Tentative fix for NVIDIA 470.256.02 driver for Linux 6.15-rc1 (EXPERIMENTAL)

nvidia-470xx-fix-linux-6.15.patch

From c8980466f837b6c1e961bcfef3cfb70435394736 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Joan=20Bruguera=20Mic=C3=B3?= <[email protected]>
Date: Sat, 12 Apr 2025 22:46:59 +0000
Subject: [PATCH] Tentative fix for NVIDIA 470.256.02 driver for Linux 6.15-rc1

This is an experimental patch! Use at your own risk!
TODO: Verify that the fix for vm_flags_(set|clear) is correct.
EDIT 20250426: I don't believe that this patch is clean: the VMA flags
               piggybacks on previous calls to acquire the VMA lock,
               which is very hacky, and most likely doesn't happen 100%
               of the time and in all code paths.

Replacing EXTRA_CFLAGS with ccflags-y is rel. commit "kbuild: remove
EXTRA_*FLAGS support" (Masahiro Yamada, 6 Feb 2025), according to which
they have been deprecated since 2007, so no need to add any fallback.

Thanks to Willy Frissen and Satadru Pramanik, who shared patches for
various of the necessary changes.
---
 Kbuild                                | 25 +++++++++----------------
 common/inc/nv-mm.h                    | 17 +++++++++++++++++
 common/inc/nv-timer.h                 | 10 ++++++++++
 nvidia-drm/nvidia-drm-connector.c     |  8 ++++++++
 nvidia-drm/nvidia-drm-linux.c         |  1 +
 nvidia-modeset/nvidia-modeset-linux.c |  5 +++--
 nvidia-uvm/uvm.c                      |  1 +
 nvidia/nv-frontend.c                  |  1 +
 nvidia/nv.c                           |  4 ++--
 9 files changed, 52 insertions(+), 20 deletions(-)

diff --git a/Kbuild b/Kbuild
index eadd8b2..f333e0b 100644
--- a/Kbuild
+++ b/Kbuild
@@ -59,27 +59,20 @@ $(foreach _module, $(NV_KERNEL_MODULES), \
  $(eval include $(src)/$(_module)/$(_module).Kbuild))
 
 
-#
-# Define CFLAGS that apply to all the NVIDIA kernel modules. EXTRA_CFLAGS
-# is deprecated since 2.6.24 in favor of ccflags-y, but we need to support
-# older kernels which do not have ccflags-y. Newer kernels append
-# $(EXTRA_CFLAGS) to ccflags-y for compatibility.
-#
-
-EXTRA_CFLAGS += -I$(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 -DNV_VERSION_STRING=\"470.256.02\" -Wno-unused-function -Wuninitialized -fno-strict-aliasing -mno-red-zone -mcmodel=kernel -DNV_UVM_ENABLE
-EXTRA_CFLAGS += $(call cc-option,-Werror=undef,)
-EXTRA_CFLAGS += -DNV_SPECTRE_V2=$(NV_SPECTRE_V2)
-EXTRA_CFLAGS += -DNV_KERNEL_INTERFACE_LAYER
+ccflags-y += -I$(src)/common/inc
+ccflags-y += -I$(src)
+ccflags-y += -Wall -MD $(DEFINES) $(INCLUDES) -Wno-cast-qual -Wno-error -Wno-format-extra-args
+ccflags-y += -D__KERNEL__ -DMODULE -DNVRM -DNV_VERSION_STRING=\"470.256.02\" -Wno-unused-function -Wuninitialized -fno-strict-aliasing -mno-red-zone -mcmodel=kernel -DNV_UVM_ENABLE
+ccflags-y += $(call cc-option,-Werror=undef,)
+ccflags-y += -DNV_SPECTRE_V2=$(NV_SPECTRE_V2)
+ccflags-y += -DNV_KERNEL_INTERFACE_LAYER
 
 #
 # Detect SGI UV systems and apply system-specific optimizations.
 #
 
 ifneq ($(wildcard /proc/sgi_uv),)
- EXTRA_CFLAGS += -DNV_CONFIG_X86_UV
+ ccflags-y += -DNV_CONFIG_X86_UV
 endif
 
 
@@ -107,7 +100,7 @@ NV_CONFTEST_CMD := /bin/sh $(NV_CONFTEST_SCRIPT) \
 
 NV_CFLAGS_FROM_CONFTEST := $(shell $(NV_CONFTEST_CMD) build_cflags)
 
-NV_CONFTEST_CFLAGS = $(NV_CFLAGS_FROM_CONFTEST) $(EXTRA_CFLAGS) -fno-pie
+NV_CONFTEST_CFLAGS = $(NV_CFLAGS_FROM_CONFTEST) $(ccflags-y) -fno-pie
 
 NV_CONFTEST_COMPILE_TEST_HEADERS := $(obj)/conftest/macros.h
 NV_CONFTEST_COMPILE_TEST_HEADERS += $(obj)/conftest/functions.h
diff --git a/common/inc/nv-mm.h b/common/inc/nv-mm.h
index da5065d..1960ce5 100644
--- a/common/inc/nv-mm.h
+++ b/common/inc/nv-mm.h
@@ -31,6 +31,7 @@ typedef int vm_fault_t;
 
 #include <linux/mm.h>
 #include <linux/sched.h>
+#include <linux/version.h>
 /* get_user_pages
  *
  * The 8-argument version of get_user_pages was deprecated by commit
@@ -248,12 +249,28 @@ static inline struct rw_semaphore *nv_mmap_get_lock(struct mm_struct *mm)
 #if defined(NV_VM_AREA_STRUCT_HAS_CONST_VM_FLAGS)
 static inline void nv_vm_flags_set(struct vm_area_struct *vma, vm_flags_t flags)
 {
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(6, 15, 0)
+    // Rel. commit "mm: uninline the main body of vma_start_write()" (Suren Baghdasaryan, 13 Feb 2025)
+    // Since Linux 6.15, vm_flags_set and vm_flags_clear call a GPL-only symbol
+    // for locking (__vma_start_write), which can't be called from non-GPL code.
+    // However, it appears all uses on the driver are on VMAs being initially
+    // mapped / which are already locked, so we can use vm_flags_reset, which
+    // doesn't lock the VMA, but rather just asserts it is already write-locked.
+    vm_flags_reset(vma, vma->vm_flags | flags);
+#else
     vm_flags_set(vma, flags);
+#endif
 }
 
 static inline void nv_vm_flags_clear(struct vm_area_struct *vma, vm_flags_t flags)
 {
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(6, 15, 0)
+    // Rel. commit "mm: uninline the main body of vma_start_write()" (Suren Baghdasaryan, 13 Feb 2025)
+    // See above
+    vm_flags_reset(vma, vma->vm_flags & ~flags);
+#else
     vm_flags_clear(vma, flags);
+#endif
 }
 #else
 static inline void nv_vm_flags_set(struct vm_area_struct *vma, unsigned long flags)
diff --git a/common/inc/nv-timer.h b/common/inc/nv-timer.h
index 18df6ea..396e121 100644
--- a/common/inc/nv-timer.h
+++ b/common/inc/nv-timer.h
@@ -25,6 +25,7 @@
 
 #include <linux/timer.h>
 #include <linux/kernel.h> // For container_of
+#include <linux/version.h>
 
 #include "conftest.h"
 
@@ -63,4 +64,13 @@ static inline void nv_timer_setup(struct nv_timer *nv_timer,
 #endif
 }
 
+#if LINUX_VERSION_CODE < KERNEL_VERSION(6, 2, 0)
+// Rel. commit "treewide: Switch/rename to timer_delete[_sync]()" (Thomas Gleixner, 5 Apr 2025)
+// This provides a shim for ancient kernels before timer_delete_sync was introduced
+static inline int timer_delete_sync(struct timer_list *timer)
+{
+    return del_timer_sync(timer);
+}
+#endif
+
 #endif // __NV_TIMER_H__
diff --git a/nvidia-drm/nvidia-drm-connector.c b/nvidia-drm/nvidia-drm-connector.c
index fe838ef..c35c9ec 100644
--- a/nvidia-drm/nvidia-drm-connector.c
+++ b/nvidia-drm/nvidia-drm-connector.c
@@ -44,6 +44,8 @@
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_edid.h>
 
+#include <linux/version.h>
+
 static void nv_drm_connector_destroy(struct drm_connector *connector)
 {
     struct nv_drm_connector *nv_connector = to_nv_connector(connector);
@@ -302,8 +304,14 @@ static int nv_drm_connector_get_modes(struct drm_connector *connector)
     return count;
 }
 
+// Rel. commit. "drm/connector: make mode_valid take a const struct drm_display_mode" (Dmitry Baryshkov, 14 Dec 2024)
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(6, 15, 0)
+static int nv_drm_connector_mode_valid(struct drm_connector          *connector,
+                                       const struct drm_display_mode *mode)
+#else
 static int nv_drm_connector_mode_valid(struct drm_connector    *connector,
                                        struct drm_display_mode *mode)
+#endif
 {
     struct drm_device *dev = connector->dev;
     struct nv_drm_device *nv_dev = to_nv_device(dev);
diff --git a/nvidia-drm/nvidia-drm-linux.c b/nvidia-drm/nvidia-drm-linux.c
index 94e463e..2319640 100644
--- a/nvidia-drm/nvidia-drm-linux.c
+++ b/nvidia-drm/nvidia-drm-linux.c
@@ -182,6 +182,7 @@ static void __exit nv_linux_drm_exit(void)
 module_init(nv_linux_drm_init);
 module_exit(nv_linux_drm_exit);
 
+MODULE_DESCRIPTION("NVIDIA GPU DRM kernel module");
 #if defined(MODULE_LICENSE)
 
 
diff --git a/nvidia-modeset/nvidia-modeset-linux.c b/nvidia-modeset/nvidia-modeset-linux.c
index de0c748..f6810d2 100644
--- a/nvidia-modeset/nvidia-modeset-linux.c
+++ b/nvidia-modeset/nvidia-modeset-linux.c
@@ -718,7 +718,7 @@ static void nvkms_kthread_q_callback(void *arg)
      * pending timers and than waiting for workqueue callbacks.
      */
     if (timer->kernel_timer_created) {
-        del_timer_sync(&timer->kernel_timer);
+        timer_delete_sync(&timer->kernel_timer);
     }
 
     /*
@@ -1740,7 +1740,7 @@ restart:
              * completion, and we wait for queue completion with
              * nv_kthread_q_stop below.
              */
-            if (del_timer_sync(&timer->kernel_timer) == 1) {
+            if (timer_delete_sync(&timer->kernel_timer) == 1) {
                 /*  We've deactivated timer so we need to clean after it */
                 list_del(&timer->timers_list);
 
@@ -1778,6 +1778,7 @@ restart:
 module_init(nvkms_init);
 module_exit(nvkms_exit);
 
+MODULE_DESCRIPTION("NVIDIA GPU modeset kernel module");
 #if defined(MODULE_LICENSE)
 
 
diff --git a/nvidia-uvm/uvm.c b/nvidia-uvm/uvm.c
index 73ceb70..c1e734f 100644
--- a/nvidia-uvm/uvm.c
+++ b/nvidia-uvm/uvm.c
@@ -1129,6 +1129,7 @@ static void __exit uvm_exit_entry(void)
 module_init(uvm_init_entry);
 module_exit(uvm_exit_entry);
 
+MODULE_DESCRIPTION("NVIDIA GPU UVM kernel module");
 MODULE_LICENSE("Dual MIT/GPL");
 MODULE_INFO(supported, "external");
 
diff --git a/nvidia/nv-frontend.c b/nvidia/nv-frontend.c
index 4d6d8af..ad82d6b 100644
--- a/nvidia/nv-frontend.c
+++ b/nvidia/nv-frontend.c
@@ -13,6 +13,7 @@
 #include "nv-reg.h"
 #include "nv-frontend.h"
 
+MODULE_DESCRIPTION("NVIDIA GPU frontend kernel module");
 #if defined(MODULE_LICENSE)
 
 
diff --git a/nvidia/nv.c b/nvidia/nv.c
index 9d7da39..57b5024 100644
--- a/nvidia/nv.c
+++ b/nvidia/nv.c
@@ -3838,7 +3838,7 @@ int NV_API_CALL nv_stop_rc_timer(
 
     nv_printf(NV_DBG_INFO, "NVRM: stopping rc timer\n");
     nv->rc_timer_enabled = 0;
-    del_timer_sync(&nvl->rc_timer.kernel_timer);
+    timer_delete_sync(&nvl->rc_timer.kernel_timer);
     nv_printf(NV_DBG_INFO, "NVRM: rc timer stopped\n");
 
     return 0;
@@ -3882,7 +3882,7 @@ void NV_API_CALL nv_stop_snapshot_timer(void)
     NV_SPIN_UNLOCK_IRQRESTORE(&nvl->snapshot_timer_lock, flags);
 
     if (timer_active)
-        del_timer_sync(&nvl->snapshot_timer.kernel_timer);
+        timer_delete_sync(&nvl->snapshot_timer.kernel_timer);
 }
 
 void NV_API_CALL nv_flush_snapshot_timer(void)
-- 
2.49.0