This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: [PATCH][AArch64] Implement -m{cpu,tune,arch}=native using only /proc/cpuinfo


Hi Kyrill, 

I checked this patch on AMD seattle board and it bootstrapped cleanly with  BOOT_CFLAGS=-O2 -mcpu=native and CFLAGS_FOR_TARGET=-O2 -mcpu=native.

With -mcpu=cortex-57,  I get 

 .cpu cortex-a57+fp+simd+crc
options passed:  test.c -mcpu=cortex-a57 -mlittle-endian -mabi=lp64 -auxbase-strip

With âmcpu=native, I get 

.cpu cortex-a57+fp+simd+crypto+crc
 options passed:  test.c -mlittle-endian -mabi=lp64 -mcpu=cortex-a57+fp+simd+crypto+crc  -auxbase-strip

It matches the features.   evtstrm is not relevant  to compiler I believe. 

Another thing is detecting cache sizes automatically.   But I think as of now we have not yet included cache parameters in aarch64 tune tables.  


Regards,
Venkat.

  -----Original Message-----
From: gcc-patches-owner@gcc.gnu.org [mailto:gcc-patches-owner@gcc.gnu.org] On Behalf Of Kyrill Tkachov
Sent: Wednesday, April 22, 2015 8:38 PM
To: GCC Patches
Cc: Marcus Shawcroft; Richard Earnshaw; James Greenhalgh; Evandro Menezes; Andrew Pinski
Subject: Re: [PATCH][AArch64] Implement -m{cpu,tune,arch}=native using only /proc/cpuinfo


On 22/04/15 12:46, Kyrill Tkachov wrote:
> [Sorry for resending twice. My mail client glitched]
>
> On 20/04/15 16:47, Kyrill Tkachov wrote:
>> Hi all,
>>
>> This is an attempt to add native CPU detection to AArch64 GNU/Linux targets.
>> Similar to other ports we use SPEC rewriting to rewrite 
>> -m{cpu,tune,arch}=native options into the appropriate 
>> CPU/architecture and the architecture extension options when appropriate (i.e. +crypto/+crc etc).
>>
>> For CPU/architecture detection it gets a bit involved, especially 
>> when running on a big.LITTLE system. My proposed approach is to look 
>> at /proc/cpuinfo/ and search for the implementer id and part number 
>> fields that uniquely identify each core (appropriate identifying 
>> information is added to aarch64-cores.def). If we find two types of 
>> core we have a big.LITTLE system, so search through the core 
>> definitions extracted from aarch64-cores.def to find if we support such a combination (currently only cortex-a57.cortex-a53 and cortex-a72.cortex-a53) and make sure that the implementer id field matches up.
>>
>> I tested this on a 4xCortex-A53 + 2xCortex-A57 big.LITTLE Ubuntu GNU/Linux system.
>> There are two formats for /proc/cpuinfo/ that I'm aware of. The first (old) one has the format:
>> --------------------------------------
>> processor    : 0
>> processor    : 1
>> processor    : 2
>> processor    : 3
>> processor    : 4
>> processor    : 5
>> Features    : fp asimd evtstrm aes pmull sha1 sha2 crc32
>> CPU implementer    : 0x41
>> CPU architecture: AArch64
>> CPU variant    : 0x0
>> CPU part    : 0xd03
>> --------------------------------------
>>
>> In this format it lists the 6 cores but the CPU part it reports is 
>> only the one for the core from which /proc/cpuinfo was read from (!), in this case one of the Cortex-A53 cores.
>> This means we detect a different CPU depending on which core GCC was 
>> invoked on. Not ideal really, but there's no more information that we can extract.
>> Given the /proc/cpuinfo above, this patch will rewrite -mcpu=native 
>> into -mcpu=cortex-a53+fp+simd+crypto+crc
>>
>> The newer /proc/cpuinfo format proposed at 
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commi
>> t/?id=44b82b7700d05a52cd983799d3ecde1a976b3bed
>> looks like this:
>>
>> --------------------------------------------------------------
>> processor       : 0
>> Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32
>> CPU implementer : 0x41
>> CPU architecture: 8
>> CPU variant     : 0x0
>> CPU part        : 0xd03
>> CPU revision    : 0
>>
>> processor       : 1
>> Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32
>> CPU implementer : 0x41
>> CPU architecture: 8
>> CPU variant     : 0x0
>> CPU part        : 0xd03
>> CPU revision    : 0
>>
>> processor       : 2
>> Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32
>> CPU implementer : 0x41
>> CPU architecture: 8
>> CPU variant     : 0x0
>> CPU part        : 0xd03
>> CPU revision    : 0
>>
>> processor       : 3
>> Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32
>> CPU implementer : 0x41
>> CPU architecture: 8
>> CPU variant     : 0x0
>> CPU part        : 0xd03
>> CPU revision    : 0
>>
>> processor       : 4
>> Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32
>> CPU implementer : 0x41
>> CPU architecture: 8
>> CPU variant     : 0x0
>> CPU part        : 0xd07
>> CPU revision    : 0
>>
>> processor       : 5
>> Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32
>> CPU implementer : 0x41
>> CPU architecture: 8
>> CPU variant     : 0x0
>> CPU part        : 0xd07
>> CPU revision    : 0
>> --------------------------------------------------------------
>>
>> The Features field is used to detect the architectural features that 
>> we map to GCC option extensions i.e. +fp,+crypto,+simd,+crc etc.
>>
>> Similarly, -march=native would be rewritten into 
>> -march=armv8-a+fp+simd+crypto+crc while -mtune=native into 
>> -march=cortex-a57.cortex-a53 (the arch extension options are not valid for -mtune).
>>
>> If it detects more than one implementer ID or the implementer IDs not 
>> matching up somewhere or some other weirdness /proc/cpuinfo or fails 
>> to recognise the CPU it will bail out and ignore the option entirely (similarly to other ports).
>>
>> The patch works fine with both /proc/cpuinfo formats although, as 
>> mentioned above, it will not be able to detect the big.LITTLE combination from the first format.
>>
>> I've filled in the implementer ID and part numbers for the 
>> Cortex-A57, Cortex-A53, Cortex-A72, X-Gene 1 cores, but I don't have 
>> that info for thunderx or exynosm1. Could someone from Cavium and Samsung help me out here? At present this patch has some false dummy values that I'd like to fill out before committing this.
> Thanks Andrew and Evandro for the info.
> I've added the numbers to the patch, so it should work on those systems.
> I'm attaching the final patch here for review.

And resending here with a minor whitespace change in aarch64-cores.def to make thunderx line up with the other entries. Thanks Evandro for pointing it out!

Kyrill

>
> Thanks,
> Kyrill
>
>
> 2014-04-22  Kyrylo Tkachov <kyrylo.tkachov@arm.com>
>
>        * config.host (case ${host}): Add aarch64*-*-linux case.
>        * config/aarch64/aarch64-cores.def: Add IMPLEMENTER_ID and PART_NUMBER
>        fields to all the cores.
>        * config/aarch64/aarch64-elf.h (DRIVER_SELF_SPECS): Add
>        MCPU_MTUNE_NATIVE_SPECS.
>        * config/aarch64/aarch64-option-extensions.def: Add FEAT_STRING field to all
>        extensions.
>        * config/aarch64/aarch64-opts.h: Adjust definition of AARCH64_CORE.
>        * config/aarch64/aarch64.c: Adjust definition of AARCH64_CORE.
>        Adjust definition of AARCH64_OPT_EXTENSION.
>        * config/aarch64/aarch64.h: Adjust definition of AARCH64_CORE.
>        (MCPU_MTUNE_NATIVE_SPECS): Define.
>        * config/aarch64/driver-aarch64.c: New file.
>        * config/aarch64/x-arch64: New file.
>        * doc/invoke.texi (AArch64 Options): Document native value for -mcpu,
>        -mtune and -march.
>
>> I've bootstrapped this on the system mentioned above with -mcpu=native in the BOOT_CFLAGS and regtested as well.
>> For the bootstrap I've used the 2nd /proc/cpuinfo format.
>>
>> I've also tested it on AArch64 hardware from ARM Ltd. and the ecosystem.
>>
>> If using the first format the bootstrap fails the comparison because, 
>> depending on the OS scheduling, some files are compiled with 
>> Cortex-A57 tuning and some with Cortex-A53 tuning and this is practically non-deterministic across stage2 and stage3!
>>
>> What do people think of this approach?
>>
>> 2014-04-20  Kyrylo Tkachov <kyrylo.tkachov@arm.com>
>>
>>        * config.host (case ${host}): Add aarch64*-*-linux case.
>>        * config/aarch64/aarch64-cores.def: Add IMPLEMENTER_ID and PART_NUMBER
>>        fields to all the cores.
>>        * config/aarch64/aarch64-elf.h (DRIVER_SELF_SPECS): Add
>>        MCPU_MTUNE_NATIVE_SPECS.
>>        * config/aarch64/aarch64-option-extensions.def: Add FEAT_STRING field to all
>>        extensions.
>>        * config/aarch64/aarch64-opts.h: Adjust definition of AARCH64_CORE.
>>        * config/aarch64/aarch64.c: Adjust definition of AARCH64_CORE.
>>        Adjust definition of AARCH64_OPT_EXTENSION.
>>        * config/aarch64/aarch64.h: Adjust definition of AARCH64_CORE.
>>        (MCPU_MTUNE_NATIVE_SPECS): Define.
>>        * config/aarch64/driver-aarch64.c: New file.
>>        * config/aarch64/x-arch64: New file.
>>        * doc/invoke.texi (AArch64 Options): Document native value for -mcpu,
>>        -mtune and -march.
>
> aarch64-native.patch
>
>
> commit bfdf31b9d71620afac43b15ebf31502022a9bc63
> Author: Kyrylo Tkachov<kyrylo.tkachov@arm.com>
> Date:   Fri Apr 10 16:39:27 2015 +0100
>
>       [AArch64] Implement -m{tune,cpu,arch}=native on AArch64 
> GNU/Linux
>
> diff --git a/gcc/config.host b/gcc/config.host index b0f5940..a8896d1 
> 100644
> --- a/gcc/config.host
> +++ b/gcc/config.host
> @@ -99,6 +99,14 @@ case ${host} in
>    esac
>
>    case ${host} in
> +  aarch64*-*-linux*)
> +    case ${target} in
> +      aarch64*-*-*)
> +       host_extra_gcc_objs="driver-aarch64.o"
> +       host_xmake_file="${host_xmake_file} aarch64/x-aarch64"
> +       ;;
> +    esac
> +    ;;
>      arm*-*-freebsd* | arm*-*-linux*)
>        case ${target} in
>          arm*-*-*)
> diff --git a/gcc/config/aarch64/aarch64-cores.def 
> b/gcc/config/aarch64/aarch64-cores.def
> index e46d91b..7e6cb73 100644
> --- a/gcc/config/aarch64/aarch64-cores.def
> +++ b/gcc/config/aarch64/aarch64-cores.def
> @@ -21,7 +21,7 @@
>
>       Before using #include to read this file, define a macro:
>
> -      AARCH64_CORE(CORE_NAME, CORE_IDENT, SCHEDULER_IDENT, ARCH, FLAGS, COSTS)
> +      AARCH64_CORE(CORE_NAME, CORE_IDENT, SCHEDULER_IDENT, ARCH, 
> + FLAGS, COSTS, IMP, PART)
>
>       The CORE_NAME is the name of the core, represented as a string constant.
>       The CORE_IDENT is the name of the core, represented as an identifier.
> @@ -30,18 +30,23 @@
>       ARCH is the architecture revision implemented by the chip.
>       FLAGS are the bitwise-or of the traits that apply to that core.
>       This need not include flags implied by the architecture.
> -   COSTS is the name of the rtx_costs routine to use.  */
> +   COSTS is the name of the rtx_costs routine to use.
> +   IMP is the implementer ID of the CPU vendor.  On a GNU/Linux system it can
> +   be found in /proc/cpuinfo.
> +   PART is the part number of the CPU.  On a GNU/Linux system it can be found
> +   in /proc/cpuinfo.  For big.LITTLE systems this should have the form at of
> +   "<big core part number>.<LITTLE core part number>".  */
>
>    /* V8 Architecture Processors.  */
>
> -AARCH64_CORE("cortex-a53",  cortexa53, cortexa53, 8,  
> AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa53) 
> -AARCH64_CORE("cortex-a57",  cortexa57, cortexa57, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa57) -AARCH64_CORE("cortex-a72",  cortexa72, cortexa57, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa57)
> -AARCH64_CORE("exynos-m1",   exynosm1,  cortexa57, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC | AARCH64_FL_CRYPTO, cortexa57)
> -AARCH64_CORE("thunderx",    thunderx,  thunderx, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC | AARCH64_FL_CRYPTO, thunderx)
> -AARCH64_CORE("xgene1",      xgene1,    xgene1,    8,  AARCH64_FL_FOR_ARCH8, xgene1)
> +AARCH64_CORE("cortex-a53",  cortexa53, cortexa53, 8,  
> +AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa53, "0x41", "0xd03") 
> +AARCH64_CORE("cortex-a57",  cortexa57, cortexa57, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa57, "0x41", "0xd07") AARCH64_CORE("cortex-a72",  cortexa72, cortexa57, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa57, "0x41", "0xd08")
> +AARCH64_CORE("exynos-m1",   exynosm1,  cortexa57, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC | AARCH64_FL_CRYPTO, cortexa57, "0x53", "0x001")
> +AARCH64_CORE("thunderx",    thunderx,  thunderx, 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC | AARCH64_FL_CRYPTO, thunderx, "0x43", "0x0a1")
> +AARCH64_CORE("xgene1",      xgene1,    xgene1,    8,  AARCH64_FL_FOR_ARCH8, xgene1, "0x50", "0x000")
>
>    /* V8 big.LITTLE implementations.  */
>
> -AARCH64_CORE("cortex-a57.cortex-a53",  cortexa57cortexa53, cortexa53, 
> 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa57) 
> -AARCH64_CORE("cortex-a72.cortex-a53",  cortexa72cortexa53, cortexa53, 
> 8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa57)
> +AARCH64_CORE("cortex-a57.cortex-a53",  cortexa57cortexa53, cortexa53, 
> +8,  AARCH64_FL_FOR_ARCH8 | AARCH64_FL_CRC, cortexa57, "0x41", 
> +"0xd07.0xd03") AARCH64_CORE("cortex-a72.cortex-a53",  
> +cortexa72cortexa53, cortexa53, 8,  AARCH64_FL_FOR_ARCH8 | 
> +AARCH64_FL_CRC, cortexa57, "0x41", "0xd08.0xd03")
> diff --git a/gcc/config/aarch64/aarch64-elf.h 
> b/gcc/config/aarch64/aarch64-elf.h
> index a5ec8cb..1ce6343 100644
> --- a/gcc/config/aarch64/aarch64-elf.h
> +++ b/gcc/config/aarch64/aarch64-elf.h
> @@ -132,7 +132,8 @@
>    #undef DRIVER_SELF_SPECS
>    #define DRIVER_SELF_SPECS \
>      " %{!mbig-endian:%{!mlittle-endian:" ENDIAN_SPEC "}}" \
> -  " %{!mabi=*:" ABI_SPEC "}"
> +  " %{!mabi=*:" ABI_SPEC "}" \
> +  MCPU_MTUNE_NATIVE_SPECS
>
>    #ifdef HAVE_AS_MABI_OPTION
>    #define ASM_MABI_SPEC "%{mabi=*:-mabi=%*}"
> diff --git a/gcc/config/aarch64/aarch64-option-extensions.def 
> b/gcc/config/aarch64/aarch64-option-extensions.def
> index 6ec3ed6..f296296 100644
> --- a/gcc/config/aarch64/aarch64-option-extensions.def
> +++ b/gcc/config/aarch64/aarch64-option-extensions.def
> @@ -21,18 +21,25 @@
>
>       Before using #include to read this file, define a macro:
>
> -      AARCH64_OPT_EXTENSION(EXT_NAME, FLAGS_ON, FLAGS_OFF)
> +      AARCH64_OPT_EXTENSION(EXT_NAME, FLAGS_ON, FLAGS_OFF, 
> + FEATURE_STRING)
>
>       EXT_NAME is the name of the extension, represented as a string constant.
>       FLAGS_ON are the bitwise-or of the features that the extension adds.
> -   FLAGS_OFF are the bitwise-or of the features that the extension removes.  */
> +   FLAGS_OFF are the bitwise-or of the features that the extension removes.
> +   FEAT_STRING is a string containing the entries in the 'Features' field of
> +   /proc/cpuinfo on a GNU/Linux system that correspond to this architecture
> +   extension being available.  Sometimes multiple entries are needed to enable
> +   the extension (for example, the 'crypto' extension depends on four
> +   entries: aes, pmull, sha1, sha2 being present).  In that case this field
> +   should contain a whitespace-separated list of the strings in 'Features'
> +   that are required.  Their order is not important.  */
>
>    /* V8 Architecture Extensions.
>       This list currently contains example extensions for CPUs that implement
>       AArch64, and therefore serves as a template for adding more CPUs in the
>       future.  */
>
> -AARCH64_OPT_EXTENSION("fp",    AARCH64_FL_FP,  AARCH64_FL_FPSIMD | AARCH64_FL_CRYPTO)
> -AARCH64_OPT_EXTENSION("simd",  AARCH64_FL_FPSIMD,      AARCH64_FL_SIMD | AARCH64_FL_CRYPTO)
> -AARCH64_OPT_EXTENSION("crypto",        AARCH64_FL_CRYPTO | AARCH64_FL_FPSIMD,  AARCH64_FL_CRYPTO)
> -AARCH64_OPT_EXTENSION("crc",   AARCH64_FL_CRC, AARCH64_FL_CRC)
> +AARCH64_OPT_EXTENSION("fp",    AARCH64_FL_FP,                          AARCH64_FL_FPSIMD | AARCH64_FL_CRYPTO, "fp")
> +AARCH64_OPT_EXTENSION("simd",  AARCH64_FL_FPSIMD,                      AARCH64_FL_SIMD | AARCH64_FL_CRYPTO,   "asimd")
> +AARCH64_OPT_EXTENSION("crypto",        AARCH64_FL_CRYPTO | AARCH64_FL_FPSIMD,  AARCH64_FL_CRYPTO,                     "aes pmull sha1 sha2")
> +AARCH64_OPT_EXTENSION("crc",   AARCH64_FL_CRC,                         AARCH64_FL_CRC,                        "crc32")
> diff --git a/gcc/config/aarch64/aarch64-opts.h 
> b/gcc/config/aarch64/aarch64-opts.h
> index f88ae5b..ea64cf4 100644
> --- a/gcc/config/aarch64/aarch64-opts.h
> +++ b/gcc/config/aarch64/aarch64-opts.h
> @@ -25,7 +25,7 @@
>    /* The various cores that implement AArch64.  */
>    enum aarch64_processor
>    {
> -#define AARCH64_CORE(NAME, INTERNAL_IDENT, SCHED, ARCH, FLAGS, COSTS) 
> \
> +#define AARCH64_CORE(NAME, INTERNAL_IDENT, SCHED, ARCH, FLAGS, COSTS, 
> +IMP, PART) \
>      INTERNAL_IDENT,
>    #include "aarch64-cores.def"
>    #undef AARCH64_CORE
> diff --git a/gcc/config/aarch64/aarch64.c 
> b/gcc/config/aarch64/aarch64.c index 954e110..ea6020f 100644
> --- a/gcc/config/aarch64/aarch64.c
> +++ b/gcc/config/aarch64/aarch64.c
> @@ -441,7 +441,7 @@ struct processor
>    /* Processor cores implementing AArch64.  */
>    static const struct processor all_cores[] =
>    {
> -#define AARCH64_CORE(NAME, IDENT, SCHED, ARCH, FLAGS, COSTS) \
> +#define AARCH64_CORE(NAME, IDENT, SCHED, ARCH, FLAGS, COSTS, IMP, 
> +PART) \
>      {NAME, SCHED, #ARCH, ARCH, FLAGS, &COSTS##_tunings},
>    #include "aarch64-cores.def"
>    #undef AARCH64_CORE
> @@ -478,7 +478,7 @@ struct aarch64_option_extension
>    /* ISA extensions in AArch64.  */
>    static const struct aarch64_option_extension all_extensions[] =
>    {
> -#define AARCH64_OPT_EXTENSION(NAME, FLAGS_ON, FLAGS_OFF) \
> +#define AARCH64_OPT_EXTENSION(NAME, FLAGS_ON, FLAGS_OFF, 
> +FEATURE_STRING) \
>      {NAME, FLAGS_ON, FLAGS_OFF},
>    #include "aarch64-option-extensions.def"
>    #undef AARCH64_OPT_EXTENSION
> diff --git a/gcc/config/aarch64/aarch64.h 
> b/gcc/config/aarch64/aarch64.h index bf59e40..1f7187b 100644
> --- a/gcc/config/aarch64/aarch64.h
> +++ b/gcc/config/aarch64/aarch64.h
> @@ -506,7 +506,7 @@ enum reg_class
>
>    enum target_cpus
>    {
> -#define AARCH64_CORE(NAME, INTERNAL_IDENT, SCHED, ARCH, FLAGS, COSTS) 
> \
> +#define AARCH64_CORE(NAME, INTERNAL_IDENT, SCHED, ARCH, FLAGS, COSTS, 
> +IMP, PART) \
>      TARGET_CPU_##INTERNAL_IDENT,
>    #include "aarch64-cores.def"
>    #undef AARCH64_CORE
> @@ -929,11 +929,24 @@ extern const char *aarch64_rewrite_mcpu (int argc, const char **argv);
>    #define BIG_LITTLE_CPU_SPEC_FUNCTIONS \
>      { "rewrite_mcpu", aarch64_rewrite_mcpu },
>
> +#if defined(__aarch64__)
> +extern const char *host_detect_local_cpu (int argc, const char **argv);
> +# define EXTRA_SPEC_FUNCTIONS                                          \
> +  { "local_cpu_detect", host_detect_local_cpu },                       \
> +  BIG_LITTLE_CPU_SPEC_FUNCTIONS
> +
> +# define MCPU_MTUNE_NATIVE_SPECS                                       \
> +   " %{march=native:%<march=native %:local_cpu_detect(arch)}"          \
> +   " %{mcpu=native:%<mcpu=native %:local_cpu_detect(cpu)}"             \
> +   " %{mtune=native:%<mtune=native %:local_cpu_detect(tune)}"
> +#else
> +# define MCPU_MTUNE_NATIVE_SPECS ""
> +# define EXTRA_SPEC_FUNCTIONS BIG_LITTLE_CPU_SPEC_FUNCTIONS #endif
> +
>    #define ASM_CPU_SPEC \
>       BIG_LITTLE_SPEC
>
> -#define EXTRA_SPEC_FUNCTIONS BIG_LITTLE_CPU_SPEC_FUNCTIONS
> -
>    #define EXTRA_SPECS                                           \
>      { "asm_cpu_spec",           ASM_CPU_SPEC }
>
> diff --git a/gcc/config/aarch64/driver-aarch64.c 
> b/gcc/config/aarch64/driver-aarch64.c
> new file mode 100644
> index 0000000..da10a4c
> --- /dev/null
> +++ b/gcc/config/aarch64/driver-aarch64.c
> @@ -0,0 +1,307 @@
> +/* Native CPU detection for aarch64.
> +   Copyright (C) 2014 Free Software Foundation, Inc.
> +
> +   This file is part of GCC.
> +
> +   GCC is free software; you can redistribute it and/or modify
> +   it under the terms of the GNU General Public License as published by
> +   the Free Software Foundation; either version 3, or (at your option)
> +   any later version.
> +
> +   GCC is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +   GNU General Public License for more details.
> +
> +   You should have received a copy of the GNU General Public License
> +   along with GCC; see the file COPYING3.  If not see 
> +<http://www.gnu.org/licenses/>.  */
> +
> +#include "config.h"
> +#include "system.h"
> +
> +struct arch_extension
> +{
> +  const char *ext;
> +  const char *feat_string;
> +};
> +
> +#define AARCH64_OPT_EXTENSION(EXT_NAME, FLAGS_ON, FLAGS_OFF, 
> +FEATURE_STRING) \
> +  { EXT_NAME, FEATURE_STRING },
> +static struct arch_extension ext_to_feat_string[] = { #include 
> +"aarch64-option-extensions.def"
> +};
> +#undef AARCH64_OPT_EXTENSION
> +
> +
> +struct aarch64_core_data
> +{
> +  const char* name;
> +  const char* arch;
> +  const char* implementer_id;
> +  const char* part_no;
> +};
> +
> +#define AARCH64_CORE(CORE_NAME, CORE_IDENT, SCHED, ARCH, FLAGS, 
> +COSTS, IMP, PART) \
> +  { CORE_NAME, #ARCH, IMP, PART },
> +
> +static struct aarch64_core_data cpu_data [] = { #include 
> +"aarch64-cores.def"
> +  { NULL, NULL, NULL, NULL }
> +};
> +
> +#undef AARCH64_CORE
> +
> +struct aarch64_arch
> +{
> +  const char* id;
> +  const char* name;
> +};
> +
> +#define AARCH64_ARCH(NAME, CORE, ARCH, FLAGS) \
> +  { #ARCH, NAME  },
> +
> +static struct aarch64_arch aarch64_arches [] = { #include 
> +"aarch64-arches.def"
> +  {NULL, NULL}
> +};
> +
> +#undef AARCH64_ARCH
> +
> +/* Return the full architecture name string corresponding to the
> +   identifier ID.  */
> +
> +static const char*
> +get_arch_name_from_id (const char* id) {
> +  unsigned int i = 0;
> +
> +  for (i = 0; aarch64_arches[i].id != NULL; i++)
> +    {
> +      if (strcmp (id, aarch64_arches[i].id) == 0)
> +        return aarch64_arches[i].name;
> +    }
> +
> +  return NULL;
> +}
> +
> +
> +/* Check wether the string CORE contains the same CPU part numbers
> +   as BL_STRING.  For example CORE="{0xd03, 0xd07}" and BL_STRING="0xd07.0xd03"
> +   should return true.  */
> +
> +static bool
> +valid_bL_string_p (const char** core, const char* bL_string) {
> +  return strstr (bL_string, core[0]) != NULL
> +         && strstr (bL_string, core[1]) != NULL; }
> +
> +/*  Return true iff ARR contains STR in one of its two elements.  */
> +
> +static bool
> +contains_string_p (const char** arr, const char* str) {
> +  bool res = false;
> +
> +  if (arr[0] != NULL)
> +    {
> +      res = strstr (arr[0], str) != NULL;
> +      if (res)
> +        return res;
> +
> +      if (arr[1] != NULL)
> +        return strstr (arr[1], str) != NULL;
> +    }
> +
> +  return false;
> +}
> +
> +/* This will be called by the spec parser in gcc.c when it sees
> +   a %:local_cpu_detect(args) construct.  Currently it will be called
> +   with either "arch", "cpu" or "tune" as argument depending on if
> +   -march=native, -mcpu=native or -mtune=native is to be substituted.
> +
> +   It returns a string containing new command line parameters to be
> +   put at the place of the above two options, depending on what CPU
> +   this is executed.  E.g. "-march=armv8-a" on a Cortex-A57 for
> +   -march=native.  If the routine can't detect a known processor,
> +   the -march or -mtune option is discarded.
> +
> +   For -mtune and -mcpu arguments it attempts to detect the CPU or
> +   a big.LITTLE system.
> +   ARGC and ARGV are set depending on the actual arguments given
> +   in the spec.  */
> +
> +const char *
> +host_detect_local_cpu (int argc, const char **argv) {
> +  const char *arch_id = NULL;
> +  const char *res = NULL;
> +  static const int num_exts = ARRAY_SIZE (ext_to_feat_string);
> +  char buf[128];
> +  FILE *f = NULL;
> +  bool arch = false;
> +  bool tune = false;
> +  bool cpu = false;
> +  unsigned int i = 0;
> +  unsigned int core_idx = 0;
> +  const char* imps[2] = { NULL, NULL };
> +  const char* cores[2] = { NULL, NULL };
> +  unsigned int n_cores = 0;
> +  unsigned int n_imps = 0;
> +  bool processed_exts = false;
> +  const char *ext_string = "";
> +
> +  gcc_assert (argc);
> +
> +  if (!argv[0])
> +    goto not_found;
> +
> +  /* Are we processing -march, mtune or mcpu?  */  arch = strcmp 
> + (argv[0], "arch") == 0;  if (!arch)
> +    tune = strcmp (argv[0], "tune") == 0;
> +
> +  if (!arch && !tune)
> +    cpu = strcmp (argv[0], "cpu") == 0;
> +
> +  if (!arch && !tune && !cpu)
> +    goto not_found;
> +
> +  f = fopen ("/proc/cpuinfo", "r");
> +
> +  if (f == NULL)
> +    goto not_found;
> +
> +  /* Look through /proc/cpuinfo to determine the implementer
> +     and then the part number that identifies a particular core.  */  
> + while (fgets (buf, sizeof (buf), f) != NULL)
> +    {
> +      if (strstr (buf, "implementer") != NULL)
> +       {
> +         for (i = 0; cpu_data[i].name != NULL; i++)
> +           if (strstr (buf, cpu_data[i].implementer_id) != NULL
> +                && !contains_string_p (imps, cpu_data[i].implementer_id))
> +             {
> +                if (n_imps == 2)
> +                  goto not_found;
> +
> +                imps[n_imps++] = cpu_data[i].implementer_id;
> +
> +                break;
> +             }
> +          continue;
> +       }
> +
> +      if (strstr (buf, "part") != NULL)
> +       {
> +         for (i = 0; cpu_data[i].name != NULL; i++)
> +           if (strstr (buf, cpu_data[i].part_no) != NULL
> +                && !contains_string_p (cores, cpu_data[i].part_no))
> +             {
> +                if (n_cores == 2)
> +                  goto not_found;
> +
> +                cores[n_cores++] = cpu_data[i].part_no;
> +               core_idx = i;
> +               arch_id = cpu_data[i].arch;
> +               break;
> +             }
> +          continue;
> +        }
> +      if (!tune && !processed_exts && strstr (buf, "Features") != NULL)
> +        {
> +          for (i = 0; i < num_exts; i++)
> +            {
> +              bool enabled = true;
> +              char *p = NULL;
> +              char *feat_string = concat 
> + (ext_to_feat_string[i].feat_string, NULL);
> +
> +              p = strtok (feat_string, " ");
> +
> +              while (p != NULL)
> +                {
> +                  if (strstr (buf, p) == NULL)
> +                    {
> +                      enabled = false;
> +                      break;
> +                    }
> +                  p = strtok (NULL, " ");
> +                }
> +              ext_string = concat (ext_string, "+", enabled ? "" : "no",
> +                                   ext_to_feat_string[i].ext, NULL);
> +            }
> +          processed_exts = true;
> +        }
> +    }
> +
> +  fclose (f);
> +  f = NULL;
> +
> +  /* Weird cpuinfo format that we don't know how to handle.  */  if 
> + (n_cores == 0 || n_cores > 2 || n_imps != 1)
> +    goto not_found;
> +
> +  if (arch && !arch_id)
> +    goto not_found;
> +
> +  if (arch)
> +    {
> +      const char* arch_name = get_arch_name_from_id (arch_id);
> +
> +      /* We got some arch indentifier that's not in aarch64-arches.def?  */
> +      if (!arch_name)
> +        goto not_found;
> +
> +      res = concat ("-march=", arch_name, NULL);
> +    }
> +  /* We have big.LITTLE.  */
> +  else if (n_cores == 2)
> +    {
> +      for (i = 0; cpu_data[i].name != NULL; i++)
> +        {
> +          if (strchr (cpu_data[i].part_no, '.') != NULL
> +              && strncmp (cpu_data[i].implementer_id, imps[0], strlen (imps[0]) - 1) == 0
> +              && valid_bL_string_p (cores, cpu_data[i].part_no))
> +            {
> +              res = concat ("-m", cpu ? "cpu" : "tune", "=", cpu_data[i].name, NULL);
> +              break;
> +            }
> +        }
> +      if (!res)
> +        goto not_found;
> +    }
> +  /* The simple, non-big.LITTLE case.  */  else
> +    {
> +      if (strncmp (cpu_data[core_idx].implementer_id, imps[0],
> +                   strlen (imps[0]) - 1) != 0)
> +        goto not_found;
> +
> +      res = concat ("-m", cpu ? "cpu" : "tune", "=",
> +                      cpu_data[core_idx].name, NULL);
> +    }
> +
> +  if (tune)
> +    return res;
> +
> +  res = concat (res, ext_string, NULL);
> +
> +  return res;
> +
> +not_found:
> +  {
> +   /* If detection fails we ignore the option.
> +      Clean up and return empty string.  */
> +
> +    if (f)
> +      fclose (f);
> +
> +    return "";
> +  }
> +}
> +
> diff --git a/gcc/config/aarch64/x-aarch64 
> b/gcc/config/aarch64/x-aarch64 new file mode 100644 index 
> 0000000..8c09e04
> --- /dev/null
> +++ b/gcc/config/aarch64/x-aarch64
> @@ -0,0 +1,3 @@
> +driver-aarch64.o: $(srcdir)/config/aarch64/driver-aarch64.c \
> +  $(CONFIG_H) $(SYSTEM_H)
> +       $(COMPILER) -c $(ALL_COMPILERFLAGS) $(ALL_CPPFLAGS) 
> +$(INCLUDES) $<
> diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi index 
> e2918cb..5787524 100644
> --- a/gcc/doc/invoke.texi
> +++ b/gcc/doc/invoke.texi
> @@ -12318,8 +12318,12 @@ This involves inserting a NOP instruction between memory instructions and
>    Specify the name of the target architecture, optionally suffixed by one or
>    more feature modifiers.  This option has the form
>    @option{-march=@var{arch}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}, 
> where the -only permissible value for @var{arch} is @samp{armv8-a}.  
> The permissible -values for @var{feature} are documented in the sub-section below.
> +only permissible value for @var{arch} is @samp{armv8-a}.
> +The permissible values for @var{feature} are documented in the 
> +sub-section below.  Additionally on native AArch64 GNU/Linux systems 
> +the value @samp{native} is available.  This option causes the 
> +compiler to pick the architecture of the host system.  If the 
> +compiler is unable to recognize the architecture of the host system this option has no effect.
>
>    Where conflicting feature modifiers are specified, the right-most feature is
>    used.
> @@ -12343,6 +12347,13 @@ Additionally, this option can specify that GCC should tune the performance
>    of the code for a big.LITTLE system.  Permissible values for this
>    option are: @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53}.
>
> +Additionally on native AArch64 GNU/Linux systems the value 
> +@samp{native} is available.
> +This option causes the compiler to pick the architecture of and tune 
> +the performance of the code for the processor of the host system.
> +If the compiler is unable to recognize the processor of the host 
> +system this option has no effect.
> +
>    Where none of @option{-mtune=}, @option{-mcpu=} or @option{-march=}
>    are specified, the code is tuned to perform well across a range
>    of target processors.
> @@ -12355,7 +12366,11 @@ Specify the name of the target processor, optionally suffixed by one or more
>    feature modifiers.  This option has the form
>    @option{-mcpu=@var{cpu}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}, where the
>    permissible values for @var{cpu} are the same as those available 
> for -@option{-mtune}.
> +@option{-mtune}.  Additionally on native AArch64 GNU/Linux systems 
> +the value @samp{native} is available.
> +This option causes the compiler to tune the performance of the code 
> +for the processor of the host system.  If the compiler is unable to 
> +recognize the processor of the host system this option has no effect.
>
>    The permissible values for @var{feature} are documented in the sub-section
>    below.
>


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]