Re: [PATCH v7 00/20] ILP32 for ARM64

From: Andrew Pinski
Date: Sun Feb 12 2017 - 08:08:11 EST

Next message: Willy Tarreau: "Re: [PATCH] lz4: fix performance regressions"
Previous message: Hoeun Ryu: "[PATCH v6] fork: free vmapped stacks in cache when cpus are offline"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Mon, Jan 9, 2017 at 3:29 AM, Yury Norov <ynorov@xxxxxxxxxxxxxxxxxx> wrote:
> This series enables aarch64 with ilp32 mode.
>
> As supporting work, it introduces ARCH_32BIT_OFF_T configuration
> option that is enabled for existing 32-bit architectures but disabled
> for new arches (so 64-bit off_t is is used by new userspace). Also it
> deprecates getrlimit and setrlimit syscalls prior to prlimit64.
>
> This version is based on linux-next from 2017-01-09. It works with
> glibc-2.25, and tested with LTP, glibc testsuite, trinity, lmbench,
> CPUSpec.
>
> This is not RFC anymore. I believe that all ABI and implementation
> issues are resolved now. The way that kernel clears registers top
> halves is probably the last question, and because there's no objection
> for current approach for more that 6 month, I think, community agrees
> with it.
>
> Patches 1, 2, 3 and 8 are general, and may be applied separately.
>
> Current version does not introduce ABI changes comparing to RFC3.
> Kernel and GLIBC trees:
> https://github.com/norov/linux/tree/ilp32-2017-01-09
> https://github.com/norov/glibc/tree/dev9
>
> (GLIBC patches are managed by Steve Ellsey, so my tree is only for
> reference.)

For folks concerned about performance, here is what we get for SPEC
CPU 2006 on ThunderX 2 CN99xx.
Positive means ILP32 is faster than LP64. This core does not have
AARCH32 so I can't compare that.
Also my LP64 scores don't change with and without the patches.

Options:
-Ofast -flto=32 -mcpu=native -fno-aggressive-loop-optimizations
-funroll-loops -fprefetch-loop-arrays
GCC 7.0.1 r245361 with ilp32 multi-arch patch applied.
4.10rc2 Plus ILP32 patches

SPEC CPU 2006 INT ILP32/LP64
400.perlbench 5.23%
401.bzip2 7.83%
403.gcc 6.22%
429.mcf 14.25%
445.gobmk -1.33%
456.hmmer -0.61%
458.sjeng 0.00%
462.libquantum -7.38%
464.h264ref 10.86%
471.omnetpp 13.53%
473.astar 1.38%
483.xalancbmk 3.73%
Score 4.29%

Rate (32):
400.perlbench 6.10%
401.bzip2 7.10%
403.gcc 6.71%
429.mcf 57.29%
445.gobmk -0.87%
456.hmmer -0.19%
458.sjeng 0.22%
462.libquantum 0.00%
464.h264ref 11.19%
471.omnetpp 11.80%
473.astar -0.29%
483.xalancbmk 8.87%
Score 8.12%

Thanks,
Andrew Pinski

>
> Changes:
> v3: https://lkml.org/lkml/2014/9/3/704
> v4: https://lkml.org/lkml/2015/4/13/691
> v5: https://lkml.org/lkml/2015/9/29/911
> v6: https://lkml.org/lkml/2016/5/23/661
> v7: RFC nowrap: https://lkml.org/lkml/2016/6/17/990
> v7: RFC2 nowrap: https://lkml.org/lkml/2016/8/17/245
> v7: RFC3 nowrap: https://lkml.org/lkml/2016/10/21/883
> v7: - 32-bit off_t deprecation is splitted for compat
> and native 32-bit arches, as it was initially
> done (patches 1, 2);
> - getrlimit() and setrlimit() syscalls deprecated for
> aarch64/ilp32 and all new architectures;
> - documentation is cleaned up (patch 4);
> - compat-related definitions moved from
> aarch64/include/elf.h to binfmt_elf32.c (patch 11)
> - for ptrace, execution mode detection is performed
> at runtime, as it was in v4 (patch 18)
>
> Andrew Pinski (6):
> arm64: rename COMPAT to AARCH32_EL0 in Kconfig
> arm64: ensure the kernel is compiled for LP64
> arm64:uapi: set __BITS_PER_LONG correctly for ILP32 and LP64
> arm64: ilp32: add sys_ilp32.c and a separate table (in entry.S) to use
> it
> arm64: ilp32: introduce ilp32-specific handlers for sigframe and
> ucontext
> arm64:ilp32: add ARM64_ILP32 to Kconfig
>
> Philipp Tomsich (1):
> arm64:ilp32: add vdso-ilp32 and use for signal return
>
> Yury Norov (13):
> compat ABI: use non-compat openat and open_by_handle_at variants
> 32-bit ABI: introduce ARCH_32BIT_OFF_T config option
> asm-generic: Drop getrlimit and setrlimit syscalls from default list
> arm64: ilp32: add documentation on the ILP32 ABI for ARM64
> thread: move thread bits accessors to separated file
> arm64: introduce is_a32_task and is_a32_thread (for AArch32 compat)
> arm64: ilp32: add is_ilp32_compat_{task,thread} and TIF_32BIT_AARCH64
> arm64: introduce binfmt_elf32.c
> arm64: ilp32: introduce binfmt_ilp32.c
> arm64: ilp32: share aarch32 syscall handlers
> arm64: signal: share lp64 signal routines to ilp32
> arm64: signal32: move ilp32 and aarch32 common code to separated file
> arm64: ptrace: handle ptrace_request differently for aarch32 and ilp32
>
> Documentation/arm64/ilp32.txt | 45 +++++++
> arch/Kconfig | 4 +
> arch/arc/Kconfig | 1 +
> arch/arc/include/uapi/asm/unistd.h | 1 +
> arch/arm/Kconfig | 1 +
> arch/arm64/Kconfig | 19 ++-
> arch/arm64/Makefile | 5 +
> arch/arm64/include/asm/compat.h | 19 +--
> arch/arm64/include/asm/elf.h | 32 ++---
> arch/arm64/include/asm/fpsimd.h | 2 +-
> arch/arm64/include/asm/ftrace.h | 2 +-
> arch/arm64/include/asm/hwcap.h | 6 +-
> arch/arm64/include/asm/is_compat.h | 90 ++++++++++++++
> arch/arm64/include/asm/memory.h | 5 +-
> arch/arm64/include/asm/processor.h | 11 +-
> arch/arm64/include/asm/ptrace.h | 2 +-
> arch/arm64/include/asm/seccomp.h | 2 +-
> arch/arm64/include/asm/signal32.h | 9 +-
> arch/arm64/include/asm/signal32_common.h | 27 ++++
> arch/arm64/include/asm/signal_common.h | 33 +++++
> arch/arm64/include/asm/signal_ilp32.h | 38 ++++++
> arch/arm64/include/asm/syscall.h | 2 +-
> arch/arm64/include/asm/thread_info.h | 4 +-
> arch/arm64/include/asm/unistd.h | 8 +-
> arch/arm64/include/asm/vdso.h | 6 +
> arch/arm64/include/uapi/asm/bitsperlong.h | 9 +-
> arch/arm64/include/uapi/asm/unistd.h | 13 ++
> arch/arm64/kernel/Makefile | 18 ++-
> arch/arm64/kernel/asm-offsets.c | 9 +-
> arch/arm64/kernel/binfmt_elf32.c | 32 +++++
> arch/arm64/kernel/binfmt_ilp32.c | 98 +++++++++++++++
> arch/arm64/kernel/cpufeature.c | 8 +-
> arch/arm64/kernel/cpuinfo.c | 20 +--
> arch/arm64/kernel/entry.S | 34 +++++-
> arch/arm64/kernel/entry32.S | 80 ------------
> arch/arm64/kernel/entry32_common.S | 107 ++++++++++++++++
> arch/arm64/kernel/entry_ilp32.S | 22 ++++
> arch/arm64/kernel/head.S | 2 +-
> arch/arm64/kernel/hw_breakpoint.c | 8 +-
> arch/arm64/kernel/perf_regs.c | 2 +-
> arch/arm64/kernel/process.c | 7 +-
> arch/arm64/kernel/ptrace.c | 80 ++++++++++--
> arch/arm64/kernel/signal.c | 102 ++++++++++------
> arch/arm64/kernel/signal32.c | 107 ----------------
> arch/arm64/kernel/signal32_common.c | 135 ++++++++++++++++++++
> arch/arm64/kernel/signal_ilp32.c | 170 ++++++++++++++++++++++++++
> arch/arm64/kernel/sys_ilp32.c | 100 +++++++++++++++
> arch/arm64/kernel/traps.c | 5 +-
> arch/arm64/kernel/vdso-ilp32/.gitignore | 2 +
> arch/arm64/kernel/vdso-ilp32/Makefile | 74 +++++++++++
> arch/arm64/kernel/vdso-ilp32/vdso-ilp32.S | 33 +++++
> arch/arm64/kernel/vdso-ilp32/vdso-ilp32.lds.S | 95 ++++++++++++++
> arch/arm64/kernel/vdso.c | 70 +++++++++--
> arch/arm64/kernel/vdso/gettimeofday.S | 20 ++-
> arch/arm64/kernel/vdso/vdso.S | 6 +-
> arch/blackfin/Kconfig | 1 +
> arch/c6x/include/uapi/asm/unistd.h | 1 +
> arch/cris/Kconfig | 1 +
> arch/frv/Kconfig | 1 +
> arch/h8300/Kconfig | 1 +
> arch/h8300/include/uapi/asm/unistd.h | 1 +
> arch/hexagon/Kconfig | 1 +
> arch/hexagon/include/uapi/asm/unistd.h | 1 +
> arch/m32r/Kconfig | 1 +
> arch/m68k/Kconfig | 1 +
> arch/metag/Kconfig | 1 +
> arch/metag/include/uapi/asm/unistd.h | 1 +
> arch/microblaze/Kconfig | 1 +
> arch/mips/Kconfig | 1 +
> arch/mn10300/Kconfig | 1 +
> arch/nios2/Kconfig | 1 +
> arch/nios2/include/uapi/asm/unistd.h | 1 +
> arch/openrisc/Kconfig | 1 +
> arch/openrisc/include/uapi/asm/unistd.h | 1 +
> arch/parisc/Kconfig | 1 +
> arch/powerpc/Kconfig | 1 +
> arch/score/Kconfig | 1 +
> arch/score/include/uapi/asm/unistd.h | 1 +
> arch/sh/Kconfig | 1 +
> arch/sparc/Kconfig | 1 +
> arch/tile/Kconfig | 1 +
> arch/tile/include/uapi/asm/unistd.h | 1 +
> arch/tile/kernel/compat.c | 3 +
> arch/unicore32/Kconfig | 1 +
> arch/unicore32/include/uapi/asm/unistd.h | 1 +
> arch/x86/Kconfig | 1 +
> arch/x86/um/Kconfig | 1 +
> arch/xtensa/Kconfig | 1 +
> drivers/clocksource/arm_arch_timer.c | 2 +-
> include/linux/fcntl.h | 2 +-
> include/linux/thread_bits.h | 63 ++++++++++
> include/linux/thread_info.h | 54 +-------
> include/uapi/asm-generic/unistd.h | 10 +-
> 93 files changed, 1600 insertions(+), 408 deletions(-)
> create mode 100644 Documentation/arm64/ilp32.txt
> create mode 100644 arch/arm64/include/asm/is_compat.h
> create mode 100644 arch/arm64/include/asm/signal32_common.h
> create mode 100644 arch/arm64/include/asm/signal_common.h
> create mode 100644 arch/arm64/include/asm/signal_ilp32.h
> create mode 100644 arch/arm64/kernel/binfmt_elf32.c
> create mode 100644 arch/arm64/kernel/binfmt_ilp32.c
> create mode 100644 arch/arm64/kernel/entry32_common.S
> create mode 100644 arch/arm64/kernel/entry_ilp32.S
> create mode 100644 arch/arm64/kernel/signal32_common.c
> create mode 100644 arch/arm64/kernel/signal_ilp32.c
> create mode 100644 arch/arm64/kernel/sys_ilp32.c
> create mode 100644 arch/arm64/kernel/vdso-ilp32/.gitignore
> create mode 100644 arch/arm64/kernel/vdso-ilp32/Makefile
> create mode 100644 arch/arm64/kernel/vdso-ilp32/vdso-ilp32.S
> create mode 100644 arch/arm64/kernel/vdso-ilp32/vdso-ilp32.lds.S
> create mode 100644 include/linux/thread_bits.h
>
> --
> 2.7.4
>

Next message: Willy Tarreau: "Re: [PATCH] lz4: fix performance regressions"
Previous message: Hoeun Ryu: "[PATCH v6] fork: free vmapped stacks in cache when cpus are offline"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]