515.65.01

CHANGELOG.md

@@ -2,6 +2,13 @@
 
 ## Release 515 Entries
 
+### [515.65.01] 2022-08-02
+
+#### Fixed
+
+- Collection of minor fixes to issues, [#6](https://github.com/NVIDIA/open-gpu-kernel-modules/pull/61) by @Joshua-Ashton
+- Remove unnecessary use of acpi_bus_get_device().
+
 ### [515.57] 2022-06-28
 
 #### Fixed

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 515.57.
+version 515.65.01.
 
 
 ## How to Build
@@ -17,7 +17,7 @@ as root:
 
 Note that the kernel modules built here must be used with gsp.bin
 firmware and user-space NVIDIA GPU driver components from a corresponding
-515.57 driver release.  This can be achieved by installing
+515.65.01 driver release.  This can be achieved by installing
 the NVIDIA GPU driver from the .run file using the `--no-kernel-modules`
 option.  E.g.,
 
@@ -167,15 +167,15 @@ for the target kernel.
 ## Compatible GPUs
 
 The open-gpu-kernel-modules can be used on any Turing or later GPU
-(see the table below). However, in the 515.57 release,
+(see the table below). However, in the 515.65.01 release,
 GeForce and Workstation support is still considered alpha-quality.
 
 To enable use of the open kernel modules on GeForce and Workstation GPUs,
 set the "NVreg_OpenRmEnableUnsupportedGpus" nvidia.ko kernel module
 parameter to 1. For more details, see the NVIDIA GPU driver end user
 README here:
 
-https://us.download.nvidia.com/XFree86/Linux-x86_64/515.57/README/kernel_open.html
+https://us.download.nvidia.com/XFree86/Linux-x86_64/515.65.01/README/kernel_open.html
 
 In the below table, if three IDs are listed, the first is the PCI Device 
 ID, the second is the PCI Subsystem Vendor ID, and the third is the PCI
@@ -704,6 +704,8 @@ Subsystem Device ID.
 | NVIDIA RTX A5500                                | 2233 10DE 165A |
 | NVIDIA RTX A5500                                | 2233 17AA 165A |
 | NVIDIA A40                                      | 2235 10DE 145A |
+| NVIDIA A10                                      | 2236 10DE 1482 |
+| NVIDIA A10G                                     | 2237 10DE 152F |
 | NVIDIA A10M                                     | 2238 10DE 1677 |
 | NVIDIA GeForce RTX 3060 Ti                      | 2414           |
 | NVIDIA GeForce RTX 3080 Ti Laptop GPU           | 2420           |

kernel-open/Kbuild

@@ -72,7 +72,7 @@ EXTRA_CFLAGS += -I$(src)/common/inc
 EXTRA_CFLAGS += -I$(src)
 EXTRA_CFLAGS += -Wall -MD $(DEFINES) $(INCLUDES) -Wno-cast-qual -Wno-error -Wno-format-extra-args
 EXTRA_CFLAGS += -D__KERNEL__ -DMODULE -DNVRM
-EXTRA_CFLAGS += -DNV_VERSION_STRING=\"515.57\"
+EXTRA_CFLAGS += -DNV_VERSION_STRING=\"515.65.01\"
 
 EXTRA_CFLAGS += -Wno-unused-function
 

kernel-open/common/inc/nv.h

@@ -989,6 +989,7 @@ NvBool     NV_API_CALL rm_disable_iomap_wc(void);
 
 void       NV_API_CALL rm_init_dynamic_power_management(nvidia_stack_t *, nv_state_t *, NvBool);
 void       NV_API_CALL rm_cleanup_dynamic_power_management(nvidia_stack_t *, nv_state_t *);
+void       NV_API_CALL rm_enable_dynamic_power_management(nvidia_stack_t *, nv_state_t *);
 NV_STATUS  NV_API_CALL rm_ref_dynamic_power(nvidia_stack_t *, nv_state_t *, nv_dynamic_power_mode_t);
 void       NV_API_CALL rm_unref_dynamic_power(nvidia_stack_t *, nv_state_t *, nv_dynamic_power_mode_t);
 NV_STATUS  NV_API_CALL rm_transition_dynamic_power(nvidia_stack_t *, nv_state_t *, NvBool);

kernel-open/conftest.sh

@@ -5288,22 +5288,6 @@ compile_test() {
             rm -f conftest$$.c
         ;;
 
-        acpi_bus_get_device)
-            #
-            # Determine if the acpi_bus_get_device() function is present
-            #
-            # acpi_bus_get_device() was removed by commit ac2a3feefad5
-            # ("ACPI: bus: Eliminate acpi_bus_get_device()") in
-            # v5.18-rc2 (2022-04-05).
-            #
-            CODE="
-            #include <linux/acpi.h>
-            int conftest_acpi_bus_get_device(void) {
-                return acpi_bus_get_device();
-            }"
-            compile_check_conftest "$CODE" "NV_ACPI_BUS_GET_DEVICE_PRESENT" "" "functions"
-        ;;
-
         dma_resv_add_fence)
             #
             # Determine if the dma_resv_add_fence() function is present.
@@ -5365,6 +5349,23 @@ compile_test() {
             compile_check_conftest "$CODE" "NV_RESERVATION_OBJECT_RESERVE_SHARED_HAS_NUM_FENCES_ARG" "" "types"
         ;;
 
+        get_task_ioprio)
+            #
+            # Determine if the __get_task_ioprio() function is present.
+            #
+            # __get_task_ioprio was added by commit 893e5d32d583
+            # ("block: Generalize get_current_ioprio() for any task") for
+            # v5.20 linux-next (2022-06-23).
+            #
+            CODE="
+            #include <linux/ioprio.h>
+            void conftest_get_task_ioprio(void) {
+                __get_task_ioprio();
+            }"
+
+            compile_check_conftest "$CODE" "NV_GET_TASK_IOPRIO_PRESENT" "" "functions"
+        ;;
+
         # When adding a new conftest entry, please use the correct format for
         # specifying the relevant upstream Linux kernel commit.
         #

kernel-open/nvidia/nv-acpi.c

@@ -715,28 +715,18 @@ NV_STATUS NV_API_CALL nv_acpi_ddc_method(
 )
 {
     acpi_status status;
-    struct acpi_device *device = NULL;
     union acpi_object *ddc = NULL;
-    struct list_head *node, *next;
     NvU32 i, largestEdidSize;
     acpi_handle dev_handle  = NULL;
     acpi_handle lcd_dev_handle  = NULL;
+    acpi_handle handle = NULL;
 
     if (!nv_acpi_get_device_handle(nv, &dev_handle))
         return NV_ERR_NOT_SUPPORTED;
 
     if (!dev_handle)
         return NV_ERR_INVALID_ARGUMENT;
 
-#if defined(NV_ACPI_BUS_GET_DEVICE_PRESENT)
-    status = acpi_bus_get_device(dev_handle, &device);
-#else
-    return NV_ERR_NOT_SUPPORTED;
-#endif
-
-    if (ACPI_FAILURE(status) || !device)
-        return NV_ERR_INVALID_ARGUMENT;
-
     if (!NV_MAY_SLEEP())
     {
 #if defined(DEBUG)
@@ -747,16 +737,16 @@ NV_STATUS NV_API_CALL nv_acpi_ddc_method(
         return NV_ERR_NOT_SUPPORTED;
     }
 
-    list_for_each_safe(node, next, &device->children)
+    while (lcd_dev_handle == NULL)
     {
         unsigned long long device_id = 0;
-        struct acpi_device *dev =
-            list_entry(node, struct acpi_device, node);
 
-        if (!dev)
-            continue;
+        status = acpi_get_next_object(ACPI_TYPE_DEVICE, dev_handle,
+                                      handle, &handle);
+        if (ACPI_FAILURE(status) || (handle == NULL))
+            break;
 
-        status = acpi_evaluate_integer(dev->handle, "_ADR", NULL, &device_id);
+        status = acpi_evaluate_integer(handle, "_ADR", NULL, &device_id);
         if (ACPI_FAILURE(status))
             /* Couldnt query device_id for this device */
             continue;
@@ -766,16 +756,13 @@ NV_STATUS NV_API_CALL nv_acpi_ddc_method(
             case 0x0118:
             case 0x0400:
             case 0xA420:
-                lcd_dev_handle = dev->handle;
+                lcd_dev_handle = handle;
                 nv_printf(NV_DBG_INFO, "NVRM: %s Found LCD: %x\n",
                           __FUNCTION__, device_id);
                 break;
             default:
                 break;
         }
-
-        if (lcd_dev_handle != NULL)
-            break;
     }
 
     if (lcd_dev_handle == NULL)
@@ -1125,15 +1112,14 @@ NV_STATUS NV_API_CALL nv_acpi_mux_method(
 )
 {
     acpi_status status;
-    struct acpi_device *device    = NULL;
     struct acpi_buffer output     = { ACPI_ALLOCATE_BUFFER, NULL };
     union acpi_object *mux        = NULL;
     union acpi_object mux_arg     = { ACPI_TYPE_INTEGER };
     struct acpi_object_list input = { 1, &mux_arg };
     acpi_handle dev_handle        = NULL;
     acpi_handle mux_dev_handle    = NULL;
+    acpi_handle handle            = NULL;
     unsigned long long device_id  = 0;
-    struct list_head *node, *next;
 
     if ((strcmp(pMethodName, "MXDS") != 0)
         && (strcmp(pMethodName, "MXDM") != 0))
@@ -1154,16 +1140,6 @@ NV_STATUS NV_API_CALL nv_acpi_mux_method(
     if (!dev_handle)
         return NV_ERR_INVALID_ARGUMENT;
 
-#if defined(NV_ACPI_BUS_GET_DEVICE_PRESENT)
-    status = acpi_bus_get_device(dev_handle, &device);
-#else
-    return NV_ERR_NOT_SUPPORTED;
-#endif
-
-
-    if (ACPI_FAILURE(status) || !device)
-        return NV_ERR_INVALID_ARGUMENT;
-
     if (!NV_MAY_SLEEP())
     {
 #if defined(DEBUG)
@@ -1172,23 +1148,16 @@ NV_STATUS NV_API_CALL nv_acpi_mux_method(
         return NV_ERR_NOT_SUPPORTED;
     }
 
-    list_for_each_safe(node, next, &device->children)
+    while (mux_dev_handle == NULL)
     {
-        struct acpi_device *dev = list_entry(node, struct acpi_device, node);
-
-        if (!dev)
-            continue;
-
-        status = acpi_evaluate_integer(dev->handle, "_ADR", NULL, &device_id);
-        if (ACPI_FAILURE(status))
-            /* Could not query device_id for this device */
-            continue;
-
-        if (device_id == muxAcpiId)
-        {
-            mux_dev_handle = dev->handle;
+        status = acpi_get_next_object(ACPI_TYPE_DEVICE, dev_handle,
+                                      handle, &handle);
+        if (ACPI_FAILURE(status) || (handle == NULL))
             break;
-        }
+
+        status = acpi_evaluate_integer(handle, "_ADR", NULL, &device_id);
+        if (ACPI_SUCCESS(status) && (device_id == muxAcpiId))
+            mux_dev_handle = handle;
     }
 
     if (mux_dev_handle == NULL)

kernel-open/nvidia/nv-dmabuf.c

@@ -797,12 +797,12 @@ nv_dma_buf_reuse(
     }
 
 
-
-
-
-
-
-
+    if (buf->ops != &nv_dma_buf_ops)
+    {
+        nv_printf(NV_DBG_ERRORS, "NVRM: Invalid dma-buf fd\n");
+        status = NV_ERR_INVALID_ARGUMENT;
+        goto cleanup_dmabuf;
+    }
 
 
     priv = buf->priv;

kernel-open/nvidia/nv-pci.c

@@ -608,6 +608,13 @@ nv_pci_probe
     dev_pm_set_driver_flags(nvl->dev, DPM_FLAG_NEVER_SKIP);
 #endif
 
+    /*
+     * Dynamic power management should be enabled as the last step.
+     * Kernel runtime power management framework can put the device
+     * into the suspended state. Hardware register access should not be done
+     * after enabling dynamic power management.
+     */
+    rm_enable_dynamic_power_management(sp, nv);
     nv_kmem_cache_free_stack(sp);
 
     return 0;
@@ -683,8 +690,9 @@ nv_pci_remove(struct pci_dev *pci_dev)
     if ((NV_ATOMIC_READ(nvl->usage_count) != 0) && !(nv->is_external_gpu))
     {
         nv_printf(NV_DBG_ERRORS,
-                  "NVRM: Attempting to remove minor device %u with non-zero usage count!\n",
-                  nvl->minor_num);
+                  "NVRM: Attempting to remove device %04x:%02x:%02x.%x with non-zero usage count!\n",
+                  NV_PCI_DOMAIN_NUMBER(pci_dev), NV_PCI_BUS_NUMBER(pci_dev),
+                  NV_PCI_SLOT_NUMBER(pci_dev), PCI_FUNC(pci_dev->devfn));
 
         /*
          * We can't return from this function without corrupting state, so we wait for
@@ -708,9 +716,10 @@ nv_pci_remove(struct pci_dev *pci_dev)
             if (!nvl)
             {
                 /* The device was not found, which should not happen */
-                nv_printf(NV_DBG_ERRORS, "NVRM: Failed removal of minor device! %04x:%02x:%02x.%x\n",
+                nv_printf(NV_DBG_ERRORS,
+                          "NVRM: Failed removal of device %04x:%02x:%02x.%x!\n",
                           NV_PCI_DOMAIN_NUMBER(pci_dev), NV_PCI_BUS_NUMBER(pci_dev),
-                          NV_PCI_SLOT_NUMBER(pci_dev), PCI_FUNC(pci_dev->devfn));
+                          NV_PCI_SLOT_NUMBER(pci_dev), PCI_FUNC(pci_dev->devfn));  
                 WARN_ON(1);
                 goto done;
             }
@@ -719,8 +728,9 @@ nv_pci_remove(struct pci_dev *pci_dev)
         }
 
         nv_printf(NV_DBG_ERRORS,
-                  "NVRM: Continuing with GPU removal for minor device %u\n",
-                  nvl->minor_num);
+                  "NVRM: Continuing with GPU removal for device %04x:%02x:%02x.%x\n",
+                  NV_PCI_DOMAIN_NUMBER(pci_dev), NV_PCI_BUS_NUMBER(pci_dev),
+                  NV_PCI_SLOT_NUMBER(pci_dev), PCI_FUNC(pci_dev->devfn));
     }
 
     rm_check_for_gpu_surprise_removal(sp, nv);

kernel-open/nvidia/nv.c

@@ -5443,7 +5443,19 @@ NvBool NV_API_CALL nv_s2idle_pm_configured(void)
         return NV_FALSE;
     }
 
+    /*
+     * init_sync_kiocb() internally uses GPL licensed __get_task_ioprio() from
+     * v5.20-rc1.
+     */
+#if defined(NV_GET_TASK_IOPRIO_PRESENT)
+    memset(&kiocb, 0, sizeof(kiocb));
+    kiocb.ki_filp = file;
+    kiocb.ki_flags = iocb_flags(file);
+    kiocb.ki_ioprio = IOPRIO_DEFAULT;
+#else
     init_sync_kiocb(&kiocb, file);
+#endif
+
     kiocb.ki_pos = 0;
     iov_iter_kvec(&iter, READ, &iov, 1, sizeof(buf));
 

kernel-open/nvidia/nvidia.Kbuild

@@ -199,7 +199,7 @@ NV_CONFTEST_FUNCTION_COMPILE_TESTS += dma_buf_ops_has_map_atomic
 NV_CONFTEST_FUNCTION_COMPILE_TESTS += dma_buf_has_dynamic_attachment
 NV_CONFTEST_FUNCTION_COMPILE_TESTS += dma_buf_attachment_has_peer2peer
 NV_CONFTEST_FUNCTION_COMPILE_TESTS += dma_set_mask_and_coherent
-NV_CONFTEST_FUNCTION_COMPILE_TESTS += acpi_bus_get_device
+NV_CONFTEST_FUNCTION_COMPILE_TESTS += get_task_ioprio
 
 NV_CONFTEST_SYMBOL_COMPILE_TESTS += is_export_symbol_gpl_of_node_to_nid
 NV_CONFTEST_SYMBOL_COMPILE_TESTS += is_export_symbol_gpl_sme_active