Go to:
Gentoo Home
Documentation
Forums
Lists
Bugs
Planet
Store
Wiki
Get Gentoo!
Gentoo's Bugzilla – Attachment 27001 Details for
Bug 29336
NVIDIA driver version 6106 corrupts the frame buffer console when using bootsplash
Home
|
New
–
[Ex]
|
Browse
|
Search
|
Privacy Policy
|
[?]
|
Reports
|
Requests
|
Help
|
New Account
|
Log In
[x]
|
Forgot Password
Login:
[x]
[patch]
NVIDIA kernel 4363 patch (used by the ebuild)
NVIDIA_kernel-1.0-4363-2.6.diff (text/plain), 107.27 KB, created by
Sander de Leeuw
on 2004-03-07 13:35:36 UTC
(
hide
)
Description:
NVIDIA kernel 4363 patch (used by the ebuild)
Filename:
MIME Type:
Creator:
Sander de Leeuw
Created:
2004-03-07 13:35:36 UTC
Size:
107.27 KB
patch
obsolete
>diff -ruN NVIDIA_kernel-1.0-4363/Makefile NVIDIA_kernel-1.0-4363-2.6/Makefile >--- NVIDIA_kernel-1.0-4363/Makefile 2003-04-20 03:57:20.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/Makefile 1970-01-01 01:00:00.000000000 +0100 >@@ -1,169 +0,0 @@ >-# This Makefile is automatically generated; do not edit >-# Generated on 'builder3.nvidia.com' on Sat Apr 19 17:46:47 PDT 2003 >- >-LINUX_MODULE=nv-linux.o >-DEFINES=-D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DNTRM -D_GNU_SOURCE -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=4363 -DNV_UNIX -DNV_LINUX -DNV_INT64_OK -DNVCPU_X86 >-INCLUDES=-I. >- >-OBJECTS=nv.o os-agp.o os-interface.o os-registry.o >-HEADERS=os-agp.h os-interface.h nv-linux.h nv-misc.h nv.h rmretval.h nvtypes.h $(VERSION_HDR) >- >-CFLAGS=-Wall -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith -Wcast-qual -Wno-multichar -O -MD $(DEFINES) $(INCLUDES) -Wno-cast-qual >- >-RESMAN_KERNEL_MODULE=nv-kernel.o >- >-KERNDIR=/lib/modules/$(shell uname -r) >- >-# check for newer paths. if found, use them, otherwise use old paths >-# these wouldn't work with the gnu make included with rh6.2 >-# KERNINC=$(if -d $(KERNDIR)/build, $(KERNDIR)/build/include, /usr/src/linux/include) >-# INSTALLDIR=$(if -d $(KERNDIR)/kernel, $(KERNDIR)/kernel/drivers/video, $(KERNDIR)/video) >- >-# this is slightly more brain-dead, but works.. >-ifeq ($(shell if test -d $(KERNDIR)/build; then echo yes; fi),yes) >-KERNINC=$(KERNDIR)/build/include >-else >-KERNINC=/usr/src/linux/include >-endif >- >-ifeq ($(shell if test -d $(KERNDIR)/kernel; then echo yes; fi),yes) >-INSTALLDIR=$(KERNDIR)/kernel/drivers/video >-else >-INSTALLDIR=$(KERNDIR)/video >-endif >- >-ifeq ($(shell echo $(NVDEBUG)),1) >- ifeq ($(shell test -z $(RMDEBUG) && echo yes),yes) >- RMDEBUG=1 >- endif >-endif >- >-ifeq ($(shell echo $(RMDEBUG)),1) >-CFLAGS += -DDEBUG -g -fno-common >-endif >- >-# this is just plain wrong, get rid of it >-BROKENDIR=$(KERNDIR)/kernel/video >- >-INSTALL=$(shell which install) >- >-# determine gcc versions used (kernel, nvidia.o) >-module_cc:=$(shell $(CC) -v 2>&1 | tail -1) >-version="^Linux version [^(]* (.*@.*) (\(.*\)) .*" >-kernel_cc:=$(shell cat /proc/version | sed "s/"$(version)"/\1/") >- >-module_cc:=$(shell echo "$(module_cc)" | cut -d ' ' -f 3) >-kernel_cc:=$(shell echo "$(kernel_cc)" | cut -d ' ' -f 3) >- >-# allow specification of alternate include file tree on command line and extra defines >-ifdef SYSINCLUDE >-INCLUDES += -I$(SYSINCLUDE) >-else >-INCLUDES += -I$(KERNINC) >-endif >- >-ifeq ($(shell sh conftest.sh remap_page_range $(INCLUDES)), 5) >- DEFINES += -DREMAP_PAGE_RANGE_5 >-endif >- >-ifeq ($(shell sh conftest.sh remap_page_range $(INCLUDES)), 4) >- DEFINES += -DREMAP_PAGE_RANGE_4 >-endif >- >-DEFINES+=$(EXTRA_DEFINES) >- >-# allow build parameters to be passed in through the environment >-ifdef BUILD_PARAMS >- DEFINES+=-D$(BUILD_PARAMS) >-endif >- >-VERSION_HDR=nv_compiler.h >- >-all: install >- >-install: package-install >- >-package-install: nvidia.o rmmod-check >- @if [ `id -ur` != 0 ]; then \ >- echo Please run \"make install\" as root.; \ >- else \ >- if [ -d $(BROKENDIR) ]; then \ >- rm -f $(BROKENDIR)/NVdriver; \ >- rmdir --ignore-fail-on-non-empty $(BROKENDIR); \ >- fi && \ >- mkdir -p $(INSTALLDIR) && \ >- rm -f $(INSTALLDIR)/NVdriver && \ >- $(INSTALL) -m 0664 -o root -g root nvidia.o $(INSTALLDIR)/nvidia.o$(O) && \ >- /sbin/depmod -a && \ >- /sbin/modprobe nvidia && \ >- sh makedevices.sh && \ >- echo "nvidia.o installed successfully."; \ >- fi >- >-RMMOD_ERROR=\ >- echo ""; \ >- echo "Unable to remove existing NVIDIA kernel module."; \ >- echo "Please be sure you have exited X before attempting"; \ >- echo "to install the NVIDIA kernel module."; \ >- echo ""; \ >- echo -en "\033[1;31m"; \ >- echo -e "*** Failed rmmod sanity check. Bailing out! ***"; \ >- echo -en "\033[0m"; \ >- exit 1; >- >-rmmod-check: >- @if /sbin/lsmod | grep nvidia > /dev/null; then \ >- if ! /sbin/rmmod nvidia > /dev/null; then $(RMMOD_ERROR) fi \ >- fi; \ >- if /sbin/lsmod | grep NVdriver > /dev/null; then \ >- if ! /sbin/rmmod NVdriver > /dev/null; then $(RMMOD_ERROR) fi \ >- fi >- >-gcc-check: >- @if [ -z $(IGNORE_CC_MISMATCH) ]; then \ >- if [ $(kernel_cc) != $(module_cc) ]; then \ >- echo " "; \ >- echo "You appear to be compiling the NVIDIA kernel module with "; \ >- echo "a compiler different from the one that was used to compile "; \ >- echo "the running kernel. This may be perfectly fine, but there "; \ >- echo "are cases where this can lead to unexpected behaviour and "; \ >- echo "system crashes. "; \ >- echo " "; \ >- echo "If you know what you are doing and want to override this "; \ >- echo "check, you can do so by setting IGNORE_CC_MISMATCH. "; \ >- echo " "; \ >- echo "In any other case, set the CC environment variable to the "; \ >- echo "name of the compiler that was used to compile the kernel. "; \ >- echo " "; \ >- echo -en "\033[1;31m"; \ >- echo -e "*** Failed cc sanity check. Bailing out! ***"; \ >- echo -en "\033[0m"; \ >- exit 1; \ >- fi \ >- fi >- >-nvidia.o: gcc-check $(LINUX_MODULE) $(RESMAN_KERNEL_MODULE) >- ld -r -o $@ $(LINUX_MODULE) $(RESMAN_KERNEL_MODULE) >- >-$(VERSION_HDR): >- echo \#define NV_COMPILER \"`$(CC) -v 2>&1 | tail -1`\" > $@ >- >-$(LINUX_MODULE): $(OBJECTS) >- ld -r -o $@ $(OBJECTS) >- >-%.o: %.c $(HEADERS) >- $(CC) -c $(CFLAGS) $< >- >-# debug tool to preprocess the file and leave .i to make it easier to untangle #defines >-%.i: %.c >- $(CC) $(CFLAGS) -E $< > $@ >- >-%.s: %.c >- $(CC) $(CFLAGS) -S $< > $@ >- >- >-clean: >- $(RM) $(OBJECTS) $(LINUX_MODULE) $(VERSION_HDR) *.d NVdriver nvidia.o >- >- >--include $(OBJECTS:%.o=%.d) >diff -ruN NVIDIA_kernel-1.0-4363/Makefile.kbuild NVIDIA_kernel-1.0-4363-2.6/Makefile.kbuild >--- NVIDIA_kernel-1.0-4363/Makefile.kbuild 1970-01-01 01:00:00.000000000 +0100 >+++ NVIDIA_kernel-1.0-4363-2.6/Makefile.kbuild 2004-01-15 13:42:25.000000000 +0100 >@@ -0,0 +1,249 @@ >+# >+# KBUILD Makefile for the NVIDIA Linux kernel module. >+# >+# The motivation for replacing the original Makefile is the hope that this >+# version will simplify the build and installation process. In the past, >+# many architectural and cosmetic changes to the Linux kernel have made it >+# difficult to maintain compatibility or required duplication of logic. >+# >+# Linux 2.6 introduces numerous such changes, many of which impact modules. >+# Relying on KBUILD, some aspects of the build system otherwise difficult >+# to support (for example, module versioning) are abstracted away and dealt >+# with elsewhere, making life significantly easier here. >+# >+# The new approach currently has its own share of problems, some of which >+# are architectural difficulties with KBUILD, others minor annoyances. For >+# this reason, an improved version of the NVIDIA Makefile is available to >+# those experiencing difficulties. >+# >+# Please report any problems you may be experiencing with this experimental >+# Makefile to either one (or, preferably, both) of us: >+# >+# Alistair J Strachan (alistair@devzero.co.uk) (first pass, enhancements) >+# Christian Zander (zander@mail.minion.de) (enhancements) >+# >+ >+all: install >+install: package-install >+ >+# >+# The NVIDIA kernel module base name and static file names. KBUILD will go >+# ahead and append ".o" or ".ko" to form the final module name. >+# >+ >+MODULE_NAME := nvidia >+VERSION_HEADER := nv_compiler.h >+ >+# >+# List of object files to link into NVIDIA kernel module; make sure KBUILD >+# understands that we want a module. >+# >+ >+RESMAN_CORE_OBJS := nv-kernel.o >+RESMAN_GLUE_OBJS := nv.o os-agp.o os-interface.o os-registry.o >+ >+$(MODULE_NAME)-objs := $(RESMAN_CORE_OBJS) $(RESMAN_GLUE_OBJS) >+ >+# >+# The precompiled kernel module build process requires a separation of the >+# closed source and open source object files. >+# >+ >+KERNEL_GLUE_NAME := nv-linux.o >+KERNEL_GLUE_OBJS := $(RESMAN_GLUE_OBJS) $(MODULE_NAME).mod.o >+ >+# >+# A bug in KBUILD 2.4 means that leaving obj-m set in top-level context >+# will cause Rules.make to call pathdown.sh, which is wrong. So, we only >+# set this conditional of a kernel-level instance. >+# >+ >+ifdef TOPDIR >+obj-m := $(MODULE_NAME).o >+endif >+ >+# >+# Include local source directory in $(CC)'s include path and set disable any >+# warning types that are of little interest to us. >+# >+ >+EXTRA_CFLAGS += -I$(src) >+EXTRA_CFLAGS += -Wno-cast-qual -Wno-strict-prototypes >+ >+# >+# We rely on these two definitions below; if they aren't set, we set them to >+# reasonable defaults (Linux 2.4's KBUILD, and top-level passes will not set >+# these). >+# >+ >+src ?= . >+obj ?= . >+ >+# >+# Determine location of the Linux kernel source tree. Allow users to override >+# the default (i.e. automatically determined) kernel source location with the >+# SYSSRC directive; this new directive replaces NVIDIA's SYSINCLUDE. >+# >+ >+KERNEL_UNAME := $(shell SYSSRC=$(SYSSRC) sh $(src)/conftest.sh get_uname) >+MODULE_ROOT := /lib/modules/$(KERNEL_UNAME)/kernel/drivers >+ >+ifdef SYSSRC >+ KERNEL_SOURCES := $(SYSSRC) >+ KERNEL_HEADERS := -I$(KERNEL_SOURCES)/include >+else >+ KERNEL_SOURCES := /lib/modules/$(KERNEL_UNAME)/build >+ KERNEL_HEADERS := -I$(KERNEL_SOURCES)/include >+endif >+ >+# >+# Sets any internal variables left unset by KBUILD (e.g. this happens during >+# a top-level run). >+# >+ >+TOPDIR ?= $(KERNEL_SOURCES) >+PATCHLEVEL ?= $(shell sh $(src)/conftest.sh kernel_patch_level $(TOPDIR)) >+ >+# >+# Linux 2.4 uses the .o module extension. Linux 2.6, however, uses the .ko >+# module extension. Handle these gracefully. >+# >+ >+ifeq ($(PATCHLEVEL), 4) >+ MODULE_OBJECT := $(MODULE_NAME).o >+else >+ MODULE_OBJECT := $(MODULE_NAME).ko >+endif >+ >+# >+# NVIDIA specific CFLAGS and #define's. The remap_page_range check has become >+# necessary with the introduction of the five argument version to Linux 2.4 >+# distribution kernels; this conflicting change cannot be detected at compile >+# time. >+# >+ >+EXTRA_CFLAGS += -D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DNTRM -D_GNU_SOURCE -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=4363 -DNV_UNIX -DNV_LINUX -DNV_INT64_OK -DNVCPU_X86 >+ >+ifeq ($(shell echo $(NVDEBUG)),1) >+ ifeq ($(shell test -z $(RMDEBUG) && echo yes),yes) >+ RMDEBUG=1 >+ endif >+endif >+ >+ifeq ($(shell echo $(RMDEBUG)),1) >+ EXTRA_CFLAGS += -DDEBUG -g -fno-common >+endif >+ >+ifeq ($(shell sh $(src)/conftest.sh remap_page_range $(KERNEL_HEADERS)), 5) >+ EXTRA_CFLAGS += -DREMAP_PAGE_RANGE_5 >+endif >+ >+ifeq ($(shell sh $(src)/conftest.sh remap_page_range $(KERNEL_HEADERS)), 4) >+ EXTRA_CFLAGS += -DREMAP_PAGE_RANGE_4 >+endif >+ >+# >+# NVIDIA binary object file includes .common section. >+# >+ >+EXTRA_LDFLAGS := -d >+ >+# >+# Miscellaneous NVIDIA kernel module build support targets. They are needed >+# to satisfy KBUILD requirements and to support NVIDIA specifics. >+# >+ >+$(obj)/$(RESMAN_CORE_OBJS): >+ cp $(src)/$(RESMAN_CORE_OBJS) $(obj)/$(RESMAN_CORE_OBJS) >+ >+$(obj)/$(VERSION_HEADER): >+ echo \#define NV_COMPILER \"`$(CC) -v 2>&1 | tail -n 1`\" > $@ >+ >+$(obj)/nv.o: $(obj)/$(VERSION_HEADER) >+ >+# >+# More quirks for Linux 2.4 KBUILD, which doesn't link automatically. >+# >+ >+ifeq ($(PATCHLEVEL), 4) >+$(obj)/$(MODULE_NAME).o: $($(MODULE_NAME)-objs) >+ $(LD) $(EXTRA_LDFLAGS) -r -o $@ $($(MODULE_NAME)-objs) >+endif >+ >+# >+# KBUILD build parameters. >+# >+ >+KBUILD_PARAMS := -C $(KERNEL_SOURCES) SUBDIRS=$(PWD) >+ >+# >+# NVIDIA sanity checks. >+# >+ >+suser-sanity-check: >+ @if ! sh conftest.sh suser_sanity_check nvidia; then \ >+ exit 1; \ >+ fi >+ >+rmmod-sanity-check: >+ @if ! sh conftest.sh rmmod_sanity_check nvidia; then \ >+ exit 1; \ >+ fi >+ >+cc-sanity-check: >+ @if ! sh conftest.sh cc_sanity_check full_output $(CC); then \ >+ exit 1; \ >+ fi >+ >+# >+# Build the NVIDIA kernel module using Linux KBUILD. This target is used by >+# the "package-install" target below. >+# >+ >+module: cc-sanity-check >+ @make CC=$(CC) $(KBUILD_PARAMS) modules >+ >+# >+# Build the NVIDIA kernel module with KBUILD. Verify that the user posesses >+# sufficient privileges. Rebuild the module dependency file. >+# >+ >+module-install: suser-sanity-check module >+ @mkdir -p $(MODULE_ROOT)/video; \ >+ install -m 0664 -o root -g root $(MODULE_OBJECT) $(MODULE_ROOT)/video; \ >+ PATH="$(PATH):/bin:/sbin" /sbin/depmod -ae; >+ >+# >+# This target builds, then installs, then creates device nodes and inserts >+# the module, if successful. >+# >+ >+package-install: module-install rmmod-sanity-check >+ @sh makedevices.sh; \ >+ PATH="$(PATH):/bin:/sbin" /sbin/modprobe $(MODULE_NAME) >+ >+# >+# Build an object file suitable for further processing by the installer and >+# inclusion as a precompiled kernel interface file. >+# >+ >+$(KERNEL_GLUE_NAME): module >+ @$(LD) $(EXTRA_LDFLAGS) -r -o $(KERNEL_GLUE_NAME) $(KERNEL_GLUE_OBJS) >+ >+# >+# Support hack, KBUILD isn't prepared to clean up after external modules. >+# >+ >+clean: >+ @$(RM) -f $(RESMAN_GLUE_OBJS) $(KERNEL_GLUE_OBJS); \ >+ $(RM) -f built-in.o nv-linux.o *.d .*.{cmd,flags}; \ >+ $(RM) -f $(MODULE_NAME).{o,ko,mod.{o,c}} $(VERSION_HEADER) *~ >+ >+# >+# Linux 2.4 KBUILD requires the inclusion of Rules.make; Linux 2.6's KBUILD >+# includes dependencies automatically. >+# >+ >+ifeq ($(PATCHLEVEL), 4) >+include $(KERNEL_SOURCES)/Rules.make >+endif >diff -ruN NVIDIA_kernel-1.0-4363/Makefile.nvidia NVIDIA_kernel-1.0-4363-2.6/Makefile.nvidia >--- NVIDIA_kernel-1.0-4363/Makefile.nvidia 1970-01-01 01:00:00.000000000 +0100 >+++ NVIDIA_kernel-1.0-4363-2.6/Makefile.nvidia 2003-07-31 19:54:17.000000000 +0200 >@@ -0,0 +1,154 @@ >+# This Makefile is automatically generated; do not edit >+# Generated on 'builder3.nvidia.com' on Mon Mar 24 16:36:51 PST 2003 >+ >+LINUX_MODULE=nv-linux.o >+DEFINES=-D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DKBUILD_MODNAME="nvidia" -DNTRM -D_GNU_SOURCE -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=4348 -DNV_UNIX -DNV_LINUX -DNV_INT64_OK -DNVCPU_X86 >+INCLUDES=-I. >+ >+OBJECTS=nv.o os-agp.o os-interface.o os-registry.o >+HEADERS=os-agp.h os-interface.h nv-linux.h nv-misc.h nv.h rmretval.h nvtypes.h $(VERSION_HDR) >+ >+CFLAGS=-Wall -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-subscripts -Wparentheses -Wcast-qual -Wno-multichar -O -MD $(DEFINES) $(INCLUDES) -Wno-cast-qual >+ >+RESMAN_KERNEL_MODULE=nv-kernel.o >+ >+KERNDIR=/lib/modules/$(shell uname -r) >+ >+# check for newer paths. if found, use them, otherwise use old paths >+# these wouldn't work with the gnu make included with rh6.2 >+# KERNINC=$(if -d $(KERNDIR)/build, $(KERNDIR)/build/include, /usr/src/linux/include) >+# INSTALLDIR=$(if -d $(KERNDIR)/kernel, $(KERNDIR)/kernel/drivers/video, $(KERNDIR)/video) >+ >+# this is slightly more brain-dead, but works.. >+ifeq ($(shell if test -d $(KERNDIR)/build; then echo yes; fi),yes) >+KERNINC=$(KERNDIR)/build/include >+else >+KERNINC=/usr/src/linux/include >+endif >+ >+ifeq ($(shell if test -d $(KERNDIR)/kernel; then echo yes; fi),yes) >+INSTALLDIR=$(KERNDIR)/kernel/drivers/video >+else >+INSTALLDIR=$(KERNDIR)/video >+endif >+ >+ifeq ($(shell echo $(NVDEBUG)),1) >+ ifeq ($(shell test -z $(RMDEBUG) && echo yes),yes) >+ RMDEBUG=1 >+ endif >+endif >+ >+ifeq ($(shell echo $(RMDEBUG)),1) >+CFLAGS += -DDEBUG -g -fno-common >+endif >+ >+# this is just plain wrong, get rid of it >+BROKENDIR=$(KERNDIR)/kernel/video >+ >+INSTALL=$(shell which install) >+ >+# allow specification of alternate include file tree on command line and extra defines >+ifdef SYSINCLUDE >+INCLUDES += -I$(SYSINCLUDE) >+INCLUDES += -I$(SYSINCLUDE)/asm/mach-default >+else >+INCLUDES += -I$(KERNINC) >+INCLUDES += -I$(KERNINC)/asm/mach-default >+endif >+ >+ifeq ($(shell sh conftest.sh remap_page_range $(INCLUDES)), 5) >+ DEFINES += -DREMAP_PAGE_RANGE_5 >+endif >+ >+ifeq ($(shell sh conftest.sh remap_page_range $(INCLUDES)), 4) >+ DEFINES += -DREMAP_PAGE_RANGE_4 >+endif >+ >+DEFINES+=$(EXTRA_DEFINES) >+ >+# allow build parameters to be passed in through the environment >+ifdef BUILD_PARAMS >+ DEFINES+=-D$(BUILD_PARAMS) >+endif >+ >+VERSION_HDR=nv_compiler.h >+ >+all: install >+ >+install: package-install >+ >+suser-sanity-check: >+ @if ! sh conftest.sh suser_sanity_check; then \ >+ echo; \ >+ echo "You have insufficient privileges for this operation. Please "; \ >+ echo "run \"make install\" as root! "; \ >+ echo; \ >+ exit 1; \ >+ fi >+ >+rmmod-sanity-check: >+ @if ! sh conftest.sh rmmod_sanity_check nvidia; then \ >+ echo; \ >+ echo "Unable to unload the currently loaded NVIDIA kernel module! "; \ >+ echo "Please be certain that you have exited X before attempting "; \ >+ echo "to install this version. "; \ >+ echo; \ >+ exit 1; \ >+ fi >+ >+cc-sanity-check: >+ @if ! sh conftest.sh cc_sanity_check $(CC); then \ >+ echo; \ >+ echo "You appear to be building the NVIDIA kernel module with a "; \ >+ echo "compiler different from the one that was used to build the "; \ >+ echo "running kernel. This may be perfectly fine, but there are "; \ >+ echo "cases where this can lead to unexpected behaviour and "; \ >+ echo "system crashes. "; \ >+ echo; \ >+ echo "If you know what you are doing and want to override this "; \ >+ echo "check, you can do so by setting IGNORE_CC_MISMATCH. "; \ >+ echo; \ >+ echo "In any other case, set the CC environment variable to the "; \ >+ echo "name of the compiler that was used to build the kernel. "; \ >+ echo; \ >+ exit 1; \ >+ fi >+ >+package-install: suser-sanity-check nvidia.o rmmod-sanity-check >+ if [ -d $(BROKENDIR) ]; then \ >+ rm -f $(BROKENDIR)/NVdriver; \ >+ rmdir --ignore-fail-on-non-empty $(BROKENDIR); \ >+ fi && \ >+ mkdir -p $(INSTALLDIR) && \ >+ rm -f $(INSTALLDIR)/NVdriver && \ >+ $(INSTALL) -m 0664 -o root -g root nvidia.o $(INSTALLDIR)/nvidia.o$(O) && \ >+ /sbin/depmod -a && \ >+ /sbin/modprobe nvidia && \ >+ sh makedevices.sh && \ >+ echo "nvidia.o installed successfully."; \ >+ >+nvidia.o: cc-sanity-check $(LINUX_MODULE) $(RESMAN_KERNEL_MODULE) >+ ld -d -r -o $@ $(LINUX_MODULE) $(RESMAN_KERNEL_MODULE) >+ >+$(VERSION_HDR): >+ echo \#define NV_COMPILER \"`$(CC) -v 2>&1 | tail -n 1`\" > $@ >+ >+$(LINUX_MODULE): $(OBJECTS) >+ ld -r -o $@ $(OBJECTS) >+ >+%.o: %.c $(HEADERS) >+ $(CC) -c $(CFLAGS) $< >+ >+# debug tool to preprocess the file and leave .i to make it easier to untangle #defines >+%.i: %.c >+ $(CC) $(CFLAGS) -E $< > $@ >+ >+%.s: %.c >+ $(CC) $(CFLAGS) -S $< > $@ >+ >+ >+clean: >+ $(RM) $(OBJECTS) $(LINUX_MODULE) $(VERSION_HDR) *.d NVdriver nvidia.o >+ >+ >+-include $(OBJECTS:%.o=%.d) >diff -ruN NVIDIA_kernel-1.0-4363/conftest.sh NVIDIA_kernel-1.0-4363-2.6/conftest.sh >--- NVIDIA_kernel-1.0-4363/conftest.sh 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/conftest.sh 2004-01-15 13:16:10.000000000 +0100 >@@ -1,5 +1,11 @@ > #!/bin/bash > >+# make sure we are in the directory containing this script >+SCRIPTDIR=`dirname $0` >+cd $SCRIPTDIR >+ >+PATH=$PATH:/bin:/sbin >+ > case "$1" in > remap_page_range) > shift >@@ -37,5 +43,166 @@ > exit 1 > fi > ;; >+ >+ cc_sanity_check) >+ shift >+ # >+ # Verify that the same compiler is used for the kernel and kernel >+ # module. >+ # >+ if test -n "$IGNORE_CC_MISMATCH" -o -n "$SYSSRC" -o -n "$SYSINCLUDE"; then >+ # >+ # The user chose to disable this test; this may or may not be >+ # a wise decision... >+ # >+ exit 0 >+ fi >+ >+ # usage: conftest.sh [cc_sanity_check] [full_output|just_msg] [$(CC)] >+ >+ OUTPUT=$1 >+ shift >+ >+ rm -f gcc-version-check >+ $* gcc-version-check.c -o gcc-version-check > /dev/null 2>&1 >+ if [ -f gcc-version-check ]; then >+ PROC_VERSION=`cat /proc/version` >+ MSG=`./gcc-version-check "$PROC_VERSION"` >+ RET=$? >+ rm -f gcc-version-check >+ else >+ MSG="Could not compile gcc-version-check.c" >+ RET=1 >+ fi >+ >+ if [ "$RET" != "0" ]; then >+ # >+ # The gcc version check failed >+ # >+ >+ if [ "$OUTPUT" = "full_output" ]; then >+ echo ""; >+ echo "gcc-version-check failed:"; >+ echo ""; >+ echo "$MSG" | fmt -w 60 >+ echo ""; >+ echo "If you know what you are doing and want to override"; >+ echo "the gcc version check, you can do so by setting the"; >+ echo "IGNORE_CC_MISMATCH environment variable to \"1\"."; >+ echo ""; >+ echo "In any other case, set the CC environment variable"; >+ echo "to the name of the compiler that was used to compile"; >+ echo "the kernel."; >+ echo "" >+ echo -en "\033[1;31m"; >+ echo -e "*** Failed cc sanity check. Bailing out! ***"; >+ echo -en "\033[0m"; >+ echo ""; >+ else >+ echo "$MSG" >+ fi >+ exit 1; >+ else >+ exit 0 >+ fi >+ ;; >+ >+ kernel_patch_level) >+ shift >+ # >+ # Determine the kernel's major patch level; this is only done if we >+ # aren't told by KBUILD. >+ # >+ echo $(cat "$1"/Makefile | grep "PATCHLEVEL =" | cut -d " " -f 3) >+ exit 0 >+ ;; >+ >+ suser_sanity_check) >+ shift >+ # >+ # Determine the caller's user id to determine if we have sufficient >+ # privileges for the requested operation. >+ # >+ if test $(id -ur) != 0; then >+ echo ""; >+ echo "Please run \"make install\" as root."; >+ echo ""; >+ echo -en "\033[1;31m"; >+ echo -e "*** Failed super-user sanity check. Bailing out! ***"; >+ echo -en "\033[0m"; >+ exit 1 >+ else >+ exit 0 >+ fi >+ ;; >+ >+ rmmod_sanity_check) >+ shift >+ # >+ # Make sure that any currently loaded NVIDIA kernel module can be >+ # unloaded. >+ # >+ if test -n "$SYSSRC" -o -n "$SYSINCLUDE"; then >+ # >+ # Don't attempt to remove the kernel module if we're not >+ # building against the running kernel. >+ # >+ exit 0 >+ fi >+ >+ if lsmod | grep -q "$1"; then >+ rmmod "$1" >& /dev/null >+ fi >+ >+ if lsmod | grep -q "$1"; then >+ # >+ # The NVIDIA kernel module is still loaded, most likely because >+ # it is busy. >+ # >+ echo ""; >+ echo "Unable to remove existing NVIDIA kernel module."; >+ echo "Please be sure you have exited X before attempting"; >+ echo "to install the NVIDIA kernel module."; >+ echo ""; >+ echo -en "\033[1;31m"; >+ echo -e "*** Failed rmmod sanity check. Bailing out! ***"; >+ echo -en "\033[0m"; >+ exit 1 >+ else >+ exit 0 >+ fi >+ ;; >+ >+ get_uname) >+ # >+ # print UTS_RELEASE from the kernel sources that we are building >+ # against; if an error occurs, give up and just return `uname -r` >+ # >+ shift >+ >+ if [ "$SYSSRC" ]; then >+ KERNEL_INCLUDE=$SYSSRC/include >+ elif [ "$SYSINCLUDE" ]; then >+ KERNEL_INCLUDE=$SYSINCLUDE >+ else >+ KERNEL_INCLUDE="/lib/modules/`uname -r`/build/include" >+ fi >+ >+ echo "#include \"linux/version.h\" >+ int main() { printf(\"%s\n\", UTS_RELEASE); return 0; }" \ >+ > conftest.c >+ >+ gcc conftest.c -o conftest -nostdinc \ >+ -I $KERNEL_INCLUDE > /dev/null 2>&1 >+ >+ if [ -f conftest ]; then >+ ./conftest >+ rm -f conftest conftest.c >+ exit 0 >+ else >+ uname -r >+ exit 1 >+ fi >+ ;; > esac > >diff -ruN NVIDIA_kernel-1.0-4363/gcc-version-check.c NVIDIA_kernel-1.0-4363-2.6/gcc-version-check.c >--- NVIDIA_kernel-1.0-4363/gcc-version-check.c 1970-01-01 01:00:00.000000000 +0100 >+++ NVIDIA_kernel-1.0-4363-2.6/gcc-version-check.c 2003-12-17 23:55:00.000000000 +0100 >@@ -0,0 +1,200 @@ >+#include <stdio.h> >+#include <string.h> >+#include <ctype.h> >+ >+/* >+ * Check that the compiler used to compile this source file matches >+ * the major and minor version of the compiler used to compile the >+ * kernel specified on the commandline. >+ * >+ * The /proc/version string for the kernel in question is specified on >+ * the commandline. >+ * >+ * Returns 0 if the major and minor versions match, and returns 1 if >+ * they do not match. >+ * >+ * The program always prints one line of output to stdout, describing >+ * why the return status was returned; this is so that the caller may >+ * use that output when explaining what happened. >+ * >+ * >+ * Some gcc version strings that have proven problematic for parsing >+ * in the past: >+ * >+ * gcc.real (GCC) 3.3 (Debian) >+ * gcc-Version 3.3 (Debian) >+ * gcc (GCC) 3.1.1 20020606 (Debian prerelease) >+ */ >+ >+ >+ >+#define STATE_BEFORE 0 >+#define STATE_IN_MAJOR 1 >+#define STATE_DOT 2 >+#define STATE_IN_MINOR 3 >+ >+#define ctoi(c) (c - '0') >+ >+ >+int main(int argc, char *argv[]) >+{ >+ char *str, *s, c; >+ int gcc_major = __GNUC__; >+ int gcc_minor = __GNUC_MINOR__; >+ int major = 0; >+ int minor = 0; >+ int state; >+ >+ /* >+ * must have exactly one argument: the kernel's proc version >+ * string to check >+ */ >+ >+ if (argc != 2) { >+ printf("No /proc/version string specified.\n"); >+ return 1; >+ } >+ >+ /* >+ * skip ahead in the /proc/version string to the first instance of >+ * "(gcc" >+ */ >+ >+ str = argv[1]; >+ s = strstr(str, "(gcc"); >+ if (!s) { >+ printf("The string \"(gcc\" was not found in the /proc/version " >+ "string: \"%s\"; please report this error to " >+ "linux-bugs@nvidia.com\n", str); >+ return 1; >+ } >+ >+ /* >+ * Parser state machine: once we have found "gcc" in the >+ * /proc/version string, scan for the next instances of "MMM.mmm"; >+ * in other words: we are looking for some number of digits (the >+ * gcc major version number), followed by a dot, followed by some >+ * number of digits (the gcc minor version number). >+ */ >+ >+ state = STATE_BEFORE; >+ >+ while(*s) { >+ c = *s; >+ switch (state) { >+ >+ /* >+ * If we are in STATE_BEFORE and find a digit, then we >+ * might have found the start of the major version number: >+ * transition to STATE_IN_MAJOR and start recording >+ * major. >+ * >+ * If we find a non-digit, then remain in STATE_BEFORE. >+ */ >+ >+ case STATE_BEFORE: >+ if (isdigit(c)) { >+ state = STATE_IN_MAJOR; >+ major = ctoi(c); >+ } else { >+ state = STATE_BEFORE; >+ major = minor = 0; >+ } >+ break; >+ >+ /* >+ * If we are in STATE_IN_MAJOR and find a digit, then stay >+ * in STATE_IN_MAJOR and continue to accumulate >+ * major. >+ * >+ * If we find a dot, then transition to STATE_DOT. >+ * >+ * If we find anything else, then this was not really the >+ * major version number: transition back to STATE_BEFORE >+ * and clear out the major/minor variables. >+ */ >+ >+ case STATE_IN_MAJOR: >+ if (isdigit(c)) { >+ state = STATE_IN_MAJOR; >+ major = (major * 10) + ctoi(c); >+ } else if (c == '.') { >+ state = STATE_DOT; >+ } else { >+ state = STATE_BEFORE; >+ major = minor = 0; >+ } >+ break; >+ >+ /* >+ * If we are in STATE_DOT and find a digit, then this is >+ * the start of the minor version number: transition to >+ * STATE_IN_MINOR, and start recording minor. >+ * >+ * If we find a non-digit, then this was not really the >+ * version: transition back to STATE_BEFORE and clear out >+ * the major/minor variables. >+ */ >+ >+ case STATE_DOT: >+ if (isdigit(c)) { >+ state = STATE_IN_MINOR; >+ minor = ctoi(c); >+ } else { >+ state = STATE_BEFORE; >+ major = minor = 0; >+ } >+ break; >+ >+ /* >+ * If we are in STATE_IN_MINOR and find a digit, then stay >+ * in STATE_IN_MINOR and continue to accumulate >+ * minor. >+ * >+ * If we find a non-digit, then we are done. >+ */ >+ >+ case STATE_IN_MINOR: >+ if (isdigit(c)) { >+ state = STATE_IN_MINOR; >+ minor = (minor * 10) + ctoi(c); >+ } else { >+ /* we are now done. */ >+ goto done; >+ } >+ break; >+ } >+ s++; >+ } >+ >+ done: >+ >+ if (state == STATE_IN_MINOR) { >+ if ((major == gcc_major) && (minor == gcc_minor)) { >+ printf("The compiler used to compile the kernel matches the " >+ "current compiler (gcc %d.%d)\n", major, minor); >+ return 0; >+ } else if ((major == gcc_major) && (gcc_major > 2)) { >+ printf("The compiler used to compile the kernel (gcc %d.%d) does " >+ "not exactly match the current compiler (gcc %d.%d), but " >+ "it is close enough.\n", >+ major, minor, gcc_major, gcc_minor); >+ return 0; >+ } else { >+ >+ printf("You appear to be compiling the NVIDIA kernel module with " >+ "a different compiler than the one that was used to " >+ "compile the running kernel. This may be fine, " >+ "but there are cases where this can lead to instability. " >+ "The compiler used to " >+ "compile the kernel was gcc %d.%d; the current " >+ "compiler is gcc %d.%d.\n", >+ major, minor, gcc_major, gcc_minor); >+ return 1; >+ } >+ } >+ >+ printf("No gcc version found in /proc/version string: \"%s\"; please " >+ "report this error to linux-bugs@nvidia.com\n", str); >+ return 1; >+} >diff -ruN NVIDIA_kernel-1.0-4363/makedevices.sh NVIDIA_kernel-1.0-4363-2.6/makedevices.sh >--- NVIDIA_kernel-1.0-4363/makedevices.sh 2003-04-20 03:57:20.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/makedevices.sh 2003-11-23 20:37:25.000000000 +0100 >@@ -1,56 +1,61 @@ >-#!/bin/bash >+#!/bin/sh > >-function error { >+error() { > echo "$(basename $0): fatal error, $1 failed" > exit 1 > } > >-modconf="" >-mktemp=`which mktemp || echo touch` >-tmp=`$mktemp -q /tmp/nvidia.XXXXXX` || error "mktemp" >- >-paths=" /etc/modutils/aliases \ >- /etc/modules.conf \ >- /etc/conf.modules " >+# create a temporary file >+tmp=`(mktemp -q /tmp/nvidia.XXXXXX) 2> /dev/null` || tmp="/tmp/nvidia.$$" > >-for path in $paths; do >- test -f $path && modconf=$path >+modconfs=" /etc/modutils/aliases \ >+ /etc/modules.conf \ >+ /etc/modprobe.d/aliases \ >+ /etc/modprobe.conf \ >+ /etc/conf.modules " >+ >+# >+# Pass the -p option to cp to preserve the permissions of $modconf and >+# /etc/rc.d/rc.modules across the update. Update all relevant files >+# including those eventually regenerated to avoid the need for awkward >+# logic here. >+# >+for modconf in $modconfs; do >+ if [ -f "$modconf" ]; then >+ cp -p $modconf $tmp >+ sed '/^alias.*\(NVdriver\|nvidia\)/d' < $modconf > $tmp >+ if [ -c /dev/.devfsd ]; then >+ echo "alias /dev/nvidia* nvidia" >> $tmp >+ else >+ if [ `uname -r | cut -d '.' -f 2` = 6 ]; then >+ echo "alias char-major-195* nvidia" >> $tmp >+ else >+ echo "alias char-major-195 nvidia" >> $tmp >+ fi >+ fi >+ mv -f $tmp $modconf >+ test -x /sbin/update-modules && /sbin/update-modules >+ fi > done > >-if [ ! -z $modconf ]; then >- # Initialize the permissions on $tmp so that we don't change the >- # permissions of $modconf when we mv it into place >- cp -p $modconf $tmp >- sed '/^alias.*\(NVdriver\|nvidia\)/d' < $modconf > $tmp >- if [ -c /dev/.devfsd ]; then >- echo "alias /dev/nvidia* nvidia" >> $tmp >- else >- echo "alias char-major-195 nvidia" >> $tmp >- fi >- mv -f $tmp $modconf >- test -x /sbin/update-modules && /sbin/update-modules >-else >- if [ -f /etc/rc.d/rc.modules ]; then >- # Initialize the permissions on $tmp so that we don't change the >- # permissions of /etc/rc.d/rc.modules when we mv it into place >- cp -p /etc/rc.d/rc.modules $tmp >- sed '/.*\(NVdriver\|nvidia\).*/d' < /etc/rc.d/rc.modules > $tmp >- echo -e "\n/sbin/modprobe nvidia" >> $tmp >- mv -f $tmp /etc/rc.d/rc.modules >- fi >+if [ -f /etc/rc.d/rc.modules ]; then >+ cp -p /etc/rc.d/rc.modules $tmp >+ sed '/.*\(NVdriver\|nvidia\).*/d' < /etc/rc.d/rc.modules > $tmp >+ echo -e "\n/sbin/modprobe nvidia" >> $tmp >+ mv -f $tmp /etc/rc.d/rc.modules > fi > > if [ ! -c /dev/.devfsd ]; then > for i in 0 1 2 3 4 5 6 7; do >- node="/dev/nvidia$i" >- rm -f $node >- mknod $node c 195 $i || error "mknod \"$node\"" >- chmod 0666 $node || error "chmod \"$node\"" >+ node="/dev/nvidia$i" >+ rm -f $node >+ mknod $node c 195 $i || error "mknod \"$node\"" >+ chmod 0666 $node || error "chmod \"$node\"" > done > node="/dev/nvidiactl" > rm -f $node >- mknod $node c 195 255 || error "mknod \"$node\"" >- chmod 0666 $node || error "chmod \"$node\"" >+ mknod $node c 195 255 || error "mknod \"$node\"" >+ chmod 0666 $node || error "chmod \"$node\"" > fi > > exit 0 >diff -ruN NVIDIA_kernel-1.0-4363/nv-linux.h NVIDIA_kernel-1.0-4363-2.6/nv-linux.h >--- NVIDIA_kernel-1.0-4363/nv-linux.h 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/nv-linux.h 2004-01-13 14:40:01.000000000 +0100 >@@ -20,35 +20,20 @@ > # define MODVERSIONS > #endif > >-#if defined (MODVERSIONS) >-#include <linux/modversions.h> >-#endif >- >-#include <linux/kernel.h> >-#include <linux/module.h> > #include <linux/version.h> > >-#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 2, 12) >-# error This driver does not support 2.2.11 or earlier kernels! >-#elif LINUX_VERSION_CODE < KERNEL_VERSION(2, 3, 0) >-# define KERNEL_2_2 >-# warning NVIDIA is considering dropping support for linux-2.2 >-# warning kernels. While end users are free to maintain their >-# warning own patches, or stick with current drivers, our new >-# warning drivers will not work "out of the box." If you are >-# warning concerned about lack of support for 2.2 kernels, >-# warning please let us know at linux-bugs@nvidia.com; we would >-# warning like to know how many users would be seriously >-# warning impacted by this decision. >+#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 3, 0) >+# error This driver does not support 2.2.x kernels! > #elif LINUX_VERSION_CODE < KERNEL_VERSION(2, 4, 0) >-# error This driver does not support 2.3.x development kernels! >+# error This driver does not support 2.3.x kernels! > #elif LINUX_VERSION_CODE < KERNEL_VERSION(2, 5, 0) > # define KERNEL_2_4 > #elif LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 0) >-# error This driver does not support 2.5.x development kernels! >-# define KERNEL_2_5 >+# error This driver does not support 2.5.x kernels! >+#elif LINUX_VERSION_CODE < KERNEL_VERSION(2, 7, 0) >+# define KERNEL_2_6 > #else >-# error This driver does not support 2.6.x or newer kernels! >+# error This driver does not support development kernels! > #endif > > #if defined (__ia64) >@@ -61,12 +46,19 @@ > #define __SMP__ > #endif > >+#if defined (MODVERSIONS) && !defined (KERNEL_2_6) >+#include <linux/modversions.h> >+#endif >+ >+#include <linux/kernel.h> >+#include <linux/module.h> >+ >+#include <linux/init.h> /* module_init, module_exit */ > #include <linux/types.h> /* pic_t, size_t, __u32, etc */ > #include <linux/errno.h> /* error codes */ > #include <linux/list.h> /* circular linked list */ > #include <linux/stddef.h> /* NULL, offsetof */ > #include <linux/wait.h> /* wait queues */ >-#include <linux/tqueue.h> /* struct tq_struct */ > > #include <linux/slab.h> /* kmalloc, kfree, etc */ > #include <linux/vmalloc.h> /* vmalloc, vfree, etc */ >@@ -74,9 +66,15 @@ > #include <linux/poll.h> /* poll_wait */ > #include <linux/delay.h> /* mdelay, udelay */ > >+#ifdef KERNEL_2_6 >+#include <linux/sched.h> /* suser(), capable() replacement */ >+#include <linux/moduleparam.h> /* module_param() */ >+#include <linux/smp_lock.h> /* kernel_locked */ >+#include <asm/kmap_types.h> /* page table entry lookup */ >+#endif >+ > #include <linux/pci.h> /* pci_find_class, etc */ >-#include <linux/wrapper.h> /* mem_map_reserve */ >-#include <linux/interrupt.h> /* mark_bh, init_bh, remove_bh */ >+#include <linux/interrupt.h> /* tasklets, interrupt helpers */ > #include <linux/timer.h> > > #include <asm/system.h> /* cli, sli, save_flags */ >@@ -85,14 +83,9 @@ > #include <asm/page.h> /* PAGE_OFFSET */ > #include <asm/pgtable.h> /* pte bit definitions */ > >-#if !defined (KERNEL_2_2) > #include <linux/spinlock.h> > #include <asm/semaphore.h> > #include <linux/highmem.h> >-#else >-#include <asm/spinlock.h> >-#include <asm/semaphore.h> >-#endif > > #ifdef CONFIG_PROC_FS > #include <linux/proc_fs.h> >@@ -132,70 +125,172 @@ > for (pos = (head)->next; pos != (head); pos = (pos)->next) > #endif > >-#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 2, 18) >-#define vmalloc_32 vmalloc >+#define GET_MODULE_SYMBOL(mod,sym) (const void *) inter_module_get(sym) >+#define PUT_MODULE_SYMBOL(sym) inter_module_put((char *) sym) >+ >+#ifdef KERNEL_2_6 >+# define NV_IS_SUSER() capable(CAP_SYS_ADMIN) >+# define NV_PCI_DEVICE_NAME(x) ((x)->pretty_name) >+# define NV_CLI() local_irq_disable() >+# define NV_SAVE_FLAGS(x) local_save_flags(x) >+# define NV_RESTORE_FLAGS(x) local_irq_restore(x) >+# define NV_MAY_SLEEP() (!in_interrupt() && !in_atomic()) >+# define NV_SMP_NUM_CPUS num_online_cpus() >+# define NV_MODULE_PARAMETER(x) module_param(x, int, 0) >+#else >+# define NV_IS_SUSER() suser() >+# define NV_PCI_DEVICE_NAME(x) ((x)->name) >+# define NV_CLI() __cli() >+# define NV_SAVE_FLAGS(x) __save_flags(x) >+# define NV_RESTORE_FLAGS(x) __restore_flags(x) >+# define NV_MAY_SLEEP() (!in_interrupt()) >+# define NV_SMP_NUM_CPUS smp_num_cpus >+# define NV_MODULE_PARAMETER(x) MODULE_PARM(x, "i") > #endif > >-#if !defined (KERNEL_2_2) >-# define LINUX_VMA_OFFS(vma) (((vma)->vm_pgoff) << PAGE_SHIFT) >-# define GET_MODULE_SYMBOL(mod,sym) (const void *) inter_module_get(sym) >-# define PUT_MODULE_SYMBOL(sym) inter_module_put((char *) sym) >-# define GET_MAP_NR(phys_page) virt_to_page(__va(phys_page)) >-# define MEM_MAP_READ_COUNT(map_nr) (atomic_read(&(map_nr)->count)) >-# define MEM_MAP_INC_COUNT(map_nr) (atomic_inc(&(map_nr)->count)) >-# define MEM_MAP_DEC_COUNT(map_nr) (atomic_dec(&(map_nr)->count)) >-# define GET_EVENT_QUEUE(nv) ((struct __wait_queue_head *) ((nv)->event_queue)) >-# define VMA_PRIVATE(vma) ((vma)->vm_private_data) >-#else >-# define in_irq() (local_irq_count[smp_processor_id()]) >-# define LINUX_VMA_OFFS(vma) ((vma)->vm_offset) >-# define GET_MODULE_SYMBOL(mod, sym) (void*) get_module_symbol((mod), (sym)) >-# define PUT_MODULE_SYMBOL(sym) >-# define GET_MAP_NR(phys_page) MAP_NR(__va(phys_page)) >-# define MEM_MAP_READ_COUNT(map_nr) (atomic_read(&mem_map[map_nr].count)) >-# define MEM_MAP_INC_COUNT(map_nr) (atomic_inc(&mem_map[map_nr].count)) >-# define MEM_MAP_DEC_COUNT(map_nr) (atomic_dec(&mem_map[map_nr].count)) >-# define GET_EVENT_QUEUE(nv) ((struct wait_queue **) &((nv)->event_queue)) >-# define VMA_PRIVATE(vma) ((void*)((vma)->vm_pte)) >+#define NV_DEVICE_NUMBER(x) minor((x)->i_rdev) >+#define NV_VMA_OFFSET(x) (((x)->vm_pgoff) << PAGE_SHIFT) >+#define NV_VMA_PRIVATE(x) ((x)->vm_private_data) >+#define NV_IS_CONTROL_DEVICE(x) (minor((x)->i_rdev) == 255) >+ >+#ifndef minor >+/* >+ * XXX Is this correct for all possible target kernels? We need this to >+ * support older 2.4 and some 2.6 kernels. >+ */ >+# define minor(x) MINOR(x) > #endif > >+#ifndef KERNEL_2_6 >+#ifndef IRQ_HANDLED >+ typedef void irqreturn_t; >+#define IRQ_HANDLED >+#endif >+#else >+#ifdef AGPGART >+ typedef struct agp_kern_info agp_kern_info; >+ typedef struct agp_memory agp_memory; >+#endif >+ typedef void* devfs_handle_t; >+#endif >+ >+#ifdef KERNEL_2_6 >+#define NV_DEVFS_REGISTER(_name, _minor) \ >+({ \ >+ devfs_handle_t __handle = NULL; \ >+ devfs_mk_cdev(MKDEV(NV_MAJOR_DEVICE_NUMBER, _minor), \ >+ S_IFCHR | S_IRUGO | S_IWUGO, _name); \ >+ __handle; \ >+}) >+ >+#define NV_DEVFS_REMOVE_DEVICE(i) devfs_remove("nvidia%d", i) >+#define NV_DEVFS_REMOVE_CONTROL() devfs_remove("nvidiactl") >+#else >+#define NV_DEVFS_REGISTER(_name, _minor) \ >+({ \ >+ devfs_handle_t __handle = \ >+ devfs_register(NULL, _name, DEVFS_FL_DEFAULT, \ >+ NV_MAJOR_DEVICE_NUMBER, _minor, \ >+ S_IFCHR | S_IRUGO | S_IWUGO, &nv_fops, NULL); \ >+ __handle; \ >+ }) >+ >+#define NV_DEVFS_REMOVE_DEVICE(i) devfs_unregister(nv_devfs_handles[i+1]) >+#define NV_DEVFS_REMOVE_CONTROL() devfs_unregister(nv_devfs_handles[0]) >+#endif >+ >+/* >+ * Linux 2.5 introduced the five argument version of remap_page_range, all >+ * relevant releases to date use it. This version was backported to 2.4 by >+ * RedHat without means to identify the change, hence this hack. >+ */ >+#ifdef KERNEL_2_6 >+#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(vma, a, ## b) >+#else > #if defined(REMAP_PAGE_RANGE_5) >-#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(vma, a, ## b) >+#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(vma, a, ## b) > #elif defined(REMAP_PAGE_RANGE_4) >-#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(a, ## b) >+#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(a, ## b) > #else >-#error "Couldn't determine number of arguments expected by remap_page_range!" >+#warning "conftest.sh failed, assuming old Linux 2.4 remap_page_range(4)!" >+ >+#define NV_REMAP_PAGE_RANGE(a, b...) remap_page_range(a, ## b) > #endif >+#endif /* KERNEL_2_6 */ > >-#if defined(pte_offset_atomic) >-#define NV_PTE_OFFSET(addres, pg_mid_dir, pte) \ >+#if defined(pmd_offset_map) >+#define NV_PMD_OFFSET(address, pgd, pmd) \ > { \ >- pte_t *pte__ = pte_offset_atomic(pg_mid_dir, address); \ >- pte = *pte__; \ >- pte_kunmap(pte__); \ >+ pmd = pmd_offset_map(pgd, address); \ >+ } >+#define NV_PMD_UNMAP(pmd) \ >+ { \ >+ pmd_unmap(pmd); \ > } >-#elif defined(pte_offset) >-#define NV_PTE_OFFSET(addres, pg_mid_dir, pte) \ >- pte = *pte_offset(pg_mid_dir, address) > #else >-#define NV_PTE_OFFSET(addres, pg_mid_dir, pte) \ >+#define NV_PMD_OFFSET(address, pgd, pmd) \ > { \ >- pte_t *pte__ = pte_offset_map(pg_mid_dir, address); \ >- pte = *pte__; \ >- pte_unmap(pte__); \ >+ pmd = pmd_offset(pgd, address); \ > } >+#define NV_PMD_UNMAP(pmd) > #endif > >-#define NV_PAGE_ALIGN(addr) ( ((addr) + PAGE_SIZE - 1) / PAGE_SIZE) >-#define NV_MASK_OFFSET(addr) ( (addr) & (PAGE_SIZE - 1) ) >+#define NV_PMD_PRESENT(pmd) \ >+ ({ \ >+ if (pmd) { \ >+ if (pmd_none(*pmd)) { \ >+ NV_PMD_UNMAP(pmd); pmd = NULL; \ >+ } \ >+ } pmd != NULL; \ >+ }) > >-#ifndef MAXMEM /* temporary define for 2.2 kernels */ >-#define MAXMEM (-PAGE_OFFSET - (64 * 1024 * 1024)) >+#if defined (pte_offset_atomic) >+#define NV_PTE_OFFSET(address, pmd, pte) \ >+ { \ >+ pte = pte_offset_atomic(pmd, address); \ >+ NV_PMD_UNMAP(pmd); \ >+ } >+#define NV_PTE_UNMAP(pte) \ >+ { \ >+ pte_kunmap(pte); \ >+ } >+#elif defined (pte_offset) >+#define NV_PTE_OFFSET(address, pmd, pte) \ >+ { \ >+ pte = pte_offset(pmd, address); \ >+ NV_PMD_UNMAP(pmd); \ >+ } >+#define NV_PTE_UNMAP(pte) >+#else >+#define NV_PTE_OFFSET(address, pmd, pte) \ >+ { \ >+ pte = pte_offset_map(pmd, address); \ >+ NV_PMD_UNMAP(pmd); \ >+ } >+#define NV_PTE_UNMAP(pte) \ >+ { \ >+ pte_unmap(pte); \ >+ } > #endif > >-#ifndef NV01_ROOT >-#define NV01_ROOT 0x00000000 >-#endif >+#define NV_PTE_PRESENT(pte) \ >+ ({ \ >+ if (pte) { \ >+ if (!pte_present(*pte)) { \ >+ NV_PTE_UNMAP(pte); pte = NULL; \ >+ } \ >+ } pte != NULL; \ >+ }) >+ >+#define NV_PTE_VALUE(pte) \ >+ ({ \ >+ unsigned long __pte_value = pte_val(*pte); \ >+ NV_PTE_UNMAP(pte); \ >+ __pte_value; \ >+ }) >+ >+#define NV_PAGE_ALIGN(addr) ( ((addr) + PAGE_SIZE - 1) / PAGE_SIZE) >+#define NV_MASK_OFFSET(addr) ( (addr) & (PAGE_SIZE - 1) ) > > #if defined(NVCPU_IA64) > #define NV_GFP_HW (GFP_KERNEL | __GFP_DMA) >@@ -252,7 +347,7 @@ > static inline int NV_IRQL_IS_RAISED() > { > unsigned long int eflags; >- __save_flags(eflags); >+ NV_SAVE_FLAGS(eflags); > return !(eflags & NV_CPU_INTERRUPT_FLAGS_BIT); > } > >@@ -310,16 +405,14 @@ > > nv_alloc_t *alloc_queue; > >- // bottom half interrupt handler info; per device >- /* keep track of any pending bottom-halves */ >- struct tq_struct *bh; >- atomic_t bh_count; >+ /* keep track of any pending bottom halfes */ >+ struct tasklet_struct tasklet; > > U032 vblank_notifier; > U032 waiting_for_vblank; > > /* queue for for NV's OS events */ >- void *event_queue; >+ wait_queue_head_t waitqueue; > > /* get a timer callback every second */ > struct timer_list rc_timer; >diff -ruN NVIDIA_kernel-1.0-4363/nv-misc.h NVIDIA_kernel-1.0-4363-2.6/nv-misc.h >--- NVIDIA_kernel-1.0-4363/nv-misc.h 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/nv-misc.h 2003-12-31 15:30:32.000000000 +0100 >@@ -53,5 +53,11 @@ > > typedef void *PHWINFO; > >+/* >+ * XXX: Make sure that arguments to and from the core resource >+ * manager are passed and expected on the stack. Is this a good >+ * place for this? >+ */ >+#define NV_API_CALL __attribute__((regparm(0))) > > #endif /* _NV_MISC_H_ */ >diff -ruN NVIDIA_kernel-1.0-4363/nv.c NVIDIA_kernel-1.0-4363-2.6/nv.c >--- NVIDIA_kernel-1.0-4363/nv.c 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/nv.c 2004-01-13 14:28:47.000000000 +0100 >@@ -16,11 +16,14 @@ > #include "os-agp.h" > > >+MODULE_LICENSE("NVIDIA"); >+ > /* > * our global state; one per device > */ > > nv_linux_state_t nv_linux_devices[NV_MAX_DEVICES] = { { { 0 } } }; >+static int nv_num_devices; > > #ifdef CONFIG_PM > /* XXX PM do we only need one, or one for each device? */ >@@ -36,27 +39,15 @@ > // keep track of opened clients and their process id so they > // can be free'd up on abnormal close > nv_client_t nv_clients[NV_MAX_CLIENTS]; >-struct tq_struct nv_bottom_halves[NV_MAX_CLIENTS]; > > #ifdef CONFIG_PROC_FS > struct proc_dir_entry *proc_nvidia; > #endif > > #ifdef CONFIG_DEVFS_FS >-devfs_handle_t nv_dev_handle[NV_MAX_DEVICES]; >-devfs_handle_t nv_ctl_handle; >+devfs_handle_t nv_devfs_handles[NV_MAX_DEVICES+1]; > #endif > >-/* >- * pick apart our minor device number >- * low 3 bits is NV device >- * if 255, then its the control device >- */ >- >-#define NV_DEVICE_NUMBER(_minor) ((_minor) & 0x0f) >-#define NV_DEVICE_IS_CONTROL_DEVICE(_minor) \ >- (((_minor) & 0xFF) == 0xFF) >- > // #define NV_DBG_MEM 1 > #undef NV_DBG_MEM > >@@ -107,10 +98,6 @@ > *** EXPORTS to Linux Kernel > ***/ > >-/* linux module interface functions (called by linux kernel) */ >-int init_module(void); >-void cleanup_module(void); >- > /* nv_kern_ functions, interfaces used by linux kernel */ > void nv_kern_vma_open(struct vm_area_struct *vma); > void nv_kern_vma_release(struct vm_area_struct *vma); >@@ -120,8 +107,8 @@ > int nv_kern_mmap(struct file *, struct vm_area_struct *); > unsigned int nv_kern_poll(struct file *, poll_table *); > int nv_kern_ioctl(struct inode *, struct file *, unsigned int, unsigned long); >-void nv_kern_bh(void *); >-void nv_kern_isr(int, void *, struct pt_regs *); >+void nv_kern_isr_bh(unsigned long); >+irqreturn_t nv_kern_isr(int, void *, struct pt_regs *); > void nv_kern_rc_timer(unsigned long); > #ifdef CONFIG_PM > int nv_kern_pm(struct pm_dev *dev, pm_request_t rqst, void *data); >@@ -144,6 +131,7 @@ > /* character driver entry points */ > > static struct file_operations nv_fops = { >+ owner: THIS_MODULE, > poll: nv_kern_poll, > ioctl: nv_kern_ioctl, > mmap: nv_kern_mmap, >@@ -298,8 +286,8 @@ > { > nv_printf(NV_DBG_MEMINFO, " 0x%x: count %d flags 0x%x\n", > *page_ptr, >- (GET_MAP_NR(*page_ptr))->count, >- (GET_MAP_NR(*page_ptr))->flags); >+ (virt_to_page(__va(*page_ptr)))->count, >+ (virt_to_page(__va(*page_ptr)))->flags); > num_pages--; > page_ptr++; > } >@@ -370,7 +358,7 @@ > phys_addr = virt_to_phys((void *) virt_addr); > > /* lock the page for dma purposes */ >- mem_map_reserve(GET_MAP_NR(phys_addr)); >+ SetPageReserved(virt_to_page(__va(phys_addr))); > > *page_ptr++ = phys_addr; > pages_needed--; >@@ -383,7 +371,7 @@ > while (page_ptr != (unsigned long *) page_list) > { > page_ptr--; >- mem_map_unreserve(GET_MAP_NR(*page_ptr)); >+ ClearPageReserved(virt_to_page(__va(*page_ptr))); > free_page((unsigned long) phys_to_virt(*page_ptr)); > } > >@@ -403,7 +391,7 @@ > > while (pages_left) > { >- mem_map_unreserve(GET_MAP_NR(*page_list)); >+ ClearPageReserved(virt_to_page(__va(*page_list))); > page_list++; > pages_left--; > } >@@ -569,11 +557,6 @@ > nv_linux_state_t *nvl; > nv_linux_state_t *nv_max_devices; > >-#if defined (KERNEL_2_2) >- struct proc_dir_entry *proc_root_driver; >- proc_root_driver = create_proc_entry("driver", flags, &proc_root); >-#endif >- > proc_nvidia = create_proc_entry("nvidia", flags, proc_root_driver); > proc_nvidia_cards = create_proc_entry("cards", flags, proc_nvidia); > proc_nvidia_agp = create_proc_entry("agp", flags, proc_nvidia); >@@ -635,9 +618,6 @@ > { > #ifdef CONFIG_PROC_FS > nvos_proc_remove_all(proc_nvidia); >-#if defined (KERNEL_2_2) >- remove_proc_entry("driver", &proc_root); >-#endif > #endif > } > >@@ -726,71 +706,75 @@ > *** EXPORTS to Linux Kernel > ***/ > >-int init_module(void) >+static int __init nvidia_init_module(void) > { >- nv_linux_state_t *nvl; >- int rc; >- int num_devices; >+ int i, rc; > > memset(nv_linux_devices, 0, sizeof(nv_linux_devices)); >- num_devices = nvos_probe_devices(); >+ nv_num_devices = nvos_probe_devices(); > >- if (num_devices == 0) { >+ if (nv_num_devices == 0) { > nv_printf(NV_DBG_ERRORS, "nvidia: no NVIDIA graphics adapter found\n"); > return -ENODEV; > } > > nv_printf(NV_DBG_ERRORS, "nvidia: loading %s\n", pNVRM_ID); > >-#ifdef CONFIG_DEVFS_FS >+#if defined(CONFIG_DEVFS_FS) && !defined(KERNEL_2_6) > rc = devfs_register_chrdev(nv_major, "nvidia", &nv_fops); > #else > rc = register_chrdev(nv_major, "nvidia", &nv_fops); > #endif > > if (rc < 0) { >- nv_printf(NV_DBG_ERRORS, "init_module: register failed\n"); >+ nv_printf(NV_DBG_ERRORS, "nvidia_init_module: register failed\n"); > return rc; > } > > #ifdef CONFIG_DEVFS_FS >- memset(nv_dev_handle, 0, sizeof(devfs_handle_t) * NV_MAX_DEVICES); > do { >+ /* >+ * XXX This code isn't pretty, but neither is dealing with the >+ * various Linux devfs implemenation(s). While Linux APIs are >+ * known to be anything but stable and oftentimes anything but >+ * well designed, the devfs interface has been more painful to >+ * deal with than most other APIs. >+ */ > char name[10]; >- int i; > >- nv_ctl_handle = devfs_register(NULL, "nvidiactl", >- DEVFS_FL_DEFAULT, nv_major, 255, >- S_IFCHR | S_IRUGO | S_IWUGO, >- &nv_fops, NULL); >+ nv_devfs_handles[0] = NV_DEVFS_REGISTER("nvidiactl", 255); > >- for (i = 0; i < num_devices; i++) { >+ for (i = 0; i < nv_num_devices; i++) { > sprintf(name, "nvidia%d", i); >- nv_dev_handle[i] = devfs_register(NULL, name, >- DEVFS_FL_DEFAULT, nv_major, i, >- S_IFCHR | S_IRUGO | S_IWUGO, >- &nv_fops, NULL); >+ nv_devfs_handles[i+1] = NV_DEVFS_REGISTER(name, i); > } > } while(0); > #endif > >- nv_printf(NV_DBG_INFO, "init_module: major number %d\n", nv_major); >+ nv_printf(NV_DBG_INFO, "nvidia_init_module: major number %d\n", nv_major); > >- // init all the bottom half structures >- for (nvl = nv_linux_devices; nvl < nv_linux_devices + NV_MAX_DEVICES; nvl++) >- { >- nvl->bh = &nv_bottom_halves[nvl - nv_linux_devices]; >- nvl->bh->routine = nv_kern_bh; >- nvl->bh->data = (void *) nvl; >- nvl->bh->sync = 0; >+ /* instantiate tasklets */ >+ for (i = 0; i < NV_MAX_DEVICES; i++) { >+ /* >+ * We keep one tasklet per card to avoid latency issues with more >+ * than one device; no two instances of a single tasklet are ever >+ * executed concurrently. >+ */ >+ atomic_set(&nv_linux_devices[i].tasklet.count, 1); >+ >+ /* >+ * Initialize the event queue for this device. This only needs to >+ * happen once for every device. >+ */ >+ init_waitqueue_head(&nv_linux_devices[i].waitqueue); > } > > // init the nvidia control device > { > nv_state_t *nv_ctl = NV_STATE_PTR(&nv_ctl_device); >- nv_ctl_device.event_queue = NULL; > nv_ctl->os_state = (void *) &nv_ctl_device; > nv_lock_init_locks(nv_ctl); >+ init_waitqueue_head(&nv_ctl_device.waitqueue); > } > > #ifdef CONFIG_PM >@@ -815,14 +799,14 @@ > /* create /proc/driver/nvidia */ > nvos_proc_create(); > >-#if defined(DEBUG) && !defined(KERNEL_2_2) >+#if defined(DEBUG) > inter_module_register("nv_linux_devices", THIS_MODULE, nv_linux_devices); > #endif > > return 0; > > failed: >-#ifdef CONFIG_DEVFS_FS >+#if defined(CONFIG_DEVFS_FS) && !defined(KERNEL_2_6) > devfs_unregister_chrdev(nv_major, "nvidia"); > #else > unregister_chrdev(nv_major, "nvidia"); >@@ -830,18 +814,18 @@ > return rc; > } > >-void cleanup_module(void) >+static void __exit nvidia_exit_module(void) > { >- int rc; > nv_linux_state_t *nvl; > nv_linux_state_t *max_devices; >+ int rc; > > /* remove /proc/driver/nvidia */ > nvos_proc_remove(); > >- nv_printf(NV_DBG_INFO, "cleanup_module\n"); >+ nv_printf(NV_DBG_INFO, "nvidia_exit_module\n"); > >-#if defined(DEBUG) && !defined(KERNEL_2_2) >+#if defined(DEBUG) > inter_module_unregister("nv_linux_devices"); > #endif > >@@ -866,33 +850,34 @@ > continue; > > nv_printf(NV_DBG_ERRORS, >- "still have vm que at cleanup_module(): 0x%x to 0x%x\n", >+ "still have vm que at nvidia_exit_module(): 0x%x to 0x%x\n", > nvl->alloc_queue->vma->vm_start, > nvl->alloc_queue->vma->vm_end); > } > } > >-#ifdef CONFIG_DEVFS_FS >+#if defined(CONFIG_DEVFS_FS) && !defined(KERNEL_2_6) > rc = devfs_unregister_chrdev(nv_major, "nvidia"); > #else > rc = unregister_chrdev(nv_major, "nvidia"); > #endif > > if (rc < 0) { >- nv_printf(NV_DBG_ERRORS, "cleanup_module: unregister nv failed\n"); >+ nv_printf(NV_DBG_ERRORS, "nvidia_exit_module: unregister nv failed\n"); > } > > #ifdef CONFIG_DEVFS_FS > do { > int i; >- for (i = 0; nv_dev_handle[i] != 0; i++) { >- devfs_unregister(nv_dev_handle[i]); >- } >- } while(0); >- devfs_unregister(nv_ctl_handle); >+ NV_DEVFS_REMOVE_CONTROL(); >+ for (i = 0; i < nv_num_devices; i++) >+ NV_DEVFS_REMOVE_DEVICE(i); >+ } while (0); > #endif > } > >+module_init(nvidia_init_module); >+module_exit(nvidia_exit_module); > > /* this is only called when the vmas are duplicated. > * this appears to only happen when the process is cloned to create >@@ -906,11 +891,11 @@ > nv_kern_vma_open(struct vm_area_struct *vma) > { > nv_printf(NV_DBG_MEMINFO, "vma_open for 0x%x - 0x%x, offset 0x%x\n", >- vma->vm_start, vma->vm_end, LINUX_VMA_OFFS(vma)); >+ vma->vm_start, vma->vm_end, NV_VMA_OFFSET(vma)); > >- if (VMA_PRIVATE(vma)) >+ if (NV_VMA_PRIVATE(vma)) > { >- nv_alloc_t *at = (nv_alloc_t *) VMA_PRIVATE(vma); >+ nv_alloc_t *at = (nv_alloc_t *) NV_VMA_PRIVATE(vma); > at->usage_count++; > > nv_printf(NV_DBG_MEMINFO, " at 0x%x, usage count %d, page_table 0x%x\n", >@@ -918,8 +903,6 @@ > > nvos_list_page_count(at->page_table, at->num_pages); > } >- >- MOD_INC_USE_COUNT; > } > > >@@ -927,11 +910,11 @@ > nv_kern_vma_release(struct vm_area_struct *vma) > { > nv_printf(NV_DBG_MEMINFO, "vma_release for 0x%x - 0x%x, offset 0x%x\n", >- vma->vm_start, vma->vm_end, LINUX_VMA_OFFS(vma)); >+ vma->vm_start, vma->vm_end, NV_VMA_OFFSET(vma)); > >- if (VMA_PRIVATE(vma)) >+ if (NV_VMA_PRIVATE(vma)) > { >- nv_alloc_t *at = (nv_alloc_t *) VMA_PRIVATE(vma); >+ nv_alloc_t *at = (nv_alloc_t *) NV_VMA_PRIVATE(vma); > > at->usage_count--; > >@@ -949,35 +932,26 @@ > if (at->page_table) > nvos_unlock_pages(at->page_table, at->num_pages); > nvos_free_alloc(at); >- VMA_PRIVATE(vma) = NULL; >+ NV_VMA_PRIVATE(vma) = NULL; > } > } >- >- MOD_DEC_USE_COUNT; > } > > >-/* at this point, this code just plain won't work with 2.2 kernels. >- * additionally, only ia64 & the 460GX need a nopage handler, and 2.2 doesn't >- * work on ia64 anyways. It's expected that at some point other agp chipsets >- * will work similar to the 460GX (AGP 3.0 spec), so pre-emptively make sure >- * this works on our standard ia32 driver. >- */ >-#if !defined(KERNEL_2_2) >- > /* AGP allocations under the 460GX are not mapped to the aperture > * addresses by the CPU. This nopage handler will fault on CPU > * accesses to AGP memory and map the address to the correct page. > */ > struct page *nv_kern_vma_nopage(struct vm_area_struct *vma, unsigned long address, int write_access) > { >+#if defined(NVCPU_IA64) && (LINUX_VERSION_CODE >= KERNEL_VERSION(2, 4, 9)) > nv_alloc_t *at, *tmp; > nv_linux_state_t *nvl; > nv_state_t *nv; > struct page *page_ptr; > int rm_status, index; > >- at = VMA_PRIVATE(vma); >+ at = NV_VMA_PRIVATE(vma); > if (at == NULL) > { > nv_printf(NV_DBG_ERRORS, "NVRM: nopage handler called without an at: " >@@ -1016,7 +990,7 @@ > // far again > nv_printf(NV_DBG_ERRORS, "NVRM: nopage handler called on a freed" > "address: vm_start 0x%x, at 0x%x\n", vma->vm_start, at); >- VMA_PRIVATE(vma) = NULL; >+ NV_VMA_PRIVATE(vma) = NULL; > return NOPAGE_SIGBUS; > } > >@@ -1037,18 +1011,16 @@ > at->page_table[index] = (void *) ((page_ptr - mem_map) << PAGE_SHIFT); > > return page_ptr; >-} > #endif >+ return NOPAGE_SIGBUS; >+} > > struct vm_operations_struct nv_vm_ops = { > nv_kern_vma_open, > nv_kern_vma_release, /* "close" */ >-#if !defined(KERNEL_2_2) > nv_kern_vma_nopage, >-#endif > }; > >- > /* > ** nv_kern_open > ** >@@ -1076,20 +1048,17 @@ > > /* for control device, just jump to its open routine */ > /* after setting up the private data */ >- if (NV_DEVICE_IS_CONTROL_DEVICE(inode->i_rdev)) >+ if (NV_IS_CONTROL_DEVICE(inode)) > return nv_kern_ctl_open(inode, file); > > /* what device are we talking about? */ >- devnum = NV_DEVICE_NUMBER(inode->i_rdev); >+ devnum = NV_DEVICE_NUMBER(inode); > if (devnum >= NV_MAX_DEVICES) > { > rc = -ENODEV; > goto failed; > } > >- >- MOD_INC_USE_COUNT; >- > nvl = &nv_linux_devices[devnum]; > nv = NV_STATE_PTR(nvl); > >@@ -1137,16 +1106,9 @@ > goto failed; > } > >-#if !defined (KERNEL_2_2) >- NV_KMALLOC(nvl->event_queue, sizeof(struct __wait_queue_head)); >- if (nvl->event_queue == NULL) >- goto failed; >- memset(nvl->event_queue, 0, sizeof(struct __wait_queue_head)); >- >- init_waitqueue_head(GET_EVENT_QUEUE(nvl)); >-#else >- nvl->event_queue = NULL; >-#endif >+ nvl->tasklet.func = nv_kern_isr_bh; >+ nvl->tasklet.data = (unsigned long) nv->pdev; >+ tasklet_enable(&nvl->tasklet); > > nv->flags |= NV_FLAG_OPEN; > } >@@ -1157,7 +1119,6 @@ > return rc; > > failed: >- MOD_DEC_USE_COUNT; > nv_unlock_ldata(nv); > return rc; > } >@@ -1179,38 +1140,32 @@ > > /* for control device, just jump to its open routine */ > /* after setting up the private data */ >- if (NV_DEVICE_IS_CONTROL_DEVICE(inode->i_rdev)) >+ if (NV_IS_CONTROL_DEVICE(inode)) > return nv_kern_ctl_close(inode, file); > >- nv_printf(NV_DBG_INFO, "nv_kern_close on device %d\n", NV_DEVICE_NUMBER(inode->i_rdev)); >+ nv_printf(NV_DBG_INFO, "nv_kern_close on device %d\n", NV_DEVICE_NUMBER(inode)); > > rm_free_unused_clients(nv, current->pid, (void *) file); > > nv_lock_ldata(nv); > if (--nv->usage_count == 0) > { >- int counter = 0; >- >- /* turn off interrupts. >- ** be careful to make sure any pending bottom half gets run >- ** or disabled before calling rm_shutdown_adapter() since >- ** it will free up the pdev. This is hard to see on single >- ** cpu systems, but easy on dual cpu :-) >- */ >+ /* >+ * The usage count for this device has dropped to zero, it can be shut >+ * down safely; disable its interrupts. >+ */ > rm_disable_adapter(nv); > >- /* give it a moment to allow any bottom half to run */ >- >-#define MAX_BH_TASKS 10 >- while (NV_ATOMIC_READ(nvl->bh_count) && (counter < MAX_BH_TASKS)) >- { >- current->state = TASK_INTERRUPTIBLE; >- schedule_timeout(HZ/50); >- counter++; >- } >+ /* >+ * Disable this device's tasklet to make sure that no bottom half will >+ * run with undefined device state. >+ */ >+ tasklet_disable(&nvl->tasklet); > >- /* free the irq, which may block until any pending interrupts */ >- /* are done being processed. */ >+ /* >+ * Free the IRQ, which may block until all pending interrupt processing >+ * has completed. >+ */ > free_irq(nv->interrupt_line, (void *) nv); > > rm_shutdown_adapter(nv); >@@ -1230,12 +1185,6 @@ > } > } > >-#if !defined (KERNEL_2_2) >- /* this only needs to be freed on 2.4 and later kernels */ >- NV_KFREE(nvl->event_queue); >- nvl->event_queue = NULL; >-#endif >- > /* leave INIT flag alone so we don't reinit every time */ > nv->flags &= ~(NV_FLAG_OPEN | NV_FLAG_WAITING); > } >@@ -1246,8 +1195,6 @@ > NV_KFREE(file->private_data); > file->private_data = (void *) 0; > >- MOD_DEC_USE_COUNT; >- > return 0; > } > >@@ -1264,10 +1211,10 @@ > nv_printf(NV_DBG_INFO, "mmap([0x%lx-0x%lx] off=0x%lx)\n", > vma->vm_start, > vma->vm_end, >- LINUX_VMA_OFFS(vma)); >+ NV_VMA_OFFSET(vma)); > > // be a bit paranoid for now >- if ((NV_MASK_OFFSET(LINUX_VMA_OFFS(vma))) || >+ if ((NV_MASK_OFFSET(NV_VMA_OFFSET(vma))) || > (NV_MASK_OFFSET(vma->vm_start)) || > (NV_MASK_OFFSET(vma->vm_end))) > { >@@ -1285,7 +1232,7 @@ > > > /* NV reg space */ >- if (IS_REG_OFFSET(nv, LINUX_VMA_OFFS(vma), vma->vm_end - vma->vm_start)) >+ if (IS_REG_OFFSET(nv, NV_VMA_OFFSET(vma), vma->vm_end - vma->vm_start)) > { > /* truncate to size of registers */ > if (pages > nv->regs->size / PAGE_SIZE) >@@ -1293,7 +1240,7 @@ > > vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); > if (NV_REMAP_PAGE_RANGE(vma->vm_start, >- LINUX_VMA_OFFS(vma), >+ NV_VMA_OFFSET(vma), > vma->vm_end - vma->vm_start, > vma->vm_page_prot)) > return -EAGAIN; >@@ -1303,7 +1250,7 @@ > } > > /* NV fb space */ >- else if (IS_FB_OFFSET(nv, LINUX_VMA_OFFS(vma), vma->vm_end - vma->vm_start)) >+ else if (IS_FB_OFFSET(nv, NV_VMA_OFFSET(vma), vma->vm_end - vma->vm_start)) > { > > /* truncate to size of framebuffer */ >@@ -1312,7 +1259,7 @@ > > vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); > if (NV_REMAP_PAGE_RANGE(vma->vm_start, >- LINUX_VMA_OFFS(vma), >+ NV_VMA_OFFSET(vma), > vma->vm_end - vma->vm_start, > vma->vm_page_prot)) > return -EAGAIN; >@@ -1322,10 +1269,10 @@ > } > > /* AGP allocator */ >- else if (IS_AGP_OFFSET(nv, LINUX_VMA_OFFS(vma), vma->vm_end - vma->vm_start)) >+ else if (IS_AGP_OFFSET(nv, NV_VMA_OFFSET(vma), vma->vm_end - vma->vm_start)) > { > nv_lock_at(nv); >- at = nvl_find_alloc(nvl, LINUX_VMA_OFFS(vma), NV_ALLOC_TYPE_AGP); >+ at = nvl_find_alloc(nvl, NV_VMA_OFFSET(vma), NV_ALLOC_TYPE_AGP); > > if (at == NULL) > { >@@ -1343,7 +1290,7 @@ > } > > at->vma = vma; >- VMA_PRIVATE(vma) = at; >+ NV_VMA_PRIVATE(vma) = at; > at->usage_count++; > > if (NV_OSAGP_ENABLED(nv)) >@@ -1366,13 +1313,13 @@ > } > > /* Magic allocator */ >- else // if (LINUX_VMA_OFFS(vma) == NV_MMAP_ALLOCATION_OFFSET) >+ else // if (NV_VMA_OFFSET(vma) == NV_MMAP_ALLOCATION_OFFSET) > { > unsigned long page = 0, pos, start; > int i = 0; > > nv_lock_at(nv); >- at = nvl_find_alloc(nvl, LINUX_VMA_OFFS(vma), NV_ALLOC_TYPE_PCI); >+ at = nvl_find_alloc(nvl, NV_VMA_OFFSET(vma), NV_ALLOC_TYPE_PCI); > > if (at == NULL) > { >@@ -1390,7 +1337,7 @@ > } > > at->vma = vma; >- VMA_PRIVATE(vma) = at; >+ NV_VMA_PRIVATE(vma) = at; > at->usage_count++; > > nv_printf(NV_DBG_INFO, "remapping %d system pages for at 0x%x\n", pages, at); >@@ -1412,9 +1359,6 @@ > > vma->vm_file = file; > >- /* just increment usage count, rather than calling vma_open */ >- MOD_INC_USE_COUNT; >- > return 0; > } > >@@ -1439,7 +1383,7 @@ > return nv_kern_ctl_poll (file, wait); > > // add us to the list >- poll_wait(file, GET_EVENT_QUEUE(nvl), wait); >+ poll_wait(file, &nvl->waitqueue, wait); > > nv_lock_ldata(nv); > >@@ -1519,12 +1463,13 @@ > > switch (_IOC_NR(cmd)) > { >+#if !defined(KERNEL_2_6) > /* debug tool; zap the module use count so we can unload driver */ > /* even if it is confused */ > case _IOC_NR(NV_IOCTL_MODULE_RESET): > atomic_set(&__this_module.uc.usecount, 1); > break; >- >+#endif > /* pass out info about the card */ > case _IOC_NR(NV_IOCTL_CARD_INFO): > { >@@ -1651,7 +1596,7 @@ > * driver receives an interrupt > * if someone waiting, then hand it off. > */ >-void nv_kern_isr( >+irqreturn_t nv_kern_isr( > int irq, > void *arg, > struct pt_regs *regs >@@ -1664,21 +1609,20 @@ > rm_isr(nv->device_number, &need_to_run_bottom_half); > if (need_to_run_bottom_half) > { >- NV_ATOMIC_INC(nvl->bh_count); >- queue_task(nvl->bh, &tq_immediate); >- mark_bh(IMMEDIATE_BH); >+ tasklet_schedule(&nvl->tasklet); > } >+ >+ return IRQ_HANDLED; > } > >-void nv_kern_bh( >- void *data >-) >+void nv_kern_isr_bh(unsigned long data) > { >- nv_linux_state_t *nvl = (nv_linux_state_t *) data; >- nv_state_t *nv = NV_STATE_PTR(nvl); >- >- NV_ATOMIC_DEC(nvl->bh_count); >- rm_isr_bh(nv->pdev); >+ /* >+ * XXX: This level of indirection is necessary to work around >+ * problems with Linux kernels using a non-standard calling >+ * convention, i.e. Arjan van de Ven's/RedHat's 2.6.0 kernels. >+ */ >+ rm_isr_bh((VOID *) data); > } > > void nv_kern_rc_timer( >@@ -1760,26 +1704,6 @@ > /* save the nv away in file->private_data */ > NV_HIDE_IN_FILEP(file, nv); > >- MOD_INC_USE_COUNT; >- >- /* if this is the first time the control device has been opened, >- * allocate the wait queue >- */ >- >- if (! nvl->event_queue) { >- >-#if !defined (KERNEL_2_2) >- NV_KMALLOC(nvl->event_queue, sizeof(struct __wait_queue_head)); >- if (nvl->event_queue == NULL) >- return -ENOMEM; >- memset(nvl->event_queue, 0, sizeof(struct __wait_queue_head)); >- >- init_waitqueue_head(GET_EVENT_QUEUE(nvl)); >-#else >- nvl->event_queue = NULL; >-#endif >- } >- > nv->flags |= NV_FLAG_OPEN + NV_FLAG_CONTROL; > > /* turn off the hotkey occurred bit */ >@@ -1807,15 +1731,7 @@ > > nv_lock_ldata(nv); > if (--nv->usage_count == 0) >- { >-#if !defined (KERNEL_2_2) >- nv_linux_state_t *nvl = NV_GET_NVL_FROM_NV_STATE(nv); >- /* this only needs to be freed on 2.4 and later kernels */ >- NV_KFREE(nvl->event_queue); >- nvl->event_queue = 0; >-#endif > nv->flags = 0; >- } > nv_unlock_ldata(nv); > > rm_free_unused_clients(nv, current->pid, (void *) file); >@@ -1825,8 +1741,6 @@ > NV_KFREE(file->private_data); > file->private_data = (void *) 0; > >- MOD_DEC_USE_COUNT; >- > return 0; > } > >@@ -1851,7 +1765,7 @@ > if (file->f_flags & O_NONBLOCK) > return -EAGAIN; > >- poll_wait(file, GET_EVENT_QUEUE(nvl), wait); >+ poll_wait(file, &nvl->waitqueue, wait); > > nv_lock_ldata(nv); > >@@ -1881,7 +1795,7 @@ > nv_ctl_device.nv_state.flags |= NV_FLAG_HOTKEY_OCCURRED; > nv_unlock_ldata(&(nv_ctl_device.nv_state)); > >- wake_up_interruptible(GET_EVENT_QUEUE(&nv_ctl_device)); >+ wake_up_interruptible(&nv_ctl_device.waitqueue); > } > > int nv_kern_read_cardinfo(char *page, char **start, off_t off, >@@ -1960,7 +1874,7 @@ > len += sprintf(page+len, "Host Bridge: \t "); > > #if defined(CONFIG_PCI_NAMES) >- len += sprintf(page+len, "%s\n", dev->name); >+ len += sprintf(page+len, "%s\n", NV_PCI_DEVICE_NAME(dev)); > #else > len += sprintf(page+len, "PCI device %04x:%04x\n", > dev->vendor, dev->device); >@@ -2088,18 +2002,21 @@ > { > unsigned long retaddr = (unsigned long) at->page_table[i]; > >- if (retaddr <= MAXMEM) >- { >- return __va((retaddr + offset)); >- } >- >- // if we've allocated via vmalloc on a highmem system, the >- // physical address may not be accessible via PAGE_OFFSET, >- // that's ok, we have a simple linear pointer already. >+ /* >+ * XXX This routine should be more straight-forward; as it >+ * is, it returns a pointer into the kernel linear mapping if >+ * the incoming address belongs to a vmalloc() allocation or >+ * the kernel logical mapping of the corresponding page if it >+ * was allocated with alloc_pages(); third case below. >+ */ > if (at->flags & NV_ALLOC_TYPE_VMALLOC) > { > return (void *)((unsigned char *) at->key_mapping + (i << PAGE_SHIFT) + offset); > } >+ else if (!(at->flags & NV_ALLOC_TYPE_CONTIG)) >+ { >+ return __va((retaddr + offset)); >+ } > > // ?? this may be a contiguous allocation, fall through > // to below? or should I just check at->flag here? >@@ -2169,9 +2086,9 @@ > unsigned long > nv_get_phys_address(unsigned long address) > { >- pgd_t *pg_dir; >- pmd_t *pg_mid_dir; >- pte_t pte; >+ pgd_t *pgd; >+ pmd_t *pmd; >+ pte_t *pte; > > #if defined(NVCPU_IA64) > if (address > __IA64_UNCACHED_OFFSET) >@@ -2184,23 +2101,24 @@ > return __pa(address); > > if (address > VMALLOC_START) >- pg_dir = pgd_offset_k(address); >+ pgd = pgd_offset_k(address); > else >- pg_dir = pgd_offset(current->mm, address); >+ pgd = pgd_offset(current->mm, address); > >- if (pgd_none(*pg_dir)) >+ if (!pgd || pgd_none(*pgd)) > goto failed; > >- pg_mid_dir = pmd_offset(pg_dir, address); >- if (pmd_none(*pg_mid_dir)) >+ NV_PMD_OFFSET(address, pgd, pmd); >+ >+ if (!NV_PMD_PRESENT(pmd)) > goto failed; > >- NV_PTE_OFFSET(address, pg_mid_dir, pte); >+ NV_PTE_OFFSET(address, pmd, pte); > >- if (!pte_present(pte)) >+ if (!NV_PTE_PRESENT(pte)) > goto failed; >- >- return ((pte_val(pte) & KERN_PAGE_MASK) | NV_MASK_OFFSET(address)); >+ >+ return (NV_PTE_VALUE(pte) & KERN_PAGE_MASK) | NV_MASK_OFFSET(address); > > failed: > return (unsigned long) NULL; >@@ -2567,7 +2485,7 @@ > > nvfp->any_fired_notifiers++; > >- wake_up_interruptible(GET_EVENT_QUEUE(nvl)); >+ wake_up_interruptible(&nvl->waitqueue); > } > > /* >@@ -2584,7 +2502,7 @@ > if (nvl->waiting_for_vblank) > nvl->vblank_notifier++; > >- wake_up_interruptible(GET_EVENT_QUEUE(nvl)); >+ wake_up_interruptible(&nvl->waitqueue); > } > > >@@ -2633,12 +2551,8 @@ > if ( (NV_AGP_DISABLED(nv)) && (config & NVOS_AGP_CONFIG_NVAGP) ) > { > /* make sure the user does not have agpgart loaded */ >-#if !defined (KERNEL_2_2) > if (inter_module_get("drm_agp")) { > inter_module_put("drm_agp"); >-#else >- if (GET_MODULE_SYMBOL(0, __MODULE_STRING(agp_enable))) { >-#endif > nv_printf(NV_DBG_WARNINGS, "NVRM: not using NVAGP, AGPGART is loaded!!\n"); > } else > status = rm_init_agp(nv); >diff -ruN NVIDIA_kernel-1.0-4363/nv.h NVIDIA_kernel-1.0-4363-2.6/nv.h >--- NVIDIA_kernel-1.0-4363/nv.h 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/nv.h 2003-12-31 15:33:31.000000000 +0100 >@@ -255,29 +255,29 @@ > * --------------------------------------------------------------------------- > */ > >-VOID* nv_find_kernel_mapping (nv_state_t *, unsigned long); >-VOID* nv_find_agp_kernel_mapping (nv_state_t *, unsigned long); >-ULONG nv_get_phys_address (ULONG); >- >-VOID nv_lock_rm (nv_state_t *); >-VOID nv_unlock_rm (nv_state_t *); >-VOID nv_lock_bh (nv_state_t *); >-VOID nv_unlock_bh (nv_state_t *); >- >-VOID nv_post_event (VOID *); >-VOID nv_post_vblank (nv_state_t *); >-VOID nv_set_hotkey_occurred_flag (VOID); >-S032 nv_int10h_call (nv_state_t *, U032 *, U032 *, U032 *, U032 *, VOID *); >- >-S032 nv_alloc_pages (nv_state_t *, VOID **, U032, U032, U032, U032, U032, U032, VOID **); >-S032 nv_free_pages (nv_state_t *, VOID **, U032, U032, VOID **); >- >-S032 nv_agp_init (nv_state_t *, VOID **, VOID **, VOID *, U032); >-S032 nv_agp_teardown (nv_state_t *); >-S032 nv_agp_translate_address (nv_state_t *, VOID *, U032, U032 *); >+VOID* NV_API_CALL nv_find_kernel_mapping (nv_state_t *, unsigned long); >+VOID* NV_API_CALL nv_find_agp_kernel_mapping (nv_state_t *, unsigned long); >+ULONG NV_API_CALL nv_get_phys_address (ULONG); >+ >+VOID NV_API_CALL nv_lock_rm (nv_state_t *); >+VOID NV_API_CALL nv_unlock_rm (nv_state_t *); >+VOID NV_API_CALL nv_lock_bh (nv_state_t *); >+VOID NV_API_CALL nv_unlock_bh (nv_state_t *); >+ >+VOID NV_API_CALL nv_post_event (VOID *); >+VOID NV_API_CALL nv_post_vblank (nv_state_t *); >+VOID NV_API_CALL nv_set_hotkey_occurred_flag (VOID); >+S032 NV_API_CALL nv_int10h_call (nv_state_t *, U032 *, U032 *, U032 *, U032 *, VOID *); >+ >+S032 NV_API_CALL nv_alloc_pages (nv_state_t *, VOID **, U032, U032, U032, U032, U032, U032, VOID **); >+S032 NV_API_CALL nv_free_pages (nv_state_t *, VOID **, U032, U032, VOID **); >+ >+S032 NV_API_CALL nv_agp_init (nv_state_t *, VOID **, VOID **, VOID *, U032); >+S032 NV_API_CALL nv_agp_teardown (nv_state_t *); >+S032 NV_API_CALL nv_agp_translate_address (nv_state_t *, VOID *, U032, U032 *); > >-S032 nv_start_rc_timer (nv_state_t *); >-S032 nv_stop_rc_timer (nv_state_t *); >+S032 NV_API_CALL nv_start_rc_timer (nv_state_t *); >+S032 NV_API_CALL nv_stop_rc_timer (nv_state_t *); > > > /* >@@ -288,41 +288,41 @@ > * --------------------------------------------------------------------------- > */ > >-BOOL rm_init_rm (VOID); >-BOOL rm_shutdown_rm (VOID); >-BOOL rm_init_adapter (nv_state_t *); >-BOOL rm_disable_adapter (nv_state_t *); >-BOOL rm_shutdown_adapter (nv_state_t *); >-BOOL rm_ioctl (nv_state_t *, VOID *, U032, VOID *); >-BOOL rm_isr (U032, U032 *); >-VOID rm_isr_bh (VOID *); >-RM_STATUS rm_power_management (nv_state_t *, U032, U032); >-U032 rm_get_vbios_version (nv_state_t *, U032 *, U032 *, U032 *, U032 *, U032 *); >-VOID rm_free_unused_clients (nv_state_t *, U032, VOID *); >- >-VOID rm_update_agp_config (nv_state_t *); >-RM_STATUS rm_init_agp (nv_state_t *); >-RM_STATUS rm_teardown_agp (nv_state_t *); >- >-RM_STATUS rm_alloc_agp_pages (nv_state_t *, VOID **, U032, U032, VOID **, U032 *); >-RM_STATUS rm_map_agp_pages (nv_state_t *, VOID **, U032, VOID *); >-RM_STATUS rm_free_agp_pages (nv_state_t *, VOID **, VOID *); >- >-RM_STATUS rm_alloc_agp_bitmap (nv_state_t *, U032, U032 *); >-RM_STATUS rm_free_agp_bitmap (nv_state_t *, U032, U032); >-RM_STATUS rm_set_agp_bitmap (nv_state_t *, VOID *); >-RM_STATUS rm_clear_agp_bitmap (nv_state_t *, VOID **); >- >-RM_STATUS rm_load_registry (nv_parm_t *); >-RM_STATUS rm_read_registry_dword (nv_state_t *, U008 *, U008 *, U032 *); >-RM_STATUS rm_write_registry_dword (nv_state_t *, U008 *, U008 *, U032); >-RM_STATUS rm_read_registry_binary (nv_state_t *, U008 *, U008 *, U008 *, U032 *); >-RM_STATUS rm_write_registry_binary (nv_state_t *, U008 *, U008 *, U008 *, U032); >+BOOL NV_API_CALL rm_init_rm (VOID); >+BOOL NV_API_CALL rm_shutdown_rm (VOID); >+BOOL NV_API_CALL rm_init_adapter (nv_state_t *); >+BOOL NV_API_CALL rm_disable_adapter (nv_state_t *); >+BOOL NV_API_CALL rm_shutdown_adapter (nv_state_t *); >+BOOL NV_API_CALL rm_ioctl (nv_state_t *, VOID *, U032, VOID *); >+BOOL NV_API_CALL rm_isr (U032, U032 *); >+VOID NV_API_CALL rm_isr_bh (VOID *); >+RM_STATUS NV_API_CALL rm_power_management (nv_state_t *, U032, U032); >+U032 NV_API_CALL rm_get_vbios_version (nv_state_t *, U032 *, U032 *, U032 *, U032 *, U032 *); >+VOID NV_API_CALL rm_free_unused_clients (nv_state_t *, U032, VOID *); >+ >+VOID NV_API_CALL rm_update_agp_config (nv_state_t *); >+RM_STATUS NV_API_CALL rm_init_agp (nv_state_t *); >+RM_STATUS NV_API_CALL rm_teardown_agp (nv_state_t *); >+ >+RM_STATUS NV_API_CALL rm_alloc_agp_pages (nv_state_t *, VOID **, U032, U032, VOID **, U032 *); >+RM_STATUS NV_API_CALL rm_map_agp_pages (nv_state_t *, VOID **, U032, VOID *); >+RM_STATUS NV_API_CALL rm_free_agp_pages (nv_state_t *, VOID **, VOID *); >+ >+RM_STATUS NV_API_CALL rm_alloc_agp_bitmap (nv_state_t *, U032, U032 *); >+RM_STATUS NV_API_CALL rm_free_agp_bitmap (nv_state_t *, U032, U032); >+RM_STATUS NV_API_CALL rm_set_agp_bitmap (nv_state_t *, VOID *); >+RM_STATUS NV_API_CALL rm_clear_agp_bitmap (nv_state_t *, VOID **); >+ >+RM_STATUS NV_API_CALL rm_load_registry (nv_parm_t *); >+RM_STATUS NV_API_CALL rm_read_registry_dword (nv_state_t *, U008 *, U008 *, U032 *); >+RM_STATUS NV_API_CALL rm_write_registry_dword (nv_state_t *, U008 *, U008 *, U032); >+RM_STATUS NV_API_CALL rm_read_registry_binary (nv_state_t *, U008 *, U008 *, U008 *, U032 *); >+RM_STATUS NV_API_CALL rm_write_registry_binary (nv_state_t *, U008 *, U008 *, U008 *, U032); > >-RM_STATUS rm_run_rc_callback (nv_state_t *); >-RM_STATUS rm_get_device_name (U032, U032, U008*); >+RM_STATUS NV_API_CALL rm_run_rc_callback (nv_state_t *); >+RM_STATUS NV_API_CALL rm_get_device_name (U032, U032, U008*); > >-NvUI64 nv_rdtsc (VOID); >+NvUI64 NV_API_CALL nv_rdtsc (VOID); > > #endif /* NVWATCH */ > >diff -ruN NVIDIA_kernel-1.0-4363/os-agp.c NVIDIA_kernel-1.0-4363-2.6/os-agp.c >--- NVIDIA_kernel-1.0-4363/os-agp.c 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/os-agp.c 2003-06-13 23:02:14.000000000 +0200 >@@ -45,44 +45,9 @@ > int ready; > } agp_gart; > >-typedef struct { >- int (*backend_acquire)(void); >- void (*backend_release)(void); >- void (*copy_info)(agp_kern_info *); >- agp_memory * (*allocate_memory)(size_t, unsigned int); >- void (*free_memory)(agp_memory *); >- int (*bind_memory)(agp_memory *, off_t); >- int (*unbind_memory)(agp_memory *); >- void (*enable)(unsigned int); >-} agp_operations_struct; >- >-agp_operations_struct agp_ops; > agp_kern_info agpinfo; > agp_gart gart; >-#if !defined (KERNEL_2_2) > const drm_agp_t *drm_agp_p; >-#endif >- >-#if defined (KERNEL_2_2) >- #define GET_AGPGART_SYMBOL(sym, sym_string) \ >- sym = (void*) GET_MODULE_SYMBOL(0, sym_string); \ >- if (sym == NULL) \ >- { \ >- nv_printf(NV_DBG_ERRORS, \ >- "NVRM: AGPGART: unable to retrieve symbol %s\n", \ >- sym_string); \ >- return 1; \ >- } >- >- #define AGP_BACKEND_ACQUIRE_SYM __MODULE_STRING(agp_backend_acquire) >- #define AGP_BACKEND_RELEASE_SYM __MODULE_STRING(agp_backend_release) >- #define AGP_COPY_INFO_SYM __MODULE_STRING(agp_copy_info) >- #define AGP_ALLOCATE_MEMORY_SYM __MODULE_STRING(agp_allocate_memory) >- #define AGP_FREE_MEMORY_SYM __MODULE_STRING(agp_free_memory) >- #define AGP_BIND_MEMORY_SYM __MODULE_STRING(agp_bind_memory) >- #define AGP_UNBIND_MEMORY_SYM __MODULE_STRING(agp_unbind_memory) >- #define AGP_ENABLE_SYM __MODULE_STRING(agp_enable) >-#endif > > #if defined(CONFIG_MTRR) > #define MTRR_DEL(gart) if ((gart).mtrr > 0) mtrr_del((gart).mtrr, 0, 0); >@@ -105,13 +70,17 @@ > U032 agp_rate; > U032 agp_sba; > U032 agp_fw; >- char* chipset; > VOID *bitmap; > U032 bitmap_size; > > memset( (void *) &gart, 0, sizeof(agp_gart)); > >-#if !defined (KERNEL_2_2) >+ /* >+ * XXX The inter_module_* mechanism has been deprecated and replaced with >+ * a different mechanism in Linux 2.5; it will go away eventually. >+ * Also, the Linux 2.5 AGP GART driver is modularized, agpgart.o does not >+ * include backend drivers. >+ */ > if (!(drm_agp_p = inter_module_get_request("drm_agp", "agpgart"))) > { > nv_printf(NV_DBG_ERRORS, >@@ -119,48 +88,13 @@ > return 1; > } > >- agp_ops.backend_acquire = drm_agp_p->acquire; >- agp_ops.backend_release = drm_agp_p->release; >- agp_ops.allocate_memory = drm_agp_p->allocate_memory; >- agp_ops.free_memory = drm_agp_p->free_memory; >- agp_ops.bind_memory = drm_agp_p->bind_memory; >- agp_ops.unbind_memory = drm_agp_p->unbind_memory; >- agp_ops.enable = drm_agp_p->enable; >- >- // looks like some newer kernels (for example mandrake 9.0's 2.4.19-16mdk) >- // have updated copy_info to return an integer value, and of course didn't >- // bother bumping the agpgart revision up. The return value is pretty >- // harmless (backend_acquire would have already failed and caused us to >- // bail), so cast the function pointer to avoid compiler warnings. >- // we may need to revisit this in the future. >- agp_ops.copy_info = (void (*)(agp_kern_info *)) drm_agp_p->copy_info; >- >-#else >-#if defined(CONFIG_KMOD) >- if ( request_module("agpgart") ) >- { >- nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: not loading agpgart.o\n"); >- return 1; >- } >-#endif >- >- GET_AGPGART_SYMBOL(agp_ops.backend_acquire, AGP_BACKEND_ACQUIRE_SYM); >- GET_AGPGART_SYMBOL(agp_ops.backend_release, AGP_BACKEND_RELEASE_SYM); >- GET_AGPGART_SYMBOL(agp_ops.copy_info, AGP_COPY_INFO_SYM); >- GET_AGPGART_SYMBOL(agp_ops.allocate_memory, AGP_ALLOCATE_MEMORY_SYM); >- GET_AGPGART_SYMBOL(agp_ops.free_memory, AGP_FREE_MEMORY_SYM); >- GET_AGPGART_SYMBOL(agp_ops.bind_memory, AGP_BIND_MEMORY_SYM); >- GET_AGPGART_SYMBOL(agp_ops.unbind_memory, AGP_UNBIND_MEMORY_SYM); >- GET_AGPGART_SYMBOL(agp_ops.enable, AGP_ENABLE_SYM); >-#endif >- > /* NOTE: from here down, return an error code of '-1' > * that indicates that agpgart is loaded, but we failed to use it > * in some way. This is so we don't try to use nvagp and lock up > * the memory controller. > */ > >- if ( (*(agp_ops.backend_acquire))() ) >+ if (drm_agp_p->acquire()) > { > nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: backend in use\n"); > return -1; >@@ -178,44 +112,21 @@ > agp_fw = 1; > agp_fw &= 0x00000001; > >- (*(agp_ops.copy_info))(&agpinfo); >- >- switch ( agpinfo.chipset ) >- { >- case INTEL_GENERIC: chipset = "Intel"; break; >- case INTEL_LX: chipset = "Intel 440LX"; break; >- case INTEL_BX: chipset = "Intel 440BX"; break; >- case INTEL_GX: chipset = "Intel 440GX"; break; >- case INTEL_I810: chipset = "Intel i810"; break; >- case INTEL_I840: chipset = "Intel i840"; break; >-#if !defined (KERNEL_2_2) >- case INTEL_I815: chipset = "Intel i815"; break; >-#if !defined(__rh_config_h__) >- case INTEL_I850: chipset = "Intel i850"; break; >-#endif >-#endif >-#if defined(NVCPU_IA64) >- case INTEL_460GX: chipset = "Intel 460GX"; break; >-#endif >- case VIA_GENERIC: chipset = "VIA"; break; >- case VIA_VP3: chipset = "VIA VP3"; break; >- case VIA_MVP3: chipset = "VIA MVP3"; break; >- case VIA_MVP4: chipset = "VIA MVP4"; break; >-#if !defined (KERNEL_2_2) >- case VIA_APOLLO_KX133: chipset = "VIA Apollo KX133"; break; >- case VIA_APOLLO_KT133: chipset = "VIA Apollo KT133"; break; >-#endif >- case VIA_APOLLO_PRO: chipset = "VIA Apollo Pro"; break; >- case SIS_GENERIC: chipset = "SiS"; break; >- case AMD_GENERIC: chipset = "AMD"; break; >- case AMD_IRONGATE: chipset = "AMD Irongate"; break; >- case ALI_M1541: chipset = "ALi M1541"; break; >- case ALI_GENERIC: chipset = "ALi"; break; >- case NOT_SUPPORTED: chipset = "unsupported"; break; >- default: chipset = "unknown"; >+#if defined(KERNEL_2_4) >+ /* >+ * The original Linux 2.4 AGP GART driver interface declared copy_info to >+ * return nothing. This changed in Linux 2.5, which reports unsupported >+ * chipsets via this function. If this Linux 2.4 kernels behaves the same >+ * way, we have no way to know. >+ */ >+ drm_agp_p->copy_info(&agpinfo); >+#else >+ if (drm_agp_p->copy_info(&agpinfo)) { >+ drm_agp_p->release(); >+ inter_module_put("drm_agp"); >+ return -1; > } >- >- nv_printf(NV_DBG_SETUP, "NVRM: AGPGART: %s chipset\n", chipset); >+#endif > > #ifdef CONFIG_MTRR > if ((gart.mtrr = mtrr_add(agpinfo.aper_base, >@@ -229,7 +140,8 @@ > */ > nv_printf(NV_DBG_ERRORS, > "NVRM: AGPGART: unable to set MTRR write-combining\n"); >- (*(agp_ops.backend_release))(); >+ drm_agp_p->release(); >+ inter_module_put("drm_agp"); > return -1; > } > #endif >@@ -244,7 +156,8 @@ > { > nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: unable to remap aperture\n"); > MTRR_DEL(gart); >- (*(agp_ops.backend_release))(); >+ drm_agp_p->release(); >+ inter_module_put("drm_agp"); > return -1; > } > >@@ -255,7 +168,8 @@ > nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: unable to allocate bitmap\n"); > iounmap(gart.aperture); > MTRR_DEL(gart); >- (*(agp_ops.backend_release))(); >+ drm_agp_p->release(); >+ inter_module_put("drm_agp"); > return -1; > } > >@@ -266,39 +180,19 @@ > os_free_mem(bitmap); > iounmap(gart.aperture); > MTRR_DEL(gart); >- (*(agp_ops.backend_release))(); >+ drm_agp_p->release(); >+ inter_module_put("drm_agp"); > return -1; > } > >- nv_printf(NV_DBG_SETUP, >- "NVRM: AGPGART: aperture: %ldM @ 0x%08lx\n", >- (unsigned long)agpinfo.aper_size, >- (unsigned long)agpinfo.aper_base); >- >- nv_printf(NV_DBG_SETUP, >- "NVRM: AGPGART: aperture mapped from 0x%08lx to 0x%08lx\n", >- agpinfo.aper_base, >- (unsigned long) gart.aperture); >- > if (!agp_sba) agpinfo.mode &= ~0x00000200; > if (!agp_fw) agpinfo.mode &= ~0x00000010; > > if (!(agp_rate & 0x00000004)) agpinfo.mode &= ~0x00000004; > if (!(agp_rate & 0x00000002)) agpinfo.mode &= ~0x00000002; > >- (*(agp_ops.enable))(agpinfo.mode); >+ drm_agp_p->enable(agpinfo.mode); > >- if (agpinfo.mode & 0x00000200) >- nv_printf(NV_DBG_SETUP, "NVRM: AGPGART: backend supports sba\n"); >- if (agpinfo.mode & 0x00000010) >- nv_printf(NV_DBG_SETUP, "NVRM: AGPGART: backend supports fw\n"); >- if (agpinfo.mode & 0x00000004) >- nv_printf(NV_DBG_SETUP, "NVRM: AGPGART: mode 4x\n"); >- else if (agpinfo.mode & 0x00000002) >- nv_printf(NV_DBG_SETUP, "NVRM: AGPGART: mode 2x\n"); >- else if (agpinfo.mode & 0x00000001) >- nv_printf(NV_DBG_SETUP, "NVRM: AGPGART: mode 1x\n"); >- > *ap_phys_base = (void*) agpinfo.aper_base; > *ap_mapped_base = (void*) gart.aperture; > *apsize = (agpinfo.aper_size * 0x100000) - 1; >@@ -332,11 +226,13 @@ > iounmap(gart.aperture); > } > >- (*(agp_ops.backend_release))(); >+ drm_agp_p->release(); > >-#if !defined (KERNEL_2_2) >+ /* >+ * XXX Same as above; the inter_module_* mechanism will go away at some >+ * point, it has been deprecated in Linux 2.5. >+ */ > inter_module_put("drm_agp"); >-#endif > > if (rm_clear_agp_bitmap(nv, &bitmap)) > { >@@ -383,7 +279,7 @@ > return RM_ERROR; > } > >- ptr = (*agp_ops.allocate_memory)(PageCount, AGP_NORMAL_MEMORY); >+ ptr = drm_agp_p->allocate_memory(PageCount, AGP_NORMAL_MEMORY); > if (ptr == NULL) > { > *pAddress = (void*) 0; >@@ -391,7 +287,7 @@ > return RM_ERR_NO_FREE_MEM; > } > >- err = (*(agp_ops.bind_memory))(ptr, *Offset); >+ err = drm_agp_p->bind_memory(ptr, *Offset); > if (err) > { > // this happens a lot when the aperture itself fills up.. >@@ -408,7 +304,7 @@ > if (status != RM_OK) > { > nv_printf(NV_DBG_ERRORS, "NVRM: memory allocation failed\n"); >- (*(agp_ops.unbind_memory))(ptr); >+ drm_agp_p->unbind_memory(ptr); > goto fail; > } > >@@ -423,7 +319,7 @@ > return RM_OK; > > fail: >- (*(agp_ops.free_memory))(ptr); >+ drm_agp_p->free_memory(ptr); > *pAddress = (void*) 0; > > return RM_ERROR; >@@ -462,7 +358,7 @@ > { > nv_printf(NV_DBG_ERRORS, "NVRM: AGPGART: unable to remap %lu pages\n", > (unsigned long)agp_data->num_pages); >- (*(agp_ops.unbind_memory))(agp_data->ptr); >+ drm_agp_p->unbind_memory(agp_data->ptr); > goto fail; > } > >@@ -480,9 +376,6 @@ > #endif /* AGPGART */ > } > >- >-#if !defined(KERNEL_2_2) >- > RM_STATUS > KernMapAGPNopage( > VOID *address, >@@ -529,9 +422,6 @@ > #endif > } > >-#endif /* !defined(KERNEL_2_2) */ >- >- > RM_STATUS KernFreeAGPPages( > nv_state_t *nv, > VOID **pAddress, >@@ -557,8 +447,8 @@ > { > size_t pages = ptr->page_count; > >- (*(agp_ops.unbind_memory))(ptr); >- (*(agp_ops.free_memory))(ptr); >+ drm_agp_p->unbind_memory(ptr); >+ drm_agp_p->free_memory(ptr); > > nv_printf(NV_DBG_INFO, "NVRM: AGPGART: freed %ld pages\n", > (unsigned long)pages); >diff -ruN NVIDIA_kernel-1.0-4363/os-interface.c NVIDIA_kernel-1.0-4363-2.6/os-interface.c >--- NVIDIA_kernel-1.0-4363/os-interface.c 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/os-interface.c 2003-06-02 00:10:13.000000000 +0200 >@@ -56,7 +56,7 @@ > PHWINFO pDev > ) > { >- return suser(); >+ return NV_IS_SUSER(); > } > > U032 os_get_page_size(VOID) >@@ -209,6 +209,11 @@ > U032 size > ) > { >+ /* >+ * XXX This needs to be !NV_MAY_SLEEP() rather than in_interrupt(); that >+ * requires quite a bit of locking to be rearranged, however, which is why >+ * I'll leave it alone for now. >+ */ > if (in_interrupt()) { > if (size <= KMALLOC_LIMIT) { > /* >@@ -286,7 +291,7 @@ > *address = (void *) va; > > for (i = 0; i < count; i++) { >- mem_map_reserve(GET_MAP_NR(__pa((va)))); >+ SetPageReserved(virt_to_page(va)); > va += PAGE_SIZE; > } > >@@ -310,7 +315,7 @@ > unsigned long va = (unsigned long) address; > > for (i = 0; i < count; i++) { >- mem_map_unreserve(GET_MAP_NR(__pa((va)))); >+ ClearPageReserved(virt_to_page(va)); > va += PAGE_SIZE; > } > >@@ -416,7 +421,7 @@ > if (in_irq() && MilliSeconds > NV_MAX_ISR_MDELAY) > return RM_ERROR; > >- if (in_interrupt()) >+ if (!NV_MAY_SLEEP()) > { > mdelay(MilliSeconds); > return RM_OK; >@@ -514,8 +519,8 @@ > // The current debug display level (default to maximum debug level) > int cur_debuglevel = 0xaaaaaaaa; > >-MODULE_PARM(silence_nvidia_output, "1i"); > static int silence_nvidia_output = 0; >+NV_MODULE_PARAMETER(silence_nvidia_output); > > > // >@@ -698,14 +703,14 @@ > > ULONG os_cli(ULONG flags) > { >- save_flags(flags); >- cli(); >+ NV_SAVE_FLAGS(flags); >+ NV_CLI(); > return flags; > } > > ULONG os_sti(ULONG flags) > { >- restore_flags(flags); >+ NV_RESTORE_FLAGS(flags); > return flags; > } > >@@ -838,16 +843,6 @@ > vaddr = ioremap_nocache(start, size_bytes); > } > >-#if defined (KERNEL_2_2) >- if ((vaddr == NULL)) // && (mode == NV_MEMORY_DEFAULT)) >- { >- unsigned long map_nr = MAP_NR(__va(start)); >- if (map_nr < max_mapnr) { >- vaddr = __va(start); >- } >- } >-#endif >- > #ifdef DEBUG > if (mode == NV_MEMORY_WRITECOMBINED) { > nv_printf(NV_DBG_ERRORS, >@@ -868,16 +863,7 @@ > U032 size_bytes > ) > { >-#if defined (KERNEL_2_2) >- if (MAP_NR(addr) < max_mapnr) { >- // if we didn't want the memory cached, this isn't necessary >- // but we shouldn't be in a timing critical piece of code. >- asm volatile("wbinvd":::"memory"); >- } else >-#endif >- { >- iounmap(addr); >- } >+ iounmap(addr); > } > > VOID* os_map_user_space( >@@ -981,7 +967,7 @@ > > U032 os_get_cpu_count() > { >- return smp_num_cpus; >+ return NV_SMP_NUM_CPUS; > } > > >@@ -1044,12 +1030,15 @@ > if (sgi_funcs.add_barrier == NULL) > { > #if defined(TESTING_SWAP) >-#if !defined (KERNEL_2_2) >+ /* >+ * XXX The inter_module_* mechanism has been deprecated in Linux 2.5, a >+ * new mechanism is in place; this code will need to be updated at some >+ * point. >+ */ > inter_module_register(ADD_BARRIER_FUNC, THIS_MODULE, sgitest_add_barrier); > inter_module_register(REMOVE_BARRIER_FUNC, THIS_MODULE, sgitest_remove_barrier); > inter_module_register(SWAP_READY_FUNC, THIS_MODULE, sgitest_swap_ready); > #endif >-#endif > sgi_funcs.add_barrier = GET_MODULE_SYMBOL(0, ADD_BARRIER_FUNC); > sgi_funcs.remove_barrier = GET_MODULE_SYMBOL(0, REMOVE_BARRIER_FUNC); > sgi_funcs.swap_ready = GET_MODULE_SYMBOL(0, SWAP_READY_FUNC); >diff -ruN NVIDIA_kernel-1.0-4363/os-interface.h NVIDIA_kernel-1.0-4363-2.6/os-interface.h >--- NVIDIA_kernel-1.0-4363/os-interface.h 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/os-interface.h 2003-12-31 15:30:57.000000000 +0100 >@@ -67,68 +67,68 @@ > * --------------------------------------------------------------------------- > */ > >-U032 os_get_page_size (VOID); >-ULONG os_get_page_mask (VOID); >-RM_STATUS os_alloc_mem (VOID **, U032); >-VOID os_free_mem (VOID *); >-RM_STATUS os_alloc_contig_pages (VOID **, U032); >-VOID os_free_contig_pages (VOID *, U032); >-RM_STATUS os_get_current_time (U032 *, U032 *); >-RM_STATUS os_delay (U032); >-RM_STATUS os_delay_us (U032); >-U032 os_get_cpu_frequency (VOID); >-RM_STATUS os_get_current_process (U032 *); >-RM_STATUS os_kill_process (U032, U032); >-U008* os_string_copy (U008 *, const U008 *); >-RM_STATUS os_strncpy_from_user (U008 *, const U008 *, U032); >-S032 os_string_compare (const U008 *, const U008 *); >-U032 os_string_length (const U008 *); >-U008* os_mem_copy (U008 *, const U008 *, U032); >-RM_STATUS os_memcpy_from_user (VOID *, const VOID *, U032); >-RM_STATUS os_memcpy_to_user (void *, const VOID *, U032); >-VOID* os_mem_set (VOID *, U008, U032); >-S032 os_mem_cmp (const U008 *, const U008 *, U032); >-VOID* os_pci_init_handle (U008, U008, U008, U016 *, U016 *); >-U008 os_pci_read_byte (VOID *, U008); >-U016 os_pci_read_word (VOID *, U008); >-U032 os_pci_read_dword (VOID *, U008); >-VOID os_pci_write_byte (VOID *, U008, U008); >-VOID os_pci_write_word (VOID *, U008, U016); >-VOID os_pci_write_dword (VOID *, U008, U032); >-VOID* os_map_kernel_space (U032, U032, U032); >-VOID os_unmap_kernel_space (VOID *, U032); >-VOID* os_map_user_space (VOID *, VOID **, U032, U032); >-VOID os_unmap_user_space (VOID *, VOID *); >-VOID* os_map_io_space (U032, U032, VOID **, U032, U032); >-VOID os_unmap_io_space (VOID *, U032, VOID *, U032); >-RM_STATUS os_flush_cpu_cache (VOID); >-RM_STATUS os_set_mem_range (U032, U032, U032); >-RM_STATUS os_unset_mem_range (U032, U032); >-BOOL os_pci_device_present (U016, U016); >-U008 os_io_read_byte (PHWINFO, U032); >-U016 os_io_read_word (PHWINFO, U032); >-U032 os_io_read_dword (PHWINFO, U032); >-VOID os_io_write_byte (PHWINFO, U032, U008); >-VOID os_io_write_word (PHWINFO, U032, U016); >-VOID os_io_write_dword (PHWINFO, U032, U032); >-ULONG os_cli (ULONG); >-ULONG os_sti (ULONG); >-VOID* os_copy_in_ioctl_param (VOID *, VOID *, U032); >-VOID* os_copy_out_ioctl_param (VOID *, VOID *, U032); >-VOID* os_registry_lookup (PHWINFO, U008 *, U008 *); >-RM_STATUS os_swap_barrier (U032, VOID (*)(VOID *), VOID *); >-RM_STATUS os_init_swap_barrier (U032); >-RM_STATUS os_remove_swap_barrier (U032); >-BOOL os_is_administrator (PHWINFO); >-VOID os_dbg_init (VOID); >-VOID os_dbg_breakpoint (VOID); >-VOID os_dbg_set_level (U032); >-U032 os_get_cpu_count (VOID); >-RM_STATUS os_raise_smp_barrier (VOID); >-RM_STATUS os_clear_smp_barrier (VOID); >+U032 NV_API_CALL os_get_page_size (VOID); >+ULONG NV_API_CALL os_get_page_mask (VOID); >+RM_STATUS NV_API_CALL os_alloc_mem (VOID **, U032); >+VOID NV_API_CALL os_free_mem (VOID *); >+RM_STATUS NV_API_CALL os_alloc_contig_pages (VOID **, U032); >+VOID NV_API_CALL os_free_contig_pages (VOID *, U032); >+RM_STATUS NV_API_CALL os_get_current_time (U032 *, U032 *); >+RM_STATUS NV_API_CALL os_delay (U032); >+RM_STATUS NV_API_CALL os_delay_us (U032); >+U032 NV_API_CALL os_get_cpu_frequency (VOID); >+RM_STATUS NV_API_CALL os_get_current_process (U032 *); >+RM_STATUS NV_API_CALL os_kill_process (U032, U032); >+U008* NV_API_CALL os_string_copy (U008 *, const U008 *); >+RM_STATUS NV_API_CALL os_strncpy_from_user (U008 *, const U008 *, U032); >+S032 NV_API_CALL os_string_compare (const U008 *, const U008 *); >+U032 NV_API_CALL os_string_length (const U008 *); >+U008* NV_API_CALL os_mem_copy (U008 *, const U008 *, U032); >+RM_STATUS NV_API_CALL os_memcpy_from_user (VOID *, const VOID *, U032); >+RM_STATUS NV_API_CALL os_memcpy_to_user (void *, const VOID *, U032); >+VOID* NV_API_CALL os_mem_set (VOID *, U008, U032); >+S032 NV_API_CALL os_mem_cmp (const U008 *, const U008 *, U032); >+VOID* NV_API_CALL os_pci_init_handle (U008, U008, U008, U016 *, U016 *); >+U008 NV_API_CALL os_pci_read_byte (VOID *, U008); >+U016 NV_API_CALL os_pci_read_word (VOID *, U008); >+U032 NV_API_CALL os_pci_read_dword (VOID *, U008); >+VOID NV_API_CALL os_pci_write_byte (VOID *, U008, U008); >+VOID NV_API_CALL os_pci_write_word (VOID *, U008, U016); >+VOID NV_API_CALL os_pci_write_dword (VOID *, U008, U032); >+VOID* NV_API_CALL os_map_kernel_space (U032, U032, U032); >+VOID NV_API_CALL os_unmap_kernel_space (VOID *, U032); >+VOID* NV_API_CALL os_map_user_space (VOID *, VOID **, U032, U032); >+VOID NV_API_CALL os_unmap_user_space (VOID *, VOID *); >+VOID* NV_API_CALL os_map_io_space (U032, U032, VOID **, U032, U032); >+VOID NV_API_CALL os_unmap_io_space (VOID *, U032, VOID *, U032); >+RM_STATUS NV_API_CALL os_flush_cpu_cache (VOID); >+RM_STATUS NV_API_CALL os_set_mem_range (U032, U032, U032); >+RM_STATUS NV_API_CALL os_unset_mem_range (U032, U032); >+BOOL NV_API_CALL os_pci_device_present (U016, U016); >+U008 NV_API_CALL os_io_read_byte (PHWINFO, U032); >+U016 NV_API_CALL os_io_read_word (PHWINFO, U032); >+U032 NV_API_CALL os_io_read_dword (PHWINFO, U032); >+VOID NV_API_CALL os_io_write_byte (PHWINFO, U032, U008); >+VOID NV_API_CALL os_io_write_word (PHWINFO, U032, U016); >+VOID NV_API_CALL os_io_write_dword (PHWINFO, U032, U032); >+ULONG NV_API_CALL os_cli (ULONG); >+ULONG NV_API_CALL os_sti (ULONG); >+VOID* NV_API_CALL os_copy_in_ioctl_param (VOID *, VOID *, U032); >+VOID* NV_API_CALL os_copy_out_ioctl_param (VOID *, VOID *, U032); >+VOID* NV_API_CALL os_registry_lookup (PHWINFO, U008 *, U008 *); >+RM_STATUS NV_API_CALL os_swap_barrier (U032, VOID (*)(VOID *), VOID *); >+RM_STATUS NV_API_CALL os_init_swap_barrier (U032); >+RM_STATUS NV_API_CALL os_remove_swap_barrier (U032); >+BOOL NV_API_CALL os_is_administrator (PHWINFO); >+VOID NV_API_CALL os_dbg_init (VOID); >+VOID NV_API_CALL os_dbg_breakpoint (VOID); >+VOID NV_API_CALL os_dbg_set_level (U032); >+U032 NV_API_CALL os_get_cpu_count (VOID); >+RM_STATUS NV_API_CALL os_raise_smp_barrier (VOID); >+RM_STATUS NV_API_CALL os_clear_smp_barrier (VOID); > >-RM_STATUS osRaiseSmpBarrier (VOID); >-RM_STATUS osClearSmpBarrier (VOID); >+RM_STATUS NV_API_CALL osRaiseSmpBarrier (VOID); >+RM_STATUS NV_API_CALL osClearSmpBarrier (VOID); > > /* > * --------------------------------------------------------------------------- >@@ -162,8 +162,8 @@ > #define NV_DBG_ERRORS 0x4 > > >-inline void out_string(const char *str); >-int nv_printf(int debuglevel, const char *printf_format, ...); >+void NV_API_CALL out_string(const char *str); >+int NV_API_CALL nv_printf(int debuglevel, const char *printf_format, ...); > > > #define NV_MEMORY_TYPE_SYSTEM 0 >diff -ruN NVIDIA_kernel-1.0-4363/os-registry.c NVIDIA_kernel-1.0-4363-2.6/os-registry.c >--- NVIDIA_kernel-1.0-4363/os-registry.c 2003-04-20 03:57:19.000000000 +0200 >+++ NVIDIA_kernel-1.0-4363-2.6/os-registry.c 2003-06-02 00:10:13.000000000 +0200 >@@ -48,24 +48,6 @@ > * This could be changed to work on a per-device basis. > */ > >-/* >- * The 2nd argument to MODULE_PARM is used to verify parameters passed >- * to the module at load time. It should be a string in the following >- * format: >- * >- * [min[-max]]{b,h,i,l,s} >- * >- * The MIN and MAX specifiers delimit the length of the array. If MAX >- * is omitted, it defaults to MIN; if both are omitted, the default is >- * 1. The final character is a type specifier. >- * >- * b byte >- * h short >- * i int >- * l long >- * s string >- */ >- > /* > * Option: VideoMemoryTypeOverride > * >@@ -92,7 +74,7 @@ > */ > > static int NVreg_VideoMemoryTypeOverride = 1; >-MODULE_PARM(NVreg_VideoMemoryTypeOverride, "i"); >+NV_MODULE_PARAMETER(NVreg_VideoMemoryTypeOverride); > > /* > * Option: EnableVia4x >@@ -111,7 +93,7 @@ > */ > > static int NVreg_EnableVia4x = 0; >-MODULE_PARM(NVreg_EnableVia4x, "i"); >+NV_MODULE_PARAMETER(NVreg_EnableVia4x); > > /* > * Option: EnableALiAGP >@@ -134,7 +116,7 @@ > */ > > static int NVreg_EnableALiAGP = 0; >-MODULE_PARM(NVreg_EnableALiAGP, "i"); >+NV_MODULE_PARAMETER(NVreg_EnableALiAGP); > > /* > * Option: ReqAGPRate >@@ -164,7 +146,7 @@ > */ > > static int NVreg_ReqAGPRate = 7; >-MODULE_PARM(NVreg_ReqAGPRate, "i"); >+NV_MODULE_PARAMETER(NVreg_ReqAGPRate); > > /* > * Option: UpdateKernelAGP >@@ -193,7 +175,7 @@ > */ > > static int NVreg_UpdateKernelAGP = 1; >-MODULE_PARM(NVreg_UpdateKernelAGP, "i"); >+NV_MODULE_PARAMETER(NVreg_UpdateKernelAGP); > > /* > * Option: EnableAGPSBA >@@ -226,7 +208,7 @@ > static int NVreg_EnableAGPSBA = 0; > #endif > >-MODULE_PARM(NVreg_EnableAGPSBA, "i"); >+NV_MODULE_PARAMETER(NVreg_EnableAGPSBA); > > /* > * Option: EnableAGPFW >@@ -250,7 +232,7 @@ > */ > > static int NVreg_EnableAGPFW = 0; >-MODULE_PARM(NVreg_EnableAGPFW, "i"); >+NV_MODULE_PARAMETER(NVreg_EnableAGPFW); > > /* > * Option: SoftEDIDs >@@ -269,7 +251,7 @@ > */ > > static int NVreg_SoftEDIDs = 1; >-MODULE_PARM(NVreg_SoftEDIDs, "i"); >+NV_MODULE_PARAMETER(NVreg_SoftEDIDs); > > /* > * Option: Mobile >@@ -293,14 +275,14 @@ > */ > > static int NVreg_Mobile = ~0; >-MODULE_PARM(NVreg_Mobile, "i"); >+NV_MODULE_PARAMETER(NVreg_Mobile); > > > static int NVreg_ResmanDebugLevel = ~0; >-MODULE_PARM(NVreg_ResmanDebugLevel, "i"); >+NV_MODULE_PARAMETER(NVreg_ResmanDebugLevel); > > static int NVreg_FlatPanelMode = 0; >-MODULE_PARM(NVreg_FlatPanelMode, "i"); >+NV_MODULE_PARAMETER(NVreg_FlatPanelMode); > > /* > * You can enable any of the registry options disabled by default by
<b>You cannot view the attachment while viewing its details because your browser does not support IFRAMEs. <a href="attachment.cgi?id=27001">View the attachment on a separate page</a>.</b>
View Attachment As Diff
View Attachment As Raw
Actions:
View
|
Diff
Attachments on
bug 29336
:
27000
| 27001
