Tcwg-commits August 2015

tcwg-commits@lists.linaro.org

1 participants
602 discussions

[toolchain/abe] 01/01: Force verbosity and -j 1

by git

This is an automated email from the git hooks/post-receive script. yvan.roux pushed a commit to branch ryanarn/reduced_parallel in repository toolchain/abe. commit abfa1957a501c81ded6fb0fa084beade094b0a4a Author: Yvan Roux <yvan.roux(a)linaro.org> Date: Tue Aug 18 14:59:18 2015 +0200 Force verbosity and -j 1 Change-Id: I5b9d07c6816bbb5654da085d4923dbf3f2ff2156 --- lib/make.sh | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/lib/make.sh b/lib/make.sh index 0693a96..7cf2885 100755 --- a/lib/make.sh +++ b/lib/make.sh @@ -731,13 +731,13 @@ make_check() fi if test x"${override_runtestflags}" != x; then - local make_flags="${make_flags} RUNTESTFLAGS=\"${override_runtestflags}\"" + local make_flags="${make_flags} RUNTESTFLAGS=\"${override_runtestflags} -v [...] fi if test x"${parallel}" = x"yes"; then local make_flags case "${target}" in - "$build"|*"-elf"*) make_flags="${make_flags} -j ${cpus}" ;; + "$build"|*"-elf"*) make_flags="${make_flags} -j 1" ;; # Try to find out which test is crashing the aarch64-linux-gnu # boards and x86_64 builders. aarch64-linux-gnu|x86_64*) make_flags="${make_flags} -j 1" ;; -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[toolchain/abe] branch ryanarn/reduced_parallel updated (4de88af -> abfa195)

by git

This is an automated email from the git hooks/post-receive script. yvan.roux pushed a change to branch ryanarn/reduced_parallel in repository toolchain/abe. from 4de88af jenkins.sh: Turn on extra verbose dejagnu logging adds fca485b VALIDATION TESTING: new abfa195 Force verbosity and -j 1 The 1 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference. Summary of changes: jenkins.sh | 2 +- lib/make.sh | 8 ++++---- 2 files changed, 5 insertions(+), 5 deletions(-) -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[gcc] branch hjl/indbr/gcc-5-branch updated (e3597b7 -> fd51e76)

by git

This is an automated email from the git hooks/post-receive script. unknown user pushed a change to branch hjl/indbr/gcc-5-branch in repository gcc. discards e3597b7 * config/i386/i386.c (ix86_function_ok_for_sibcall): Check flag_plt. discards 2762701 * gcc.target/i386/noplt-1.c (dg-do): Fix target selector. * [...] discards 6809d8d Generate indirect branch relocation via GOT discards e3c0218 Check if x86 gas supports indirect branch via GOT discards c85dec2 Add more tests for PR target/66232 discards 8e203c6 * config/i386/i386.h (enum reg_class): Move CLOBBERED_REGS p [...] discards baa4a2c1 Properly handle -fno-plt in ix86_expand_call discards 34d3e9c Allow indirect sibcall with register arguments discards b2e3ad4 Allow indirect branch via GOT slot for x32 discards b49ae10 2015-06-04 Sriraman Tallam <tmsriram(a)google.com> discards 2f86d31 Fix PR target/65753 discards bcd83da * calls.c (prepare_call_address): Transform PLT call to GOT [...] adds dd61457 PR c++/66919 * pt.c (tsubst_copy): Pass complain to mark_us [...] adds b66dfba Daily bump. new 045ed2e * calls.c (prepare_call_address): Transform PLT call to GOT [...] new b95473c Fix PR target/65753 new fad38a0 Allow indirect branch via GOT slot for x32 new 991291d * config/i386/i386.c (ix86_function_ok_for_sibcall): Check flag_plt. new c06740d * config/i386/i386.h (enum reg_class): Move CLOBBERED_REGS p [...] new ddce735 2015-06-04 Sriraman Tallam <tmsriram(a)google.com> new 739f56d * gcc.target/i386/noplt-1.c (dg-do): Fix target selector. * [...] new 442e79d Allow indirect sibcall with register arguments new 2381bc3 Add more tests for PR target/66232 new 19894c6 Properly handle -fno-plt in ix86_expand_call new 3fd7104 Check if x86 gas supports indirect branch via GOT new fd51e76 Generate indirect branch relocation via GOT This update added new revisions after undoing existing revisions. That is to say, some revisions that were in the old version of the branch are not in the new version. This situation occurs when a user --force pushes a change and generates a repository containing something like this: * -- * -- B -- O -- O -- O (e3597b7) \ N -- N -- N refs/heads/hjl/indbr/gcc-5-branch (fd51e76) You should already have received notification emails for all of the O revisions, and so the following emails describe only the N revisions from the common base, B. Any revisions marked "omits" are not gone; other references still refer to them. Any revisions marked "discards" are gone forever. The 12 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference. Summary of changes: gcc/DATESTAMP | 2 +- gcc/config/i386/predicates.md | 7 +++++- gcc/cp/ChangeLog | 5 ++++ gcc/cp/pt.c | 3 ++- gcc/testsuite/g++.dg/cpp1y/auto-fn27.C | 42 +++++++++++++++++++++++++++++++ gcc/testsuite/gcc.target/i386/pr67215-3.c | 12 +++++++++ 6 files changed, 68 insertions(+), 3 deletions(-) create mode 100644 gcc/testsuite/g++.dg/cpp1y/auto-fn27.C create mode 100644 gcc/testsuite/gcc.target/i386/pr67215-3.c -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[gcc] 01/01: Properly handle -fno-plt in ix86_expand_call

by git

This is an automated email from the git hooks/post-receive script. unknown user pushed a commit to branch hjl/pr67215/master in repository gcc. commit 8ade1936895516ea40a89dc88daf4267df60b226 Author: H.J. Lu <hjl.tools(a)gmail.com> Date: Sun Aug 16 04:46:20 2015 -0700 Properly handle -fno-plt in ix86_expand_call prepare_call_address in calls.c is the wrong place to handle -fno-plt. We shoudn't force function address into register and hope that load function address via GOT and indirect call via register will be folded into indirect call via GOT, which doesn't always happen. Also non-PIC case can only be handled in backend. Instead, backend should expand external function call into indirect call via GOT for -fno-plt. This patch reverts -fno-plt in prepare_call_address and handles it in ix86_expand_call. Other backends may need similar changes to support -fno-plt. Alternately, we can introduce a target hook to indicate whether an external function should be called via register for -fno-plt so that i386 backend can disable it in prepare_call_address. gcc/ PR target/67215 * calls.c (prepare_call_address): Don't handle -fno-plt here. * config/i386/i386.c (ix86_expand_call): Generate indirect call via GOT for -fno-plt. Support indirect call via GOT for x32. * config/i386/predicates.md (sibcall_memory_operand): Allow GOT memory operand. gcc/testsuite/ PR target/67215 * gcc.target/i386/pr67215-1.c: New test. * gcc.target/i386/pr67215-2.c: Likewise. * gcc.target/i386/pr67215-3.c: Likewise. --- gcc/calls.c | 12 ------ gcc/config/i386/i386.c | 71 ++++++++++++++++++++++++------- gcc/config/i386/predicates.md | 7 ++- gcc/testsuite/gcc.target/i386/pr67215-1.c | 20 +++++++++ gcc/testsuite/gcc.target/i386/pr67215-2.c | 20 +++++++++ gcc/testsuite/gcc.target/i386/pr67215-3.c | 12 ++++++ 6 files changed, 113 insertions(+), 29 deletions(-) diff --git a/gcc/calls.c b/gcc/calls.c index 5636725..7cce9be 100644 --- a/gcc/calls.c +++ b/gcc/calls.c @@ -203,18 +203,6 @@ prepare_call_address (tree fndecl_or_type, rtx funexp, rtx sta [...] && targetm.small_register_classes_for_mode_p (FUNCTION_MODE)) ? force_not_mem (memory_address (FUNCTION_MODE, funexp)) : memory_address (FUNCTION_MODE, funexp)); - else if (flag_pic - && fndecl_or_type - && TREE_CODE (fndecl_or_type) == FUNCTION_DECL - && (!flag_plt - || lookup_attribute ("noplt", DECL_ATTRIBUTES (fndecl_or_type))) - && !targetm.binds_local_p (fndecl_or_type)) - { - /* This is done only for PIC code. There is no easy interface to force the - function address into GOT for non-PIC case. non-PIC case needs to be - handled specially by the backend. */ - funexp = force_reg (Pmode, funexp); - } else if (! sibcallp) { if (!NO_FUNCTION_CSE && optimize && ! flag_no_function_cse) diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c index 05fa5e1..ac9a6c4 100644 --- a/gcc/config/i386/i386.c +++ b/gcc/config/i386/i386.c @@ -25649,21 +25649,54 @@ ix86_expand_call (rtx retval, rtx fnaddr, rtx callarg1, /* Static functions and indirect calls don't need the pic register. Also, check if PLT was explicitly avoided via no-plt or "noplt" attribute, making it an indirect call. */ + rtx addr = XEXP (fnaddr, 0); if (flag_pic - && (!TARGET_64BIT - || (ix86_cmodel == CM_LARGE_PIC - && DEFAULT_ABI != MS_ABI)) - && GET_CODE (XEXP (fnaddr, 0)) == SYMBOL_REF - && !SYMBOL_REF_LOCAL_P (XEXP (fnaddr, 0)) - && flag_plt - && (SYMBOL_REF_DECL ((XEXP (fnaddr, 0))) == NULL_TREE - || !lookup_attribute ("noplt", - DECL_ATTRIBUTES (SYMBOL_REF_DECL (XEXP (fnaddr, 0)))))) + && GET_CODE (addr) == SYMBOL_REF + && !SYMBOL_REF_LOCAL_P (addr)) { - use_reg (&use, gen_rtx_REG (Pmode, REAL_PIC_OFFSET_TABLE_REGNUM)); - if (ix86_use_pseudo_pic_reg ()) - emit_move_insn (gen_rtx_REG (Pmode, REAL_PIC_OFFSET_TABLE_REGNUM), - pic_offset_table_rtx); + if (flag_plt + && (SYMBOL_REF_DECL (addr) == NULL_TREE + || !lookup_attribute ("noplt", + DECL_ATTRIBUTES (SYMBOL_REF_DECL (addr))))) + { + if (!TARGET_64BIT + || (ix86_cmodel == CM_LARGE_PIC + && DEFAULT_ABI != MS_ABI)) + { + use_reg (&use, gen_rtx_REG (Pmode, + REAL_PIC_OFFSET_TABLE_REGNUM)); + if (ix86_use_pseudo_pic_reg ()) + emit_move_insn (gen_rtx_REG (Pmode, + REAL_PIC_OFFSET_TABLE_REGNUM), + pic_offset_table_rtx); + } + } + else if (!TARGET_PECOFF && !TARGET_MACHO) + { + if (TARGET_64BIT) + { + fnaddr = gen_rtx_UNSPEC (Pmode, + gen_rtvec (1, addr), + UNSPEC_GOTPCREL); + fnaddr = gen_rtx_CONST (Pmode, fnaddr); + } + else + { + fnaddr = gen_rtx_UNSPEC (Pmode, gen_rtvec (1, addr), + UNSPEC_GOT); + fnaddr = gen_rtx_CONST (Pmode, fnaddr); + fnaddr = gen_rtx_PLUS (Pmode, pic_offset_table_rtx, + fnaddr); + } + fnaddr = gen_const_mem (Pmode, fnaddr); + /* Pmode may not be the same as word_mode for x32, which + doesn't support indirect branch via 32-bit memory slot. + Since x32 GOT slot is 64 bit with zero upper 32 bits, + indirect branch via x32 GOT slot is OK. */ + if (GET_MODE (fnaddr) != word_mode) + fnaddr = gen_rtx_ZERO_EXTEND (word_mode, fnaddr); + fnaddr = gen_rtx_MEM (QImode, fnaddr); + } } } @@ -25685,9 +25718,15 @@ ix86_expand_call (rtx retval, rtx fnaddr, rtx callarg1, && GET_CODE (XEXP (fnaddr, 0)) == SYMBOL_REF && !local_symbolic_operand (XEXP (fnaddr, 0), VOIDmode)) fnaddr = gen_rtx_MEM (QImode, construct_plt_address (XEXP (fnaddr, 0))); - else if (sibcall - ? !sibcall_insn_operand (XEXP (fnaddr, 0), word_mode) - : !call_insn_operand (XEXP (fnaddr, 0), word_mode)) + /* Since x32 GOT slot is 64 bit with zero upper 32 bits, indirect + branch via x32 GOT slot is OK. */ + else if (!(TARGET_X32 + && MEM_P (fnaddr) + && GET_CODE (XEXP (fnaddr, 0)) == ZERO_EXTEND + && GOT_memory_operand (XEXP (XEXP (fnaddr, 0), 0), Pmode)) + && (sibcall + ? !sibcall_insn_operand (XEXP (fnaddr, 0), word_mode) + : !call_insn_operand (XEXP (fnaddr, 0), word_mode))) { fnaddr = convert_to_mode (word_mode, XEXP (fnaddr, 0), 1); fnaddr = gen_rtx_MEM (QImode, copy_to_mode_reg (word_mode, fnaddr)); diff --git a/gcc/config/i386/predicates.md b/gcc/config/i386/predicates.md index a9c8623..be2df76 100644 --- a/gcc/config/i386/predicates.md +++ b/gcc/config/i386/predicates.md @@ -593,7 +593,12 @@ ;; Return true if OP is a memory operands that can be used in sibcalls. (define_predicate "sibcall_memory_operand" (and (match_operand 0 "memory_operand") - (match_test "CONSTANT_P (XEXP (op, 0))"))) + (match_test "CONSTANT_P (XEXP (op, 0)) + || (GET_CODE (XEXP (op, 0)) == PLUS + && REG_P (XEXP (XEXP (op, 0), 0)) + && GET_CODE (XEXP (XEXP (op, 0), 1)) == CONST + && GET_CODE (XEXP (XEXP (XEXP (op, 0), 1), 0)) == UNSPEC + && XINT (XEXP (XEXP (XEXP (op, 0), 1), 0), 1) == UNSPEC_GOT)"))) ;; Test for a valid operand for a call instruction. ;; Allow constant call address operands in Pmode only. diff --git a/gcc/testsuite/gcc.target/i386/pr67215-1.c b/gcc/testsuite/gcc.target/i [...] new file mode 100644 index 0000000..fd37f8e --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr67215-1.c @@ -0,0 +1,20 @@ +/* { dg-do compile { target *-*-linux* } } */ +/* { dg-options "-O2 -fpic -fno-plt" } */ + +extern char* bar (int); +extern char* arr[32]; + +void +foo (void) +{ + int i; + + for (i = 0; i < 32; i++) + arr[i] = bar (128); +} + +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOTPCREL" { target { ! ia32 } } } } */ +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "mov(l|q)\[ \t\]*.bar@GOTPCREL" { target { ! ia [...] +/* { dg-final { scan-assembler-not "movl\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "call\[ \t\]*.bar@PLT" } } */ diff --git a/gcc/testsuite/gcc.target/i386/pr67215-2.c b/gcc/testsuite/gcc.target/i [...] new file mode 100644 index 0000000..ebf2919 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr67215-2.c @@ -0,0 +1,20 @@ +/* { dg-do compile { target *-*-linux* } } */ +/* { dg-options "-O2 -fpic" } */ + +extern char* bar (int) __attribute__ ((noplt)); +extern char* arr[32]; + +void +foo (void) +{ + int i; + + for (i = 0; i < 32; i++) + arr[i] = bar (128); +} + +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOTPCREL" { target { ! ia32 } } } } */ +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "mov(l|q)\[ \t\]*.bar@GOTPCREL" { target { ! ia [...] +/* { dg-final { scan-assembler-not "movl\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "call\[ \t\]*.bar@PLT" } } */ diff --git a/gcc/testsuite/gcc.target/i386/pr67215-3.c b/gcc/testsuite/gcc.target/i [...] new file mode 100644 index 0000000..dbd9a2f --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr67215-3.c @@ -0,0 +1,12 @@ +/* { dg-do compile { target *-*-linux* } } */ +/* { dg-options "-O2 -fpic -fno-plt -fdump-rtl-expand" } */ + +extern int bar (void); + +int +foo (void) +{ + return bar (); +} + +/* { dg-final { scan-rtl-dump "\\(call \\(mem:QI \\(mem/u/c:" "expand" } } */ -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[gcc] branch hjl/pr67215/master updated (04258b4 -> 8ade193)

by git

This is an automated email from the git hooks/post-receive script. unknown user pushed a change to branch hjl/pr67215/master in repository gcc. discards 04258b4 Properly handle -fno-plt in ix86_expand_call adds 75c2bf3 * da.po: Update. adds ca9d1f7 * except.c (check_noexcept_r): Assert that fn is POINTER_TYPE_P. adds a02b42f * constexpr.c (cxx_eval_store_expression): Don't set CONSTR [...] adds 7963b19 PR c++/67104 * constexpr.c (array_index_cmp, find_array_cto [...] adds 59246f9 PR c++/67244 * pt.c (tsubst_copy_and_build): Call insert_pe [...] adds 61fba55 /cp 2015-08-17 Paolo Carlini <paolo.carlini(a)oracle.com> adds 65b3995 mips.exp: Redefine isa variable which is referenced in later code. adds 8578096 Daily bump. adds 863356e xtensa: reimplement register spilling adds 8a596a7 xtensa: use unwind-dw2-fde-dip instead of unwind-dw2-fde adds 15b21d2 xtensa: fix _Unwind_GetCFA adds 1eb4f3d * tree-ssa-loop-niter.c (refine_value_range_using_guard): Ne [...] new 8ade193 Properly handle -fno-plt in ix86_expand_call This update added new revisions after undoing existing revisions. That is to say, some revisions that were in the old version of the branch are not in the new version. This situation occurs when a user --force pushes a change and generates a repository containing something like this: * -- * -- B -- O -- O -- O (04258b4) \ N -- N -- N refs/heads/hjl/pr67215/master (8ade193) You should already have received notification emails for all of the O revisions, and so the following emails describe only the N revisions from the common base, B. Any revisions marked "omits" are not gone; other references still refer to them. Any revisions marked "discards" are gone forever. The 1 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference. Summary of changes: gcc/ChangeLog | 6 + gcc/DATESTAMP | 2 +- gcc/config/i386/predicates.md | 7 +- gcc/cp/ChangeLog | 21 ++ gcc/cp/call.c | 19 +- gcc/cp/constexpr.c | 191 ++++++++--- gcc/cp/except.c | 4 +- gcc/cp/pt.c | 2 + gcc/po/ChangeLog | 4 + gcc/po/da.po | 360 +++++++-------------- gcc/testsuite/ChangeLog | 21 ++ gcc/testsuite/g++.dg/cpp0x/constexpr-array12.C | 8 + gcc/testsuite/g++.dg/cpp0x/lambda/lambda-nested5.C | 29 ++ gcc/testsuite/g++.dg/cpp0x/nullptr34.C | 17 + gcc/testsuite/g++.dg/cpp0x/pr51313.C | 2 +- gcc/testsuite/g++.dg/warn/Wconversion2.C | 3 +- gcc/testsuite/g++.dg/warn/Wnull-conversion-1.C | 9 +- gcc/testsuite/g++.old-deja/g++.other/null3.C | 3 +- .../gcc.dg/tree-ssa/{scev-9.c => loop-bound-1.c} | 6 +- .../gcc.dg/tree-ssa/{scev-9.c => loop-bound-3.c} | 8 +- gcc/testsuite/gcc.dg/tree-ssa/loop-bound-5.c | 22 ++ gcc/testsuite/gcc.target/i386/pr67215-3.c | 12 + gcc/testsuite/gcc.target/mips/mips.exp | 1 + gcc/tree-ssa-loop-niter.c | 312 ++++++++++++++++-- libgcc/ChangeLog | 17 + libgcc/config/xtensa/lib2funcs.S | 30 +- libgcc/config/xtensa/t-windowed | 2 +- libgcc/config/xtensa/unwind-dw2-xtensa.c | 2 +- 28 files changed, 771 insertions(+), 349 deletions(-) create mode 100644 gcc/testsuite/g++.dg/cpp0x/constexpr-array12.C create mode 100644 gcc/testsuite/g++.dg/cpp0x/lambda/lambda-nested5.C create mode 100644 gcc/testsuite/g++.dg/cpp0x/nullptr34.C copy gcc/testsuite/gcc.dg/tree-ssa/{scev-9.c => loop-bound-1.c} (58%) copy gcc/testsuite/gcc.dg/tree-ssa/{scev-9.c => loop-bound-3.c} (51%) create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/loop-bound-5.c create mode 100644 gcc/testsuite/gcc.target/i386/pr67215-3.c -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[toolchain/abe] 01/01: VALIDATION TESTING:

by git

This is an automated email from the git hooks/post-receive script. yvan.roux pushed a commit to branch yvan/validation-testing in repository toolchain/abe. commit fca485b0afe0344c80e4aad20e32efbc708dce28 Author: Yvan Roux <yvan.roux(a)linaro.org> Date: Tue Aug 18 13:21:45 2015 +0200 VALIDATION TESTING: Switch make check to -j 1 for x86_64 native builds jenkins.sh: even moree dejagnu verbosity Change-Id: I076d518964014795b7987deac46c816b04e27c11 --- jenkins.sh | 2 +- lib/make.sh | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/jenkins.sh b/jenkins.sh index 661c74a..e7aee89 100755 --- a/jenkins.sh +++ b/jenkins.sh @@ -343,7 +343,7 @@ fi (cd $user_snapshots/gcc.git; git fetch origin '+refs/changes/*:refs/remotes/gerrit [...] # Turn on extra verbose dejagnu logging -export RUNTESTFLAGS="-v -v -d" +export RUNTESTFLAGS="-v -v -v -v -d" # Now we build the cross compiler, for a native compiler this becomes # the stage2 bootstrap build. diff --git a/lib/make.sh b/lib/make.sh index c86989e..0693a96 100755 --- a/lib/make.sh +++ b/lib/make.sh @@ -739,8 +739,8 @@ make_check() case "${target}" in "$build"|*"-elf"*) make_flags="${make_flags} -j ${cpus}" ;; # Try to find out which test is crashing the aarch64-linux-gnu - # boards. - aarch64-linux-gnu) make_flags="${make_flags} -j 1" ;; + # boards and x86_64 builders. + aarch64-linux-gnu|x86_64*) make_flags="${make_flags} -j 1" ;; # Double parallelization when running tests on remote boards # to avoid host idling when waiting for the board. *) make_flags="${make_flags} -j $((2*${cpus}))" ;; -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[toolchain/abe] branch yvan/validation-testing created (now fca485b)

by git

This is an automated email from the git hooks/post-receive script. yvan.roux pushed a change to branch yvan/validation-testing in repository toolchain/abe. at fca485b VALIDATION TESTING: This branch includes the following new commits: new fca485b VALIDATION TESTING: The 1 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference. -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[gcc] 01/01: Properly handle -fno-plt in ix86_expand_call

by git

This is an automated email from the git hooks/post-receive script. unknown user pushed a commit to branch hjl/pr67215/master in repository gcc. commit 04258b418d2ea105249d371a06805122d8953816 Author: H.J. Lu <hjl.tools(a)gmail.com> Date: Sun Aug 16 04:46:20 2015 -0700 Properly handle -fno-plt in ix86_expand_call prepare_call_address in calls.c is the wrong place to handle -fno-plt. We shoudn't force function address into register and hope that load function address via GOT and indirect call via register will be folded into indirect call via GOT, which doesn't always happen. Also non-PIC case can only be handled in backend. Instead, backend should expand external function call into indirect call via GOT for -fno-plt. This patch reverts -fno-plt in prepare_call_address and handles it in ix86_expand_call. Other backends may need similar changes to support -fno-plt. Alternately, we can introduce a target hook to indicate whether an external function should be called via register for -fno-plt so that i386 backend can disable it in prepare_call_address. gcc/ PR target/67215 * calls.c (prepare_call_address): Don't handle -fno-plt here. * config/i386/i386.c (ix86_expand_call): Generate indirect call via GOT for -fno-plt. Support indirect call via GOT for x32. gcc/testsuite/ PR target/67215 * gcc.target/i386/pr67215-1.c: New test. * gcc.target/i386/pr67215-2.c: Likewise. --- gcc/calls.c | 12 ------ gcc/config/i386/i386.c | 71 ++++++++++++++++++++++++------- gcc/testsuite/gcc.target/i386/pr67215-1.c | 20 +++++++++ gcc/testsuite/gcc.target/i386/pr67215-2.c | 20 +++++++++ 4 files changed, 95 insertions(+), 28 deletions(-) diff --git a/gcc/calls.c b/gcc/calls.c index 5636725..7cce9be 100644 --- a/gcc/calls.c +++ b/gcc/calls.c @@ -203,18 +203,6 @@ prepare_call_address (tree fndecl_or_type, rtx funexp, rtx sta [...] && targetm.small_register_classes_for_mode_p (FUNCTION_MODE)) ? force_not_mem (memory_address (FUNCTION_MODE, funexp)) : memory_address (FUNCTION_MODE, funexp)); - else if (flag_pic - && fndecl_or_type - && TREE_CODE (fndecl_or_type) == FUNCTION_DECL - && (!flag_plt - || lookup_attribute ("noplt", DECL_ATTRIBUTES (fndecl_or_type))) - && !targetm.binds_local_p (fndecl_or_type)) - { - /* This is done only for PIC code. There is no easy interface to force the - function address into GOT for non-PIC case. non-PIC case needs to be - handled specially by the backend. */ - funexp = force_reg (Pmode, funexp); - } else if (! sibcallp) { if (!NO_FUNCTION_CSE && optimize && ! flag_no_function_cse) diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c index 05fa5e1..ac9a6c4 100644 --- a/gcc/config/i386/i386.c +++ b/gcc/config/i386/i386.c @@ -25649,21 +25649,54 @@ ix86_expand_call (rtx retval, rtx fnaddr, rtx callarg1, /* Static functions and indirect calls don't need the pic register. Also, check if PLT was explicitly avoided via no-plt or "noplt" attribute, making it an indirect call. */ + rtx addr = XEXP (fnaddr, 0); if (flag_pic - && (!TARGET_64BIT - || (ix86_cmodel == CM_LARGE_PIC - && DEFAULT_ABI != MS_ABI)) - && GET_CODE (XEXP (fnaddr, 0)) == SYMBOL_REF - && !SYMBOL_REF_LOCAL_P (XEXP (fnaddr, 0)) - && flag_plt - && (SYMBOL_REF_DECL ((XEXP (fnaddr, 0))) == NULL_TREE - || !lookup_attribute ("noplt", - DECL_ATTRIBUTES (SYMBOL_REF_DECL (XEXP (fnaddr, 0)))))) + && GET_CODE (addr) == SYMBOL_REF + && !SYMBOL_REF_LOCAL_P (addr)) { - use_reg (&use, gen_rtx_REG (Pmode, REAL_PIC_OFFSET_TABLE_REGNUM)); - if (ix86_use_pseudo_pic_reg ()) - emit_move_insn (gen_rtx_REG (Pmode, REAL_PIC_OFFSET_TABLE_REGNUM), - pic_offset_table_rtx); + if (flag_plt + && (SYMBOL_REF_DECL (addr) == NULL_TREE + || !lookup_attribute ("noplt", + DECL_ATTRIBUTES (SYMBOL_REF_DECL (addr))))) + { + if (!TARGET_64BIT + || (ix86_cmodel == CM_LARGE_PIC + && DEFAULT_ABI != MS_ABI)) + { + use_reg (&use, gen_rtx_REG (Pmode, + REAL_PIC_OFFSET_TABLE_REGNUM)); + if (ix86_use_pseudo_pic_reg ()) + emit_move_insn (gen_rtx_REG (Pmode, + REAL_PIC_OFFSET_TABLE_REGNUM), + pic_offset_table_rtx); + } + } + else if (!TARGET_PECOFF && !TARGET_MACHO) + { + if (TARGET_64BIT) + { + fnaddr = gen_rtx_UNSPEC (Pmode, + gen_rtvec (1, addr), + UNSPEC_GOTPCREL); + fnaddr = gen_rtx_CONST (Pmode, fnaddr); + } + else + { + fnaddr = gen_rtx_UNSPEC (Pmode, gen_rtvec (1, addr), + UNSPEC_GOT); + fnaddr = gen_rtx_CONST (Pmode, fnaddr); + fnaddr = gen_rtx_PLUS (Pmode, pic_offset_table_rtx, + fnaddr); + } + fnaddr = gen_const_mem (Pmode, fnaddr); + /* Pmode may not be the same as word_mode for x32, which + doesn't support indirect branch via 32-bit memory slot. + Since x32 GOT slot is 64 bit with zero upper 32 bits, + indirect branch via x32 GOT slot is OK. */ + if (GET_MODE (fnaddr) != word_mode) + fnaddr = gen_rtx_ZERO_EXTEND (word_mode, fnaddr); + fnaddr = gen_rtx_MEM (QImode, fnaddr); + } } } @@ -25685,9 +25718,15 @@ ix86_expand_call (rtx retval, rtx fnaddr, rtx callarg1, && GET_CODE (XEXP (fnaddr, 0)) == SYMBOL_REF && !local_symbolic_operand (XEXP (fnaddr, 0), VOIDmode)) fnaddr = gen_rtx_MEM (QImode, construct_plt_address (XEXP (fnaddr, 0))); - else if (sibcall - ? !sibcall_insn_operand (XEXP (fnaddr, 0), word_mode) - : !call_insn_operand (XEXP (fnaddr, 0), word_mode)) + /* Since x32 GOT slot is 64 bit with zero upper 32 bits, indirect + branch via x32 GOT slot is OK. */ + else if (!(TARGET_X32 + && MEM_P (fnaddr) + && GET_CODE (XEXP (fnaddr, 0)) == ZERO_EXTEND + && GOT_memory_operand (XEXP (XEXP (fnaddr, 0), 0), Pmode)) + && (sibcall + ? !sibcall_insn_operand (XEXP (fnaddr, 0), word_mode) + : !call_insn_operand (XEXP (fnaddr, 0), word_mode))) { fnaddr = convert_to_mode (word_mode, XEXP (fnaddr, 0), 1); fnaddr = gen_rtx_MEM (QImode, copy_to_mode_reg (word_mode, fnaddr)); diff --git a/gcc/testsuite/gcc.target/i386/pr67215-1.c b/gcc/testsuite/gcc.target/i [...] new file mode 100644 index 0000000..fd37f8e --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr67215-1.c @@ -0,0 +1,20 @@ +/* { dg-do compile { target *-*-linux* } } */ +/* { dg-options "-O2 -fpic -fno-plt" } */ + +extern char* bar (int); +extern char* arr[32]; + +void +foo (void) +{ + int i; + + for (i = 0; i < 32; i++) + arr[i] = bar (128); +} + +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOTPCREL" { target { ! ia32 } } } } */ +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "mov(l|q)\[ \t\]*.bar@GOTPCREL" { target { ! ia [...] +/* { dg-final { scan-assembler-not "movl\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "call\[ \t\]*.bar@PLT" } } */ diff --git a/gcc/testsuite/gcc.target/i386/pr67215-2.c b/gcc/testsuite/gcc.target/i [...] new file mode 100644 index 0000000..ebf2919 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr67215-2.c @@ -0,0 +1,20 @@ +/* { dg-do compile { target *-*-linux* } } */ +/* { dg-options "-O2 -fpic" } */ + +extern char* bar (int) __attribute__ ((noplt)); +extern char* arr[32]; + +void +foo (void) +{ + int i; + + for (i = 0; i < 32; i++) + arr[i] = bar (128); +} + +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOTPCREL" { target { ! ia32 } } } } */ +/* { dg-final { scan-assembler "call\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "mov(l|q)\[ \t\]*.bar@GOTPCREL" { target { ! ia [...] +/* { dg-final { scan-assembler-not "movl\[ \t\]*.bar@GOT\\(" { target ia32 } } } */ +/* { dg-final { scan-assembler-not "call\[ \t\]*.bar@PLT" } } */ -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[gcc] branch hjl/pr67215/master updated (09e56e8 -> 04258b4)

by git

This is an automated email from the git hooks/post-receive script. unknown user pushed a change to branch hjl/pr67215/master in repository gcc. discards 09e56e8 Add comments. omits 96865ec Properly handle -fno-plt in ix86_expand_call adds 9be7a73 PR fortran/41387 * gfortran.texi: New section "File operati [...] adds d351aaf * trans-intrinsic.c (conv_intrinsic_ieee_is_negative): Use t [...] adds 6de19d8 PR fortran/54656 adds 6e79e8c Daily bump. adds 5702129 2015-08-17 Richard Biener <rguenther(a)suse.de> Naveen H [...] adds 6645b02 [ARM] arm memcpy of aligned data adds 0436eae * config/arm/arm.c (neon_element_bits): Replace call [...] adds 29aa549 2015-08-17 Richard Biener <rguenther(a)suse.de> adds 22723c2 * c-pretty-print.c (pp_c_cv_qualifiers): Remove code dealing [...] adds 316a001 * config/i386/i386.c: Remove include of fibheap.h. new 04258b4 Properly handle -fno-plt in ix86_expand_call This update added new revisions after undoing existing revisions. That is to say, some revisions that were in the old version of the branch are not in the new version. This situation occurs when a user --force pushes a change and generates a repository containing something like this: * -- * -- B -- O -- O -- O (09e56e8) \ N -- N -- N refs/heads/hjl/pr67215/master (04258b4) You should already have received notification emails for all of the O revisions, and so the following emails describe only the N revisions from the common base, B. Any revisions marked "omits" are not gone; other references still refer to them. Any revisions marked "discards" are gone forever. The 1 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference. Summary of changes: gcc/ChangeLog | 46 ++++++++++++++++++++++ gcc/DATESTAMP | 2 +- gcc/c-family/ChangeLog | 5 +++ gcc/c-family/c-pretty-print.c | 31 +++------------ gcc/config/arm/arm.c | 12 ++++-- gcc/config/arm/neon.md | 12 +++--- gcc/config/i386/i386.c | 16 ++++---- gcc/config/rs6000/rs6000.c | 3 +- gcc/config/spu/spu.c | 2 +- gcc/expmed.c | 8 ++-- gcc/expr.c | 2 +- gcc/fold-const.c | 2 +- gcc/fortran/ChangeLog | 22 +++++++++++ gcc/fortran/f95-lang.c | 35 ++++++++-------- gcc/fortran/gfortran.texi | 28 +++++++++++++ gcc/fortran/mathbuiltins.def | 1 - gcc/fortran/trans-intrinsic.c | 20 ++++++---- gcc/match.pd | 9 +++++ gcc/optabs.c | 2 +- gcc/simplify-rtx.c | 2 +- gcc/testsuite/ChangeLog | 17 ++++++++ gcc/testsuite/gcc.dg/pr16107.c | 16 ++++++++ gcc/testsuite/gcc.dg/torture/pr67221.c | 31 +++++++++++++++ gcc/testsuite/gcc.target/arm/memcpy-aligned-1.c | 12 ++++++ .../gfortran.dg/{norm_4.f90 => norm2_4.f90} | 0 gcc/testsuite/gfortran.dg/norm2_5.f90 | 23 +++++++++++ gcc/tree-cfg.c | 6 +-- gcc/tree-ssa-sccvn.c | 19 ++++++++- gcc/tree-vect-patterns.c | 13 +++--- 29 files changed, 304 insertions(+), 93 deletions(-) create mode 100644 gcc/testsuite/gcc.dg/pr16107.c create mode 100644 gcc/testsuite/gcc.dg/torture/pr67221.c create mode 100644 gcc/testsuite/gcc.target/arm/memcpy-aligned-1.c rename gcc/testsuite/gfortran.dg/{norm_4.f90 => norm2_4.f90} (100%) create mode 100644 gcc/testsuite/gfortran.dg/norm2_5.f90 -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

[toolchain/abe] branch bernie/benchmarking updated (cf75e63 -> 5adbe64)

by git

This is an automated email from the git hooks/post-receive script. bernie.ogden pushed a change to branch bernie/benchmarking in repository toolchain/abe. discards cf75e63 Add gfortran to host dependencies discards 7923e9f Detect failure to scp tarball new 4b002a9 Detect failure to scp tarball new 5adbe64 Add gfortran to host dependencies This update added new revisions after undoing existing revisions. That is to say, some revisions that were in the old version of the branch are not in the new version. This situation occurs when a user --force pushes a change and generates a repository containing something like this: * -- * -- B -- O -- O -- O (cf75e63) \ N -- N -- N refs/heads/bernie/benchmarking (5adbe64) You should already have received notification emails for all of the O revisions, and so the following emails describe only the N revisions from the common base, B. Any revisions marked "omits" are not gone; other references still refer to them. Any revisions marked "discards" are gone forever. The 2 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference. Summary of changes: scripts/Benchmark.job | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) -- To stop receiving notification emails like this one, please contact the administrator of this repository.

9 years, 3 months

← Newer
1
...
24
25
26
27
28
29
30
...
61
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

Tcwg-commits August 2015