555.42.02

(cherry picked from commit 5a1c474)

Author: niv

CHANGELOG.md

@@ -1,10 +1,10 @@
 # Changelog
 
-## Release 550 Entries
+## Release 555 Entries
 
-### [550.100] 2024-07-09
+### [555.42.02] 2024-05-21
 
-### [550.90.07] 2024-06-04
+## Release 550 Entries
 
 ### [550.78] 2024-04-25
 

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 550.100.
+version 555.42.02.
 
 
 ## 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
-550.100 driver release.  This can be achieved by installing
+555.42.02 driver release.  This can be achieved by installing
 the NVIDIA GPU driver from the .run file using the `--no-kernel-modules`
 option.  E.g.,
 
@@ -74,7 +74,7 @@ kernel.
 
 The NVIDIA open kernel modules support the same range of Linux kernel
 versions that are supported with the proprietary NVIDIA kernel modules.
-This is currently Linux kernel 3.10 or newer.
+This is currently Linux kernel 4.15 or newer.
 
 
 ## How to Contribute
@@ -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_64/550.100/README/kernel_open.html
+https://us.download.nvidia.com/XFree86/Linux-x86_64/555.42.02/README/kernel_open.html
 
 For vGPU support, please refer to the README.vgpu packaged in the vGPU Host
 Package for more details.
@@ -757,8 +757,6 @@ Subsystem Device ID.
 | NVIDIA H100 80GB HBM3                           | 2330 10DE 16C0 |
 | NVIDIA H100 80GB HBM3                           | 2330 10DE 16C1 |
 | NVIDIA H100 PCIe                                | 2331 10DE 1626 |
-| NVIDIA H200                                     | 2335 10DE 18BE |
-| NVIDIA H200                                     | 2335 10DE 18BF |
 | NVIDIA H100                                     | 2339 10DE 17FC |
 | NVIDIA H800 NVL                                 | 233A 10DE 183A |
 | NVIDIA GH200 120GB                              | 2342 10DE 16EB |
@@ -858,6 +856,7 @@ Subsystem Device ID.
 | NVIDIA RTX A500 Embedded GPU                    | 25FB           |
 | NVIDIA GeForce RTX 4090                         | 2684           |
 | NVIDIA GeForce RTX 4090 D                       | 2685           |
+| NVIDIA GeForce RTX 4070 Ti SUPER                | 2689           |
 | NVIDIA RTX 6000 Ada Generation                  | 26B1 1028 16A1 |
 | NVIDIA RTX 6000 Ada Generation                  | 26B1 103C 16A1 |
 | NVIDIA RTX 6000 Ada Generation                  | 26B1 10DE 16A1 |
@@ -875,7 +874,6 @@ Subsystem Device ID.
 | NVIDIA L40S                                     | 26B9 10DE 1851 |
 | NVIDIA L40S                                     | 26B9 10DE 18CF |
 | NVIDIA L20                                      | 26BA 10DE 1957 |
-| NVIDIA L20                                      | 26BA 10DE 1990 |
 | NVIDIA GeForce RTX 4080 SUPER                   | 2702           |
 | NVIDIA GeForce RTX 4080                         | 2704           |
 | NVIDIA GeForce RTX 4070 Ti SUPER                | 2705           |

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=\"550.100\"
+EXTRA_CFLAGS += -DNV_VERSION_STRING=\"555.42.02\"
 
 ifneq ($(SYSSRCHOST1X),)
  EXTRA_CFLAGS += -I$(SYSSRCHOST1X)
@@ -118,7 +118,7 @@ ifeq ($(ARCH),x86_64)
 endif
 
 ifeq ($(ARCH),powerpc)
- EXTRA_CFLAGS += -mlittle-endian -mno-strict-align -mno-altivec
+ EXTRA_CFLAGS += -mlittle-endian -mno-strict-align
 endif
 
 EXTRA_CFLAGS += -DNV_UVM_ENABLE
@@ -172,6 +172,7 @@ NV_CFLAGS_FROM_CONFTEST := $(shell $(NV_CONFTEST_CMD) build_cflags)
 NV_CONFTEST_CFLAGS = $(NV_CFLAGS_FROM_CONFTEST) $(EXTRA_CFLAGS) -fno-pie
 NV_CONFTEST_CFLAGS += $(call cc-disable-warning,pointer-sign)
 NV_CONFTEST_CFLAGS += $(call cc-option,-fshort-wchar,)
+NV_CONFTEST_CFLAGS += $(call cc-option,-Werror=incompatible-pointer-types,)
 NV_CONFTEST_CFLAGS += -Wno-error
 
 NV_CONFTEST_COMPILE_TEST_HEADERS := $(obj)/conftest/macros.h

kernel-open/Makefile

@@ -28,7 +28,7 @@ else
   else
     KERNEL_UNAME ?= $(shell uname -r)
     KERNEL_MODLIB := /lib/modules/$(KERNEL_UNAME)
-    KERNEL_SOURCES := $(shell test -d $(KERNEL_MODLIB)/source && echo $(KERNEL_MODLIB)/source || echo $(KERNEL_MODLIB)/build)
+    KERNEL_SOURCES := $(shell ((test -d $(KERNEL_MODLIB)/source && echo $(KERNEL_MODLIB)/source) || (test -d $(KERNEL_MODLIB)/build/source && echo $(KERNEL_MODLIB)/build/source)) || echo $(KERNEL_MODLIB)/build)
   endif
 
   KERNEL_OUTPUT := $(KERNEL_SOURCES)
@@ -42,7 +42,11 @@ else
   else
     KERNEL_UNAME ?= $(shell uname -r)
     KERNEL_MODLIB := /lib/modules/$(KERNEL_UNAME)
-    ifeq ($(KERNEL_SOURCES), $(KERNEL_MODLIB)/source)
+    # $(filter patter...,text) - Returns all whitespace-separated words in text that
+    # do match any of the pattern words, removing any words that do not match.
+    # Set the KERNEL_OUTPUT only if either $(KERNEL_MODLIB)/source or
+    # $(KERNEL_MODLIB)/build/source path matches the KERNEL_SOURCES.
+    ifneq ($(filter $(KERNEL_SOURCES),$(KERNEL_MODLIB)/source $(KERNEL_MODLIB)/build/source),)
       KERNEL_OUTPUT := $(KERNEL_MODLIB)/build
       KBUILD_PARAMS := KBUILD_OUTPUT=$(KERNEL_OUTPUT)
     endif

kernel-open/common/inc/nv-linux.h

@@ -58,14 +58,10 @@
 #include <linux/version.h>
 #include <linux/utsname.h>
 
-#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 32)
-#error "This driver does not support kernels older than 2.6.32!"
-#elif LINUX_VERSION_CODE < KERNEL_VERSION(2, 7, 0)
-#  define KERNEL_2_6
-#elif LINUX_VERSION_CODE >= KERNEL_VERSION(3, 0, 0)
-#  define KERNEL_3
-#else
-#error "This driver does not support development kernels!"
+#if LINUX_VERSION_CODE == KERNEL_VERSION(4, 4, 0)
+// Version 4.4 is allowed, temporarily, although not officially supported.
+#elif LINUX_VERSION_CODE < KERNEL_VERSION(4, 15, 0)
+#error "This driver does not support kernels older than Linux 4.15!"
 #endif
 
 #if defined (CONFIG_SMP) && !defined (__SMP__)
@@ -836,16 +832,16 @@ static inline dma_addr_t nv_phys_to_dma(struct device *dev, NvU64 pa)
 #define NV_PRINT_AT(nv_debug_level,at)                                           \
     {                                                                            \
         nv_printf(nv_debug_level,                                                \
-            "NVRM: VM: %s:%d: 0x%p, %d page(s), count = %d, flags = 0x%08x, "    \
+            "NVRM: VM: %s:%d: 0x%p, %d page(s), count = %d, "                    \
             "page_table = 0x%p\n",  __FUNCTION__, __LINE__, at,                  \
             at->num_pages, NV_ATOMIC_READ(at->usage_count),                      \
-            at->flags, at->page_table);                                          \
+            at->page_table);                                                     \
     }
 
 #define NV_PRINT_VMA(nv_debug_level,vma)                                                 \
     {                                                                                    \
         nv_printf(nv_debug_level,                                                        \
-            "NVRM: VM: %s:%d: 0x%lx - 0x%lx, 0x%08x bytes @ 0x%016llx, 0x%p, 0x%p\n",    \
+            "NVRM: VM: %s:%d: 0x%lx - 0x%lx, 0x%08lx bytes @ 0x%016llx, 0x%p, 0x%p\n",    \
             __FUNCTION__, __LINE__, vma->vm_start, vma->vm_end, NV_VMA_SIZE(vma),        \
             NV_VMA_OFFSET(vma), NV_VMA_PRIVATE(vma), NV_VMA_FILE(vma));                  \
     }
@@ -1078,6 +1074,8 @@ static inline void nv_kmem_ctor_dummy(void *arg)
         kmem_cache_destroy(kmem_cache);     \
     }
 
+#define NV_KMEM_CACHE_ALLOC_ATOMIC(kmem_cache)     \
+    kmem_cache_alloc(kmem_cache, GFP_ATOMIC)
 #define NV_KMEM_CACHE_ALLOC(kmem_cache)     \
     kmem_cache_alloc(kmem_cache, GFP_KERNEL)
 #define NV_KMEM_CACHE_FREE(ptr, kmem_cache) \
@@ -1104,6 +1102,23 @@ static inline void *nv_kmem_cache_zalloc(struct kmem_cache *k, gfp_t flags)
 #endif
 }
 
+static inline int nv_kmem_cache_alloc_stack_atomic(nvidia_stack_t **stack)
+{
+    nvidia_stack_t *sp = NULL;
+#if defined(NVCPU_X86_64)
+    if (rm_is_altstack_in_use())
+    {
+        sp = NV_KMEM_CACHE_ALLOC_ATOMIC(nvidia_stack_t_cache);
+        if (sp == NULL)
+            return -ENOMEM;
+        sp->size = sizeof(sp->stack);
+        sp->top = sp->stack + sp->size;
+    }
+#endif
+    *stack = sp;
+    return 0;
+}
+
 static inline int nv_kmem_cache_alloc_stack(nvidia_stack_t **stack)
 {
     nvidia_stack_t *sp = NULL;

kernel-open/common/inc/nv-mm.h

@@ -29,17 +29,17 @@
 typedef int vm_fault_t;
 #endif
 
-/* pin_user_pages
- * Presence of pin_user_pages() also implies the presence of unpin-user_page().
- * Both were added in the v5.6-rc1
+/*
+ * pin_user_pages()
  *
- * pin_user_pages() was added by commit eddb1c228f7951d399240
- * ("mm/gup: introduce pin_user_pages*() and FOLL_PIN") in v5.6-rc1 (2020-01-30)
+ * Presence of pin_user_pages() also implies the presence of unpin-user_page().
+ * Both were added in the v5.6.
  *
- * Removed vmas parameter from pin_user_pages() by commit 40896a02751
- * ("mm/gup: remove vmas parameter from pin_user_pages()")
- * in linux-next, expected in v6.5-rc1 (2023-05-17)
+ * pin_user_pages() was added by commit eddb1c228f79
+ * ("mm/gup: introduce pin_user_pages*() and FOLL_PIN") in v5.6.
  *
+ * Removed vmas parameter from pin_user_pages() by commit 4c630f307455
+ * ("mm/gup: remove vmas parameter from pin_user_pages()") in v6.5.
  */
 
 #include <linux/mm.h>
@@ -63,25 +63,28 @@ typedef int vm_fault_t;
     #define NV_UNPIN_USER_PAGE put_page
 #endif // NV_PIN_USER_PAGES_PRESENT
 
-/* get_user_pages
+/*
+ * get_user_pages()
  *
- * The 8-argument version of get_user_pages was deprecated by commit
- * (2016 Feb 12: cde70140fed8429acf7a14e2e2cbd3e329036653)for the non-remote case
+ * The 8-argument version of get_user_pages() was deprecated by commit
+ * cde70140fed8 ("mm/gup: Overload get_user_pages() functions") in v4.6-rc1.
  * (calling get_user_pages with current and current->mm).
  *
- * Completely moved to the 6 argument version of get_user_pages -
- * 2016 Apr 4: c12d2da56d0e07d230968ee2305aaa86b93a6832
+ * Completely moved to the 6 argument version of get_user_pages() by
+ * commit c12d2da56d0e ("mm/gup: Remove the macro overload API migration
+ * helpers from the get_user*() APIs") in v4.6-rc4.
  *
- * write and force parameters were replaced with gup_flags by -
- * 2016 Oct 12: 768ae309a96103ed02eb1e111e838c87854d8b51
+ * write and force parameters were replaced with gup_flags by
+ * commit 768ae309a961 ("mm: replace get_user_pages() write/force parameters
+ * with gup_flags") in v4.9.
  *
  * A 7-argument version of get_user_pages was introduced into linux-4.4.y by
- * commit 8e50b8b07f462ab4b91bc1491b1c91bd75e4ad40 which cherry-picked the
- * replacement of the write and force parameters with gup_flags
+ * commit 8e50b8b07f462 ("mm: replace get_user_pages() write/force parameters
+ * with gup_flags") which cherry-picked the replacement of the write and
+ * force parameters with gup_flags.
  *
- * Removed vmas parameter from get_user_pages() by commit 7bbf9c8c99
- * ("mm/gup: remove unused vmas parameter from get_user_pages()")
- * in linux-next, expected in v6.5-rc1 (2023-05-17)
+ * Removed vmas parameter from get_user_pages() by commit 54d020692b34
+ * ("mm/gup: remove unused vmas parameter from get_user_pages()") in v6.5.
  *
  */
 
@@ -112,18 +115,19 @@ typedef int vm_fault_t;
     }
 #endif // NV_GET_USER_PAGES_HAS_ARGS_FLAGS
 
-/* pin_user_pages_remote
+/*
+ * pin_user_pages_remote()
  *
- * pin_user_pages_remote() was added by commit eddb1c228f7951d399240
- * ("mm/gup: introduce pin_user_pages*() and FOLL_PIN") in v5.6 (2020-01-30)
+ * pin_user_pages_remote() was added by commit eddb1c228f79
+ * ("mm/gup: introduce pin_user_pages*() and FOLL_PIN") in v5.6.
  *
  * pin_user_pages_remote() removed 'tsk' parameter by commit
- * 64019a2e467a ("mm/gup: remove task_struct pointer for  all gup code")
- * in v5.9-rc1 (2020-08-11). *
+ * 64019a2e467a ("mm/gup: remove task_struct pointer for all gup code")
+ * in v5.9.
  *
  * Removed unused vmas parameter from pin_user_pages_remote() by commit
- * 83bcc2e132("mm/gup: remove unused vmas parameter from pin_user_pages_remote()")
- * in linux-next, expected in v6.5-rc1 (2023-05-14)
+ * 0b295316b3a9 ("mm/gup: remove unused vmas parameter from
+ * pin_user_pages_remote()") in v6.5.
  *
  */
 
@@ -143,7 +147,7 @@ typedef int vm_fault_t;
 
 /*
  * get_user_pages_remote() was added by commit 1e9877902dc7
- * ("mm/gup: Introduce get_user_pages_remote()") in v4.6 (2016-02-12).
+ * ("mm/gup: Introduce get_user_pages_remote()") in v4.6.
  *
  * Note that get_user_pages_remote() requires the caller to hold a reference on
  * the task_struct (if non-NULL and if this API has tsk argument) and the mm_struct.
@@ -153,19 +157,17 @@ typedef int vm_fault_t;
  *
  * get_user_pages_remote() write/force parameters were replaced
  * with gup_flags by commit 9beae1ea8930 ("mm: replace get_user_pages_remote()
- * write/force parameters with gup_flags") in v4.9 (2016-10-13).
+ * write/force parameters with gup_flags") in v4.9.
  *
  * get_user_pages_remote() added 'locked' parameter by commit 5b56d49fc31d
- * ("mm: add locked parameter to get_user_pages_remote()") in
- * v4.10 (2016-12-14).
+ * ("mm: add locked parameter to get_user_pages_remote()") in v4.10.
  *
  * get_user_pages_remote() removed 'tsk' parameter by
  * commit 64019a2e467a ("mm/gup: remove task_struct pointer for
- * all gup code") in v5.9-rc1 (2020-08-11).
+ * all gup code") in v5.9.
  *
- * Removed vmas parameter from get_user_pages_remote() by commit a4bde14d549 
- * ("mm/gup: remove vmas parameter from get_user_pages_remote()")
- * in linux-next, expected in v6.5-rc1 (2023-05-14)
+ * Removed vmas parameter from get_user_pages_remote() by commit ca5e863233e8
+ * ("mm/gup: remove vmas parameter from get_user_pages_remote()") in v6.5.
  *
  */
 

kernel-open/common/inc/nv.h

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 1999-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-FileCopyrightText: Copyright (c) 1999-2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
  * SPDX-License-Identifier: MIT
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -609,6 +609,15 @@ typedef enum
     NV_POWER_STATE_RUNNING
 } nv_power_state_t;
 
+typedef struct
+{
+    const char *vidmem_power_status;
+    const char *dynamic_power_status;
+    const char *gc6_support;
+    const char *gcoff_support;
+    const char *s0ix_status;
+} nv_power_info_t;
+
 #define NV_PRIMARY_VGA(nv)      ((nv)->primary_vga)
 
 #define NV_IS_CTL_DEVICE(nv)    ((nv)->flags & NV_FLAG_CONTROL)
@@ -778,7 +787,7 @@ nv_state_t*  NV_API_CALL  nv_get_ctl_state       (void);
 
 void   NV_API_CALL  nv_set_dma_address_size      (nv_state_t *, NvU32 );
 
-NV_STATUS  NV_API_CALL  nv_alias_pages           (nv_state_t *, NvU32, NvU32, NvU32, NvU64, NvU64 *, void **);
+NV_STATUS  NV_API_CALL  nv_alias_pages           (nv_state_t *, NvU32, NvU64, NvU32, NvU32, NvU64, NvU64 *, void **);
 NV_STATUS  NV_API_CALL  nv_alloc_pages           (nv_state_t *, NvU32, NvU64, NvBool, NvU32, NvBool, NvBool, NvS32, NvU64 *, void **);
 NV_STATUS  NV_API_CALL  nv_free_pages            (nv_state_t *, NvU32, NvBool, NvU32, void *);
 
@@ -822,6 +831,7 @@ void   NV_API_CALL  nv_acpi_methods_init         (NvU32 *);
 void   NV_API_CALL  nv_acpi_methods_uninit       (void);
 
 NV_STATUS  NV_API_CALL  nv_acpi_method           (NvU32, NvU32, NvU32, void *, NvU16, NvU32 *, void *, NvU16 *);
+NV_STATUS  NV_API_CALL  nv_acpi_d3cold_dsm_for_upstream_port (nv_state_t *, NvU8 *, NvU32, NvU32, NvU32 *);
 NV_STATUS  NV_API_CALL  nv_acpi_dsm_method       (nv_state_t *, NvU8 *, NvU32, NvBool, NvU32, void *, NvU16, NvU32 *, void *, NvU16 *);
 NV_STATUS  NV_API_CALL  nv_acpi_ddc_method       (nv_state_t *, void *, NvU32 *, NvBool);
 NV_STATUS  NV_API_CALL  nv_acpi_dod_method       (nv_state_t *, NvU32 *, NvU32 *);
@@ -990,10 +1000,10 @@ NV_STATUS  NV_API_CALL  rm_p2p_init_mapping       (nvidia_stack_t *, NvU64, NvU6
 NV_STATUS  NV_API_CALL  rm_p2p_destroy_mapping    (nvidia_stack_t *, NvU64);
 NV_STATUS  NV_API_CALL  rm_p2p_get_pages          (nvidia_stack_t *, NvU64, NvU32, NvU64, NvU64, NvU64 *, NvU32 *, NvU32 *, NvU32 *, NvU8 **, void *);
 NV_STATUS  NV_API_CALL  rm_p2p_get_gpu_info       (nvidia_stack_t *, NvU64, NvU64, NvU8 **, void **);
-NV_STATUS  NV_API_CALL  rm_p2p_get_pages_persistent (nvidia_stack_t *,  NvU64, NvU64, void **, NvU64 *, NvU32 *, void *, void *);
+NV_STATUS  NV_API_CALL  rm_p2p_get_pages_persistent (nvidia_stack_t *,  NvU64, NvU64, void **, NvU64 *, NvU32 *, void *, void *, void **);
 NV_STATUS  NV_API_CALL  rm_p2p_register_callback  (nvidia_stack_t *, NvU64, NvU64, NvU64, void *, void (*)(void *), void *);
 NV_STATUS  NV_API_CALL  rm_p2p_put_pages          (nvidia_stack_t *, NvU64, NvU32, NvU64, void *);
-NV_STATUS  NV_API_CALL  rm_p2p_put_pages_persistent(nvidia_stack_t *, void *, void *);
+NV_STATUS  NV_API_CALL  rm_p2p_put_pages_persistent(nvidia_stack_t *, void *, void *, void *);
 NV_STATUS  NV_API_CALL  rm_p2p_dma_map_pages      (nvidia_stack_t *, nv_dma_device_t *, NvU8 *, NvU64, NvU32, NvU64 *, void **);
 NV_STATUS  NV_API_CALL  rm_dma_buf_dup_mem_handle (nvidia_stack_t *, nv_state_t *, NvHandle, NvHandle, NvHandle, NvHandle, void *, NvHandle, NvU64, NvU64, NvHandle *, void **);
 void       NV_API_CALL  rm_dma_buf_undup_mem_handle(nvidia_stack_t *, nv_state_t *, NvHandle, NvHandle);
@@ -1027,9 +1037,7 @@ void       NV_API_CALL rm_enable_dynamic_power_management(nvidia_stack_t *, nv_s
 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, NvBool *);
-const char* NV_API_CALL rm_get_vidmem_power_status(nvidia_stack_t *, nv_state_t *);
-const char* NV_API_CALL rm_get_dynamic_power_management_status(nvidia_stack_t *, nv_state_t *);
-const char* NV_API_CALL rm_get_gpu_gcx_support(nvidia_stack_t *, nv_state_t *, NvBool);
+void       NV_API_CALL rm_get_power_info(nvidia_stack_t *, nv_state_t *, nv_power_info_t *);
 
 void       NV_API_CALL rm_acpi_notify(nvidia_stack_t *, nv_state_t *, NvU32);
 void       NV_API_CALL rm_acpi_nvpcf_notify(nvidia_stack_t *);