From patchwork Fri Aug 26 18:18:30 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ulrich Weigand X-Patchwork-Id: 14953 Received: (qmail 125669 invoked by alias); 26 Aug 2016 18:18:50 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Delivered-To: mailing list gdb-patches@sourceware.org Received: (qmail 125655 invoked by uid 89); 26 Aug 2016 18:18:49 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.3 required=5.0 tests=BAYES_00, KAM_ASCII_DIVIDERS, RCVD_IN_DNSWL_LOW, RCVD_IN_SEMBACKSCATTER, SPF_PASS autolearn=no version=3.3.2 spammy=19700101, 1970-01-01, 128bit, 128-bit X-HELO: mx0a-001b2d01.pphosted.com Received: from mx0a-001b2d01.pphosted.com (HELO mx0a-001b2d01.pphosted.com) (148.163.156.1) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Fri, 26 Aug 2016 18:18:39 +0000 Received: from pps.filterd (m0098410.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.16.0.11/8.16.0.11) with SMTP id u7QIEASR087398 for ; Fri, 26 Aug 2016 14:18:38 -0400 Received: from e06smtp07.uk.ibm.com (e06smtp07.uk.ibm.com [195.75.94.103]) by mx0a-001b2d01.pphosted.com with ESMTP id 252b0dswad-1 (version=TLSv1.2 cipher=AES256-SHA bits=256 verify=NOT) for ; Fri, 26 Aug 2016 14:18:38 -0400 Received: from localhost by e06smtp07.uk.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Fri, 26 Aug 2016 19:18:34 +0100 Received: from d06dlp02.portsmouth.uk.ibm.com (9.149.20.14) by e06smtp07.uk.ibm.com (192.168.101.137) with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted; Fri, 26 Aug 2016 19:18:33 +0100 X-IBM-Helo: d06dlp02.portsmouth.uk.ibm.com X-IBM-MailFrom: uweigand@de.ibm.com X-IBM-RcptTo: gdb-patches@sourceware.org Received: from b06cxnps3074.portsmouth.uk.ibm.com (d06relay09.portsmouth.uk.ibm.com [9.149.109.194]) by d06dlp02.portsmouth.uk.ibm.com (Postfix) with ESMTP id D8B4E219004D for ; Fri, 26 Aug 2016 19:17:55 +0100 (BST) Received: from d06av03.portsmouth.uk.ibm.com (d06av03.portsmouth.uk.ibm.com [9.149.37.213]) by b06cxnps3074.portsmouth.uk.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id u7QIIW9U18940104 for ; Fri, 26 Aug 2016 18:18:32 GMT Received: from d06av03.portsmouth.uk.ibm.com (localhost [127.0.0.1]) by d06av03.portsmouth.uk.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id u7QIIWTw013914 for ; Fri, 26 Aug 2016 12:18:32 -0600 Received: from oc7340732750.ibm.com (icon-9-167-240-195.megacenter.de.ibm.com [9.167.240.195]) by d06av03.portsmouth.uk.ibm.com (8.14.4/8.14.4/NCO v10.0 AVin) with ESMTP id u7QIIVQt013911; Fri, 26 Aug 2016 12:18:31 -0600 Received: by oc7340732750.ibm.com (Postfix, from userid 500) id CE9991309C; Fri, 26 Aug 2016 20:18:30 +0200 (CEST) Subject: [9/9 v2] Support 128-bit IEEE floating-point types on Intel and Power To: joseph@codesourcery.com (Joseph Myers) Date: Fri, 26 Aug 2016 20:18:30 +0200 (CEST) From: "Ulrich Weigand" Cc: gdb-patches@sourceware.org In-Reply-To: from "Joseph Myers" at Aug 25, 2016 11:27:45 PM MIME-Version: 1.0 X-TM-AS-MML: disable X-Content-Scanned: Fidelis XPS MAILER x-cbid: 16082618-0028-0000-0000-000002081E11 X-IBM-AV-DETECTION: SAVI=unused REMOTE=unused XFE=unused x-cbparentid: 16082618-0029-0000-0000-000020126DCE Message-Id: <20160826181830.CE9991309C@oc7340732750.ibm.com> X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10432:, , definitions=2016-08-26_13:, , signatures=0 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 spamscore=0 suspectscore=2 malwarescore=0 phishscore=0 adultscore=0 bulkscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.0.1-1604210000 definitions=main-1608260236 Joseph Myers wrote: > On Thu, 25 Aug 2016, Ulrich Weigand wrote: > > > Since at this point we do not yet have any indication in the debug info to > > distinguish different floating-point formats of the same length, we simply > > use the type name as hint. Types named "__float128" get the IEEE format. > > Note that now (GCC 7) __float128 is a built-in typedef (i.e. a name > registered with the register_builtin_type langhook) for _Float128, and > _Float128 is the primary name used when the type is created. If I build a > program with GCC 7 that uses __float128, the debug info I get references > _Float128 (and not __float128 as far as I can see). > > Obviously debugging should work whether someone uses _Float128 or > __float128 (and in C++, right now they can only use __float128, as the > keyword _Float128 is C-specific). Then on Power, _Float64x is also > binary128 when available (but is a distinct type at the C level from > _Float128), but on x86 it has the Intel extended format (but is a distinct > type from long double). Ah, good point, I hadn't tested against current GCC mainline. The version below adds support to handle _Float128 (on Power and Intel) and _Float64x (on Power only) as 128-bit IEEE format. Because the rest of the new types already match what GDB will choose by default for a type of their size, this means all of them should now work correctly. I've also added a test case to verify that all the supported _FloatN and _FloatNx types work correctly. This passes when using GCC 7 on both Power and Intel. (The __float128 test still passes when using GCC 6.) Tested on x64_64-linux and powerpc64le-linux. Bye, Ulrich ChangeLog: * i386-tdep.c (i386_floatformat_for_type): New function. (i386_gdbarch_init): Install it. * ppc-linux-tdep.c (ppc_floatformat_for_type): New function. (ppc_linux_init_abi): Install it. testsuite/ChangeLog: * gdb.base/float128.c: New file. * gdb.base/float128.exp: Likewise. * gdb.base/floatn.c: Likewise. * gdb.base/floatn.exp: Likewise. Index: binutils-gdb/gdb/i386-tdep.c =================================================================== --- binutils-gdb.orig/gdb/i386-tdep.c 2016-08-25 13:11:56.778869254 +0200 +++ binutils-gdb/gdb/i386-tdep.c 2016-08-26 19:47:23.246447990 +0200 @@ -8161,6 +8161,23 @@ i386_fast_tracepoint_valid_at (struct gd } } +/* Return a floating-point format for a floating-point variable of + length LEN in bits. If non-NULL, NAME is the name of its type. + If no suitable type is found, return NULL. */ + +const struct floatformat ** +i386_floatformat_for_type (struct gdbarch *gdbarch, + const char *name, int len) +{ + if (len == 128 && name) + if (strcmp (name, "__float128") == 0 + || strcmp (name, "_Float128") == 0 + || strcmp (name, "complex _Float128") == 0) + return floatformats_ia64_quad; + + return default_floatformat_for_type (gdbarch, name, len); +} + static int i386_validate_tdesc_p (struct gdbarch_tdep *tdep, struct tdesc_arch_data *tdesc_data) @@ -8370,6 +8387,9 @@ i386_gdbarch_init (struct gdbarch_info i alignment. */ set_gdbarch_long_double_bit (gdbarch, 96); + /* Support for floating-point data type variants. */ + set_gdbarch_floatformat_for_type (gdbarch, i386_floatformat_for_type); + /* Register numbers of various important registers. */ set_gdbarch_sp_regnum (gdbarch, I386_ESP_REGNUM); /* %esp */ set_gdbarch_pc_regnum (gdbarch, I386_EIP_REGNUM); /* %eip */ Index: binutils-gdb/gdb/ppc-linux-tdep.c =================================================================== --- binutils-gdb.orig/gdb/ppc-linux-tdep.c 2016-08-25 13:11:56.779869258 +0200 +++ binutils-gdb/gdb/ppc-linux-tdep.c 2016-08-26 19:14:45.210254378 +0200 @@ -1628,6 +1628,25 @@ ppc_init_linux_record_tdep (struct linux record_tdep->ioctl_FIOQSIZE = 0x40086680; } +/* Return a floating-point format for a floating-point variable of + length LEN in bits. If non-NULL, NAME is the name of its type. + If no suitable type is found, return NULL. */ + +const struct floatformat ** +ppc_floatformat_for_type (struct gdbarch *gdbarch, + const char *name, int len) +{ + if (len == 128 && name) + if (strcmp (name, "__float128") == 0 + || strcmp (name, "_Float128") == 0 + || strcmp (name, "_Float64x") == 0 + || strcmp (name, "complex _Float128") == 0 + || strcmp (name, "complex _Float64x") == 0) + return floatformats_ia64_quad; + + return default_floatformat_for_type (gdbarch, name, len); +} + static void ppc_linux_init_abi (struct gdbarch_info info, struct gdbarch *gdbarch) @@ -1651,6 +1670,9 @@ ppc_linux_init_abi (struct gdbarch_info set_gdbarch_long_double_bit (gdbarch, 16 * TARGET_CHAR_BIT); set_gdbarch_long_double_format (gdbarch, floatformats_ibm_long_double); + /* Support for floating-point data type variants. */ + set_gdbarch_floatformat_for_type (gdbarch, ppc_floatformat_for_type); + /* Handle inferior calls during interrupted system calls. */ set_gdbarch_write_pc (gdbarch, ppc_linux_write_pc); Index: binutils-gdb/gdb/testsuite/gdb.base/float128.c =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ binutils-gdb/gdb/testsuite/gdb.base/float128.c 2016-08-26 19:14:45.213254397 +0200 @@ -0,0 +1,30 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2016 Free Software Foundation, Inc. + + This program 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 of the License, or + (at your option) any later version. + + This program 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 this program. If not, see . */ + +#include +#include + +long double ld; +__float128 f128; + +int main() +{ + ld = 1.375l; + f128 = 2.375q; + + return 0; +} Index: binutils-gdb/gdb/testsuite/gdb.base/float128.exp =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ binutils-gdb/gdb/testsuite/gdb.base/float128.exp 2016-08-26 19:14:45.217254424 +0200 @@ -0,0 +1,76 @@ +# Copyright 2016 Free Software Foundation, Inc. + +# This program 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 of the License, or +# (at your option) any later version. +# +# This program 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 this program. If not, see . + +# This file is part of the gdb testsuite. It is intended to test that +# gdb could correctly handle floating point constant with a suffix. + +standard_testfile .c + +proc do_compile { {opts {}} } { + global srcdir subdir srcfile binfile + set ccopts {debug quiet} + foreach opt $opts {lappend ccopts "additional_flags=$opt"} + gdb_compile "${srcdir}/${subdir}/${srcfile}" "$binfile" executable $ccopts +} + +if { [do_compile] != "" && [do_compile {-mfloat128}] != "" } { + untested "compiler can't handle __float128 type?" + return -1 +} + +clean_restart ${binfile} + +if ![runto_main] then { + perror "couldn't run to breakpoint" + continue +} + +# Run to the breakpoint at return. +gdb_breakpoint [gdb_get_line_number "return"] +gdb_continue_to_breakpoint "return" + +# Print the original value of ld and f128 +gdb_test "print ld" ".* = 1\\.375.*" "The original value of ld is 1.375" +gdb_test "print f128" ".* = 2\\.375.*" "The original value of f128 is 2.375" + +# Test that gdb could correctly recognize float constant expression with a suffix. +# FIXME: gdb does not yet recognize the GNU extension 'q' suffix for __float128 constants. +gdb_test "print ld=-1.375l" ".* = -1\\.375.*" "Try to change ld to -1.375 with 'print ld=-1.375l'" +gdb_test "print f128=-2.375l" ".* = -2\\.375.*" "Try to change f128 to -2.375 with 'print f128=-2.375l'" + +# Test that gdb could handle the above correctly with "set var" command. +set test "set variable ld=10.375l" +gdb_test_multiple "set var ld=10.375l" "$test" { + -re "$gdb_prompt $" { + pass "$test" + } + -re "Invalid number.*$gdb_prompt $" { + fail "$test (do not recognize 10.375l)" + } +} + +set test "set variable f128=20.375l" +gdb_test_multiple "set var f128=20.375l" "$test" { + -re "$gdb_prompt $" { + pass "$test" + } + -re "Invalid number.*$gdb_prompt $" { + fail "$test (do not recognize 20.375l)" + } +} + +gdb_test "print ld" ".* = 10\\.375.*" "The value of ld is changed to 10.375" +gdb_test "print f128" ".* = 20\\.375.*" "The value of f128 is changed to 20.375" + Index: binutils-gdb/gdb/testsuite/gdb.base/floatn.c =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ binutils-gdb/gdb/testsuite/gdb.base/floatn.c 2016-08-26 19:14:45.221254450 +0200 @@ -0,0 +1,48 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2016 Free Software Foundation, Inc. + + This program 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 of the License, or + (at your option) any later version. + + This program 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 this program. If not, see . */ + +#include +#include + +_Float32 f32; +_Float64 f64; +_Float128 f128; +_Float32x f32x; +_Float64x f64x; + +_Complex _Float32 c32; +_Complex _Float64 c64; +_Complex _Float128 c128; +_Complex _Float32x c32x; +_Complex _Float64x c64x; + +int main() +{ + f32 = 1.5f32; + f64 = 2.25f64; + f128 = 3.375f128; + f32x = 10.5f32x; + f64x = 20.25f64x; + + c32 = 1.5f32 + 1.0if; + c64 = 2.25f64 + 1.0if; + c128 = 3.375f128 + 1.0if; + c32x = 10.5f32x + 1.0if; + c64x = 20.25f64x + 1.0if; + + return 0; +} Index: binutils-gdb/gdb/testsuite/gdb.base/floatn.exp =================================================================== --- /dev/null 1970-01-01 00:00:00.000000000 +0000 +++ binutils-gdb/gdb/testsuite/gdb.base/floatn.exp 2016-08-26 19:14:45.225254476 +0200 @@ -0,0 +1,124 @@ +# Copyright 2016 Free Software Foundation, Inc. + +# This program 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 of the License, or +# (at your option) any later version. +# +# This program 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 this program. If not, see . + +# This file is part of the gdb testsuite. It is intended to test that +# gdb could correctly handle floating point constant with a suffix. + +standard_testfile .c + +proc do_compile { {opts {}} } { + global srcdir subdir srcfile binfile + set ccopts {debug quiet} + foreach opt $opts {lappend ccopts "additional_flags=$opt"} + gdb_compile "${srcdir}/${subdir}/${srcfile}" "$binfile" executable $ccopts +} + +if { [do_compile] != "" && [do_compile {-mfloat128}] != "" } { + untested "compiler can't handle _FloatN/_FloatNx types?" + return -1 +} + +clean_restart ${binfile} + +if ![runto_main] then { + perror "couldn't run to breakpoint" + continue +} + +# Run to the breakpoint at return. +gdb_breakpoint [gdb_get_line_number "return"] +gdb_continue_to_breakpoint "return" + +# Print the original values of f32, f64, f128, f32x, f64x. +gdb_test "print f32" ".* = 1\\.5.*" "The original value of f32 is 1.5" +gdb_test "print f64" ".* = 2\\.25.*" "The original value of f64 is 2.25" +gdb_test "print f128" ".* = 3\\.375.*" "The original value of f128 is 3.375" +gdb_test "print f32x" ".* = 10\\.5.*" "The original value of f32x is 10.5" +gdb_test "print f64x" ".* = 20\\.25.*" "The original value of f64x is 20.25" + +# Test that gdb could correctly recognize float constant expression with a suffix. +# FIXME: gdb does not yet recognize the suffix for _FloatN/_FloatNx types. +gdb_test "print f32=-1.5" ".* = -1\\.5.*" "Try to change f32 to -1.5 with 'print f32=-1.5'" +gdb_test "print f64=-2.25" ".* = -2\\.25.*" "Try to change f64 to -2.25 with 'print f64=-2.25'" +gdb_test "print f128=-3.375" ".* = -3\\.375.*" "Try to change f128 to -3.375 with 'print f128=-3.375'" +gdb_test "print f32x=-10.5" ".* = -10\\.5.*" "Try to change f32x to -10.5 with 'print f32=-1.5x'" +gdb_test "print f64x=-20.25" ".* = -20\\.25.*" "Try to change f64x to -20.25 with 'print f64=-2.25x'" + +# Test that gdb could handle the above correctly with "set var" command. +set test "set variable f32 = 10.5" +gdb_test_multiple "set var f32=10.5" "$test" { + -re "$gdb_prompt $" { + pass "$test" + } + -re "Invalid number.*$gdb_prompt $" { + fail "$test (do not recognize 10.5)" + } +} + +set test "set variable f64 = 20.25" +gdb_test_multiple "set var f64=20.25" "$test" { + -re "$gdb_prompt $" { + pass "$test" + } + -re "Invalid number.*$gdb_prompt $" { + fail "$test (do not recognize 20.25)" + } +} + +set test "set variable f128 = 30.375" +gdb_test_multiple "set var f128=30.375" "$test" { + -re "$gdb_prompt $" { + pass "$test" + } + -re "Invalid number.*$gdb_prompt $" { + fail "$test (do not recognize 30.375)" + } +} + +set test "set variable f32x = 100.5" +gdb_test_multiple "set var f32x=100.5" "$test" { + -re "$gdb_prompt $" { + pass "$test" + } + -re "Invalid number.*$gdb_prompt $" { + fail "$test (do not recognize 100.5)" + } +} + +set test "set variable f64x = 200.25" +gdb_test_multiple "set var f64x=200.25" "$test" { + -re "$gdb_prompt $" { + pass "$test" + } + -re "Invalid number.*$gdb_prompt $" { + fail "$test (do not recognize 200.25)" + } +} + +gdb_test "print f32" ".* = 10\\.5.*" "The value of f32 is changed to 10.5" +gdb_test "print f64" ".* = 20\\.25.*" "The value of f64 is changed to 20.25" +gdb_test "print f128" ".* = 30\\.375.*" "The value of f128 is changed to 30.375" +gdb_test "print f32x" ".* = 100\\.5.*" "The value of f32x is changed to 100.5" +gdb_test "print f64x" ".* = 200\\.25.*" "The value of f64x is changed to 200.25" + +# Print the original values of c32, c64, c128, c32x, c64x. +gdb_test "print c32" ".* = 1\\.5 \\+ 1 \\* I.*" "The original value of c32 is 1.5 + 1 * I" +gdb_test "print c64" ".* = 2\\.25 \\+ 1 \\* I.*" "The original value of c64 is 2.25 + 1 * I" +gdb_test "print c128" ".* = 3\\.375 \\+ 1 \\* I.*" "The original value of c128 is 3.375 + 1 * I" +gdb_test "print c32x" ".* = 10\\.5 \\+ 1 \\* I.*" "The original value of c32x is 10.5 + 1 * I" +gdb_test "print c64x" ".* = 20\\.25 \\+ 1 \\* I.*" "The original value of c64x is 20.25 + 1 * I" + +# FIXME: GDB cannot parse non-trivial complex constants yet. +