mode_lib->vba.DCFCLKState[i][j] and possibly other values in mode_lib->vba makes sense?
To check this patch, I submitted it to our Internal CI, where we ran a couple of IGT tests in multiple ASICs, and I conducted a simple smoke test using 5600XT and a Raven system. Everything was fine.
Finally, I checked Dmytro's opinion about this change, and he agreed that your patch is fine.
Great, thanks.
dri-devel@lists.freedesktop.org
-
Michel Dänzer -
Michel Dänzer -
Rodrigo Siqueira Jordao