Files
linux-nvgpu/userspace/Makefile
Thomas Fleury 5d6536b48f gpu: nvgpu: unit: libnvgpu-fifo for host makefiles
channel and tsg units used to build object files from
upper fifo directory. But this was causing build
dependencies issues and stale .o files could be used.
Fixed this by building a shared library for nvgpu-fifo
for host makefiles.

Jira NVGPU-3480

Change-Id: I32655c207f48baa5cde40846ae5a276c7c8fa6fd
Signed-off-by: Thomas Fleury <tfleury@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2150370
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vedashree Vidwans <vvidwans@nvidia.com>
Tested-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-by: Adeel Raza <araza@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
2019-07-12 11:25:10 -07:00

163 lines
5.2 KiB
Makefile

# Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved.
#
# Permission is hereby granted, free of charge, to any person obtaining a
# copy of this software and associated documentation files (the "Software"),
# to deal in the Software without restriction, including without limitation
# the rights to use, copy, modify, merge, publish, distribute, sublicense,
# and/or sell copies of the Software, and to permit persons to whom the
# Software is furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in
# all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
# DEALINGS IN THE SOFTWARE.
# TODO:
# - Separate rule for nvgpu_unit shared library
# - Proper header dependency checking.
# Turn off suffix rules. They are deprecated.
.SUFFIXES:
# Full paths. Makes submakefiles easier. That said this make file _must_ be run
# from <NVGPU>/userspace/.
TWD=$(CURDIR)
# Top level out dir.
OUT=$(TWD)/build
# Core unit test framework.
CORE_SRC=$(TWD)/src
CORE_OUT=$(OUT)/nvgpu_unit_core
# Nvgpu driver code.
NVGPU_SRC=$(TWD)/../drivers/gpu/nvgpu
NVGPU_OUT=$(OUT)/libnvgpu
# Unit tests themselves.
UNIT_SRC=$(TWD)/units
UNIT_OUT=$(OUT)/units
INCLUDES= \
-I$(NVGPU_SRC) \
-I$(NVGPU_SRC)/include \
-I$(TWD)/../include \
-I$(TWD)/../include/uapi \
-I$(TWD)/include
# This is safety build by default.
NV_BUILD_CONFIGURATION_IS_SAFETY=1
# This Makefile is only for host POSIX builds
NVGPU_POSIX=1
# Enable fault injection for unit tests
NVGPU_FAULT_INJECTION_ENABLEMENT=1
# Linux configs. We want these so that we can mirror builds from the actual
# Linux kernel.
# include Makefile.configs
CONFIGS := -D__NVGPU_POSIX__ \
-DNVGPU_UNITTEST_FAULT_INJECTION_ENABLEMENT \
-D__NVGPU_UNIT_TEST__
# For target "libnvgpu", safety profile should be used
# safety-debug profile is used otherwise
ifeq ( ,$(findstring libnvgpu, '$(MAKECMDGOALS)'))
NVGPU_FORCE_DEBUG_PROFILE := 1
else
NVGPU_FORCE_DEBUG_PROFILE := 0
endif
NVGPU_FORCE_SAFETY_PROFILE := 1
include $(NVGPU_SRC)/Makefile.shared.configs
CONFIGS+=$(NVGPU_COMMON_CFLAGS)
# Compiler, c-flags, etc.
# CC = clang
CC = gcc
CFLAGS = -Wall -Wextra -ggdb -Werror -Wno-unused-parameter \
-Wno-missing-field-initializers -Wformat -Wchar-subscripts \
-Wparentheses -Wtrigraphs -Wpointer-arith -Wmissing-declarations \
-Wmissing-prototypes -Wredundant-decls -Wmain -Wreturn-type \
-Wmultichar -Wunused -Wmissing-braces -Wstrict-aliasing \
-Wsign-compare -Waddress -Wno-unused-local-typedefs -fPIC \
-Wno-maybe-uninitialized $(INCLUDES) $(CONFIGS)
LIB_PATHS = -L$(OUT) -L$(UNIT_OUT)
LIBS = -lpthread -pthread -lgcov -ldl
# Source files. We expect $(OBJS) and $(HEADERS) to get filled in here.
include Makefile.sources
all: $(OUT)/nvgpu_unit $(UNITS)
# Convenience targets.
.PHONY: libnvgpu core units
libnvgpu: $(OUT)/libnvgpu-drv.so
core: $(OUT)/nvgpu_unit
units: $(UNITS)
# Note the weird libnvgpu_unit.so file: this is a bit of a hack. It lets the
# unit tests link back against the nvgpu_unit executable so that they can call
# functions (like unit_info()) directly. This shared library isn't actually
# used for anything beyond that.
#
# Also it really should have its own rule...
$(OUT)/nvgpu_unit: $(OUT)/libnvgpu-drv.so $(CORE_OBJS)
$(CC) -shared -o $(OUT)/libnvgpu_unit.so \
$(CORE_OBJS) $(LIB_PATHS) $(LIBS)
$(CC) --coverage \
-o $(OUT)/nvgpu_unit $(CORE_OBJS) $(LIB_PATHS) $(LIBS)
$(OUT)/libnvgpu-drv.so: $(OBJS)
$(CC) -shared -Wl,--no-undefined -o $(OUT)/libnvgpu-drv.so $(OBJS) -lgcov -rdynamic -lpthread
# Default build target for all the nvgpu driver object files we want to build in
# userspace. These get bundled into libnvgpu-drv.so.
$(NVGPU_OUT)/%.o : $(NVGPU_SRC)/%.c $(HEADERS)
@if [ ! -d $(dir $@) ] ; then \
mkdir -p $(dir $@) ; \
fi
$(CC) --coverage $(CFLAGS) -c -o $@ $<
# Build target for unit test files. These are not part of the libnvgpu-drv.so.
# These comprise the unit test framework.
$(CORE_OUT)/%.o : $(CORE_SRC)/%.c $(CORE_HEADERS)
@if [ ! -d $(dir $@) ] ; then \
mkdir -p $(dir $@) ; \
fi
$(CC) --coverage $(CFLAGS) -c -o $@ $<
# Certain variables should be exported to the unit test module builds.
export TWD INCLUDES CONFIGS UNIT_SRC UNIT_OUT
export CC CFLAGS LIB_PATHS LIBS
.PHONY: $(UNITS)
$(UNITS): $(OUT)/libnvgpu-drv.so
@echo "Building unit module: $@"
@+$(MAKE) --no-print-directory -C $@
.PHONY: clean nvgpu_clean core_clean unit_clean
clean: nvgpu_clean core_clean unit_clean
rm -rf $(OUT)
nvgpu_clean:
rm -rf $(OUT)/libnvgpu*
core_clean:
rm -rf $(OUT)/nvgpu_unit*
unit_clean:
@for d in $(UNITS); do \
echo Cleaning $$d; \
$(MAKE) --no-print-directory -C $$d clean; \
done
rm -rf $(OUT)/units