LA: rework atomic implemention #58
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* config.gcc: Add LoongArch tuples. * config.host: Likewise. * config/host-linux.c: Likewise. * configure: Likewise. * configure.ac: Likewise. * common/config/loongarch/loongarch-common.c: New file. * config/loongarch/constraints.md: Likewise. * config/loongarch/driver-native.c: Likewise. * config/loongarch/generic.md: Likewise. * config/loongarch/genopt.sh: Likewise. * config/loongarch/gnu-user.h: Likewise. * config/loongarch/gs464v.md: Likewise. * config/loongarch/larchintrin.h: Likewise. * config/loongarch/linux.h: Likewise. * config/loongarch/loongarch-builtins.c: Likewise. * config/loongarch/loongarch-c.c: Likewise. * config/loongarch/loongarch-cpus.def: Likewise. * config/loongarch/loongarch-ftypes.def: Likewise. * config/loongarch/loongarch-modes.def: Likewise. * config/loongarch/loongarch-opts.h: Likewise. * config/loongarch/loongarch-protos.h: Likewise. * config/loongarch/loongarch-tables.opt: Likewise. * config/loongarch/loongarch.c: Likewise. * config/loongarch/loongarch.h: Likewise. * config/loongarch/loongarch.md: Likewise. * config/loongarch/loongarch.opt: Likewise. * config/loongarch/predicates.md: Likewise. * config/loongarch/sync.md: Likewise. * config/loongarch/t-linux: Likewise. * config/loongarch/t-loongarch: Likewise. * config/loongarch/x-native: Likewise.
* config.host: Add LoongArch tuples. * configure: Likewise. * configure.ac: Likewise. * config/loongarch/crtfastmath.c: New file. * config/loongarch/crti.S: Likewise. * config/loongarch/crtn.S: Likewise. * config/loongarch/lib2funcs.c: Likewise. * config/loongarch/linux-unwind.h: Likewise. * config/loongarch/sfp-machine.h: Likewise. * config/loongarch/t-crtstuff: Likewise. * config/loongarch/t-elf: Likewise. * config/loongarch/t-loongarch: Likewise. * config/loongarch/t-loongarch64: Likewise. * config/loongarch/t-softfp-tf: Likewise.
* configure.tgt: Add LoongArch tuple.
* lib/target-supports.exp: Define the LoongArch target. * g++.dg/cpp0x/constexpr-rom.C: Skip on LoongArch. * g++.old-deja/g++.abi/ptrmem.C: Likewise. * gcc.dg/loop-8.c: Likewise. * gcc.dg/torture/stackalign/builtin-apply-2.c: Likewise. * gcc.dg/tree-ssa/ssa-fre-3.c: Likewise. * g++.old-deja/g++.pt/ptrmem6.C: Add support LoongArch. * gcc.dg/20020312-2.c: Likewise. * gcc.target/loongarch/loongarch.exp: New file. * gcc.target/loongarch/tst-asm-const.c: Likewise.
* config/picflag.m4: Add LoongArch tuples. * configure: Likewise. * configure.ac: Likewise.
gcc/
* config/loongarch/loongarch-builtins.c: Code style.
* config/loongarch/loongarch.c: Likewise.
* config/loongarch/loongarch.md: Likewise
gcc/config/loongarch * loongarch-cpus.def: Modify the value of ISA in loongarch_cpu_info_table.
Also add helpful error messages.
gcc/
* config.gcc: Modify LoongArch config.
…e use of the resolver attribute. gcc/ config.gcc: Indirect function support.
gcc/ * config.gcc: Removed '-march=loongarch'. gcc/config/loongarch/ * loongarch-cpus.def: * loongarch-tables.opt: add PROCESSOR_LOONGARCH32. * loongarch.c: Likewise. * loongarch.md: Likewise.
gcc/config/loongarch * loongarch.c: attribute interrupt.
Co-authored-by: WÁNG Xuěruì <1175567+xen0n@users.noreply.github.com>
Co-authored-by: WÁNG Xuěruì <1175567+xen0n@users.noreply.github.com>
…ls'. gcc/config/loongarch: * loongarch.c: Delete hook TARGET_MACHINE_DEPENDENT_REORG. * loongarch.opt: Delete mrelax-pic-calls.
gcc/config/loongarch: * loongarch-c.c: Delete no-float support. * loongarch.c: Likewise. * loongarch.opt: Likewise.
gcc/config/loongarch/ * loongarch.c: Remove some unused code.
gcc/config/loongarch/ * loongarch.c: Change movmemsi to cpymemsi. * loongarch.h: Likewise. * loongarch.md: Likewise.
Commit message shamelessly copied from 1777beb by jakub: This function, because it is sometimes called even outside of function bodies, uses create_tmp_var_raw rather than create_tmp_var. But in order for that to work, when first referenced, the VAR_DECLs need to appear in a TARGET_EXPR so that during gimplification the var gets the right DECL_CONTEXT and is added to local decls. gcc/ * config/loongarch/loongarch-builtins.c: Use TARGET_EXPR instead of MODIFY_EXPR.
* gcc/common/config/loongarch/loongarch-common.c: Added to compile parameter '-mmemcpy' processing * gcc/config.gcc: Add arch gs464v support. gcc/config/loongarch/ * gnu-user.h: Delete LINUX_DRIVER_SELF_SPECS. * loongarch-cpus.def: Modify gs464v's isa from 0 to 2. * loongarch-opts.h: * loongarch.c: Reimplementing handlers loongarch_option_override. * loongarch.h: Delete BASE_DRIVER_SELF_SPECS. * loongarch.opt: Define loongarch_isa_flags.
gcc/config/loongarch/ * loongarch-protos.h: Delete some unused codes. * loongarch.c: Likewise. * loongarch.h: Likewise. * loongarch.md: Likewise. * predicates.md: Likewise.
gcc/config/loongarch/ * loongarch.c: Remove DBX support. * loongarch.h: Likewise.
* gcc/config.gcc: Remove loongarch32 support. gcc/config/loongarch/ * larchintrin.h: Remove loongarch32 support. * loongarch-c.c: Likewise. * loongarch-cpus.def: Likewise. * loongarch-tables.opt: Likewise. * loongarch.c: Likewise. * loongarch.h: Likewise. * loongarch.md: Likewise. * loongarch.opt: Likewise.
libgcc/config/loongarch/ * linux-unwind.h: Replace 'struct _sig_ucontext_t' with 'struct ucontext_t'.
gcc/
* config.gcc: Revised '--with' handling logic.
* config.host: Likewise.
gcc/config/loongarch/
* driver-native.c: Removed.
* gnu-user.h: Reorganized .h definitions for readability.
* linux.h: Likewise.
* loongarch.h: Likewise.
* loongarch-opts.c: Revised '-m' handling logic.
* loongarch-opts.h: Likewise.
* loongarch.opt: Likewise
* loongarch-driver.c: Likewise.
* loongarch-driver.h: Likewise.
* loongarch.c: Likewise.
* t-loongarch: Likewise.
* x-native: Likewise, removed.
* genopt.sh: Likewise, removed.
* loongarch-cpus.c: Made processor property tables mostly static.
* loongarch-cpus.h: Likewise.
* loongarch-rtx-cost.h: Likewise.
* loongarch-cpus.def: Likewise, Removed.
* loongarch-tables.opt: Likeise, Removed.
* gs464v.md: Replaced condition macros.
* loongarch-c.c: Likewise.
* loongarch.md: Likewise.
* t-linux: Added basic multilib support (int/fp ABI).
commit 9069e94 Author: Kito Cheng <kito.cheng@sifive.com> Date: Mon Feb 24 10:54:21 2020 -0600 RISC-V: Adjust floating point code gen for LTGT compare - Using gcc.dg/torture/pr91323.c as testcase, so no new testcase introduced. - According latest GCC document LTGT and discussion on pr91323 LTGT should signals on NaNs, like GE/GT/LE/LT. - So we expand (LTGT a b) to ((LT a b) | (GT a b)) for fit the document. gcc/config/loongarch/ * loongarch.md: Change the code gen for LTGT.
Rationales: 1. No upstreamed GCC ports use "native" as the default. I don't think the upstream will accept this. 2. Using "native" as the default can cause severe trouble for distro maintainers. 3. Explicitly setting --with-arch=native when cross compiling actually *makes* sense: you can use a cross compiler (A) to cross compile a native compiler (B) for loongarch64 which uses -march=native by default. Compiler B runs on loongarch64 so it can detect the loongarch64 CPU feature at runtime. OTOH, the cross compiler, A, can't use -march=native because it generally does not run on a loongarch64 CPU. gcc/ * config.gcc: Use loongarch64 instead of native as the default --with-arch setting. And, accept --with-arch=native while cross compiling, but reject it for cross compiler.
…get triplet.
gcc/
* config.gcc: Use loongarch64 instead of native as the default
--with-arch setting.
gcc/config/loongarch
* loongarch-driver.c: print a message if the current ABI
and library/startfile search path do not match.
gcc/
* config/loongarch/loongarch-protos.h:
Modify loongarch_output_probe_stack_range declarations.
* config/loongarch/lloongarch.c
(loongarch_emit_probe_stack_range): Fix the condition for stack
probe loop unrolling, to avoid a dead loop causing segfault.
* config/loongarch/lloongarch.h: Define LARCH_PROLOGUE_TEMP3_REGNUM.
* config/loongarch/lloongarch.md
(probe_stack_range_<P:mode>): Add an input operand.
gcc/config/loongarch/
* loongarch.h: modified ASM_SPEC to include "%{mfloat-abi=*}".
+. change style. +. use hint 0x700 in dbar for future extension. +. add shortcut for dbar when ll.*/sc.* pair succeed, so remove TARGET_FIX_LOONGSON3_LLSC
xen0n
pushed a commit
to xen0n/gcc
that referenced
this pull request
Nov 27, 2021
* LoongArch: conform to LoongArch toolchain conventions. * LoongArch: toolchain convention conformance fix (patch 2) * LoongArch: toolchain convention conformance fix (patch 3) * Update loongarch.h * LoongArch: edit GCC documentation for command-line arguments. * fix coding style. * Update linux.h * bug fix: allow native-FPU configuration. * modify code style. * documents the "configure" interface for loongarch. * modify code style (patch 2). Co-authored-by: Yang Yujie <yangyujie@loongson.cn> Co-authored-by: ChengLulu <88083425+chenglulu326@users.noreply.github.com>
hubot
pushed a commit
that referenced
this pull request
Apr 8, 2024
In PR 113907 comment #58, Honza found a case where ICF thinks bodies of functions are equivalent but becaise of difference in aliases in a memory access, different aggregate jump functions are associated with supposedly equivalent call statements. This patch adds a way to compare jump functions and plugs it into ICF to avoid the issue. gcc/ChangeLog: 2024-03-20 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * ipa-prop.h (class ipa_vr): Declare new overload of a member function equal_p. (ipa_jump_functions_equivalent_p): Declare. * ipa-prop.cc (ipa_vr::equal_p): New function. (ipa_agg_pass_through_jf_equivalent_p): Likewise. (ipa_agg_jump_functions_equivalent_p): Likewise. (ipa_jump_functions_equivalent_p): Likewise. * ipa-cp.h (values_equal_for_ipcp_p): Declare. * ipa-cp.cc (values_equal_for_ipcp_p): Make function public. * ipa-icf-gimple.cc: Include alloc-pool.h, symbol-summary.h, sreal.h, ipa-cp.h and ipa-prop.h. (func_checker::compare_gimple_call): Comapre jump functions. gcc/testsuite/ChangeLog: 2024-03-20 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * gcc.dg/lto/pr113907_0.c: New. * gcc.dg/lto/pr113907_1.c: Likewise. * gcc.dg/lto/pr113907_2.c: Likewise.
hubot
pushed a commit
that referenced
this pull request
May 7, 2024
vsetvli local eliminate is only consider the current demand instead of full demand, and it will use that incomplete info to remove vsetvli. Give following example from PR114747: vsetvli a5,a1,e8,m4,ta,mu # 57, ratio=2, sew=8, lmul=4 vsetvli zero,a5,e16,m8,ta,ma # 58, ratio=2, sew=16, lmul=8 vle8.v v8,0(a0) # 13, demand ratio=2 vzext.vf2 v24,v8 # 14, demand sew=16 and lmul=8 Insn #58 will removed because #57 has satisfied demand of #13, but it's not consider #14. It should doing more demand analyze, but this bug only present in GCC 13 branch, and we should not change too much on this release branch, so the best way is make the check more conservative - remove only if the target vsetvl_discard_result having same SEW and LMUL as the source vsetvli. gcc/ChangeLog: PR target/114747 * config/riscv/riscv-vsetvl.cc (local_eliminate_vsetvl_insn): Check target vsetvl_discard_result and source vsetvli has same SEW and LMUL. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/vsetvl/pr114747.c: New.
hubot
pushed a commit
that referenced
this pull request
May 14, 2024
This is a manual backport of r14-9840-g1162861439fd3c from master. Manual because the bits and value range representation in jump functions have changes during the gcc 14 development cycle. In PR 113907 comment #58, Honza found a case where ICF thinks bodies of functions are equivalent but becaise of difference in aliases in a memory access, different aggregate jump functions are associated with supposedly equivalent call statements. This patch adds a way to compare jump functions and plugs it into ICF to avoid the issue. gcc/ChangeLog: 2024-05-14 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * ipa-prop.h (ipa_jump_functions_equivalent_p): Declare. (values_equal_for_ipcp_p): Likewise. * ipa-prop.cc (ipa_agg_pass_through_jf_equivalent_p): New function. (ipa_agg_jump_functions_equivalent_p): Likewise. (ipa_jump_functions_equivalent_p): Likewise. * ipa-cp.cc (values_equal_for_ipcp_p): Make function public. * ipa-icf-gimple.cc: Include alloc-pool.h, symbol-summary.h, sreal.h, ipa-cp.h and ipa-prop.h. (func_checker::compare_gimple_call): Comapre jump functions. gcc/testsuite/ChangeLog: 2024-05-10 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * gcc.dg/lto/pr113907_0.c: New. * gcc.dg/lto/pr113907_1.c: Likewise. * gcc.dg/lto/pr113907_2.c: Likewise.
Frosty515
pushed a commit
to FrostyOS-dev/gcc
that referenced
this pull request
May 23, 2024
vsetvli local eliminate is only consider the current demand instead of full demand, and it will use that incomplete info to remove vsetvli. Give following example from PR114747: vsetvli a5,a1,e8,m4,ta,mu # 57, ratio=2, sew=8, lmul=4 vsetvli zero,a5,e16,m8,ta,ma # 58, ratio=2, sew=16, lmul=8 vle8.v v8,0(a0) # 13, demand ratio=2 vzext.vf2 v24,v8 # 14, demand sew=16 and lmul=8 Insn gcc-mirror#58 will removed because gcc-mirror#57 has satisfied demand of gcc-mirror#13, but it's not consider gcc-mirror#14. It should doing more demand analyze, but this bug only present in GCC 13 branch, and we should not change too much on this release branch, so the best way is make the check more conservative - remove only if the target vsetvl_discard_result having same SEW and LMUL as the source vsetvli. gcc/ChangeLog: PR target/114747 * config/riscv/riscv-vsetvl.cc (local_eliminate_vsetvl_insn): Check target vsetvl_discard_result and source vsetvli has same SEW and LMUL. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/vsetvl/pr114747.c: New.
Frosty515
pushed a commit
to FrostyOS-dev/gcc
that referenced
this pull request
May 23, 2024
This is a manual backport of r14-9840-g1162861439fd3c from master. Manual because the bits and value range representation in jump functions have changes during the gcc 14 development cycle. In PR 113907 comment gcc-mirror#58, Honza found a case where ICF thinks bodies of functions are equivalent but becaise of difference in aliases in a memory access, different aggregate jump functions are associated with supposedly equivalent call statements. This patch adds a way to compare jump functions and plugs it into ICF to avoid the issue. gcc/ChangeLog: 2024-05-14 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * ipa-prop.h (ipa_jump_functions_equivalent_p): Declare. (values_equal_for_ipcp_p): Likewise. * ipa-prop.cc (ipa_agg_pass_through_jf_equivalent_p): New function. (ipa_agg_jump_functions_equivalent_p): Likewise. (ipa_jump_functions_equivalent_p): Likewise. * ipa-cp.cc (values_equal_for_ipcp_p): Make function public. * ipa-icf-gimple.cc: Include alloc-pool.h, symbol-summary.h, sreal.h, ipa-cp.h and ipa-prop.h. (func_checker::compare_gimple_call): Comapre jump functions. gcc/testsuite/ChangeLog: 2024-05-10 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * gcc.dg/lto/pr113907_0.c: New. * gcc.dg/lto/pr113907_1.c: Likewise. * gcc.dg/lto/pr113907_2.c: Likewise.
hubot
pushed a commit
that referenced
this pull request
May 28, 2024
This is a manual backport of r14-9840-g1162861439fd3c from master. Manual because the bits and value range representation in jump functions have changes during the gcc 14 development cycle. In PR 113907 comment #58, Honza found a case where ICF thinks bodies of functions are equivalent but becaise of difference in aliases in a memory access, different aggregate jump functions are associated with supposedly equivalent call statements. This patch adds a way to compare jump functions and plugs it into ICF to avoid the issue. gcc/ChangeLog: 2024-05-14 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * ipa-prop.h (ipa_jump_functions_equivalent_p): Declare. (values_equal_for_ipcp_p): Likewise. * ipa-prop.cc (ipa_agg_pass_through_jf_equivalent_p): New function. (ipa_agg_jump_functions_equivalent_p): Likewise. (ipa_jump_functions_equivalent_p): Likewise. * ipa-cp.cc (values_equal_for_ipcp_p): Make function public. * ipa-icf-gimple.cc: Include alloc-pool.h, symbol-summary.h, sreal.h, ipa-cp.h and ipa-prop.h. (func_checker::compare_gimple_call): Comapre jump functions. gcc/testsuite/ChangeLog: 2024-05-10 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * gcc.dg/lto/pr113907_0.c: New. * gcc.dg/lto/pr113907_1.c: Likewise. * gcc.dg/lto/pr113907_2.c: Likewise. (cherry picked from commit 1db45e8)
MBeijer
pushed a commit
to AmigaPorts/gcc
that referenced
this pull request
Jun 6, 2025
vsetvli local eliminate is only consider the current demand instead of full demand, and it will use that incomplete info to remove vsetvli. Give following example from PR114747: vsetvli a5,a1,e8,m4,ta,mu # 57, ratio=2, sew=8, lmul=4 vsetvli zero,a5,e16,m8,ta,ma # 58, ratio=2, sew=16, lmul=8 vle8.v v8,0(a0) # 13, demand ratio=2 vzext.vf2 v24,v8 # 14, demand sew=16 and lmul=8 Insn gcc-mirror#58 will removed because gcc-mirror#57 has satisfied demand of gcc-mirror#13, but it's not consider gcc-mirror#14. It should doing more demand analyze, but this bug only present in GCC 13 branch, and we should not change too much on this release branch, so the best way is make the check more conservative - remove only if the target vsetvl_discard_result having same SEW and LMUL as the source vsetvli. gcc/ChangeLog: PR target/114747 * config/riscv/riscv-vsetvl.cc (local_eliminate_vsetvl_insn): Check target vsetvl_discard_result and source vsetvli has same SEW and LMUL. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/vsetvl/pr114747.c: New.
MBeijer
pushed a commit
to AmigaPorts/gcc
that referenced
this pull request
Jun 6, 2025
This is a manual backport of r14-9840-g1162861439fd3c from master. Manual because the bits and value range representation in jump functions have changes during the gcc 14 development cycle. In PR 113907 comment gcc-mirror#58, Honza found a case where ICF thinks bodies of functions are equivalent but becaise of difference in aliases in a memory access, different aggregate jump functions are associated with supposedly equivalent call statements. This patch adds a way to compare jump functions and plugs it into ICF to avoid the issue. gcc/ChangeLog: 2024-05-14 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * ipa-prop.h (ipa_jump_functions_equivalent_p): Declare. (values_equal_for_ipcp_p): Likewise. * ipa-prop.cc (ipa_agg_pass_through_jf_equivalent_p): New function. (ipa_agg_jump_functions_equivalent_p): Likewise. (ipa_jump_functions_equivalent_p): Likewise. * ipa-cp.cc (values_equal_for_ipcp_p): Make function public. * ipa-icf-gimple.cc: Include alloc-pool.h, symbol-summary.h, sreal.h, ipa-cp.h and ipa-prop.h. (func_checker::compare_gimple_call): Comapre jump functions. gcc/testsuite/ChangeLog: 2024-05-10 Martin Jambor <mjambor@suse.cz> PR ipa/113907 * gcc.dg/lto/pr113907_0.c: New. * gcc.dg/lto/pr113907_1.c: Likewise. * gcc.dg/lto/pr113907_2.c: Likewise.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.