From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 2063) id 55191385C332; Fri, 30 Sep 2022 12:18:31 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 55191385C332 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1664540311; bh=Q01YT0WvnPWsJEKoQaeeATjvg7Q0Fg8aw0tbLTA8l/s=; h=From:To:Subject:Date:From; b=HHl+YqqmfIbVYF/IHnAFFMG4nhhtO4E6KWd+uUMBLv5sz+5roMYd606pSTo1NsWVv JR7I0Ddhei75gyedHAP/y9sQVrusMakx1rISPvLQ1glDQLLVwLUDdox2kVclCMna5o nHm+WsHYDfR4Puhl8dJp/HZlczrnamZ9wPIv3Gco= MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="utf-8" From: Kewen Lin To: gcc-cvs@gcc.gnu.org Subject: [gcc r13-2984] rs6000: Rework ELFv2 support for -fpatchable-function-entry* [PR99888] X-Act-Checkin: gcc X-Git-Author: Kewen Lin X-Git-Refname: refs/heads/master X-Git-Oldrev: c28957176dc0adff6e4b8fcbe6b91f5b82900563 X-Git-Newrev: c23b5006d3ffeda1a9edf5fd817765a6da3696ca Message-Id: <20220930121831.55191385C332@sourceware.org> Date: Fri, 30 Sep 2022 12:18:31 +0000 (GMT) List-Id: https://gcc.gnu.org/g:c23b5006d3ffeda1a9edf5fd817765a6da3696ca commit r13-2984-gc23b5006d3ffeda1a9edf5fd817765a6da3696ca Author: Kewen Lin Date: Fri Sep 30 07:16:49 2022 -0500 rs6000: Rework ELFv2 support for -fpatchable-function-entry* [PR99888] As PR99888 and its related show, the current support for -fpatchable-function-entry on powerpc ELFv2 doesn't work well with global entry existence. For example, with one command line option -fpatchable-function-entry=3,2, it got below w/o this patch: .LPFE1: nop nop .type foo, @function foo: nop .LFB0: .cfi_startproc .LCF0: 0: addis 2,12,.TOC.-.LCF0@ha addi 2,2,.TOC.-.LCF0@l .localentry foo,.-foo , the assembly is unexpected since the patched nops have no effects when being entered from local entry. This patch is to update the nops patched before and after local entry, it looks like: .type foo, @function foo: .LFB0: .cfi_startproc .LCF0: 0: addis 2,12,.TOC.-.LCF0@ha addi 2,2,.TOC.-.LCF0@l nop nop .localentry foo,.-foo nop PR target/99888 PR target/105649 gcc/ChangeLog: * doc/invoke.texi (option -fpatchable-function-entry): Adjust the documentation for PowerPC ELFv2 ABI dual entry points. * config/rs6000/rs6000-internal.h (rs6000_print_patchable_function_entry): New function declaration. * config/rs6000/rs6000-logue.cc (rs6000_output_function_prologue): Support patchable-function-entry by emitting nops before and after local entry for the function that needs global entry. * config/rs6000/rs6000.cc (rs6000_print_patchable_function_entry): Skip the function that needs global entry till global entry has been emitted. * config/rs6000/rs6000.h (struct machine_function): New bool member global_entry_emitted. gcc/testsuite/ChangeLog: * gcc.target/powerpc/pr99888-1.c: New test. * gcc.target/powerpc/pr99888-2.c: New test. * gcc.target/powerpc/pr99888-3.c: New test. * gcc.target/powerpc/pr99888-4.c: New test. * gcc.target/powerpc/pr99888-5.c: New test. * gcc.target/powerpc/pr99888-6.c: New test. * c-c++-common/patchable_function_entry-default.c: Adjust for powerpc_elfv2 to avoid compilation error. Diff: --- gcc/config/rs6000/rs6000-internal.h | 4 ++ gcc/config/rs6000/rs6000-logue.cc | 32 ++++++++++++++++ gcc/config/rs6000/rs6000.cc | 10 ++++- gcc/config/rs6000/rs6000.h | 4 ++ gcc/doc/invoke.texi | 8 +++- .../patchable_function_entry-default.c | 3 ++ gcc/testsuite/gcc.target/powerpc/pr99888-1.c | 43 ++++++++++++++++++++++ gcc/testsuite/gcc.target/powerpc/pr99888-2.c | 43 ++++++++++++++++++++++ gcc/testsuite/gcc.target/powerpc/pr99888-3.c | 11 ++++++ gcc/testsuite/gcc.target/powerpc/pr99888-4.c | 13 +++++++ gcc/testsuite/gcc.target/powerpc/pr99888-5.c | 13 +++++++ gcc/testsuite/gcc.target/powerpc/pr99888-6.c | 14 +++++++ 12 files changed, 194 insertions(+), 4 deletions(-) diff --git a/gcc/config/rs6000/rs6000-internal.h b/gcc/config/rs6000/rs6000-internal.h index b9e82c0468d..e75b8d5c7e8 100644 --- a/gcc/config/rs6000/rs6000-internal.h +++ b/gcc/config/rs6000/rs6000-internal.h @@ -182,6 +182,10 @@ extern tree rs6000_fold_builtin (tree fndecl ATTRIBUTE_UNUSED, tree *args ATTRIBUTE_UNUSED, bool ignore ATTRIBUTE_UNUSED); +extern void rs6000_print_patchable_function_entry (FILE *, + unsigned HOST_WIDE_INT, + bool); + extern bool rs6000_passes_float; extern bool rs6000_passes_long_double; extern bool rs6000_passes_vector; diff --git a/gcc/config/rs6000/rs6000-logue.cc b/gcc/config/rs6000/rs6000-logue.cc index a11d020ccd0..3621cb501c7 100644 --- a/gcc/config/rs6000/rs6000-logue.cc +++ b/gcc/config/rs6000/rs6000-logue.cc @@ -4009,11 +4009,43 @@ rs6000_output_function_prologue (FILE *file) fprintf (file, "\tadd 2,2,12\n"); } + unsigned short patch_area_size = crtl->patch_area_size; + unsigned short patch_area_entry = crtl->patch_area_entry; + /* Need to emit the patching area. */ + if (patch_area_size > 0) + { + cfun->machine->global_entry_emitted = true; + /* As ELFv2 ABI shows, the allowable bytes between the global + and local entry points are 0, 4, 8, 16, 32 and 64 when + there is a local entry point. Considering there are two + non-prefixed instructions for global entry point prologue + (8 bytes), the count for patchable nops before local entry + point would be 2, 6 and 14. It's possible to support those + other counts of nops by not making a local entry point, but + we don't have clear use cases for them, so leave them + unsupported for now. */ + if (patch_area_entry > 0) + { + if (patch_area_entry != 2 + && patch_area_entry != 6 + && patch_area_entry != 14) + error ("unsupported number of nops before function entry (%u)", + patch_area_entry); + rs6000_print_patchable_function_entry (file, patch_area_entry, + true); + patch_area_size -= patch_area_entry; + } + } + fputs ("\t.localentry\t", file); assemble_name (file, name); fputs (",.-", file); assemble_name (file, name); fputs ("\n", file); + /* Emit the nops after local entry. */ + if (patch_area_size > 0) + rs6000_print_patchable_function_entry (file, patch_area_size, + patch_area_entry == 0); } else if (rs6000_pcrel_p ()) diff --git a/gcc/config/rs6000/rs6000.cc b/gcc/config/rs6000/rs6000.cc index bbe21eacc6b..b9496d7f268 100644 --- a/gcc/config/rs6000/rs6000.cc +++ b/gcc/config/rs6000/rs6000.cc @@ -14930,8 +14930,14 @@ rs6000_print_patchable_function_entry (FILE *file, if (!(TARGET_64BIT && DEFAULT_ABI != ABI_ELFv2) && HAVE_GAS_SECTION_LINK_ORDER) flags |= SECTION_LINK_ORDER; - default_print_patchable_function_entry_1 (file, patch_area_size, record_p, - flags); + bool global_entry_needed_p = rs6000_global_entry_point_prologue_needed_p (); + /* For a function which needs global entry point, we will emit the + patchable area before and after local entry point under the control of + cfun->machine->global_entry_emitted, see the handling in function + rs6000_output_function_prologue. */ + if (!global_entry_needed_p || cfun->machine->global_entry_emitted) + default_print_patchable_function_entry_1 (file, patch_area_size, record_p, + flags); } enum rtx_code diff --git a/gcc/config/rs6000/rs6000.h b/gcc/config/rs6000/rs6000.h index eb7b2158497..b4df22b6030 100644 --- a/gcc/config/rs6000/rs6000.h +++ b/gcc/config/rs6000/rs6000.h @@ -2435,6 +2435,10 @@ typedef struct GTY(()) machine_function bool lr_is_wrapped_separately; bool toc_is_wrapped_separately; bool mma_return_type_error; + /* Indicate global entry is emitted, only useful when the function requires + global entry. It helps to control the patchable area before and after + local entry. */ + bool global_entry_emitted; } machine_function; #endif diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi index 2ac9cfc35f9..518bfdf0867 100644 --- a/gcc/doc/invoke.texi +++ b/gcc/doc/invoke.texi @@ -16884,9 +16884,13 @@ the area size or to remove it completely on a single function. If @code{N=0}, no pad location is recorded. The NOP instructions are inserted at---and maybe before, depending on -@var{M}---the function entry address, even before the prologue. +@var{M}---the function entry address, even before the prologue. On +PowerPC with the ELFv2 ABI, for a function with dual entry points, +the local entry point is this function entry address. -The maximum value of @var{N} and @var{M} is 65535. +The maximum value of @var{N} and @var{M} is 65535. On PowerPC with the +ELFv2 ABI, for a function with dual entry points, the supported values +for @var{M} are 0, 2, 6 and 14. @end table diff --git a/gcc/testsuite/c-c++-common/patchable_function_entry-default.c b/gcc/testsuite/c-c++-common/patchable_function_entry-default.c index 7036f7bfbea..a501efccb19 100644 --- a/gcc/testsuite/c-c++-common/patchable_function_entry-default.c +++ b/gcc/testsuite/c-c++-common/patchable_function_entry-default.c @@ -1,6 +1,9 @@ /* { dg-do compile { target { ! { nvptx*-*-* visium-*-* } } } } */ /* { dg-options "-O2 -fpatchable-function-entry=3,1" } */ /* { dg-additional-options "-fno-pie" { target sparc*-*-* } } */ +/* See PR99888, one single preceding nop isn't allowed on powerpc_elfv2, + so overriding with two preceding nops to make it pass there. */ +/* { dg-additional-options "-fpatchable-function-entry=3,2" { target powerpc_elfv2 } } */ /* { dg-final { scan-assembler-times "nop|NOP|SWYM" 3 { target { ! { alpha*-*-* } } } } } */ /* { dg-final { scan-assembler-times "bis" 3 { target alpha*-*-* } } } */ diff --git a/gcc/testsuite/gcc.target/powerpc/pr99888-1.c b/gcc/testsuite/gcc.target/powerpc/pr99888-1.c new file mode 100644 index 00000000000..9370b4e7438 --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr99888-1.c @@ -0,0 +1,43 @@ +/* Verify no errors for different nops after local entry on ELFv2. */ + +extern int a; + +__attribute__ ((noipa, patchable_function_entry (1, 0))) +int test1 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (2, 0))) +int test2 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (3, 0))) +int test3 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (4, 0))) +int test4 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (5, 0))) +int test5 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (6, 0))) +int test6 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (7, 0))) +int test7 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (8, 0))) +int test8 (int b) { + return a + b; +} diff --git a/gcc/testsuite/gcc.target/powerpc/pr99888-2.c b/gcc/testsuite/gcc.target/powerpc/pr99888-2.c new file mode 100644 index 00000000000..45061712602 --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr99888-2.c @@ -0,0 +1,43 @@ +/* Verify no errors for 2, 6 and 14 nops before local entry on ELFv2. */ + +extern int a; + +__attribute__ ((noipa, patchable_function_entry (2, 2))) +int test1 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (4, 2))) +int test2 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (6, 6))) +int test3 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (8, 6))) +int test4 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (16, 6))) +int test5 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (14, 14))) +int test6 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (28, 14))) +int test7 (int b) { + return a + b; +} + +__attribute__ ((noipa, patchable_function_entry (64, 14))) +int test8 (int b) { + return a + b; +} diff --git a/gcc/testsuite/gcc.target/powerpc/pr99888-3.c b/gcc/testsuite/gcc.target/powerpc/pr99888-3.c new file mode 100644 index 00000000000..4531ae32036 --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr99888-3.c @@ -0,0 +1,11 @@ +/* { dg-options "-fpatchable-function-entry=1" } */ + +/* Verify no errors on ELFv2, using command line option instead of + function attribute. */ + +extern int a; + +int test (int b) { + return a + b; +} + diff --git a/gcc/testsuite/gcc.target/powerpc/pr99888-4.c b/gcc/testsuite/gcc.target/powerpc/pr99888-4.c new file mode 100644 index 00000000000..00a8d4d316e --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr99888-4.c @@ -0,0 +1,13 @@ +/* { dg-require-effective-target powerpc_elfv2 } */ +/* There is no global entry point prologue with pcrel. */ +/* { dg-options "-mno-pcrel -fpatchable-function-entry=1,1" } */ + +/* Verify one error emitted for unexpected 1 nop before local + entry. */ + +extern int a; + +int test (int b) { + return a + b; +} +/* { dg-error "unsupported number of nops before function entry \\(1\\)" "" { target *-*-* } .-1 } */ diff --git a/gcc/testsuite/gcc.target/powerpc/pr99888-5.c b/gcc/testsuite/gcc.target/powerpc/pr99888-5.c new file mode 100644 index 00000000000..39d3b4465f1 --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr99888-5.c @@ -0,0 +1,13 @@ +/* { dg-require-effective-target powerpc_elfv2 } */ +/* There is no global entry point prologue with pcrel. */ +/* { dg-options "-mno-pcrel -fpatchable-function-entry=7,3" } */ + +/* Verify one error emitted for unexpected 3 nops before local + entry. */ + +extern int a; + +int test (int b) { + return a + b; +} +/* { dg-error "unsupported number of nops before function entry \\(3\\)" "" { target *-*-* } .-1 } */ diff --git a/gcc/testsuite/gcc.target/powerpc/pr99888-6.c b/gcc/testsuite/gcc.target/powerpc/pr99888-6.c new file mode 100644 index 00000000000..c6c18dcc7ac --- /dev/null +++ b/gcc/testsuite/gcc.target/powerpc/pr99888-6.c @@ -0,0 +1,14 @@ +/* { dg-require-effective-target powerpc_elfv2 } */ +/* There is no global entry point prologue with pcrel. */ +/* { dg-options "-mno-pcrel" } */ + +/* Verify one error emitted for unexpected 4 nops before local + entry. */ + +extern int a; + +__attribute__ ((patchable_function_entry (20, 4))) +int test (int b) { + return a + b; +} +/* { dg-error "unsupported number of nops before function entry \\(4\\)" "" { target *-*-* } .-1 } */