550.40.63

CHANGELOG.md

@@ -2,6 +2,8 @@
 
 ## Release 550 Entries
 
+### [550.40.63] 2024-05-31
+
 ### [550.40.61] 2024-04-23
 
 ### [550.40.59] 2024-04-01

README.md

@@ -1,7 +1,7 @@
 # NVIDIA Linux Open GPU Kernel Module Source
 
 This is the source release of the NVIDIA Linux open GPU kernel modules,
-version 550.40.61.
+version 550.40.63.
 
 
 ## How to Build
@@ -17,7 +17,7 @@ as root:
 
 Note that the kernel modules built here must be used with GSP
 firmware and user-space NVIDIA GPU driver components from a corresponding
-550.40.61 driver release.  This can be achieved by installing
+550.40.63 driver release.  This can be achieved by installing
 the NVIDIA GPU driver from the .run file using the `--no-kernel-modules`
 option.  E.g.,
 
@@ -188,7 +188,7 @@ encountered specific to them.
 For details on feature support and limitations, see the NVIDIA GPU driver
 end user README here:
 
-https://us.download.nvidia.com/XFree86/Linux-x86_64/550.40.61/README/kernel_open.html
+https://us.download.nvidia.com/XFree86/Linux-x86_64/550.40.63/README/kernel_open.html
 
 For vGPU support, please refer to the README.vgpu packaged in the vGPU Host
 Package for more details.
@@ -651,6 +651,7 @@ Subsystem Device ID.
 | NVIDIA T1000 8GB                                | 1FF0 17AA 1612 |
 | NVIDIA T400 4GB                                 | 1FF2 1028 1613 |
 | NVIDIA T400 4GB                                 | 1FF2 103C 1613 |
+| NVIDIA T400E                                    | 1FF2 103C 18FF |
 | NVIDIA T400 4GB                                 | 1FF2 103C 8A80 |
 | NVIDIA T400 4GB                                 | 1FF2 10DE 1613 |
 | NVIDIA T400E                                    | 1FF2 10DE 18FF |
@@ -829,6 +830,14 @@ Subsystem Device ID.
 | NVIDIA GeForce RTX 3050 4GB Laptop GPU          | 25AB           |
 | NVIDIA GeForce RTX 3050 6GB Laptop GPU          | 25AC           |
 | NVIDIA GeForce RTX 2050                         | 25AD           |
+| NVIDIA RTX A1000                                | 25B0 1028 1878 |
+| NVIDIA RTX A1000                                | 25B0 103C 1878 |
+| NVIDIA RTX A1000                                | 25B0 10DE 1878 |
+| NVIDIA RTX A1000                                | 25B0 17AA 1878 |
+| NVIDIA RTX A400                                 | 25B2 1028 1879 |
+| NVIDIA RTX A400                                 | 25B2 103C 1879 |
+| NVIDIA RTX A400                                 | 25B2 10DE 1879 |
+| NVIDIA RTX A400                                 | 25B2 17AA 1879 |
 | NVIDIA A16                                      | 25B6 10DE 14A9 |
 | NVIDIA A2                                       | 25B6 10DE 157E |
 | NVIDIA RTX A2000 Laptop GPU                     | 25B8           |
@@ -907,8 +916,11 @@ Subsystem Device ID.
 | NVIDIA GeForce RTX 4050 Laptop GPU              | 28A1           |
 | NVIDIA RTX 2000 Ada Generation                  | 28B0 1028 1870 |
 | NVIDIA RTX 2000 Ada Generation                  | 28B0 103C 1870 |
+| NVIDIA RTX 2000E Ada Generation                 | 28B0 103C 1871 |
 | NVIDIA RTX 2000 Ada Generation                  | 28B0 10DE 1870 |
+| NVIDIA RTX 2000E Ada Generation                 | 28B0 10DE 1871 |
 | NVIDIA RTX 2000 Ada Generation                  | 28B0 17AA 1870 |
+| NVIDIA RTX 2000E Ada Generation                 | 28B0 17AA 1871 |
 | NVIDIA RTX 2000 Ada Generation Laptop GPU       | 28B8           |
 | NVIDIA RTX 1000 Ada Generation Laptop GPU       | 28B9           |
 | NVIDIA RTX 500 Ada Generation Laptop GPU        | 28BA           |

kernel-open/Kbuild

@@ -72,7 +72,7 @@ EXTRA_CFLAGS += -I$(src)/common/inc
 EXTRA_CFLAGS += -I$(src)
 EXTRA_CFLAGS += -Wall $(DEFINES) $(INCLUDES) -Wno-cast-qual -Wno-format-extra-args
 EXTRA_CFLAGS += -D__KERNEL__ -DMODULE -DNVRM
-EXTRA_CFLAGS += -DNV_VERSION_STRING=\"550.40.61\"
+EXTRA_CFLAGS += -DNV_VERSION_STRING=\"550.40.63\"
 
 ifneq ($(SYSSRCHOST1X),)
  EXTRA_CFLAGS += -I$(SYSSRCHOST1X)

kernel-open/common/inc/nvstatuscodes.h

@@ -151,6 +151,7 @@ NV_STATUS_CODE(NV_ERR_RISCV_ERROR,                     0x00000079, "Generic RISC
 NV_STATUS_CODE(NV_ERR_FABRIC_MANAGER_NOT_PRESENT,      0x0000007A, "Fabric Manager is not loaded")
 NV_STATUS_CODE(NV_ERR_ALREADY_SIGNALLED,               0x0000007B, "Semaphore Surface value already >= requested wait value")
 NV_STATUS_CODE(NV_ERR_QUEUE_TASK_SLOT_NOT_AVAILABLE,   0x0000007C, "PMU RPC error due to no queue slot available for this event")
+NV_STATUS_CODE(NV_ERR_KEY_ROTATION_IN_PROGRESS,        0x0000007D, "Operation not allowed as key rotation is in progress")
 
 // Warnings:
 NV_STATUS_CODE(NV_WARN_HOT_SWITCH,                     0x00010001, "WARNING Hot switch")

kernel-open/nvidia-uvm/uvm_channel_test.c

@@ -340,9 +340,9 @@ static NV_STATUS uvm_test_iommu_rc_for_gpu(uvm_gpu_t *gpu)
     if (!domain || !iommu_is_dma_domain(domain))
         return NV_OK;
 
-    // Only run if ATS is enabled. Otherwise the CE doesn't get response on
-    // writing to unmapped location.
-    if (!g_uvm_global.ats.enabled)
+    // Only run if ATS is enabled with 64kB base page.
+    // Otherwise the CE doesn't get response on writing to unmapped location.
+    if (!g_uvm_global.ats.enabled || PAGE_SIZE != UVM_PAGE_SIZE_64K)
         return NV_OK;
 
     status = uvm_mem_alloc_sysmem_and_map_cpu_kernel(data_size, NULL, &sysmem);

kernel-open/nvidia-uvm/uvm_va_space.c

@@ -527,6 +527,15 @@ void uvm_va_space_destroy(uvm_va_space_t *va_space)
     nv_kthread_q_flush(&g_uvm_global.global_q);
 
     for_each_gpu_in_mask(gpu, retained_gpus) {
+        // Free the processor masks allocated in uvm_va_space_register_gpu().
+        // The mask is also freed in uvm_va_space_unregister_gpu() but that
+        // function won't be called in uvm_release() and uvm_release_deferred()
+        // path.
+        uvm_processor_mask_cache_free(va_space->peers_to_release[uvm_id_value(gpu->id)]);
+
+        // Set the pointer to NULL to avoid accidental re-use and double free.
+        va_space->peers_to_release[uvm_id_value(gpu->id)] = NULL;
+
         if (!gpu->parent->isr.replayable_faults.handling) {
             UVM_ASSERT(!gpu->parent->isr.non_replayable_faults.handling);
             continue;
@@ -543,14 +552,6 @@ void uvm_va_space_destroy(uvm_va_space_t *va_space)
         if (gpu->parent->access_counters_supported)
             uvm_parent_gpu_access_counters_disable(gpu->parent, va_space);
 
-        // Free the processor masks allocated in uvm_va_space_register_gpu().
-        // The mask is also freed in uvm_va_space_unregister_gpu() but that
-        // function won't be called in uvm_release() and uvm_release_deferred()
-        // path.
-        uvm_processor_mask_cache_free(va_space->peers_to_release[uvm_id_value(gpu->id)]);
-
-        // Set the pointer to NULL to avoid accidental re-use and double free.
-        va_space->peers_to_release[uvm_id_value(gpu->id)] = NULL;
     }
 
     // Check that all CPU/GPU affinity masks are empty

kernel-open/nvidia/export_nvswitch.h

@@ -998,6 +998,22 @@ nvswitch_os_get_supported_register_events_params
     NvBool *bUserSuppliesOsData
 );
 
+/*
+ * @Brief : Is TNVL mode enabled.
+ *
+ * @Description : Returns if TNVL is enabled for the device
+ *
+ * @param[in] device        a reference to the device
+ *
+ * @returns                 NV_TRUE,  if TNVL is enabled
+ *                          NV_FALSE, if TNVL is disabled
+ */
+NvBool
+nvswitch_lib_is_tnvl_enabled
+(
+    nvswitch_device *device
+);
+
 #ifdef __cplusplus
 }
 #endif

kernel-open/nvidia/ioctl_common_nvswitch.h

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2019 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 2019-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a

kernel-open/nvidia/ioctl_nvswitch.h

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2017-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 2017-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -164,6 +164,7 @@ typedef struct
     NVSWITCH_DEVICE_FABRIC_STATE deviceState;
     NVSWITCH_DEVICE_BLACKLIST_REASON deviceReason;
     NvU32 physId;
+    NvBool bTnvlEnabled;
 
     /* See ctrl_dev_nvswitch.h for struct definition modification guidelines */
 } NVSWITCH_DEVICE_INSTANCE_INFO_V2;

kernel-open/nvidia/libspdm_aead.c

@@ -159,7 +159,14 @@ static int lkca_aead_internal(struct crypto_aead *aead,
     }
 
     if (rc != 0) {
-        pr_info("Encryption FAILED\n");
+        if (enc) {
+            pr_info("aead.c: Encryption failed with error %i\n", rc);
+        } else {
+            pr_info("aead.c: Decryption failed with error %i\n", rc);
+            if (rc == -EBADMSG) {
+                pr_info("aead.c: Authentication tag mismatch!\n");
+            }
+        }
     }
 
     *data_out_size = data_in_size;

kernel-open/nvidia/linux_nvswitch.c

@@ -1000,6 +1000,8 @@ nvswitch_ctl_get_devices_v2(NVSWITCH_GET_DEVICES_V2_PARAMS *p)
                                                  &p->info[index].deviceState,
                                                  &p->info[index].deviceReason,
                                                  &p->info[index].driverState);
+
+            p->info[index].bTnvlEnabled = nvswitch_lib_is_tnvl_enabled(nvswitch_dev->lib_device);
             mutex_unlock(&nvswitch_dev->device_mutex);
         }
         index++;

kernel-open/nvidia/os-interface.c

@@ -1061,7 +1061,15 @@ NV_STATUS NV_API_CALL os_flush_user_cache(void)
 
 void NV_API_CALL os_flush_cpu_write_combine_buffer(void)
 {
-    wmb();
+#if defined(NVCPU_X86_64)
+    asm volatile("sfence" ::: "memory");
+#elif defined(NVCPU_PPC64LE)
+    __asm__ __volatile__ ("sync" : : : "memory");
+#elif defined(NVCPU_AARCH64)
+    asm volatile("dsb st" : : : "memory");
+#else
+    mb();
+#endif
 }
 
 // override initial debug level from registry

src/common/displayport/inc/dp_connectorimpl.h

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 1993-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 1993-2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -360,6 +360,7 @@ namespace DisplayPort
         // the stale messages from previous discovery.
         //
         bool        bForceClearPendingMsg;
+        bool        bSkipFakeDeviceDpcdAccess;
 
 
         Group *perHeadAttachedGroup[NV_MAX_HEADS];

src/common/displayport/inc/dp_deviceimpl.h

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 1993-2021 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 1993-2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -199,8 +199,9 @@ namespace DisplayPort
         TriState bAsyncSDPCapable;
         bool bMSAOverMSTCapable;
         bool bDscPassThroughColorFormatWar;
+        bool bSkipFakeDeviceDpcdAccess;
 
-        DeviceImpl(DPCDHAL * hal, ConnectorImpl * connector, DeviceImpl * parent);
+        DeviceImpl(DPCDHAL * hal, ConnectorImpl * connector, DeviceImpl * parent, bool bSkipFakeDeviceDpcdAccess);
         ~DeviceImpl();
 
         virtual bool isCableOk();

src/common/displayport/inc/dp_regkeydatabase.h

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2020-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 2020-2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -83,6 +83,7 @@
 // Bug 4459839 : This regkey will enable DSC irrespective of LT status.
 //
 #define NV_DP_REGKEY_FORCE_DSC_ON_SINK                 "DP_FORCE_DSC_ON_SINK"
+#define NV_DP_REGKEY_ENABLE_SKIP_DPCD_READS_WAR        "DP_BUG_4478047_WAR"
 
 //
 // Data Base used to store all the regkey values.
@@ -119,6 +120,7 @@ struct DP_REGKEY_DATABASE
     bool  bReassessMaxLink;
     bool  bMSTPCONCapsReadDisabled;
     bool  bForceDscOnSink;
+    bool  bSkipFakeDeviceDpcdAccess;
 };
 
 #endif //INCLUDED_DP_REGKEYDATABASE_H

src/common/displayport/src/dp_connectorimpl.cpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 1993-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 1993-2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -166,15 +166,16 @@ void ConnectorImpl::applyRegkeyOverrides(const DP_REGKEY_DATABASE& dpRegkeyDatab
         this->bKeepLinkAliveMST = dpRegkeyDatabase.bOptLinkKeptAliveMst;
         this->bKeepLinkAliveSST = dpRegkeyDatabase.bOptLinkKeptAliveSst;
     }
-    this->bReportDeviceLostBeforeNew    = dpRegkeyDatabase.bReportDeviceLostBeforeNew;
-    this->maxLinkRateFromRegkey         = dpRegkeyDatabase.applyMaxLinkRateOverrides;
-    this->bEnableAudioBeyond48K         = dpRegkeyDatabase.bAudioBeyond48kEnabled;
-    this->bDisableSSC                   = dpRegkeyDatabase.bSscDisabled;
-    this->bEnableFastLT                 = dpRegkeyDatabase.bFastLinkTrainingEnabled;
-    this->bDscMstCapBug3143315          = dpRegkeyDatabase.bDscMstCapBug3143315;
-    this->bPowerDownPhyBeforeD3         = dpRegkeyDatabase.bPowerDownPhyBeforeD3;
-    this->bReassessMaxLink              = dpRegkeyDatabase.bReassessMaxLink;
-    this->bForceDscOnSink               = dpRegkeyDatabase.bForceDscOnSink;
+    this->bReportDeviceLostBeforeNew     = dpRegkeyDatabase.bReportDeviceLostBeforeNew;
+    this->maxLinkRateFromRegkey          = dpRegkeyDatabase.applyMaxLinkRateOverrides;
+    this->bEnableAudioBeyond48K          = dpRegkeyDatabase.bAudioBeyond48kEnabled;
+    this->bDisableSSC                    = dpRegkeyDatabase.bSscDisabled;
+    this->bEnableFastLT                  = dpRegkeyDatabase.bFastLinkTrainingEnabled;
+    this->bDscMstCapBug3143315           = dpRegkeyDatabase.bDscMstCapBug3143315;
+    this->bPowerDownPhyBeforeD3          = dpRegkeyDatabase.bPowerDownPhyBeforeD3;
+    this->bReassessMaxLink               = dpRegkeyDatabase.bReassessMaxLink;
+    this->bForceDscOnSink                = dpRegkeyDatabase.bForceDscOnSink;
+    this->bSkipFakeDeviceDpcdAccess      = dpRegkeyDatabase.bSkipFakeDeviceDpcdAccess;
 }
 
 void ConnectorImpl::setPolicyModesetOrderMitigation(bool enabled)
@@ -478,7 +479,7 @@ void ConnectorImpl::processNewDevice(const DiscoveryManager::Device & device,
     }
     else
     {
-        newDev = new DeviceImpl(hal, this, parent);
+        newDev = new DeviceImpl(hal, this, parent, this->bSkipFakeDeviceDpcdAccess);
     }
 
     if (parent)
@@ -4632,11 +4633,6 @@ bool ConnectorImpl::trainLinkOptimized(LinkConfiguration lConfig)
                 }
             }
 
-            //
-            // There is no point in fallback here since we are link training  
-            // to loweset link config that can support the mode.
-            //
-            lowestSelected.policy.setSkipFallBack(true);
             bLinkTrainingSuccessful = train(lowestSelected, false);
             //
             // If LT failed, check if skipLT was marked. If so, clear the flag and
@@ -7022,7 +7018,7 @@ void ConnectorImpl::createFakeMuxDevice(const NvU8 *buffer, NvU32 bufferSize)
         return;
     }
 
-    DeviceImpl *newDev = new DeviceImpl(hal, this, NULL);
+    DeviceImpl *newDev = new DeviceImpl(hal, this, NULL, this->bSkipFakeDeviceDpcdAccess);
     if (!newDev)
     {
         return;

src/common/displayport/src/dp_deviceimpl.cpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 1993-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 1993-2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -75,7 +75,7 @@ DeviceImpl::~DeviceImpl()
 }
 
 
-DeviceImpl::DeviceImpl(DPCDHAL * hal, ConnectorImpl * connector, DeviceImpl * parent)
+DeviceImpl::DeviceImpl(DPCDHAL * hal, ConnectorImpl * connector, DeviceImpl * parent, bool bSkipFakeDeviceDpcdAccess)
     : parent(parent),
       hal(hal),
       activeGroup(0),
@@ -95,7 +95,8 @@ DeviceImpl::DeviceImpl(DPCDHAL * hal, ConnectorImpl * connector, DeviceImpl * pa
       bIgnoreMsaCapCached(false),
       bSdpExtCapable(Indeterminate),
       bAsyncSDPCapable(Indeterminate),
-      bDscPassThroughColorFormatWar(false)
+      bDscPassThroughColorFormatWar(false),
+      bSkipFakeDeviceDpcdAccess(bSkipFakeDeviceDpcdAccess)
 {
     bandwidth.enum_path.dataValid = false;
     shadow.plugged = false;
@@ -375,6 +376,12 @@ AuxBus::status DeviceImpl::getDpcdData(unsigned offset, NvU8 * buffer,
                                        unsigned * sizeCompleted,
                                        unsigned  * pNakReason)
 {
+    if (this->bSkipFakeDeviceDpcdAccess && isFakedMuxDevice())
+    {
+        DP_LOG(("Device is faked, returning nack\n"));
+        return AuxBus::nack;
+    }
+
     if (!buffer || !sizeCompleted)
     {
         // default param may be NULL
@@ -403,6 +410,12 @@ AuxBus::status DeviceImpl::setDpcdData(unsigned offset, NvU8 * buffer,
                                        unsigned * sizeCompleted,
                                        unsigned  * pNakReason)
 {
+    if (this->bSkipFakeDeviceDpcdAccess && isFakedMuxDevice())
+    {
+        DP_LOG(("Device is faked, returning nack\n"));
+        return AuxBus::nack;
+    }
+
     if (!buffer || !sizeCompleted)
     {
         // default param may be NULL