Am Wed, 28 Jun 2017 19:46:30 +0300 schrieb Mike Rapoport <rppt@xxxxxxxxxxxxxxxxxx>: > Hello Michael, > > On Wed, Jun 28, 2017 at 06:11:59PM +0200, Michael Holzheu wrote: > > This patch adds the required functionality to make CRIU run on mainframe > > z Systems machines (s390x). > > > > S390x is is a 64 bit big endian CISC architecture. > > > > This patch set does not include: > > > > * Support for 31 bit compat tasks > > * Support for processes that use s390 runtime instrumentation > > > > Some parts of the patch have been contributed by: > > > > * Alice Frosi <alice@xxxxxxxxxxxxxxxxxx> > > > > Reviewed-by: Alice Frosi <alice@xxxxxxxxxxxxxxxxxx> > > Signed-off-by: Michael Holzheu <holzheu@xxxxxxxxxxxxxxxxxx> > > --- > > Makefile | 37 +- > > compel/.gitignore | 1 + > > compel/Makefile | 8 + > > compel/arch/s390/plugins/include/asm/prologue.h | 1 + > > .../arch/s390/plugins/include/asm/syscall-types.h | 34 ++ > > compel/arch/s390/plugins/std/parasite-head.S | 26 + > > .../s390/plugins/std/syscalls/Makefile.syscalls | 58 +++ > > .../plugins/std/syscalls/syscall-common-s390.S | 37 ++ > > .../s390/plugins/std/syscalls/syscall-s390.tbl | 108 ++++ > > .../arch/s390/plugins/std/syscalls/syscalls-s390.c | 26 + > > compel/arch/s390/scripts/compel-pack.lds.S | 40 ++ > > compel/arch/s390/src/lib/cpu.c | 42 ++ > > compel/arch/s390/src/lib/handle-elf-host.c | 1 + > > compel/arch/s390/src/lib/handle-elf.c | 22 + > > compel/arch/s390/src/lib/include/handle-elf.h | 13 + > > compel/arch/s390/src/lib/include/syscall.h | 8 + > > .../s390/src/lib/include/uapi/asm/breakpoints.h | 15 + > > compel/arch/s390/src/lib/include/uapi/asm/cpu.h | 10 + > > compel/arch/s390/src/lib/include/uapi/asm/fpu.h | 14 + > > .../s390/src/lib/include/uapi/asm/infect-types.h | 75 +++ > > .../src/lib/include/uapi/asm/processor-flags.h | 0 > > .../arch/s390/src/lib/include/uapi/asm/sigframe.h | 75 +++ > > compel/arch/s390/src/lib/infect.c | 559 +++++++++++++++++++++ > > compel/plugins/Makefile | 7 + > > compel/src/lib/handle-elf.c | 25 + > > compel/src/main.c | 3 + > > criu/arch/s390/Makefile | 10 + > > criu/arch/s390/cpu.c | 158 ++++++ > > criu/arch/s390/crtools.c | 341 +++++++++++++ > > criu/arch/s390/include/asm/dump.h | 12 + > > criu/arch/s390/include/asm/int.h | 6 + > > criu/arch/s390/include/asm/parasite-syscall.h | 6 + > > criu/arch/s390/include/asm/parasite.h | 7 + > > criu/arch/s390/include/asm/restore.h | 29 ++ > > criu/arch/s390/include/asm/restorer.h | 65 +++ > > criu/arch/s390/include/asm/types.h | 37 ++ > > criu/arch/s390/include/asm/vdso.h | 23 + > > criu/arch/s390/restorer.c | 37 ++ > > criu/arch/s390/sigframe.c | 20 + > > criu/arch/s390/vdso-pie.c | 65 +++ > > criu/mem.c | 19 + > > criu/pie/Makefile | 7 + > > criu/pie/Makefile.library | 7 + > > criu/pie/util-vdso.c | 25 +- > > criu/proc_parse.c | 19 + > > criu/tty.c | 11 + > > criu/util.c | 5 + > > images/Makefile | 1 + > > images/core-s390.proto | 39 ++ > > images/core.proto | 3 + > > images/cpuinfo.proto | 5 + > > include/common/arch/s390/asm/atomic.h | 67 +++ > > include/common/arch/s390/asm/atomic_ops.h | 74 +++ > > include/common/arch/s390/asm/bitops.h | 164 ++++++ > > include/common/arch/s390/asm/bitsperlong.h | 6 + > > include/common/arch/s390/asm/linkage.h | 22 + > > include/common/arch/s390/asm/page.h | 19 + > > test/zdtm/lib/arch/s390/include/asm/atomic.h | 66 +++ > > test/zdtm/lib/test.c | 5 + > > test/zdtm/static/Makefile | 1 + > > test/zdtm/static/fanotify00.c | 3 + > > test/zdtm/static/mntns-deleted-dst | 0 > > test/zdtm/static/netns-nf.desc | 2 +- > > test/zdtm/static/rtc.desc | 2 +- > > test/zdtm/static/s390x_mmap_high.c | 64 +++ > > test/zdtm/static/s390x_mmap_high.desc | 1 + > > test/zdtm/static/socket-tcp-closed-last-ack.desc | 4 +- > > test/zdtm/static/vt.c | 8 +- > > 68 files changed, 2683 insertions(+), 27 deletions(-) > > This is really big patch and it would be really hard to review it. > > I would suggest to split it into several smaller patches. > I also believe that it would be better for the review if the arch specific > bits would be separated from core functionality changes. I thought about that, but in the end almost all in this patch is s390 specific stuff. Even the few common code files have "#ifdef __s390x__". Patches 2-7 are common code changes that are not 100% s390 specific. So what exactly is your suggestion for splitting up the patch? Or perhaps could you point me to the ppc or arm patches? Michael -- To unsubscribe from this list: send the line "unsubscribe linux-s390" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
