This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [COMMITTED][arm] Revert r251800 & r251799
- From: Vidya Praveen <vidyapraveen at arm dot com>
- To: gcc-patches at gcc dot gnu dot org
- Cc: nd at arm dot com
- Date: Mon, 11 Sep 2017 15:29:09 +0100
- Subject: Re: [COMMITTED][arm] Revert r251800 & r251799
- Authentication-results: sourceware.org; auth=none
- Authentication-results: spf=none (sender IP is ) smtp.mailfrom=vidyapraveen at arm dot com;
- Nodisclaimer: True
- References: <20170911142012.GB25869@e109359-lin.cambridge.arm.com>
- Spamdiagnosticmetadata: NSPM
- Spamdiagnosticoutput: 1:99
Now with the patch :-)
VP.
On Mon, Sep 11, 2017 at 03:20:12PM +0100, Vidya Praveen wrote:
> Hello,
>
> The following two related patches need to be reverted as it causes cross-native
> builds to fail with the following message:
>
> g++ -c -DIN_GCC -DGENERATOR_FILE -I. [...] \
> -o build/genpreds.o /path/to/src/gcc/gcc/genpreds.c
> In file included from ./options.h:8:0,
> from ./tm.h:23,
> from /path/to/src/gcc/gcc/genpreds.c:26:
> /path/to/src/gcc/gcc/config/arm/arm-opts.h:29:21: fatal error: arm-isa.h: No such file or directory
> #include "arm-isa.h"
> ^
> genpreds depends on GTM_H which does not depend on options.h, or any of its
> dependencies. Nevertheless, it still tries to include options.h when reading
> tm.h, so we miss the rule to build arm-isa.h. It is unclear why it is only an
> issue with the cross-native builds.
>
> For now, in order to keep the builds going, I am reverting these patches.
>
>
> r251800 | rearnsha | 2017-09-06 14:42:54 +0100 (Wed, 06 Sep 2017) | 16 lines
>
> [arm] Improve error checking in parsecpu.awk
>
> This patch adds a bit more error checking to parsecpu.awk to ensure
> that statements are not missing arguments or have excess arguments
> beyond those permitted. It also slightly improves the handling of
> errors so that we terminate properly if parsing fails and be as
> helpful as we can while in the parsing phase.
>
> * config/arm/parsecpu.awk (fatal): Note that we've encountered an
> error. Only quit immediately if parsing is complete.
> (BEGIN): Initialize fatal_err and parse_done.
> (begin fpu, end fpu): Check number of arguments.
> (begin arch, end arch): Likewise.
> (begin cpu, end cpu): Likewise.
> (cname, tune for, tune flags, architecture, fpu, option): Likewise.
> (optalias): Likewise.
>
> r251799 | rearnsha | 2017-09-06 14:42:46 +0100 (Wed, 06 Sep 2017) | 31 lines
>
> [arm] auto-generate arm-isa.h from CPU descriptions
>
> This patch autogenerates arm-isa.h from new entries in arm-cpus.in.
> This has the primary advantage that it makes the description file more
> self-contained, but it also solves the 'array dimensioning' problem
> that Tamar recently encountered. It adds two new constructs to
> arm-cpus.in: features and fgroups. Fgroups are simply a way of naming
> a group of feature bits so that they can be referenced together. We
> follow the convention that feature bits are all lower case, while
> fgroups are (predominantly) upper case. This is helpful as in some
> contexts they share the same namespace. Most of the minor changes in
> this patch are related to adopting this new naming convention.
>
> * config.gcc (arm*-*-*): Don't add arm-isa.h to tm_p_file.
> * config/arm/arm-isa.h: Delete. Move definitions to ...
> * arm-cpus.in: ... here. Use new feature and fgroup values.
> * config/arm/arm.c (arm_option_override): Use lower case for feature
> bit names.
> * config/arm/arm.h (TARGET_HARD_FLOAT): Likewise.
> (TARGET_VFP3, TARGET_VFP5, TARGET_FMA): Likewise.
> * config/arm/parsecpu.awk (END): Add new command 'isa'.
> (isa_pfx): Delete.
> (print_isa_bits_for): New function.
> (gen_isa): New function.
> (gen_comm_data): Use print_isa_bits_for.
> (define feature): New keyword.
> (define fgroup): New keyword.
> * config/arm/t-arm (OPTIONS_H_EXTRA): Add arm-isa.h
> (arm-isa.h): Add rule to generate file.
> * common/config/arm/arm-common.c: (arm_canon_arch_option): Use lower
> case for feature bit names.
>
> Regards,
> VP.
>
>
> gcc/ChangeLog:
>
> 2017-09-11 Vidya Praveen <vidyapraveen@arm.com>
>
> Revert r251800 and r251799.
diff --git a/gcc/common/config/arm/arm-common.c b/gcc/common/config/arm/arm-common.c
index 7cb99ec..38bd3a7 100644
--- a/gcc/common/config/arm/arm-common.c
+++ b/gcc/common/config/arm/arm-common.c
@@ -574,7 +574,7 @@ arm_canon_arch_option (int argc, const char **argv)
{
/* The easiest and safest way to remove the default fpu
capabilities is to look for a '+no..' option that removes
- the base FPU bit (isa_bit_vfpv2). If that doesn't exist
+ the base FPU bit (isa_bit_VFPv2). If that doesn't exist
then the best we can do is strip out all the bits that
might be part of the most capable FPU we know about,
which is "crypto-neon-fp-armv8". */
@@ -586,7 +586,7 @@ arm_canon_arch_option (int argc, const char **argv)
++ext)
{
if (ext->remove
- && check_isa_bits_for (ext->isa_bits, isa_bit_vfpv2))
+ && check_isa_bits_for (ext->isa_bits, isa_bit_VFPv2))
{
arm_initialize_isa (fpu_isa, ext->isa_bits);
bitmap_and_compl (target_isa, target_isa, fpu_isa);
@@ -620,7 +620,7 @@ arm_canon_arch_option (int argc, const char **argv)
{
/* Clearing the VFPv2 bit is sufficient to stop any extention that
builds on the FPU from matching. */
- bitmap_clear_bit (target_isa, isa_bit_vfpv2);
+ bitmap_clear_bit (target_isa, isa_bit_VFPv2);
}
/* If we don't have a selected architecture by now, something's
@@ -692,8 +692,8 @@ arm_canon_arch_option (int argc, const char **argv)
capable FPU variant that we do support. This is sufficient for
multilib selection. */
- if (bitmap_bit_p (target_isa_unsatisfied, isa_bit_vfpv2)
- && bitmap_bit_p (fpu_isa, isa_bit_vfpv2))
+ if (bitmap_bit_p (target_isa_unsatisfied, isa_bit_VFPv2)
+ && bitmap_bit_p (fpu_isa, isa_bit_VFPv2))
{
std::list<candidate_extension *>::iterator ipoint = extensions.begin ();
diff --git a/gcc/config.gcc b/gcc/config.gcc
index 994b098..630832f 100644
--- a/gcc/config.gcc
+++ b/gcc/config.gcc
@@ -593,7 +593,7 @@ x86_64-*-*)
tm_file="vxworks-dummy.h ${tm_file}"
;;
arm*-*-*)
- tm_p_file="arm/arm-flags.h ${tm_p_file} arm/aarch-common-protos.h"
+ tm_p_file="arm/arm-flags.h arm/arm-isa.h ${tm_p_file} arm/aarch-common-protos.h"
tm_file="vxworks-dummy.h ${tm_file}"
;;
mips*-*-* | sh*-*-* | sparc*-*-*)
diff --git a/gcc/config/arm/arm-cpus.in b/gcc/config/arm/arm-cpus.in
index 07de4c9..d009a9e 100644
--- a/gcc/config/arm/arm-cpus.in
+++ b/gcc/config/arm/arm-cpus.in
@@ -40,210 +40,6 @@
# names in the final compiler. The order within each group is preserved and
# forms the order for the list within the compiler.
-# Most objects in this file support forward references. The major
-# exception is feature groups, which may only refer to previously
-# defined features or feature groups. This is done to avoid the risk
-# of feature groups recursively referencing each other and causing
-# the parser to hang.
-
-# Features - general convention: all lower case.
-
-# Extended multiply
-define feature armv3m
-
-# 26-bit mode support
-define feature mode26
-
-# 32-bit mode support
-define feature mode32
-
-# Architecture rel 4
-define feature armv4
-
-# Architecture rel 5
-define feature armv5
-
-# Thumb aware.
-define feature thumb
-
-# Architecture rel 5e.
-define feature armv5e
-
-# XScale.
-define feature xscale
-
-# Architecture rel 6.
-define feature armv6
-
-# Architecture rel 6k.
-define feature armv6k
-
-# Thumb-2.
-define feature thumb2
-
-# Instructions not present in 'M' profile.
-define feature notm
-
-# Architecture uses be8 mode in big-endian.
-define feature be8
-
-# Thumb division instructions.
-define feature tdiv
-
-# Architecture rel 7e-m.
-define feature armv7em
-
-# Architecture rel 7.
-define feature armv7
-
-# ARM division instructions.
-define feature adiv
-
-# Architecture rel 8.
-define feature armv8
-
-# ARMv8 CRC32 instructions.
-define feature crc32
-
-# XScale v2 (Wireless MMX).
-define feature iwmmxt
-
-# XScale Wireless MMX2.
-define feature iwmmxt2
-
-# Architecture rel 8.1.
-define feature armv8_1
-
-# Architecutre rel 8.2.
-define feature armv8_2
-
-# M-Profile security extensions.
-define feature cmse
-
-# Floating point and Neon extensions.
-# VFPv1 is not supported in GCC.
-
-# Vector floating point v2.
-define feature vfpv2
-
-# Vector floating point v3.
-define feature vfpv3
-
-# Vector floating point v4.
-define feature vfpv4
-
-# Floating point v5.
-define feature fpv5
-
-# ARMv7-A LPAE.
-define feature lpae
-
-# Advanced SIMD instructions.
-define feature neon
-
-# Conversions to/from fp16 (VFPv3 extension).
-define feature fp16conv
-
-# Double precision operations supported.
-define feature fp_dbl
-
-# 32 Double precision registers.
-define feature fp_d32
-
-# Crypto extension to ARMv8.
-define feature crypto
-
-# FP16 data processing (half-precision float).
-define feature fp16
-
-
-# ISA Quirks (errata?). Don't forget to add this to the fgroup
-# ALL_QUIRKS below.
-
-# No volatile memory in IT blocks.
-define feature quirk_no_volatile_ce
-
-# Previously mis-identified by GCC.
-define feature quirk_armv6kz
-
-# Cortex-M3 LDRD quirk.
-define feature quirk_cm3_ldrd
-
-# (Very) slow multiply operations. Should probably be a tuning bit.
-define feature smallmul
-
-# Feature groups. Conventionally all (or mostly) upper case.
-
-# List of all cryptographic extensions to stripout if crypto is
-# disabled. Currently, that's trivial, but we define it anyway for
-# consistency with the SIMD and FP disable lists.
-define fgroup ALL_CRYPTO crypto
-
-# List of all SIMD bits to strip out if SIMD is disabled. This does
-# strip off 32 D-registers, but does not remove support for
-# double-precision FP.
-define fgroup ALL_SIMD fp_d32 neon ALL_CRYPTO
-
-# List of all FPU bits to strip out if -mfpu is used to override the
-# default. fp16 is deliberately missing from this list.
-define fgroup ALL_FPU_INTERNAL vfpv2 vfpv3 vfpv4 fpv5 fp16conv fp_dbl ALL_SIMD
-
-# Similarly, but including fp16 and other extensions that aren't part of
-# -mfpu support.
-define fgroup ALL_FP fp16 ALL_FPU_INTERNAL
-
-define fgroup ARMv2 notm
-define fgroup ARMv3 ARMv2 mode32
-define fgroup ARMv3m ARMv3 armv3m
-define fgroup ARMv4 ARMv3m armv4
-define fgroup ARMv4t ARMv4 thumb
-define fgroup ARMv5 ARMv4 armv5
-define fgroup ARMv5t ARMv5 thumb
-define fgroup ARMv5e ARMv5 armv5e
-define fgroup ARMv5te ARMv5e thumb
-define fgroup ARMv5tej ARMv5te
-define fgroup ARMv6 ARMv5te armv6 be8
-define fgroup ARMv6j ARMv6
-define fgroup ARMv6k ARMv6 armv6k
-define fgroup ARMv6z ARMv6
-define fgroup ARMv6kz ARMv6k quirk_armv6kz
-define fgroup ARMv6zk ARMv6k
-define fgroup ARMv6t2 ARMv6 thumb2
-# This is suspect. ARMv6-m doesn't really pull in any useful features
-# from ARMv5* or ARMv6.
-define fgroup ARMv6m mode32 armv3m armv4 thumb armv5 armv5e armv6
-# This is suspect, the 'common' ARMv7 subset excludes the thumb2 'DSP' and
-# integer SIMD instructions that are in ARMv6T2. */
-define fgroup ARMv7 ARMv6m thumb2 armv7
-
-define fgroup ARMv7a ARMv7 notm armv6k
-define fgroup ARMv7ve ARMv7a adiv tdiv lpae
-define fgroup ARMv7r ARMv7a tdiv
-define fgroup ARMv7m ARMv7 tdiv
-define fgroup ARMv7em ARMv7m armv7em
-define fgroup ARMv8a ARMv7ve armv8
-define fgroup ARMv8_1a ARMv8a crc32 armv8_1
-define fgroup ARMv8_2a ARMv8_1a armv8_2
-define fgroup ARMv8m_base ARMv6m armv8 cmse tdiv
-define fgroup ARMv8m_main ARMv7m armv8 cmse
-define fgroup ARMv8r ARMv8a
-
-# Useful combinations.
-define fgroup VFPv2 vfpv2
-define fgroup VFPv3 VFPv2 vfpv3
-define fgroup VFPv4 VFPv3 vfpv4 fp16conv
-define fgroup FPv5 VFPv4 fpv5
-
-define fgroup FP_DBL fp_dbl
-define fgroup FP_D32 FP_DBL fp_d32
-define fgroup FP_ARMv8 FPv5 FP_D32
-define fgroup NEON FP_D32 neon
-define fgroup CRYPTO NEON crypto
-
-# List of all quirk bits to strip out when comparing CPU features with
-# architectures.
-define fgroup ALL_QUIRKS quirk_no_volatile_ce quirk_armv6kz quirk_cm3_ldrd
-
# Architecture entries
# format:
# begin arch <name>
@@ -259,38 +55,38 @@ begin arch armv2
tune for arm2
tune flags CO_PROC NO_MODE32
base 2
- isa ARMv2 mode26
+ isa ARMv2 bit_mode26
end arch armv2
begin arch armv2a
tune for arm2
tune flags CO_PROC NO_MODE32
base 2
- isa ARMv2 mode26
+ isa ARMv2 bit_mode26
end arch armv2a
begin arch armv3
tune for arm6
tune flags CO_PROC
base 3
- isa ARMv3 mode26
+ isa ARMv3 bit_mode26
end arch armv3
begin arch armv3m
tune for arm7m
tune flags CO_PROC
base 3M
- isa ARMv3m mode26
+ isa ARMv3m bit_mode26
end arch armv3m
begin arch armv4
tune for arm7tdmi
tune flags CO_PROC
base 4
- isa ARMv4 mode26
+ isa ARMv4 bit_mode26
end arch armv4
-# Strictly, mode26 is a permitted option for v4t, but there are no
+# Strictly, bit_mode26 is a permitted option for v4t, but there are no
# implementations that support it, so we will leave it out for now.
begin arch armv4t
tune for arm7tdmi
@@ -448,14 +244,14 @@ begin arch armv7-a
option fp add VFPv3 FP_DBL
optalias vfpv3-d16 fp
option vfpv3 add VFPv3 FP_D32
- option vfpv3-d16-fp16 add VFPv3 FP_DBL fp16conv
- option vfpv3-fp16 add VFPv3 FP_DBL FP_D32 fp16conv
+ option vfpv3-d16-fp16 add VFPv3 FP_DBL bit_fp16conv
+ option vfpv3-fp16 add VFPv3 FP_DBL FP_D32 bit_fp16conv
option vfpv4-d16 add VFPv4 FP_DBL
option vfpv4 add VFPv4 FP_D32
option simd add VFPv3 NEON
optalias neon simd
optalias neon-vfpv3 simd
- option neon-fp16 add VFPv3 NEON fp16conv
+ option neon-fp16 add VFPv3 NEON bit_fp16conv
option neon-vfpv4 add VFPv4 NEON
option nosimd remove ALL_SIMD
option nofp remove ALL_FP
@@ -470,14 +266,14 @@ begin arch armv7ve
# fp => VFPv4-d16, simd => neon-vfpv4
option vfpv3-d16 add VFPv3 FP_DBL
option vfpv3 add VFPv3 FP_D32
- option vfpv3-d16-fp16 add VFPv3 FP_DBL fp16conv
- option vfpv3-fp16 add VFPv3 FP_DBL FP_D32 fp16conv
+ option vfpv3-d16-fp16 add VFPv3 FP_DBL bit_fp16conv
+ option vfpv3-fp16 add VFPv3 FP_DBL FP_D32 bit_fp16conv
option fp add VFPv4 FP_DBL
optalias vfpv4-d16 fp
option vfpv4 add VFPv4 FP_D32
option neon add VFPv3 NEON
optalias neon-vfpv3 neon
- option neon-fp16 add VFPv3 NEON fp16conv
+ option neon-fp16 add VFPv3 NEON bit_fp16conv
option simd add VFPv4 NEON
optalias neon-vfpv4 simd
option nosimd remove ALL_SIMD
@@ -495,9 +291,9 @@ begin arch armv7-r
optalias vfpv3xd fp.sp
option fp add VFPv3 FP_DBL
optalias vfpv3-d16 fp
- option idiv add adiv
+ option idiv add bit_adiv
option nofp remove ALL_FP
- option noidiv remove adiv
+ option noidiv remove bit_adiv
end arch armv7-r
begin arch armv7-m
@@ -531,7 +327,7 @@ begin arch armv8-a
base 8A
profile A
isa ARMv8a
- option crc add crc32
+ option crc add bit_crc32
option simd add FP_ARMv8 NEON
option crypto add FP_ARMv8 CRYPTO
option nocrypto remove ALL_CRYPTO
@@ -557,7 +353,7 @@ begin arch armv8.2-a
profile A
isa ARMv8_2a
option simd add FP_ARMv8 NEON
- option fp16 add fp16 FP_ARMv8 NEON
+ option fp16 add bit_fp16 FP_ARMv8 NEON
option crypto add FP_ARMv8 CRYPTO
option nocrypto remove ALL_CRYPTO
option nofp remove ALL_FP
@@ -576,12 +372,12 @@ begin arch armv8-m.main
base 8M_MAIN
profile M
isa ARMv8m_main
- option dsp add armv7em
+ option dsp add bit_ARMv7em
# fp => FPv5-sp-d16; fp.dp => FPv5-d16
option fp add FPv5
option fp.dp add FPv5 FP_DBL
option nofp remove ALL_FP
- option nodsp remove armv7em
+ option nodsp remove bit_ARMv7em
end arch armv8-m.main
begin arch armv8-r
@@ -590,7 +386,7 @@ begin arch armv8-r
base 8R
profile R
isa ARMv8r
- option crc add crc32
+ option crc add bit_crc32
# fp.sp => fp-armv8 (d16); simd => simd + fp-armv8 + d32 + double precision
# note: no fp option for fp-armv8 (d16) + double precision at the moment
option fp.sp add FPv5
@@ -604,14 +400,14 @@ begin arch iwmmxt
tune for iwmmxt
tune flags LDSCHED STRONG XSCALE
base 5TE
- isa ARMv5te xscale iwmmxt
+ isa ARMv5te bit_xscale bit_iwmmxt
end arch iwmmxt
begin arch iwmmxt2
tune for iwmmxt2
tune flags LDSCHED STRONG XSCALE
base 5TE
- isa ARMv5te xscale iwmmxt iwmmxt2
+ isa ARMv5te bit_xscale bit_iwmmxt bit_iwmmxt2
end arch iwmmxt2
# CPU entries
@@ -987,7 +783,7 @@ end cpu arm1022e
begin cpu xscale
tune flags LDSCHED XSCALE
architecture armv5te
- isa xscale
+ isa bit_xscale
costs xscale
end cpu xscale
@@ -1165,14 +961,14 @@ begin cpu generic-armv7-a
fpu vfpv3-d16
option vfpv3-d16 add VFPv3 FP_DBL
option vfpv3 add VFPv3 FP_D32
- option vfpv3-d16-fp16 add VFPv3 FP_DBL fp16conv
- option vfpv3-fp16 add VFPv3 FP_D32 fp16conv
+ option vfpv3-d16-fp16 add VFPv3 FP_DBL bit_fp16conv
+ option vfpv3-fp16 add VFPv3 FP_D32 bit_fp16conv
option vfpv4-d16 add VFPv4 FP_DBL
option vfpv4 add VFPv4 FP_D32
option simd add VFPv3 NEON
optalias neon simd
optalias neon-vfpv3 simd
- option neon-fp16 add VFPv3 NEON fp16conv
+ option neon-fp16 add VFPv3 NEON bit_fp16conv
option neon-vfpv4 add VFPv4 NEON
option nosimd remove ALL_SIMD
option nofp remove ALL_FP
@@ -1548,7 +1344,7 @@ begin fpu vfpv3
end fpu vfpv3
begin fpu vfpv3-fp16
- isa VFPv3 FP_D32 fp16conv
+ isa VFPv3 FP_D32 bit_fp16conv
end fpu vfpv3-fp16
begin fpu vfpv3-d16
@@ -1556,7 +1352,7 @@ begin fpu vfpv3-d16
end fpu vfpv3-d16
begin fpu vfpv3-d16-fp16
- isa VFPv3 FP_DBL fp16conv
+ isa VFPv3 FP_DBL bit_fp16conv
end fpu vfpv3-d16-fp16
begin fpu vfpv3xd
@@ -1564,7 +1360,7 @@ begin fpu vfpv3xd
end fpu vfpv3xd
begin fpu vfpv3xd-fp16
- isa VFPv3 fp16conv
+ isa VFPv3 bit_fp16conv
end fpu vfpv3xd-fp16
begin fpu neon
@@ -1576,7 +1372,7 @@ begin fpu neon-vfpv3
end fpu neon-vfpv3
begin fpu neon-fp16
- isa VFPv3 NEON fp16conv
+ isa VFPv3 NEON bit_fp16conv
end fpu neon-fp16
begin fpu vfpv4
diff --git a/gcc/config/arm/arm-isa.h b/gcc/config/arm/arm-isa.h
new file mode 100644
index 0000000..dbd29ea
--- /dev/null
+++ b/gcc/config/arm/arm-isa.h
@@ -0,0 +1,172 @@
+/* ISA feature bits for ARM.
+ Copyright (C) 2016-2017 Free Software Foundation, Inc.
+ Contributed by ARM Ltd.
+
+ 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.
+
+ Under Section 7 of GPL version 3, you are granted additional
+ permissions described in the GCC Runtime Library Exception, version
+ 3.1, as published by the Free Software Foundation.
+
+ You should have received a copy of the GNU General Public License and
+ a copy of the GCC Runtime Library Exception along with this program;
+ see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
+ <http://www.gnu.org/licenses/>. */
+
+#ifndef ARM_ISA_FEATURE_H
+#define ARM_ISA_FEATURE_H
+
+enum isa_feature
+ {
+ isa_nobit, /* Must be first. */
+ isa_bit_ARMv3m, /* Extended multiply. */
+ isa_bit_mode26, /* 26-bit mode support. */
+ isa_bit_mode32, /* 32-bit mode support. */
+ isa_bit_ARMv4, /* Architecture rel 4. */
+ isa_bit_ARMv5, /* Architecture rel 5. */
+ isa_bit_thumb, /* Thumb aware. */
+ isa_bit_ARMv5e, /* Architecture rel 5e. */
+ isa_bit_xscale, /* XScale. */
+ isa_bit_ARMv6, /* Architecture rel 6. */
+ isa_bit_ARMv6k, /* Architecture rel 6k. */
+ isa_bit_thumb2, /* Thumb-2. */
+ isa_bit_notm, /* Instructions not present in 'M' profile. */
+ isa_bit_be8, /* Architecture uses be8 mode in big-endian. */
+ isa_bit_tdiv, /* Thumb division instructions. */
+ isa_bit_ARMv7em, /* Architecture rel 7e-m. */
+ isa_bit_ARMv7, /* Architecture rel 7. */
+ isa_bit_adiv, /* ARM division instructions. */
+ isa_bit_ARMv8, /* Architecture rel 8. */
+ isa_bit_crc32, /* ARMv8 CRC32 instructions. */
+ isa_bit_iwmmxt, /* XScale v2 (Wireless MMX). */
+ isa_bit_iwmmxt2, /* XScale Wireless MMX2. */
+ isa_bit_ARMv8_1, /* Architecture rel 8.1. */
+ isa_bit_ARMv8_2, /* Architecutre rel 8.2. */
+ isa_bit_cmse, /* M-Profile security extensions. */
+ /* Floating point and Neon extensions. */
+ /* VFPv1 is not supported in GCC. */
+ isa_bit_VFPv2, /* Vector floating point v2. */
+ isa_bit_VFPv3, /* Vector floating point v3. */
+ isa_bit_VFPv4, /* Vector floating point v4. */
+ isa_bit_FPv5, /* Floating point v5. */
+ isa_bit_lpae, /* ARMv7-A LPAE. */
+ isa_bit_neon, /* Advanced SIMD instructions. */
+ isa_bit_fp16conv, /* Conversions to/from fp16 (VFPv3 extension). */
+ isa_bit_fp_dbl, /* Double precision operations supported. */
+ isa_bit_fp_d32, /* 32 Double precision registers. */
+ isa_bit_crypto, /* Crypto extension to ARMv8. */
+ isa_bit_fp16, /* FP16 data processing (half-precision float). */
+
+ /* ISA Quirks (errata?). Don't forget to add this to the list of
+ all quirks below. */
+ isa_quirk_no_volatile_ce, /* No volatile memory in IT blocks. */
+ isa_quirk_ARMv6kz, /* Previously mis-identified by GCC. */
+ isa_quirk_cm3_ldrd, /* Cortex-M3 LDRD quirk. */
+
+ /* Aren't currently, but probably should be tuning bits. */
+ isa_bit_smallmul, /* Slow multiply operations. */
+
+ /* Tuning bits. Should be elsewhere. */
+ isa_tune_co_proc, /* Has co-processor bus. */
+ isa_tune_ldsched, /* Load scheduling necessary. */
+ isa_tune_strong, /* StrongARM. */
+ isa_tune_wbuf, /* Schedule for write buffer ops (ARM6 & 7 only). */
+
+ /* Must be last, used to dimension arrays. */
+ isa_num_bits
+ };
+
+/* Helper macros for use when defining CPUs and architectures.
+
+ There must be no parenthesees in these lists, since they are used
+ to initialize arrays. */
+
+#define ISA_ARMv2 isa_bit_notm
+#define ISA_ARMv3 ISA_ARMv2, isa_bit_mode32
+#define ISA_ARMv3m ISA_ARMv3, isa_bit_ARMv3m
+#define ISA_ARMv4 ISA_ARMv3m, isa_bit_ARMv4
+#define ISA_ARMv4t ISA_ARMv4, isa_bit_thumb
+#define ISA_ARMv5 ISA_ARMv4, isa_bit_ARMv5
+#define ISA_ARMv5t ISA_ARMv5, isa_bit_thumb
+#define ISA_ARMv5e ISA_ARMv5, isa_bit_ARMv5e
+#define ISA_ARMv5te ISA_ARMv5e, isa_bit_thumb
+#define ISA_ARMv5tej ISA_ARMv5te
+#define ISA_ARMv6 ISA_ARMv5te, isa_bit_ARMv6, isa_bit_be8
+#define ISA_ARMv6j ISA_ARMv6
+#define ISA_ARMv6k ISA_ARMv6, isa_bit_ARMv6k
+#define ISA_ARMv6z ISA_ARMv6
+#define ISA_ARMv6kz ISA_ARMv6k, isa_quirk_ARMv6kz
+#define ISA_ARMv6zk ISA_ARMv6k
+#define ISA_ARMv6t2 ISA_ARMv6, isa_bit_thumb2
+
+/* This is suspect. ARMv6-m doesn't really pull in any useful features
+ from ARMv5* or ARMv6. */
+#define ISA_ARMv6m isa_bit_mode32, isa_bit_ARMv3m, isa_bit_ARMv4, \
+ isa_bit_thumb, isa_bit_ARMv5, isa_bit_ARMv5e, isa_bit_ARMv6
+/* This is suspect, the 'common' ARMv7 subset excludes the thumb2 'DSP' and
+ integer SIMD instructions that are in ARMv6T2. */
+#define ISA_ARMv7 ISA_ARMv6m, isa_bit_thumb2, isa_bit_ARMv7
+#define ISA_ARMv7a ISA_ARMv7, isa_bit_notm, isa_bit_ARMv6k
+#define ISA_ARMv7ve ISA_ARMv7a, isa_bit_adiv, isa_bit_tdiv, isa_bit_lpae
+#define ISA_ARMv7r ISA_ARMv7a, isa_bit_tdiv
+#define ISA_ARMv7m ISA_ARMv7, isa_bit_tdiv
+#define ISA_ARMv7em ISA_ARMv7m, isa_bit_ARMv7em
+#define ISA_ARMv8a ISA_ARMv7ve, isa_bit_ARMv8
+#define ISA_ARMv8_1a ISA_ARMv8a, isa_bit_crc32, isa_bit_ARMv8_1
+#define ISA_ARMv8_2a ISA_ARMv8_1a, isa_bit_ARMv8_2
+#define ISA_ARMv8m_base ISA_ARMv6m, isa_bit_ARMv8, isa_bit_cmse, isa_bit_tdiv
+#define ISA_ARMv8m_main ISA_ARMv7m, isa_bit_ARMv8, isa_bit_cmse
+#define ISA_ARMv8r ISA_ARMv8a
+
+/* List of all cryptographic extensions to stripout if crypto is
+ disabled. Currently, that's trivial, but we define it anyway for
+ consistency with the SIMD and FP disable lists. */
+#define ISA_ALL_CRYPTO isa_bit_crypto
+
+/* List of all SIMD bits to strip out if SIMD is disabled. This does
+ strip off 32 D-registers, but does not remove support for
+ double-precision FP. */
+#define ISA_ALL_SIMD isa_bit_fp_d32, isa_bit_neon, ISA_ALL_CRYPTO
+
+/* List of all FPU bits to strip out if -mfpu is used to override the
+ default. isa_bit_fp16 is deliberately missing from this list. */
+#define ISA_ALL_FPU_INTERNAL \
+ isa_bit_VFPv2, isa_bit_VFPv3, isa_bit_VFPv4, isa_bit_FPv5, \
+ isa_bit_fp16conv, isa_bit_fp_dbl, ISA_ALL_SIMD
+
+/* Similarly, but including fp16 and other extensions that aren't part of
+ -mfpu support. */
+#define ISA_ALL_FP isa_bit_fp16, ISA_ALL_FPU_INTERNAL
+
+/* Useful combinations. */
+#define ISA_VFPv2 isa_bit_VFPv2
+#define ISA_VFPv3 ISA_VFPv2, isa_bit_VFPv3
+#define ISA_VFPv4 ISA_VFPv3, isa_bit_VFPv4, isa_bit_fp16conv
+#define ISA_FPv5 ISA_VFPv4, isa_bit_FPv5
+
+#define ISA_FP_DBL isa_bit_fp_dbl
+#define ISA_FP_D32 ISA_FP_DBL, isa_bit_fp_d32
+#define ISA_FP_ARMv8 ISA_FPv5, ISA_FP_D32
+#define ISA_NEON ISA_FP_D32, isa_bit_neon
+#define ISA_CRYPTO ISA_NEON, isa_bit_crypto
+
+/* List of all quirk bits to strip out when comparing CPU features with
+ architectures. */
+#define ISA_ALL_QUIRKS isa_quirk_no_volatile_ce, isa_quirk_ARMv6kz, \
+ isa_quirk_cm3_ldrd
+
+/* Helper macro so that we can concatenate multiple features together
+ with arm-*.def files, since macro substitution can't have commas within an
+ argument that lacks parenthesis. */
+#define ISA_FEAT(X) X,
+#endif
diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c
index bca8a34..f0e7788 100644
--- a/gcc/config/arm/arm.c
+++ b/gcc/config/arm/arm.c
@@ -3361,22 +3361,22 @@ arm_option_override (void)
/* Initialize boolean versions of the architectural flags, for use
in the arm.md file. */
- arm_arch3m = bitmap_bit_p (arm_active_target.isa, isa_bit_armv3m);
- arm_arch4 = bitmap_bit_p (arm_active_target.isa, isa_bit_armv4);
+ arm_arch3m = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv3m);
+ arm_arch4 = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv4);
arm_arch4t = arm_arch4 && bitmap_bit_p (arm_active_target.isa, isa_bit_thumb);
- arm_arch5 = bitmap_bit_p (arm_active_target.isa, isa_bit_armv5);
- arm_arch5e = bitmap_bit_p (arm_active_target.isa, isa_bit_armv5e);
+ arm_arch5 = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv5);
+ arm_arch5e = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv5e);
arm_arch5te = arm_arch5e
&& bitmap_bit_p (arm_active_target.isa, isa_bit_thumb);
- arm_arch6 = bitmap_bit_p (arm_active_target.isa, isa_bit_armv6);
- arm_arch6k = bitmap_bit_p (arm_active_target.isa, isa_bit_armv6k);
+ arm_arch6 = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv6);
+ arm_arch6k = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv6k);
arm_arch_notm = bitmap_bit_p (arm_active_target.isa, isa_bit_notm);
arm_arch6m = arm_arch6 && !arm_arch_notm;
- arm_arch7 = bitmap_bit_p (arm_active_target.isa, isa_bit_armv7);
- arm_arch7em = bitmap_bit_p (arm_active_target.isa, isa_bit_armv7em);
- arm_arch8 = bitmap_bit_p (arm_active_target.isa, isa_bit_armv8);
- arm_arch8_1 = bitmap_bit_p (arm_active_target.isa, isa_bit_armv8_1);
- arm_arch8_2 = bitmap_bit_p (arm_active_target.isa, isa_bit_armv8_2);
+ arm_arch7 = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv7);
+ arm_arch7em = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv7em);
+ arm_arch8 = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv8);
+ arm_arch8_1 = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv8_1);
+ arm_arch8_2 = bitmap_bit_p (arm_active_target.isa, isa_bit_ARMv8_2);
arm_arch_thumb1 = bitmap_bit_p (arm_active_target.isa, isa_bit_thumb);
arm_arch_thumb2 = bitmap_bit_p (arm_active_target.isa, isa_bit_thumb2);
arm_arch_xscale = bitmap_bit_p (arm_active_target.isa, isa_bit_xscale);
@@ -3406,9 +3406,9 @@ arm_option_override (void)
/* And finally, set up some quirks. */
arm_arch_no_volatile_ce
- = bitmap_bit_p (arm_active_target.isa, isa_bit_quirk_no_volatile_ce);
- arm_arch6kz = arm_arch6k && bitmap_bit_p (arm_active_target.isa,
- isa_bit_quirk_armv6kz);
+ = bitmap_bit_p (arm_active_target.isa, isa_quirk_no_volatile_ce);
+ arm_arch6kz
+ = arm_arch6k && bitmap_bit_p (arm_active_target.isa, isa_quirk_ARMv6kz);
/* V5 code we generate is completely interworking capable, so we turn off
TARGET_INTERWORK here to avoid many tests later on. */
@@ -3453,7 +3453,7 @@ arm_option_override (void)
else if (TARGET_HARD_FLOAT_ABI)
{
arm_pcs_default = ARM_PCS_AAPCS_VFP;
- if (!bitmap_bit_p (arm_active_target.isa, isa_bit_vfpv2))
+ if (!bitmap_bit_p (arm_active_target.isa, isa_bit_VFPv2))
error ("-mfloat-abi=hard: selected processor lacks an FPU");
}
else
@@ -3556,7 +3556,7 @@ arm_option_override (void)
/* Enable -mfix-cortex-m3-ldrd by default for Cortex-M3 cores. */
if (fix_cm3_ldrd == 2)
{
- if (bitmap_bit_p (arm_active_target.isa, isa_bit_quirk_cm3_ldrd))
+ if (bitmap_bit_p (arm_active_target.isa, isa_quirk_cm3_ldrd))
fix_cm3_ldrd = 1;
else
fix_cm3_ldrd = 0;
diff --git a/gcc/config/arm/arm.h b/gcc/config/arm/arm.h
index 2d71e8f..5fdb65b 100644
--- a/gcc/config/arm/arm.h
+++ b/gcc/config/arm/arm.h
@@ -122,7 +122,7 @@ extern tree arm_fp16_type_node;
/* Use hardware floating point instructions. */
#define TARGET_HARD_FLOAT (arm_float_abi != ARM_FLOAT_ABI_SOFT \
&& bitmap_bit_p (arm_active_target.isa, \
- isa_bit_vfpv2))
+ isa_bit_VFPv2))
#define TARGET_SOFT_FLOAT (!TARGET_HARD_FLOAT)
/* User has permitted use of FP instructions, if they exist for this
target. */
@@ -169,10 +169,10 @@ extern tree arm_fp16_type_node;
#define TARGET_VFPD32 (bitmap_bit_p (arm_active_target.isa, isa_bit_fp_d32))
/* FPU supports VFPv3 instructions. */
-#define TARGET_VFP3 (bitmap_bit_p (arm_active_target.isa, isa_bit_vfpv3))
+#define TARGET_VFP3 (bitmap_bit_p (arm_active_target.isa, isa_bit_VFPv3))
/* FPU supports FPv5 instructions. */
-#define TARGET_VFP5 (bitmap_bit_p (arm_active_target.isa, isa_bit_fpv5))
+#define TARGET_VFP5 (bitmap_bit_p (arm_active_target.isa, isa_bit_FPv5))
/* FPU only supports VFP single-precision instructions. */
#define TARGET_VFP_SINGLE (!TARGET_VFP_DOUBLE)
@@ -194,7 +194,7 @@ extern tree arm_fp16_type_node;
(TARGET_HARD_FLOAT && (TARGET_FP16 && TARGET_VFP5))
/* FPU supports fused-multiply-add operations. */
-#define TARGET_FMA (bitmap_bit_p (arm_active_target.isa, isa_bit_vfpv4))
+#define TARGET_FMA (bitmap_bit_p (arm_active_target.isa, isa_bit_VFPv4))
/* FPU supports Crypto extensions. */
#define TARGET_CRYPTO (bitmap_bit_p (arm_active_target.isa, isa_bit_crypto))
diff --git a/gcc/config/arm/parsecpu.awk b/gcc/config/arm/parsecpu.awk
index 0b4fc68..070d193 100644
--- a/gcc/config/arm/parsecpu.awk
+++ b/gcc/config/arm/parsecpu.awk
@@ -22,7 +22,6 @@
# data: Print the standard 'C' data tables for the CPUs
# common-data: Print the 'C' data for shared driver/compiler files
# headers: Print the standard 'C' headers for the CPUs
-# isa: Generate the arm-isa.h header
# md: Print the machine description fragment
# opt: Print the option tables fragment
# chkcpu <name>: Checks that <name> is a valid CPU
@@ -32,8 +31,7 @@
function fatal (m) {
print "error ("lineno"): " m > "/dev/stderr"
- fatal_err = 1
- if (parse_done) exit 1
+ exit 1
}
function toplevel () {
@@ -85,44 +83,9 @@ function tune_flag_pfx (f) {
return "TF_" f
}
-# Print out the bits for the features in FLIST, which may be a
-# mixture of fgroup and individual bits. Print each feature needed
-# exactly once. Terminate the list with isa_nobit. Prefix each line by
-# INDENT. Does not print a new line at the end.
-function print_isa_bits_for (flist, indent) {
- nbits = split (flist, bits)
-
- for (bit = 1; bit <= nbits; bit++) {
- if (bits[bit] in features) {
- pbit[bits[bit]] = 1
- } else if (bits[bit] in fgroup) {
- for (gbits in fgrp_bits) {
- split (gbits, bitsep, SUBSEP)
- if (bitsep[1] == bits[bit]) {
- pbit[bitsep[2]] = 1
- }
- }
- } else fatal("feature " bits[bit] " not declared")
- }
- zbit = ORS
- ORS = ""
- print indent "{\n" indent " "
- ORS = ", "
- count = 0
- for (bname in pbit) {
- print "isa_bit_" bname
- count++
- if (count == 4) {
- count = 0
- ORS = ""
- print "\n" indent " "
- ORS = ", "
- }
- }
- ORS = ""
- print "isa_nobit\n" indent "}"
- ORS = zbit
- delete pbit
+function isa_pfx (f) {
+ if (f ~ /^(bit|quirk)_.*/) return "isa_" f
+ return "ISA_" f
}
function gen_headers () {
@@ -162,35 +125,6 @@ function gen_headers () {
print "};"
}
-function gen_isa () {
- boilerplate("C")
- print "enum isa_feature {"
- print " isa_nobit = 0,"
- for (fbit in features) {
- print " isa_bit_" fbit ","
- }
- print " isa_num_bits"
- print "};\n"
-
- for (fgrp in fgroup) {
- print "#define ISA_"fgrp " \\"
- z = ORS
- ORS = ""
- first = 1
- for (bitcomb in fgrp_bits) {
- split (bitcomb, bitsep, SUBSEP)
- if (bitsep[1] == fgrp) {
- if (first) {
- first = 0
- } else print ", \\\n"
- print " isa_bit_" bitsep[2]
- }
- }
- ORS = z
- print "\n"
- }
-}
-
function gen_data () {
boilerplate("C")
@@ -221,6 +155,7 @@ function gen_data () {
}
print " {TARGET_CPU_arm_none, 0, NULL}"
print "};"
+
}
function gen_comm_data () {
@@ -237,8 +172,8 @@ function gen_comm_data () {
print " {"
print " \"" opts[opt] "\", " \
cpu_opt_remove[cpus[n],opts[opt]] ", false,"
- print_isa_bits_for(cpu_opt_isa[cpus[n],opts[opt]], " ")
- print "\n },"
+ print " { " cpu_opt_isa[cpus[n],opts[opt]] ", isa_nobit }"
+ print " },"
}
if (cpus[n] in cpu_optaliases) {
naliases = split (cpu_optaliases[cpus[n]], aliases)
@@ -253,8 +188,8 @@ function gen_comm_data () {
print " {"
print " \"" aliases[alias] "\", " \
cpu_opt_remove[cpus[n],equiv] ", true, "
- print_isa_bits_for(cpu_opt_isa[cpus[n],equiv], " ")
- print "\n },"
+ print " { " cpu_opt_isa[cpus[n],equiv] ", isa_nobit }"
+ print " },"
}
}
print " { NULL, false, false, {isa_nobit}}"
@@ -279,7 +214,8 @@ function gen_comm_data () {
if (! (feats[1] in arch_isa)) {
fatal("unknown arch " feats[1] " for cpu " cpus[n])
}
- all_isa_bits = arch_isa[feats[1]]
+ print " {"
+ print " " arch_isa[feats[1]] ","
for (m = 2; m <= nfeats; m++) {
if (! ((feats[1], feats[m]) in arch_opt_isa)) {
fatal("unknown feature " feats[m] " for architecture " feats[1])
@@ -287,16 +223,42 @@ function gen_comm_data () {
if (arch_opt_remove[feats[1],feats[m]] == "true") {
fatal("cannot remove features from architecture specs")
}
- all_isa_bits = all_isa_bits " " arch_opt_isa[feats[1],feats[m]]
+ # The isa_features array that is being initialized here has a length
+ # of max isa_bit_num, which is the last entry in the enum.
+ # Logically this means that the number of features is implicitly
+ # never more than the number of feature bits we have. This is only
+ # true if we don't emit duplicates here. So keep track of which
+ # options we have already emitted so we don't emit them twice.
+ nopts = split (arch_opt_isa[feats[1],feats[m]], opts, ",")
+ for (i = 1; i <= nopts; i++) {
+ if (! (opts[i] in seen)) {
+ print " " opts[i] ","
+ seen[opts[i]]
+ }
+ }
}
if (cpus[n] in cpu_fpu) {
- all_isa_bits = all_isa_bits " " fpu_isa[cpu_fpu[cpus[n]]]
+ nopts = split (fpu_isa[cpu_fpu[cpus[n]]], opts, ",")
+ for (i = 1; i <= nopts; i++) {
+ if (! (opts[i] in seen)) {
+ print " " opts[i] ","
+ seen[opts[i]]
+ }
+ }
}
if (cpus[n] in cpu_isa) {
- all_isa_bits = all_isa_bits " " cpu_isa[cpus[n]]
+ nopts = split (cpu_isa[cpus[n]], opts, ",")
+ for (i = 1; i <= nopts; i++) {
+ if (! (opts[i] in seen)) {
+ print " " opts[i] ","
+ seen[opts[i]]
+ }
+ }
}
- print_isa_bits_for(all_isa_bits, " ")
- print "\n },"
+ delete seen
+ print " isa_nobit"
+ print " }"
+ print " },"
# arch
print " TARGET_ARCH_" arch_cnames[feats[1]]
print " },"
@@ -316,8 +278,8 @@ function gen_comm_data () {
print " {"
print " \"" opts[opt] "\", " \
arch_opt_remove[archs[n],opts[opt]] ", false,"
- print_isa_bits_for(arch_opt_isa[archs[n],opts[opt]], " ")
- print "\n },"
+ print " { " arch_opt_isa[archs[n],opts[opt]] ", isa_nobit }"
+ print " },"
}
if (archs[n] in arch_optaliases) {
naliases = split (arch_optaliases[archs[n]], aliases)
@@ -332,8 +294,8 @@ function gen_comm_data () {
print " {"
print " \"" aliases[alias] "\", " \
arch_opt_remove[archs[n],equiv] ", true, "
- print_isa_bits_for(arch_opt_isa[archs[n],equiv], " ")
- print "\n },"
+ print " { " arch_opt_isa[archs[n],equiv] ", isa_nobit }"
+ print " },"
}
}
print " { NULL, false, false, {isa_nobit}}"
@@ -359,8 +321,10 @@ function gen_comm_data () {
print " arch_opttab_" arch_cnames[archs[n]] ","
} else print " NULL,"
# common.isa_bits
- print_isa_bits_for(arch_isa[archs[n]], " ")
- print ","
+ print " {"
+ print " " arch_isa[archs[n]] ","
+ print " isa_nobit"
+ print " },"
# arch, base_arch
print " \"" arch_base[archs[n]] "\", BASE_ARCH_" \
arch_base[archs[n]] ","
@@ -387,8 +351,11 @@ function gen_comm_data () {
for (n = 1; n <= nfpus; n++) {
print " {"
print " \"" fpus[n] "\","
- print_isa_bits_for(fpu_isa[fpus[n]], " ")
- print "\n },"
+ print " {"
+ print " " fpu_isa[fpus[n]] ","
+ print " isa_nobit"
+ print " }"
+ print " },"
}
print "};"
@@ -503,68 +470,25 @@ BEGIN {
arch_name = ""
fpu_name = ""
lineno = 0
- fatal_err = 0
- parse_done = 0
if (cmd == "") fatal("Usage parsecpu.awk -v cmd=<xyz>")
}
-# New line. Reset parse status and increment line count for error messages
// {
lineno++
parse_ok = 0
}
-# Comments must be on a line on their own.
/^#/ {
parse_ok = 1
}
-/^define feature / {
- if (NF != 3) fatal("syntax: define feature <name>")
- toplevel()
- fbit = $3
- if (fbit in features) fatal("feature " fbit " already defined")
- features[fbit] = 1
- parse_ok = 1
-}
-
-/^define fgroup / {
- if (NF < 4) fatal("syntax: define fgroup <name> <feature> [<feature>]*")
- toplevel()
- fgrp = $3
- if (fgrp in fgroup) fatal("feature group " fgrp " already defined")
- if (fgrp in features) fatal("feature group " fgrp " aliases a feature")
- fcount = NF
- for (n = 4; n <= fcount; n++) {
- feat = $n
- if (feat in features) {
- fgrp_bits[fgrp,feat] = 1
- } else if (feat in fgroup) {
- # fgroups may reference other fgroups, copy their bits
- # to our bits. To avoid recursion we don't set fgroup[fgrp]
- # until after we have done this, so such attempts will result
- # in an invalid group definition.
- for (bitcomb in fgrp_bits) {
- split (bitcomb, bitsep, SUBSEP)
- if (bitsep[1] == feat) {
- fgrp_bits[fgrp,bitsep[2]] = 1
- }
- }
- } else fatal("feature group member " feat " unrecognized")
- }
- fgroup[fgrp] = 1
- parse_ok = 1
-}
-
/^begin fpu / {
- if (NF != 3) fatal("syntax: begin fpu <name>")
toplevel()
fpu_name = $3
parse_ok = 1
}
/^end fpu / {
- if (NF != 3) fatal("syntax: end fpu <name>")
if (fpu_name != $3) fatal("mimatched end fpu")
if (! (fpu_name in fpu_isa)) {
fatal("fpu definition \"" fpu_name "\" lacks an \"isa\" statement")
@@ -577,28 +501,24 @@ BEGIN {
}
/^begin arch / {
- if (NF != 3) fatal("syntax: begin arch <name>")
toplevel()
arch_name = $3
parse_ok = 1
}
/^[ ]*base / {
- if (NF != 2) fatal("syntax: base <architecture-base-name>")
if (arch_name == "") fatal("\"base\" statement outside of arch block")
arch_base[arch_name] = $2
parse_ok = 1
}
/^[ ]*profile / {
- if (NF != 2) fatal("syntax: profile <profile-name>")
if (arch_name == "") fatal("\"profile\" statement outside of arch block")
arch_prof[arch_name] = $2
parse_ok = 1
}
/^end arch / {
- if (NF != 3) fatal("syntax: end arch <name>")
if (arch_name != $3) fatal("mimatched end arch")
if (! arch_name in arch_tune_for) {
fatal("arch definition lacks a \"tune for\" statement")
@@ -614,21 +534,18 @@ BEGIN {
}
/^begin cpu / {
- if (NF != 3) fatal("syntax: begin cpu <name>")
toplevel()
cpu_name = $3
parse_ok = 1
}
/^[ ]*cname / {
- if (NF != 2) fatal("syntax: cname <identifier>")
if (cpu_name == "") fatal("\"cname\" outside of cpu block")
cpu_cnames[cpu_name] = $2
parse_ok = 1
}
/^[ ]*tune for / {
- if (NF != 3) fatal("syntax: tune for <cpu-name>")
if (cpu_name != "") {
cpu_tune_for[cpu_name] = $3
} else if (arch_name != "") {
@@ -638,7 +555,6 @@ BEGIN {
}
/^[ ]*tune flags / {
- if (NF < 3) fatal("syntax: tune flags <flag> [<flag>]*")
flags=""
flag_count = NF
for (n = 3; n <= flag_count; n++) {
@@ -655,27 +571,24 @@ BEGIN {
}
/^[ ]*architecture / {
- if (NF != 2) fatal("syntax: architecture <arch-name>")
if (cpu_name == "") fatal("\"architecture\" outside of cpu block")
cpu_arch[cpu_name] = $2
parse_ok = 1
}
/^[ ]*fpu / {
- if (NF != 2) fatal("syntax: fpu <fpu-name>")
if (cpu_name == "") fatal("\"fpu\" outside of cpu block")
cpu_fpu[cpu_name] = $2
parse_ok = 1
}
/^[ ]*isa / {
- if (NF < 2) fatal("syntax: isa <feature-or-fgroup> [<feature-or-fgroup>]*")
flags=""
flag_count = NF
for (n = 2; n <= flag_count; n++) {
if (n == 2) {
- flags = $n
- } else flags = flags " " $n
+ flags = isa_pfx($n)
+ } else flags = flags "," isa_pfx($n)
}
if (cpu_name != "") {
cpu_isa[cpu_name] = flags
@@ -688,7 +601,6 @@ BEGIN {
}
/^[ ]*option / {
- if (NF < 4) fatal("syntax: option <name> add|remove <feature-or-fgroup>+")
name=$2
if ($3 == "add") {
remove = "false"
@@ -699,8 +611,8 @@ BEGIN {
flag_count = NF
for (n = 4; n <= flag_count; n++) {
if (n == 4) {
- flags = $n
- } else flags = flags " " $n
+ flags = isa_pfx($n)
+ } else flags = flags "," isa_pfx($n)
}
if (cpu_name != "") {
cpu_opts[cpu_name] = cpu_opts[cpu_name] " " name
@@ -715,7 +627,6 @@ BEGIN {
}
/^[ ]*optalias / {
- if (NF != 3) fatal("syntax: optalias <name> <option-name>")
name=$2
alias=$3
if (cpu_name != "") {
@@ -729,14 +640,12 @@ BEGIN {
}
/^[ ]*costs / {
- if (NF != 2) fatal("syntax: costs <identifier>")
if (cpu_name == "") fatal("\"costs\" outside of cpu block")
cpu_cost[cpu_name] = $2
parse_ok = 1
}
/^end cpu / {
- if (NF != 3) fatal("syntax: end cpu <name>")
if (cpu_name != $3) fatal("mimatched end cpu")
if (! (cpu_name in cpu_cnames)) {
cpu_cnames[cpu_name] = cpu_name
@@ -753,8 +662,6 @@ BEGIN {
}
END {
- parse_done = 1
- if (fatal_err) exit 1
toplevel()
if (cmd == "data") {
gen_data()
@@ -762,8 +669,6 @@ END {
gen_comm_data()
} else if (cmd == "headers") {
gen_headers()
- } else if (cmd == "isa") {
- gen_isa()
} else if (cmd == "md") {
gen_md()
} else if (cmd == "opt") {
diff --git a/gcc/config/arm/t-arm b/gcc/config/arm/t-arm
index 2cd472a..16177e0 100644
--- a/gcc/config/arm/t-arm
+++ b/gcc/config/arm/t-arm
@@ -20,7 +20,6 @@
TM_H += arm-cpu.h
GTM_H += arm-cpu.h
-OPTIONS_H_EXTRA += arm-isa.h
# All md files - except for arm.md.
# This list should be kept in alphabetical order and updated whenever an md
@@ -88,14 +87,6 @@ s-arm-cpu: $(srcdir)/config/arm/parsecpu.awk \
$(SHELL) $(srcdir)/../move-if-change tmp-arm-cpu.h arm-cpu.h
$(STAMP) s-arm-cpu
-arm-isa.h: s-arm-isa ; @true
-s-arm-isa: $(srcdir)/config/arm/parsecpu.awk \
- $(srcdir)/config/arm/arm-cpus.in
- $(AWK) -f $(srcdir)/config/arm/parsecpu.awk -v cmd=isa \
- $(srcdir)/config/arm/arm-cpus.in > tmp-arm-isa.h
- $(SHELL) $(srcdir)/../move-if-change tmp-arm-isa.h arm-isa.h
- $(STAMP) s-arm-isa
-
arm-cpu-data.h: s-arm-data ; @true
s-arm-data: $(srcdir)/config/arm/parsecpu.awk \
$(srcdir)/config/arm/arm-cpus.in