diff options
author | Liam <byteslice@airmail.cc> | 2023-10-24 18:53:53 +0200 |
---|---|---|
committer | Liam <byteslice@airmail.cc> | 2023-10-25 19:05:55 +0200 |
commit | 4df063209bebeb8737cbefd47a8958977aff9c2d (patch) | |
tree | 4b0d8b2c2b777aaeebe2a3d10b3b770bf809b824 /src/core | |
parent | nvdrv: add ioctl command serialization, convert nvhost_as_gpu (diff) | |
download | yuzu-4df063209bebeb8737cbefd47a8958977aff9c2d.tar yuzu-4df063209bebeb8737cbefd47a8958977aff9c2d.tar.gz yuzu-4df063209bebeb8737cbefd47a8958977aff9c2d.tar.bz2 yuzu-4df063209bebeb8737cbefd47a8958977aff9c2d.tar.lz yuzu-4df063209bebeb8737cbefd47a8958977aff9c2d.tar.xz yuzu-4df063209bebeb8737cbefd47a8958977aff9c2d.tar.zst yuzu-4df063209bebeb8737cbefd47a8958977aff9c2d.zip |
Diffstat (limited to 'src/core')
-rw-r--r-- | src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp | 112 | ||||
-rw-r--r-- | src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h | 28 |
2 files changed, 43 insertions, 97 deletions
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp index 6081d92e9..92e677b3d 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp @@ -6,6 +6,7 @@ #include "common/logging/log.h" #include "core/core.h" #include "core/core_timing.h" +#include "core/hle/service/nvdrv/devices/ioctl_serialization.h" #include "core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h" #include "core/hle/service/nvdrv/nvdrv.h" @@ -27,23 +28,23 @@ NvResult nvhost_ctrl_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span<const u8> case 'G': switch (command.cmd) { case 0x1: - return ZCullGetCtxSize(input, output); + return Wrap1(&nvhost_ctrl_gpu::ZCullGetCtxSize, input, output); case 0x2: - return ZCullGetInfo(input, output); + return Wrap1(&nvhost_ctrl_gpu::ZCullGetInfo, input, output); case 0x3: - return ZBCSetTable(input, output); + return Wrap1(&nvhost_ctrl_gpu::ZBCSetTable, input, output); case 0x4: - return ZBCQueryTable(input, output); + return Wrap1(&nvhost_ctrl_gpu::ZBCQueryTable, input, output); case 0x5: - return GetCharacteristics(input, output); + return Wrap1(&nvhost_ctrl_gpu::GetCharacteristics1, input, output); case 0x6: - return GetTPCMasks(input, output); + return Wrap1(&nvhost_ctrl_gpu::GetTPCMasks1, input, output); case 0x7: - return FlushL2(input, output); + return Wrap1(&nvhost_ctrl_gpu::FlushL2, input, output); case 0x14: - return GetActiveSlotMask(input, output); + return Wrap1(&nvhost_ctrl_gpu::GetActiveSlotMask, input, output); case 0x1c: - return GetGpuTime(input, output); + return Wrap1(&nvhost_ctrl_gpu::GetGpuTime, input, output); default: break; } @@ -65,9 +66,9 @@ NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> case 'G': switch (command.cmd) { case 0x5: - return GetCharacteristics(input, output, inline_output); + return Wrap3(&nvhost_ctrl_gpu::GetCharacteristics3, input, output, inline_output); case 0x6: - return GetTPCMasks(input, output, inline_output); + return Wrap3(&nvhost_ctrl_gpu::GetTPCMasks3, input, output, inline_output); default: break; } @@ -82,10 +83,8 @@ NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span<const u8> void nvhost_ctrl_gpu::OnOpen(DeviceFD fd) {} void nvhost_ctrl_gpu::OnClose(DeviceFD fd) {} -NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::GetCharacteristics1(IoctlCharacteristics& params) { LOG_DEBUG(Service_NVDRV, "called"); - IoctlCharacteristics params{}; - std::memcpy(¶ms, input.data(), input.size()); params.gc.arch = 0x120; params.gc.impl = 0xb; params.gc.rev = 0xa1; @@ -123,15 +122,13 @@ NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span<const u8> input, std::spa params.gc.gr_compbit_store_base_hw = 0x0; params.gpu_characteristics_buf_size = 0xA0; params.gpu_characteristics_buf_addr = 0xdeadbeef; // Cannot be 0 (UNUSED) - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span<const u8> input, std::span<u8> output, - std::span<u8> inline_output) { +NvResult nvhost_ctrl_gpu::GetCharacteristics3(IoctlCharacteristics& params, + std::span<u8> inline_output) { LOG_DEBUG(Service_NVDRV, "called"); - IoctlCharacteristics params{}; - std::memcpy(¶ms, input.data(), input.size()); + params.gc.arch = 0x120; params.gc.impl = 0xb; params.gc.rev = 0xa1; @@ -169,70 +166,46 @@ NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span<const u8> input, std::spa params.gc.gr_compbit_store_base_hw = 0x0; params.gpu_characteristics_buf_size = 0xA0; params.gpu_characteristics_buf_addr = 0xdeadbeef; // Cannot be 0 (UNUSED) - - std::memcpy(output.data(), ¶ms, output.size()); - std::memcpy(inline_output.data(), ¶ms.gc, inline_output.size()); + std::memcpy(inline_output.data(), ¶ms.gc, + std::min(sizeof(params.gc), inline_output.size())); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span<const u8> input, std::span<u8> output) { - IoctlGpuGetTpcMasksArgs params{}; - std::memcpy(¶ms, input.data(), input.size()); +NvResult nvhost_ctrl_gpu::GetTPCMasks1(IoctlGpuGetTpcMasksArgs& params) { LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); if (params.mask_buffer_size != 0) { params.tcp_mask = 3; } - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span<const u8> input, std::span<u8> output, - std::span<u8> inline_output) { - IoctlGpuGetTpcMasksArgs params{}; - std::memcpy(¶ms, input.data(), input.size()); +NvResult nvhost_ctrl_gpu::GetTPCMasks3(IoctlGpuGetTpcMasksArgs& params, + std::span<u8> inline_output) { LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); if (params.mask_buffer_size != 0) { params.tcp_mask = 3; } - std::memcpy(output.data(), ¶ms, output.size()); - std::memcpy(inline_output.data(), ¶ms.tcp_mask, inline_output.size()); + std::memcpy(inline_output.data(), ¶ms.tcp_mask, + std::min(sizeof(params.tcp_mask), inline_output.size())); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetActiveSlotMask(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::GetActiveSlotMask(IoctlActiveSlotMask& params) { LOG_DEBUG(Service_NVDRV, "called"); - IoctlActiveSlotMask params{}; - if (input.size() > 0) { - std::memcpy(¶ms, input.data(), input.size()); - } params.slot = 0x07; params.mask = 0x01; - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(IoctlZcullGetCtxSize& params) { LOG_DEBUG(Service_NVDRV, "called"); - - IoctlZcullGetCtxSize params{}; - if (input.size() > 0) { - std::memcpy(¶ms, input.data(), input.size()); - } params.size = 0x1; - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::ZCullGetInfo(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::ZCullGetInfo(IoctlNvgpuGpuZcullGetInfoArgs& params) { LOG_DEBUG(Service_NVDRV, "called"); - - IoctlNvgpuGpuZcullGetInfoArgs params{}; - - if (input.size() > 0) { - std::memcpy(¶ms, input.data(), input.size()); - } - params.width_align_pixels = 0x20; params.height_align_pixels = 0x20; params.pixel_squares_by_aliquots = 0x400; @@ -243,53 +216,28 @@ NvResult nvhost_ctrl_gpu::ZCullGetInfo(std::span<const u8> input, std::span<u8> params.subregion_width_align_pixels = 0x20; params.subregion_height_align_pixels = 0x40; params.subregion_count = 0x10; - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::ZBCSetTable(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::ZBCSetTable(IoctlZbcSetTable& params) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); - - IoctlZbcSetTable params{}; - std::memcpy(¶ms, input.data(), input.size()); // TODO(ogniK): What does this even actually do? - - // Prevent null pointer being passed as arg 1 - if (output.empty()) { - LOG_WARNING(Service_NVDRV, "Avoiding passing null pointer to memcpy"); - } else { - std::memcpy(output.data(), ¶ms, output.size()); - } return NvResult::Success; } -NvResult nvhost_ctrl_gpu::ZBCQueryTable(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::ZBCQueryTable(IoctlZbcQueryTable& params) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); - - IoctlZbcQueryTable params{}; - std::memcpy(¶ms, input.data(), input.size()); - // TODO : To implement properly - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::FlushL2(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::FlushL2(IoctlFlushL2& params) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); - - IoctlFlushL2 params{}; - std::memcpy(¶ms, input.data(), input.size()); - // TODO : To implement properly - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetGpuTime(std::span<const u8> input, std::span<u8> output) { +NvResult nvhost_ctrl_gpu::GetGpuTime(IoctlGetGpuTime& params) { LOG_DEBUG(Service_NVDRV, "called"); - - IoctlGetGpuTime params{}; - std::memcpy(¶ms, input.data(), input.size()); params.gpu_time = static_cast<u64_le>(system.CoreTiming().GetGlobalTimeNs().count()); - std::memcpy(output.data(), ¶ms, output.size()); return NvResult::Success; } diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h index 97995551c..e1977a6b5 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h @@ -151,21 +151,19 @@ private: }; static_assert(sizeof(IoctlGetGpuTime) == 0x10, "IoctlGetGpuTime is incorrect size"); - NvResult GetCharacteristics(std::span<const u8> input, std::span<u8> output); - NvResult GetCharacteristics(std::span<const u8> input, std::span<u8> output, - std::span<u8> inline_output); - - NvResult GetTPCMasks(std::span<const u8> input, std::span<u8> output); - NvResult GetTPCMasks(std::span<const u8> input, std::span<u8> output, - std::span<u8> inline_output); - - NvResult GetActiveSlotMask(std::span<const u8> input, std::span<u8> output); - NvResult ZCullGetCtxSize(std::span<const u8> input, std::span<u8> output); - NvResult ZCullGetInfo(std::span<const u8> input, std::span<u8> output); - NvResult ZBCSetTable(std::span<const u8> input, std::span<u8> output); - NvResult ZBCQueryTable(std::span<const u8> input, std::span<u8> output); - NvResult FlushL2(std::span<const u8> input, std::span<u8> output); - NvResult GetGpuTime(std::span<const u8> input, std::span<u8> output); + NvResult GetCharacteristics1(IoctlCharacteristics& params); + NvResult GetCharacteristics3(IoctlCharacteristics& params, std::span<u8> inline_output); + + NvResult GetTPCMasks1(IoctlGpuGetTpcMasksArgs& params); + NvResult GetTPCMasks3(IoctlGpuGetTpcMasksArgs& params, std::span<u8> inline_output); + + NvResult GetActiveSlotMask(IoctlActiveSlotMask& params); + NvResult ZCullGetCtxSize(IoctlZcullGetCtxSize& params); + NvResult ZCullGetInfo(IoctlNvgpuGpuZcullGetInfoArgs& params); + NvResult ZBCSetTable(IoctlZbcSetTable& params); + NvResult ZBCQueryTable(IoctlZbcQueryTable& params); + NvResult FlushL2(IoctlFlushL2& params); + NvResult GetGpuTime(IoctlGetGpuTime& params); EventInterface& events_interface; |