mirror of
https://github.com/checkpoint-restore/criu.git
synced 2026-07-20 16:51:37 +00:00
55a5993bc7
To support Checkpoint Restore with AMDGPUs for ROCm workloads, introduce
a new plugin to assist CRIU with the help of AMD KFD kernel driver. This
initial commit just provides the basic framework to build up further
capabilities. Like CRIU, the amdgpu plugin also uses protobuf to
serialize
and save the amdkfd data which is mostly VRAM contents with some
metadata.
We generate a data file "amdgpu-kfd-<id>.img" during the dump stage. On restore
this file is read and extracted to re-create various types of buffer
objects that belonged to the previously checkpointed process. Upon
restore the mmap page offset within a device file might change so we use
the new hook to update and adjust the mmap offsets for newly created
target process. This is needed for sys_mmap call in pie restorer phase.
Support for queues and events is added in future patches of this series.
With the current implementation (amdgpu_plugin), we support:
- Only compute workloads such (Non Gfx) are supported
- GPU visible inside a container
- AMD GPU Gfx 9 Family
- Pytorch Benchmarks such as BERT Base
amdgpu plugin dependes on libdrm and libdrm_amdgpu which are typically
installed with libdrm-dev package. We build amdgpu_plugin only when the
dependencies are met on the target system and when user intends to
install the amdgpu plugin and not by default with criu build.
Suggested-by: Felix Kuehling <felix.kuehling@amd.com>
Co-authored-by: David Yat Sin <david.yatsin@amd.com>
Signed-off-by: David Yat Sin <david.yatsin@amd.com>
Signed-off-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
102 lines
3.5 KiB
Text
102 lines
3.5 KiB
Text
include $(__nmk_dir)utils.mk
|
|
include $(__nmk_dir)msg.mk
|
|
include scripts/feature-tests.mak
|
|
|
|
ifeq ($(call try-cc,$(FEATURE_TEST_LIBBSD_DEV),-lbsd),true)
|
|
LIBS_FEATURES += -lbsd
|
|
FEATURE_DEFINES += -DCONFIG_HAS_LIBBSD
|
|
else
|
|
$(info Note: Building without setproctitle() and strlcpy() support.)
|
|
$(info $(info) To enable these features, please install libbsd-devel (RPM) / libbsd-dev (DEB).)
|
|
endif
|
|
|
|
ifeq ($(call pkg-config-check,libselinux),y)
|
|
LIBS_FEATURES += -lselinux
|
|
FEATURE_DEFINES += -DCONFIG_HAS_SELINUX
|
|
endif
|
|
|
|
ifeq ($(call pkg-config-check,libbpf),y)
|
|
LIBS_FEATURES += -lbpf
|
|
FEATURE_DEFINES += -DCONFIG_HAS_LIBBPF
|
|
export CONFIG_HAS_LIBBPF := y
|
|
endif
|
|
|
|
ifeq ($(call pkg-config-check,libdrm),y)
|
|
export CONFIG_AMDGPU := y
|
|
$(info Note: Building criu with amdgpu_plugin.)
|
|
else
|
|
$(info Note: Building criu without amdgpu_plugin.)
|
|
$(info Note: libdrm and libdrm_amdgpu are required to build amdgpu_plugin.)
|
|
endif
|
|
|
|
ifeq ($(NO_GNUTLS)x$(call pkg-config-check,gnutls),xy)
|
|
LIBS_FEATURES += -lgnutls
|
|
export CONFIG_GNUTLS := y
|
|
FEATURE_DEFINES += -DCONFIG_GNUTLS
|
|
else
|
|
$(info Note: Building without GnuTLS support)
|
|
endif
|
|
|
|
ifeq ($(call pkg-config-check,libnftables),y)
|
|
LIB_NFTABLES := $(shell $(PKG_CONFIG) --libs libnftables)
|
|
ifeq ($(call try-cc,$(FEATURE_TEST_NFTABLES_LIB_API_0),$(LIB_NFTABLES)),true)
|
|
LIBS_FEATURES += $(LIB_NFTABLES)
|
|
FEATURE_DEFINES += -DCONFIG_HAS_NFTABLES_LIB_API_0
|
|
else ifeq ($(call try-cc,$(FEATURE_TEST_NFTABLES_LIB_API_1),$(LIB_NFTABLES)),true)
|
|
LIBS_FEATURES += $(LIB_NFTABLES)
|
|
FEATURE_DEFINES += -DCONFIG_HAS_NFTABLES_LIB_API_1
|
|
else
|
|
$(warning Warn: you have libnftables installed but it has incompatible API)
|
|
$(warning Warn: Building without nftables support)
|
|
endif
|
|
else
|
|
$(warning Warn: you have no libnftables installed)
|
|
$(warning Warn: Building without nftables support)
|
|
endif
|
|
|
|
export LIBS += $(LIBS_FEATURES)
|
|
|
|
CONFIG_FILE = .config
|
|
|
|
$(CONFIG_FILE):
|
|
touch $(CONFIG_FILE)
|
|
|
|
ifeq ($(ARCH),x86)
|
|
# CONFIG_COMPAT is only for x86 now, no need for compile-test other archs
|
|
ifeq ($(call try-asm,$(FEATURE_TEST_X86_COMPAT)),true)
|
|
export CONFIG_COMPAT := y
|
|
FEATURE_DEFINES += -DCONFIG_COMPAT
|
|
else
|
|
$(info Note: Building without ia32 C/R, missed ia32 support in gcc)
|
|
$(info $(info) That may be related to missing gcc-multilib in your)
|
|
$(info $(info) distribution or you may have Debian with buggy toolchain)
|
|
$(info $(info) (issue https://github.com/checkpoint-restore/criu/issues/315))
|
|
endif
|
|
endif
|
|
|
|
export DEFINES += $(FEATURE_DEFINES)
|
|
export CFLAGS += $(FEATURE_DEFINES)
|
|
|
|
FEATURES_LIST := TCP_REPAIR STRLCPY STRLCAT PTRACE_PEEKSIGINFO \
|
|
SETPROCTITLE_INIT MEMFD TCP_REPAIR_WINDOW FSCONFIG MEMFD_CREATE
|
|
|
|
# $1 - config name
|
|
define gen-feature-test
|
|
ifeq ($$(call try-cc,$$(FEATURE_TEST_$(1)),$$(LIBS_FEATURES),$$(DEFINES)),true)
|
|
$(Q) echo '#define CONFIG_HAS_$(1)' >> $$@
|
|
$(Q) echo '' >> $$@
|
|
endif
|
|
endef
|
|
|
|
define config-header-rule
|
|
$(CONFIG_HEADER): scripts/feature-tests.mak $(CONFIG_FILE)
|
|
$(call msg-gen, $$@)
|
|
$(Q) echo '#ifndef __CR_CONFIG_H__' > $$@
|
|
$(Q) echo '#define __CR_CONFIG_H__' >> $$@
|
|
$(Q) echo '' >> $$@
|
|
$(call map,gen-feature-test,$(FEATURES_LIST))
|
|
$(Q) cat $(CONFIG_FILE) | sed -n -e '/^[^#]/s/^/#define CONFIG_/p' >> $$@
|
|
$(Q) echo '#endif /* __CR_CONFIG_H__ */' >> $$@
|
|
endef
|
|
|
|
$(eval $(config-header-rule))
|