rpms/gcc
7
0
Fork 0
Code Issues Pull Requests Releases Activity

import gcc-8.5.0-10.el8

Browse Source
This commit is contained in:
CentOS Sources 2022-05-10 03:18:55 -04:00 committed by Stepan Oksanichenko
parent a2480e2f4e
commit fafdb81fd7
6 changed files with 1843 additions and 132 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

635
SOURCES/gcc8-add-Wbidirectional.patch → SOURCES/gcc8-Wbidi-chars.patch
View File

@ -1,25 +1,122 @@
commit 51c500269bf53749b107807d84271385fad35628
Author: Marek Polacek <polacek@redhat.com>
Date: Wed Oct 6 14:33:59 2021 -0400
libcpp: Implement -Wbidi-chars for CVE-2021-42574 [PR103026]
From a link below:
"An issue was discovered in the Bidirectional Algorithm in the Unicode
Specification through 14.0. It permits the visual reordering of
characters via control sequences, which can be used to craft source code
that renders different logic than the logical ordering of tokens
ingested by compilers and interpreters. Adversaries can leverage this to
encode source code for compilers accepting Unicode such that targeted
vulnerabilities are introduced invisibly to human reviewers."
More info:
https://nvd.nist.gov/vuln/detail/CVE-2021-42574
https://trojansource.codes/
This is not a compiler bug. However, to mitigate the problem, this patch
implements -Wbidi-chars=[none|unpaired|any] to warn about possibly
misleading Unicode bidirectional control characters the preprocessor may
encounter.
The default is =unpaired, which warns about improperly terminated
bidirectional control characters; e.g. a LRE without its corresponding PDF.
The level =any warns about any use of bidirectional control characters.
This patch handles both UCNs and UTF-8 characters. UCNs designating
bidi characters in identifiers are accepted since r204886. Then r217144
enabled -fextended-identifiers by default. Extended characters in C/C++
identifiers have been accepted since r275979. However, this patch still
warns about mixing UTF-8 and UCN bidi characters; there seems to be no
good reason to allow mixing them.
We warn in different contexts: comments (both C and C++-style), string
literals, character constants, and identifiers. Expectedly, UCNs are ignored
in comments and raw string literals. The bidirectional control characters
can nest so this patch handles that as well.
I have not included nor tested this at all with Fortran (which also has
string literals and line comments).
Dave M. posted patches improving diagnostic involving Unicode characters.
This patch does not make use of this new infrastructure yet.
PR preprocessor/103026
gcc/c-family/ChangeLog:
* c.opt (Wbidi-chars, Wbidi-chars=): New option.
gcc/ChangeLog:
* doc/invoke.texi: Document -Wbidi-chars.
libcpp/ChangeLog:
* include/cpplib.h (enum cpp_bidirectional_level): New.
(struct cpp_options): Add cpp_warn_bidirectional.
(enum cpp_warning_reason): Add CPP_W_BIDIRECTIONAL.
* internal.h (struct cpp_reader): Add warn_bidi_p member
function.
* init.c (cpp_create_reader): Set cpp_warn_bidirectional.
* lex.c (bidi): New namespace.
(get_bidi_utf8): New function.
(get_bidi_ucn): Likewise.
(maybe_warn_bidi_on_close): Likewise.
(maybe_warn_bidi_on_char): Likewise.
(_cpp_skip_block_comment): Implement warning about bidirectional
control characters.
(skip_line_comment): Likewise.
(forms_identifier_p): Likewise.
(lex_identifier): Likewise.
(lex_string): Likewise.
(lex_raw_string): Likewise.
gcc/testsuite/ChangeLog:
* c-c++-common/Wbidi-chars-1.c: New test.
* c-c++-common/Wbidi-chars-2.c: New test.
* c-c++-common/Wbidi-chars-3.c: New test.
* c-c++-common/Wbidi-chars-4.c: New test.
* c-c++-common/Wbidi-chars-5.c: New test.
* c-c++-common/Wbidi-chars-6.c: New test.
* c-c++-common/Wbidi-chars-7.c: New test.
* c-c++-common/Wbidi-chars-8.c: New test.
* c-c++-common/Wbidi-chars-9.c: New test.
* c-c++-common/Wbidi-chars-10.c: New test.
* c-c++-common/Wbidi-chars-11.c: New test.
* c-c++-common/Wbidi-chars-12.c: New test.
* c-c++-common/Wbidi-chars-13.c: New test.
* c-c++-common/Wbidi-chars-14.c: New test.
* c-c++-common/Wbidi-chars-15.c: New test.
* c-c++-common/Wbidi-chars-16.c: New test.
* c-c++-common/Wbidi-chars-17.c: New test.
diff --git a/gcc/c-family/c.opt b/gcc/c-family/c.opt
index f591b39be5a..3e9a97f4ae2 100644
index f591b39be5a..cf922812198 100644
--- a/gcc/c-family/c.opt
+++ b/gcc/c-family/c.opt
@@ -334,6 +334,30 @@ Wbad-function-cast
C ObjC Var(warn_bad_function_cast) Warning
Warn about casting functions to incompatible types.
+Wbidirectional
+C ObjC C++ ObjC++ Warning Alias(Wbidirectional=,any,none)
+Wbidi-chars
+C ObjC C++ ObjC++ Warning Alias(Wbidi-chars=,any,none)
+;
+
+Wbidirectional=
+Wbidi-chars=
+C ObjC C++ ObjC++ RejectNegative Joined Warning CPP(cpp_warn_bidirectional) CppReason(CPP_W_BIDIRECTIONAL) Var(warn_bidirectional) Init(bidirectional_unpaired) Enum(cpp_bidirectional_level)
+-Wbidirectional=[none|unpaired|any] Warn about UTF-8 bidirectional characters.
+-Wbidi-chars=[none|unpaired|any] Warn about UTF-8 bidirectional control characters.
+
+; Required for these enum values.
+SourceInclude
+cpplib.h
+
+Enum
+Name(cpp_bidirectional_level) Type(int) UnknownError(argument %qs to %<-Wbidirectional%> not recognized)
+Name(cpp_bidirectional_level) Type(int) UnknownError(argument %qs to %<-Wbidi-chars%> not recognized)
+
+EnumValue
+Enum(cpp_bidirectional_level) String(none) Value(bidirectional_none)
@ -34,48 +131,50 @@ index f591b39be5a..3e9a97f4ae2 100644
C ObjC C++ ObjC++ Var(warn_bool_compare) Warning LangEnabledBy(C ObjC C++ ObjC++,Wall)
Warn about boolean expression compared with an integer value different from true/false.
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index 78ca7738df2..7342dd2e142 100644
index 78ca7738df2..cc85c53aede 100644
--- a/gcc/doc/invoke.texi
+++ b/gcc/doc/invoke.texi
@@ -264,7 +264,9 @@ Objective-C and Objective-C++ Dialects}.
@@ -264,7 +264,8 @@ Objective-C and Objective-C++ Dialects}.
-Walloc-zero -Walloc-size-larger-than=@var{n}
-Walloca -Walloca-larger-than=@var{n} @gol
-Wno-aggressive-loop-optimizations -Warray-bounds -Warray-bounds=@var{n} @gol
--Wno-attributes -Wbool-compare -Wbool-operation @gol
+-Wno-attributes @gol
+-Wbidirectional=@r{[}none@r{|}unpaired@r{|}any@r{]} @gol
+-Wno-attributes -Wbidi-chars=@r{[}none@r{|}unpaired@r{|}any@r{]} @gol
+-Wbool-compare -Wbool-operation @gol
-Wno-builtin-declaration-mismatch @gol
-Wno-builtin-macro-redefined -Wc90-c99-compat -Wc99-c11-compat @gol
-Wc++-compat -Wc++11-compat -Wc++14-compat @gol
@@ -5606,6 +5608,21 @@ Warn about declarations using the @code{alias} and similar attributes whose
@@ -5606,6 +5607,23 @@ Warn about declarations using the @code{alias} and similar attributes whose
target is incompatible with the type of the alias. @xref{Function Attributes,
,Declaring Attributes of Functions}.
+@item -Wbidirectional=@r{[}none@r{|}unpaired@r{|}any@r{]}
+@opindex Wbidirectional=
+@opindex Wbidirectional
+@opindex Wno-bidirectional
+Warn about UTF-8 bidirectional characters. Such characters can change
+left-to-right writing direction into right-to-left (and vice versa),
+which can cause confusion between the logical order and visual order.
+This may be dangerous; for instance, it may seem that a piece of code
+is not commented out, whereas it in fact is.
+@item -Wbidi-chars=@r{[}none@r{|}unpaired@r{|}any@r{]}
+@opindex Wbidi-chars=
+@opindex Wbidi-chars
+@opindex Wno-bidi-chars
+Warn about possibly misleading UTF-8 bidirectional control characters in
+comments, string literals, character constants, and identifiers. Such
+characters can change left-to-right writing direction into right-to-left
+(and vice versa), which can cause confusion between the logical order and
+visual order. This may be dangerous; for instance, it may seem that a piece
+of code is not commented out, whereas it in fact is.
+
+There are three levels of warning supported by GCC@. The default is
+@option{-Wbidirectional=unpaired}, which warns about improperly terminated
+bidi contexts. @option{-Wbidirectional=none} turns the warning off.
+@option{-Wbidirectional=any} warns about any use of bidirectional characters.
+@option{-Wbidi-chars=unpaired}, which warns about improperly terminated
+bidi contexts. @option{-Wbidi-chars=none} turns the warning off.
+@option{-Wbidi-chars=any} warns about any use of bidirectional control
+characters.
+
@item -Wbool-compare
@opindex Wno-bool-compare
@opindex Wbool-compare
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-1.c b/gcc/testsuite/c-c++-common/Wbidirectional-1.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-1.c b/gcc/testsuite/c-c++-common/Wbidi-chars-1.c
new file mode 100644
index 00000000000..750de81fdd8
index 00000000000..34f5ac19271
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-1.c
@@ -0,0 +1,11 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-1.c
@@ -0,0 +1,12 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+
+int main() {
@ -87,14 +186,15 @@ index 00000000000..750de81fdd8
+/* { dg-warning "bidirectional" "" { target *-*-* } .-1 } */
+ return 0;
+}
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-10.c b/gcc/testsuite/c-c++-common/Wbidirectional-10.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-10.c b/gcc/testsuite/c-c++-common/Wbidi-chars-10.c
new file mode 100644
index 00000000000..cd4abeeefbd
index 00000000000..3f851b69e65
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-10.c
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-10.c
@@ -0,0 +1,27 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=unpaired" } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* More nesting testing. */
+
+/* RLE‫ LRI⁦ PDF‬ PDI⁩*/
@ -107,7 +207,6 @@ index 00000000000..cd4abeeefbd
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLE_\u202b_RLI_\u2067_PDI_\u2069_PDF_\u202c;
+int FSI_\u2068_LRO_\u202d_PDI_\u2069_PDF_\u202c;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int FSI_\u2068;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int FSI_\u2068_PDI_\u2069;
@ -120,28 +219,30 @@ index 00000000000..cd4abeeefbd
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLI_\u2067_RLI_\u2067_RLI_\u2067_RLI_\u2067_FSI_\u2068_PDI_\u2069_PDI_\u2069_PDI_\u2069_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-11.c b/gcc/testsuite/c-c++-common/Wbidirectional-11.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-11.c b/gcc/testsuite/c-c++-common/Wbidi-chars-11.c
new file mode 100644
index 00000000000..89ed6e8affc
index 00000000000..44d044d82de
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-11.c
@@ -0,0 +1,8 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-11.c
@@ -0,0 +1,9 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=unpaired" } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test that we warn when mixing UCN and UTF-8. */
+
+const char *s1 = "LRE_‪_PDF_\u202c";
+/* { dg-warning "mismatch" "" { target *-*-* } .-1 } */
+const char *s2 = "LRE_\u202a_PDF_‬";
+/* { dg-warning "mismatch" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-12.c b/gcc/testsuite/c-c++-common/Wbidirectional-12.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-12.c b/gcc/testsuite/c-c++-common/Wbidi-chars-12.c
new file mode 100644
index 00000000000..20d1566401a
index 00000000000..b07eec1da91
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-12.c
@@ -0,0 +1,18 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-12.c
@@ -0,0 +1,19 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile { target { c || c++11 } } } */
+/* { dg-options "-Wbidirectional=any" } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test raw strings. */
+
+const char *s1 = R"(a b c LRE‪ 1 2 3 PDF‬ x y z)";
@ -158,14 +259,15 @@ index 00000000000..20d1566401a
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
+const char *s9 = R"(a b c PDF‬ x y z)";
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-13.c b/gcc/testsuite/c-c++-common/Wbidirectional-13.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-13.c b/gcc/testsuite/c-c++-common/Wbidi-chars-13.c
new file mode 100644
index 00000000000..08010e3b37b
index 00000000000..b2dd9fde752
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-13.c
@@ -0,0 +1,16 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-13.c
@@ -0,0 +1,17 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile { target { c || c++11 } } } */
+/* { dg-options "-Wbidirectional=unpaired" } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test raw strings. */
+
+const char *s1 = R"(a b c LRE‪ 1 2 3)";
@ -180,12 +282,190 @@ index 00000000000..08010e3b37b
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+const char *s6 = R"(a b c RLI⁧ 1 2 3)";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-2.c b/gcc/testsuite/c-c++-common/Wbidirectional-2.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-14.c b/gcc/testsuite/c-c++-common/Wbidi-chars-14.c
new file mode 100644
index 00000000000..4e04202e058
index 00000000000..ba5f75d9553
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-2.c
@@ -0,0 +1,8 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-14.c
@@ -0,0 +1,38 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test PDI handling, which also pops any subsequent LREs, RLEs, LROs,
+ or RLOs. */
+
+/* LRI_⁦_LRI_⁦_RLE_‫_RLE_‫_RLE_‫_PDI_⁩*/
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+// LRI_⁦_RLE_‫_RLE_‫_RLE_‫_PDI_⁩
+// LRI_⁦_RLO_‮_RLE_‫_RLE_‫_PDI_⁩
+// LRI_⁦_RLO_‮_RLE_‫_PDI_⁩
+// FSI_⁨_RLO_‮_PDI_⁩
+// FSI_⁨_FSI_⁨_RLO_‮_PDI_⁩
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+int LRI_\u2066_LRI_\u2066_LRE_\u202a_LRE_\u202a_LRE_\u202a_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int LRI_\u2066_LRI_\u2066_LRE_\u202a_LRE_\u202a_LRE_\u202a_PDI_\u2069_PDI_\u2069;
+int LRI_\u2066_LRI_\u2066_LRI_\u2066_LRE_\u202a_LRE_\u202a_LRE_\u202a_PDI_\u2069_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int PDI_\u2069;
+int LRI_\u2066_PDI_\u2069;
+int RLI_\u2067_PDI_\u2069;
+int LRE_\u202a_LRI_\u2066_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int LRI_\u2066_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+int LRI_\u2066_LRE_\u202a_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+int RLI_\u2067_LRI_\u2066_LRE_\u202a_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int FSI_\u2068_LRI_\u2066_LRE_\u202a_LRE_\u202a_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLO_\u202e_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int RLI_\u2067_PDI_\u2069_RLI_\u2067;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+int FSI_\u2068_PDF_\u202c_PDI_\u2069;
+int FSI_\u2068_FSI_\u2068_PDF_\u202c_PDI_\u2069;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-15.c b/gcc/testsuite/c-c++-common/Wbidi-chars-15.c
new file mode 100644
index 00000000000..a0ce8ff5e2c
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-15.c
@@ -0,0 +1,59 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test unpaired bidi control chars in multiline comments. */
+
+/*
+ * LRE‪ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * RLE‫ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * LRO‭ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * RLO‮ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * LRI⁦ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * RLI⁧ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/*
+ * FSI⁨ end
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/* LRE‪
+ PDF‬ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+/* FSI⁨
+ PDI⁩ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+
+/* LRE<‪>
+ *
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-3 } */
+
+/*
+ * LRE<‪>
+ */
+/* { dg-warning "unpaired" "" { target *-*-* } .-2 } */
+
+/*
+ *
+ * LRE<‪> */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+
+/* RLI<⁧> */ /* PDI<⁩> */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
+/* LRE<‪> */ /* PDF<‬> */
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-16.c b/gcc/testsuite/c-c++-common/Wbidi-chars-16.c
new file mode 100644
index 00000000000..baa0159861c
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-16.c
@@ -0,0 +1,26 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test LTR/RTL chars. */
+
+/* LTR<‎> */
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+// LTR<‎>
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+/* RTL<‏> */
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+// RTL<‏>
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+
+const char *s1 = "LTR<‎>";
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+const char *s2 = "LTR\u200e";
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+const char *s3 = "LTR\u200E";
+/* { dg-warning "U\\+200E" "" { target *-*-* } .-1 } */
+const char *s4 = "RTL<‏>";
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+const char *s5 = "RTL\u200f";
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
+const char *s6 = "RTL\u200F";
+/* { dg-warning "U\\+200F" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-17.c b/gcc/testsuite/c-c++-common/Wbidi-chars-17.c
new file mode 100644
index 00000000000..07cb4321f96
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-17.c
@@ -0,0 +1,30 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test LTR/RTL chars. */
+
+/* LTR<‎> */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// LTR<‎>
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* RTL<‏> */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+// RTL<‏>
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int ltr_\u200e;
+/* { dg-error "universal character " "" { target *-*-* } .-1 } */
+int rtl_\u200f;
+/* { dg-error "universal character " "" { target *-*-* } .-1 } */
+
+const char *s1 = "LTR<‎>";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s2 = "LTR\u200e";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s3 = "LTR\u200E";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s4 = "RTL<‏>";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s5 = "RTL\u200f";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+const char *s6 = "RTL\u200F";
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-2.c b/gcc/testsuite/c-c++-common/Wbidi-chars-2.c
new file mode 100644
index 00000000000..2340374f276
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-2.c
@@ -0,0 +1,9 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+
+int main() {
@ -194,12 +474,13 @@ index 00000000000..4e04202e058
+ __builtin_printf("Hello world.\n");
+ return 0;
+}
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-3.c b/gcc/testsuite/c-c++-common/Wbidirectional-3.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-3.c b/gcc/testsuite/c-c++-common/Wbidi-chars-3.c
new file mode 100644
index 00000000000..921300e94e0
index 00000000000..9dc7edb6e64
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-3.c
@@ -0,0 +1,10 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-3.c
@@ -0,0 +1,11 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+
+int main() {
@ -210,14 +491,15 @@ index 00000000000..921300e94e0
+ }
+ return 0;
+}
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-4.c b/gcc/testsuite/c-c++-common/Wbidirectional-4.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-4.c b/gcc/testsuite/c-c++-common/Wbidi-chars-4.c
new file mode 100644
index 00000000000..bdf334d9986
index 00000000000..49f856b9bfe
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-4.c
@@ -0,0 +1,149 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-4.c
@@ -0,0 +1,172 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=any -Wno-multichar -Wno-overflow" } */
+/* { dg-options "-Wbidi-chars=any -Wno-multichar -Wno-overflow" } */
+/* Test all bidi chars in various contexts (identifiers, comments,
+ string literals, character constants), both UCN and UTF-8. The bidi
+ chars here are properly terminated, except for the character constants. */
@ -263,6 +545,28 @@ index 00000000000..bdf334d9986
+// a b c PDF‬ x y z
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+
+/* Multiline comments. */
+/* a b c PDI⁩ x y z
+ */
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-2 } */
+/* a b c PDF‬ x y z
+ */
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-2 } */
+/* first
+ a b c PDI⁩ x y z
+ */
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-2 } */
+/* first
+ a b c PDF‬ x y z
+ */
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-2 } */
+/* first
+ a b c PDI⁩ x y z */
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
+/* first
+ a b c PDF‬ x y z */
+/* { dg-warning "U\\+202C" "" { target *-*-* } .-1 } */
+
+void
+g1 ()
+{
@ -365,14 +669,15 @@ index 00000000000..bdf334d9986
+/* { dg-warning "U\\+2068" "" { target *-*-* } .-1 } */
+int X\u2069;
+/* { dg-warning "U\\+2069" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-5.c b/gcc/testsuite/c-c++-common/Wbidirectional-5.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-5.c b/gcc/testsuite/c-c++-common/Wbidi-chars-5.c
new file mode 100644
index 00000000000..08b373a1bda
index 00000000000..f5776806c79
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-5.c
@@ -0,0 +1,149 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-5.c
@@ -0,0 +1,172 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=unpaired -Wno-multichar -Wno-overflow" } */
+/* { dg-options "-Wbidi-chars=unpaired -Wno-multichar -Wno-overflow" } */
+/* Test all bidi chars in various contexts (identifiers, comments,
+ string literals, character constants), both UCN and UTF-8. The bidi
+ chars here are properly terminated, except for the character constants. */
@ -418,6 +723,28 @@ index 00000000000..08b373a1bda
+// a b c PDF‬ x y z
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+/* Multiline comments. */
+/* a b c PDI⁩ x y z
+ */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* a b c PDF‬ x y z
+ */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* first
+ a b c PDI⁩ x y z
+ */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* first
+ a b c PDF‬ x y z
+ */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-2 } */
+/* first
+ a b c PDI⁩ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+/* first
+ a b c PDF‬ x y z */
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+
+void
+g1 ()
+{
@ -520,14 +847,15 @@ index 00000000000..08b373a1bda
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
+int X\u2069;
+/* { dg-bogus "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-6.c b/gcc/testsuite/c-c++-common/Wbidirectional-6.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-6.c b/gcc/testsuite/c-c++-common/Wbidi-chars-6.c
new file mode 100644
index 00000000000..c0b37c3a130
index 00000000000..a65d6faf60e
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-6.c
@@ -0,0 +1,129 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-6.c
@@ -0,0 +1,130 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=unpaired" } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test nesting of bidi chars in various contexts. */
+
+/* Terminated by the wrong char: */
@ -655,28 +983,30 @@ index 00000000000..c0b37c3a130
+int a_\u202a_\u202c_\u202a_\u202c_;
+int a_\u202a_\u202c_\u202a_;
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-7.c b/gcc/testsuite/c-c++-common/Wbidirectional-7.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-7.c b/gcc/testsuite/c-c++-common/Wbidi-chars-7.c
new file mode 100644
index 00000000000..f0f7b3ca14a
index 00000000000..d012d420ec0
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-7.c
@@ -0,0 +1,8 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-7.c
@@ -0,0 +1,9 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=any" } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test we ignore UCNs in comments. */
+
+// a b c \u202a 1 2 3
+// a b c \u202A 1 2 3
+/* a b c \u202a 1 2 3 */
+/* a b c \u202A 1 2 3 */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-8.c b/gcc/testsuite/c-c++-common/Wbidirectional-8.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-8.c b/gcc/testsuite/c-c++-common/Wbidi-chars-8.c
new file mode 100644
index 00000000000..c7d02193131
index 00000000000..4f54c5092ec
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-8.c
@@ -0,0 +1,12 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-8.c
@@ -0,0 +1,13 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=any" } */
+/* { dg-options "-Wbidi-chars=any" } */
+/* Test \u vs \U. */
+
+int a_\u202A;
@ -687,14 +1017,15 @@ index 00000000000..c7d02193131
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
+int a_\U0000202a_4;
+/* { dg-warning "U\\+202A" "" { target *-*-* } .-1 } */
diff --git a/gcc/testsuite/c-c++-common/Wbidirectional-9.c b/gcc/testsuite/c-c++-common/Wbidirectional-9.c
diff --git a/gcc/testsuite/c-c++-common/Wbidi-chars-9.c b/gcc/testsuite/c-c++-common/Wbidi-chars-9.c
new file mode 100644
index 00000000000..d029209babb
index 00000000000..e2af1b1ca97
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/Wbidirectional-9.c
@@ -0,0 +1,28 @@
+++ b/gcc/testsuite/c-c++-common/Wbidi-chars-9.c
@@ -0,0 +1,29 @@
+/* PR preprocessor/103026 */
+/* { dg-do compile } */
+/* { dg-options "-Wbidirectional=unpaired" } */
+/* { dg-options "-Wbidi-chars=unpaired" } */
+/* Test that we properly separate bidi contexts (comment/identifier/character
+ constant/string literal). */
+
@ -722,21 +1053,21 @@ index 00000000000..d029209babb
+int lre_\u202a; const char *s4 = "PDF\u202c";
+/* { dg-warning "unpaired" "" { target *-*-* } .-1 } */
diff --git a/libcpp/include/cpplib.h b/libcpp/include/cpplib.h
index 3ad52d5e01e..cecbefd59d4 100644
index 3ad52d5e01e..e0dcb7f0529 100644
--- a/libcpp/include/cpplib.h
+++ b/libcpp/include/cpplib.h
@@ -305,6 +305,17 @@ enum cpp_normalize_level {
normalized_none
};
+/* The possible bidirectional characters checking levels, from least
+/* The possible bidirectional control characters checking levels, from least
+ restrictive to most. */
+enum cpp_bidirectional_level {
+ /* No checking. */
+ bidirectional_none,
+ /* Only detect unpaired uses of bidirectional characters. */
+ /* Only detect unpaired uses of bidirectional control characters. */
+ bidirectional_unpaired,
+ /* Detect any use of bidirectional characters. */
+ /* Detect any use of bidirectional control characters. */
+ bidirectional_any
+};
+
@ -747,7 +1078,7 @@ index 3ad52d5e01e..cecbefd59d4 100644
/* True if warn about differences between C++98 and C++11. */
bool cpp_warn_cxx11_compat;
+ /* Nonzero of bidirectional characters checking is on. See enum
+ /* Nonzero if bidirectional control characters checking is on. See enum
+ cpp_bidirectional_level. */
+ unsigned char cpp_warn_bidirectional;
+
@ -776,17 +1107,35 @@ index ca3fbaa5c05..5c15da82ff8 100644
/* Default CPP arithmetic to something sensible for the host for the
benefit of dumb users like fix-header. */
diff --git a/libcpp/internal.h b/libcpp/internal.h
index 4f74f995cec..53b4c0f4af7 100644
--- a/libcpp/internal.h
+++ b/libcpp/internal.h
@@ -576,6 +576,13 @@ struct cpp_reader
/* If non-null, the lexer will use this location for the next token
instead of getting a location from the linemap. */
source_location *forced_token_location_p;
+
+ /* Returns true iff we should warn about UTF-8 bidirectional control
+ characters. */
+ bool warn_bidi_p () const
+ {
+ return CPP_OPTION (this, cpp_warn_bidirectional) != bidirectional_none;
+ }
};
/* Character classes. Based on the more primitive macros in safe-ctype.h.
diff --git a/libcpp/lex.c b/libcpp/lex.c
index a408f912c5c..3ba0a7ba818 100644
index a408f912c5c..ea7f75e842e 100644
--- a/libcpp/lex.c
+++ b/libcpp/lex.c
@@ -1164,6 +1164,284 @@ _cpp_process_line_notes (cpp_reader *pfile, int in_comment)
@@ -1164,6 +1164,324 @@ _cpp_process_line_notes (cpp_reader *pfile, int in_comment)
}
}
+namespace bidi {
+ enum kind {
+ NONE, LRE, RLE, LRO, RLO, LRI, RLI, FSI, PDF, PDI
+ NONE, LRE, RLE, LRO, RLO, LRI, RLI, FSI, PDF, PDI, LTR, RTL
+ };
+
+ /* All the UTF-8 encodings of bidi characters start with E2. */
@ -813,13 +1162,19 @@ index a408f912c5c..3ba0a7ba818 100644
+ vec.truncate (len - 1);
+ }
+
+ /* Return the context of the Ith element. */
+ kind ctx_at (unsigned int i)
+ {
+ return (vec[i] & 1) ? PDF : PDI;
+ }
+
+ /* Return which context is currently opened. */
+ kind current_ctx ()
+ {
+ unsigned int len = vec.count ();
+ if (len == 0)
+ return NONE;
+ return (vec[len - 1] & 1) ? PDF : PDI;
+ return ctx_at (len - 1);
+ }
+
+ /* Return true if the current context comes from a UCN origin, that is,
@ -847,13 +1202,27 @@ index a408f912c5c..3ba0a7ba818 100644
+ case FSI:
+ vec.push (ucn_p ? 2u : 0u);
+ break;
+ /* PDF terminates the scope of the last LRE, RLE, LRO, or RLO
+ whose scope has not yet been terminated. */
+ case PDF:
+ if (current_ctx () == PDF)
+ pop ();
+ break;
+ /* PDI terminates the scope of the last LRI, RLI, or FSI whose
+ scope has not yet been terminated, as well as the scopes of
+ any subsequent LREs, RLEs, LROs, or RLOs whose scopes have not
+ yet been terminated. */
+ case PDI:
+ if (current_ctx () == PDI)
+ pop ();
+ for (int i = vec.count () - 1; i >= 0; --i)
+ if (ctx_at (i) == PDI)
+ {
+ vec.truncate (i);
+ break;
+ }
+ break;
+ case LTR:
+ case RTL:
+ /* These aren't popped by a PDF/PDI. */
+ break;
+ [[likely]] case NONE:
+ break;
@ -885,6 +1254,10 @@ index a408f912c5c..3ba0a7ba818 100644
+ return "U+202C (POP DIRECTIONAL FORMATTING)";
+ case PDI:
+ return "U+2069 (POP DIRECTIONAL ISOLATE)";
+ case LTR:
+ return "U+200E (LEFT-TO-RIGHT MARK)";
+ case RTL:
+ return "U+200F (RIGHT-TO-LEFT MARK)";
+ default:
+ abort ();
+ }
@ -911,6 +1284,10 @@ index a408f912c5c..3ba0a7ba818 100644
+ return bidi::LRO;
+ case 0xae:
+ return bidi::RLO;
+ case 0x8e:
+ return bidi::LTR;
+ case 0x8f:
+ return bidi::RTL;
+ default:
+ break;
+ }
@ -988,6 +1365,18 @@ index a408f912c5c..3ba0a7ba818 100644
+ default:
+ break;
+ }
+ else if (p[2] == '0')
+ switch (p[3])
+ {
+ case 'e':
+ case 'E':
+ return bidi::LTR;
+ case 'f':
+ case 'F':
+ return bidi::RTL;
+ default:
+ break;
+ }
+
+ return bidi::NONE;
+}
@ -1008,7 +1397,7 @@ index a408f912c5c..3ba0a7ba818 100644
+ = linemap_position_for_column (pfile->line_table,
+ CPP_BUF_COLUMN (pfile->buffer, p));
+ cpp_warning_with_line (pfile, CPP_W_BIDIRECTIONAL, loc, 0,
+ "unpaired UTF-8 bidirectional character "
+ "unpaired UTF-8 bidirectional control character "
+ "detected");
+ }
+ /* We're done with this context. */
@ -1065,17 +1454,15 @@ index a408f912c5c..3ba0a7ba818 100644
/* Skip a C-style block comment. We find the end of the comment by
seeing if an asterisk is before every '/' we encounter. Returns
nonzero if comment terminated by EOF, zero otherwise.
@@ -1175,7 +1453,8 @@ _cpp_skip_block_comment (cpp_reader *pfile)
@@ -1175,6 +1493,7 @@ _cpp_skip_block_comment (cpp_reader *pfile)
cpp_buffer *buffer = pfile->buffer;
const uchar *cur = buffer->cur;
uchar c;
-
+ const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
+ != bidirectional_none);
+ const bool warn_bidi_p = pfile->warn_bidi_p ();
cur++;
if (*cur == '/')
cur++;
@@ -1189,7 +1468,11 @@ _cpp_skip_block_comment (cpp_reader *pfile)
@@ -1189,7 +1508,11 @@ _cpp_skip_block_comment (cpp_reader *pfile)
if (c == '/')
{
if (cur[-2] == '*')
@ -1088,7 +1475,7 @@ index a408f912c5c..3ba0a7ba818 100644
/* Warn about potential nested comments, but not if the '/'
comes immediately before the true comment delimiter.
@@ -1208,6 +1491,8 @@ _cpp_skip_block_comment (cpp_reader *pfile)
@@ -1208,6 +1531,8 @@ _cpp_skip_block_comment (cpp_reader *pfile)
{
unsigned int cols;
buffer->cur = cur - 1;
@ -1097,12 +1484,12 @@ index a408f912c5c..3ba0a7ba818 100644
_cpp_process_line_notes (pfile, true);
if (buffer->next_line >= buffer->rlimit)
return true;
@@ -1218,6 +1503,13 @@ _cpp_skip_block_comment (cpp_reader *pfile)
@@ -1218,6 +1543,13 @@ _cpp_skip_block_comment (cpp_reader *pfile)
cur = buffer->cur;
}
+ /* If this is a beginning of a UTF-8 encoding, it might be
+ a bidirectional character. */
+ a bidirectional control character. */
+ else if (__builtin_expect (c == bidi::utf8_start, 0) && warn_bidi_p)
+ {
+ bidi::kind kind = get_bidi_utf8 (cur - 1);
@ -1111,12 +1498,11 @@ index a408f912c5c..3ba0a7ba818 100644
}
buffer->cur = cur;
@@ -1233,9 +1525,32 @@ skip_line_comment (cpp_reader *pfile)
@@ -1233,9 +1565,31 @@ skip_line_comment (cpp_reader *pfile)
{
cpp_buffer *buffer = pfile->buffer;
source_location orig_line = pfile->line_table->highest_line;
+ const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
+ != bidirectional_none);
+ const bool warn_bidi_p = pfile->warn_bidi_p ();
- while (*buffer->cur != '\n')
- buffer->cur++;
@ -1146,7 +1532,7 @@ index a408f912c5c..3ba0a7ba818 100644
_cpp_process_line_notes (pfile, true);
return orig_line != pfile->line_table->highest_line;
@@ -1315,11 +1630,14 @@ warn_about_normalization (cpp_reader *pfile,
@@ -1315,11 +1669,13 @@ warn_about_normalization (cpp_reader *pfile,
/* Returns TRUE if the sequence starting at buffer->cur is invalid in
an identifier. FIRST is TRUE if this starts an identifier. */
@ -1156,12 +1542,11 @@ index a408f912c5c..3ba0a7ba818 100644
struct normalize_state *state)
{
cpp_buffer *buffer = pfile->buffer;
+ const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
+ != bidirectional_none);
+ const bool warn_bidi_p = pfile->warn_bidi_p ();
if (*buffer->cur == '$')
{
@@ -1343,6 +1661,12 @@ forms_identifier_p (cpp_reader *pfile, int first,
@@ -1343,6 +1699,12 @@ forms_identifier_p (cpp_reader *pfile, int first,
{
cppchar_t s;
buffer->cur += 2;
@ -1174,16 +1559,15 @@ index a408f912c5c..3ba0a7ba818 100644
if (_cpp_valid_ucn (pfile, &buffer->cur, buffer->rlimit, 1 + !first,
state, &s, NULL, NULL))
return true;
@@ -1450,6 +1774,8 @@ lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn,
@@ -1450,6 +1812,7 @@ lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn,
const uchar *cur;
unsigned int len;
unsigned int hash = HT_HASHSTEP (0, *base);
+ const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
+ != bidirectional_none);
+ const bool warn_bidi_p = pfile->warn_bidi_p ();
cur = pfile->buffer->cur;
if (! starts_ucn)
@@ -1472,6 +1798,8 @@ lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn,
@@ -1472,6 +1835,8 @@ lex_identifier (cpp_reader *pfile, const uchar *base, bool starts_ucn,
pfile->buffer->cur++;
}
} while (forms_identifier_p (pfile, false, nst));
@ -1192,23 +1576,21 @@ index a408f912c5c..3ba0a7ba818 100644
result = _cpp_interpret_identifier (pfile, base,
pfile->buffer->cur - base);
*spelling = cpp_lookup (pfile, base, pfile->buffer->cur - base);
@@ -1673,6 +2001,8 @@ lex_raw_string (cpp_reader *pfile, cpp_token *token, const uchar *base,
@@ -1673,6 +2038,7 @@ lex_raw_string (cpp_reader *pfile, cpp_token *token, const uchar *base,
_cpp_buff *first_buff = NULL, *last_buff = NULL;
size_t raw_prefix_start;
_cpp_line_note *note = &pfile->buffer->notes[pfile->buffer->cur_note];
+ const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
+ != bidirectional_none);
+ const bool warn_bidi_p = pfile->warn_bidi_p ();
type = (*base == 'L' ? CPP_WSTRING :
*base == 'U' ? CPP_STRING32 :
@@ -1909,8 +2239,16 @@ lex_raw_string (cpp_reader *pfile, cpp_token *token, const uchar *base,
@@ -1909,8 +2275,15 @@ lex_raw_string (cpp_reader *pfile, cpp_token *token, const uchar *base,
cur = base = pfile->buffer->cur;
note = &pfile->buffer->notes[pfile->buffer->cur_note];
}
+ else if (__builtin_expect ((unsigned char) c == bidi::utf8_start, 0)
+ && warn_bidi_p)
+ maybe_warn_bidi_on_char (pfile, cur - 1,
+ get_bidi_utf8 (cur - 1),
+ maybe_warn_bidi_on_char (pfile, cur - 1, get_bidi_utf8 (cur - 1),
+ /*ucn_p=*/false);
}
@ -1218,12 +1600,11 @@ index a408f912c5c..3ba0a7ba818 100644
if (CPP_OPTION (pfile, user_literals))
{
/* If a string format macro, say from inttypes.h, is placed touching
@@ -2005,15 +2343,28 @@ lex_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
@@ -2005,15 +2378,27 @@ lex_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
else
terminator = '>', type = CPP_HEADER_NAME;
+ const bool warn_bidi_p = (CPP_OPTION (pfile, cpp_warn_bidirectional)
+ != bidirectional_none);
+ const bool warn_bidi_p = pfile->warn_bidi_p ();
for (;;)
{
cppchar_t c = *cur++;
@ -1249,7 +1630,7 @@ index a408f912c5c..3ba0a7ba818 100644
else if (c == '\n')
{
cur--;
@@ -2030,6 +2381,11 @@ lex_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
@@ -2030,6 +2415,11 @@ lex_string (cpp_reader *pfile, cpp_token *token, const uchar *base)
}
else if (c == '\0')
saw_NUL = true;

105
SOURCES/gcc8-aarch64-mtune-neoverse-512tvb.patch Normal file
View File

@ -0,0 +1,105 @@
From 9c108bb84d3a2447dac730c455df658be0a2c751 Mon Sep 17 00:00:00 2001
From: Richard Sandiford <richard.sandiford@arm.com>
Date: Tue, 17 Aug 2021 15:15:27 +0100
Subject: [PATCH] aarch64: Add -mtune=neoverse-512tvb
To: gcc-patches@gcc.gnu.org
This patch adds an option to tune for Neoverse cores that have
a total vector bandwidth of 512 bits (4x128 for Advanced SIMD
and a vector-length-dependent equivalent for SVE). This is intended
to be a compromise between tuning aggressively for a single core like
Neoverse V1 (which can be too narrow) and tuning for AArch64 cores
in general (which can be too wide).
-mcpu=neoverse-512tvb is equivalent to -mcpu=neoverse-v1
-mtune=neoverse-512tvb.
gcc/
* doc/invoke.texi: Document -mtune=neoverse-512tvb and
-mcpu=neoverse-512tvb.
* config/aarch64/aarch64-cores.def (neoverse-512tvb): New entry.
* config/aarch64/aarch64-tune.md: Regenerate.
(cherry picked from commit 048039c49b96875144f67e7789fdea54abf7710b)
---
gcc/config/aarch64/aarch64-cores.def | 1 +
gcc/config/aarch64/aarch64-tune.md | 2 +-
gcc/doc/invoke.texi | 25 ++++++++++++++++++++++---
3 files changed, 24 insertions(+), 4 deletions(-)
diff --git a/gcc/config/aarch64/aarch64-cores.def b/gcc/config/aarch64/aarch64-cores.def
index dfb839c01cc..f348d31e22e 100644
--- a/gcc/config/aarch64/aarch64-cores.def
+++ b/gcc/config/aarch64/aarch64-cores.def
@@ -99,6 +99,7 @@ AARCH64_CORE("saphira", saphira, falkor, 8_3A, AARCH64_FL_FOR_ARCH8_3
/* ARM ('A') cores. */
AARCH64_CORE("zeus", zeus, cortexa57, 8_4A, AARCH64_FL_FOR_ARCH8_4 | AARCH64_FL_F16 | AARCH64_FL_RCPC | AARCH64_FL_SVE | AARCH64_FL_RNG, neoversev1, 0x41, 0xd40, -1)
AARCH64_CORE("neoverse-v1", neoversev1, cortexa57, 8_4A, AARCH64_FL_FOR_ARCH8_4 | AARCH64_FL_F16 | AARCH64_FL_RCPC | AARCH64_FL_SVE | AARCH64_FL_RNG, neoversev1, 0x41, 0xd40, -1)
+AARCH64_CORE("neoverse-512tvb", neoverse512tvb, cortexa57, 8_4A, AARCH64_FL_FOR_ARCH8_4 | AARCH64_FL_F16 | AARCH64_FL_RCPC | AARCH64_FL_SVE | AARCH64_FL_RNG, neoversev1, INVALID_IMP, INVALID_CORE, -1)
/* Armv8.5-A Architecture Processors. */
AARCH64_CORE("neoverse-n2", neoversen2, cortexa57, 8_4A, AARCH64_FL_FOR_ARCH8_4 | AARCH64_FL_F16 | AARCH64_FL_SVE | AARCH64_FL_RNG, neoversen2, 0x41, 0xd49, -1)
diff --git a/gcc/config/aarch64/aarch64-tune.md b/gcc/config/aarch64/aarch64-tune.md
index 2d7c9aa4740..09b76480f0b 100644
--- a/gcc/config/aarch64/aarch64-tune.md
+++ b/gcc/config/aarch64/aarch64-tune.md
@@ -1,5 +1,5 @@
;; -*- buffer-read-only: t -*-
;; Generated automatically by gentune.sh from aarch64-cores.def
(define_attr "tune"
- "cortexa35,cortexa53,cortexa57,cortexa72,cortexa73,thunderx,thunderxt88p1,thunderxt88,thunderxt81,thunderxt83,xgene1,falkor,qdf24xx,exynosm1,thunderx2t99p1,vulcan,thunderx2t99,cortexa55,cortexa75,cortexa76,ares,neoversen1,saphira,zeus,neoversev1,neoversen2,cortexa57cortexa53,cortexa72cortexa53,cortexa73cortexa35,cortexa73cortexa53,cortexa75cortexa55"
+ "cortexa35,cortexa53,cortexa57,cortexa72,cortexa73,thunderx,thunderxt88p1,thunderxt88,thunderxt81,thunderxt83,xgene1,falkor,qdf24xx,exynosm1,thunderx2t99p1,vulcan,thunderx2t99,cortexa55,cortexa75,cortexa76,ares,neoversen1,saphira,zeus,neoversev1,neoverse512tvb,neoversen2,cortexa57cortexa53,cortexa72cortexa53,cortexa73cortexa35,cortexa73cortexa53,cortexa75cortexa55"
(const (symbol_ref "((enum attr_tune) aarch64_tune)")))
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index 78ca7738df2..68fda03281a 100644
--- a/gcc/doc/invoke.texi
+++ b/gcc/doc/invoke.texi
@@ -14772,9 +14772,9 @@ performance of the code. Permissible values for this option are:
@samp{generic}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a55},
@samp{cortex-a57}, @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-a75},
@samp{cortex-a76}, @samp{ares}, @samp{neoverse-n1}, @samp{neoverse-n2},
-@samp{neoverse-v1}, @samp{zeus}, @samp{exynos-m1}, @samp{falkor},
-@samp{qdf24xx}, @samp{saphira}, @samp{xgene1}, @samp{vulcan}, @samp{thunderx},
-@samp{thunderxt88}, @samp{thunderxt88p1}, @samp{thunderxt81},
+@samp{neoverse-v1}, @samp{zeus}, @samp{neoverse-512tvb}, @samp{exynos-m1},
+@samp{falkor}, @samp{qdf24xx}, @samp{saphira}, @samp{xgene1}, @samp{vulcan},
+@samp{thunderx}, @samp{thunderxt88}, @samp{thunderxt88p1}, @samp{thunderxt81},
@samp{thunderxt83}, @samp{thunderx2t99}, @samp{cortex-a57.cortex-a53},
@samp{cortex-a72.cortex-a53}, @samp{cortex-a73.cortex-a35},
@samp{cortex-a73.cortex-a53}, @samp{cortex-a75.cortex-a55},
@@ -14785,6 +14785,15 @@ The values @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
@samp{cortex-a75.cortex-a55} specify that GCC should tune for a
big.LITTLE system.
+The value @samp{neoverse-512tvb} specifies that GCC should tune
+for Neoverse cores that (a) implement SVE and (b) have a total vector
+bandwidth of 512 bits per cycle. In other words, the option tells GCC to
+tune for Neoverse cores that can execute 4 128-bit Advanced SIMD arithmetic
+instructions a cycle and that can execute an equivalent number of SVE
+arithmetic instructions per cycle (2 for 256-bit SVE, 4 for 128-bit SVE).
+This is more general than tuning for a specific core like Neoverse V1
+but is more specific than the default tuning described below.
+
Additionally on native AArch64 GNU/Linux systems the value
@samp{native} tunes performance to the host system. This option has no effect
if the compiler is unable to recognize the processor of the host system.
@@ -14814,6 +14823,16 @@ by @option{-mtune}). Where this option is used in conjunction
with @option{-march} or @option{-mtune}, those options take precedence
over the appropriate part of this option.
+@option{-mcpu=neoverse-512tvb} is special in that it does not refer
+to a specific core, but instead refers to all Neoverse cores that
+(a) implement SVE and (b) have a total vector bandwidth of 512 bits
+a cycle. Unless overridden by @option{-march},
+@option{-mcpu=neoverse-512tvb} generates code that can run on a
+Neoverse V1 core, since Neoverse V1 is the first Neoverse core with
+these properties. Unless overridden by @option{-mtune},
+@option{-mcpu=neoverse-512tvb} tunes code in the same way as for
+@option{-mtune=neoverse-512tvb}.
+
@item -moverride=@var{string}
@opindex moverride
Override tuning decisions made by the back-end in response to a
--
2.25.1

844
SOURCES/gcc8-pch-tweaks.patch Normal file
View File

@ -0,0 +1,844 @@
commit fe7c3ecff1f9c0520090a77fa824d8c5d9dbec12
Author: Jakub Jelinek <jakub@redhat.com>
Date: Fri Dec 3 11:03:30 2021 +0100
pch: Add support for PCH for relocatable executables [PR71934]
So, if we want to make PCH work for PIEs, I'd say we can:
1) add a new GTY option, say callback, which would act like
skip for non-PCH and for PCH would make us skip it but
remember for address bias translation
2) drop the skip for tree_translation_unit_decl::language
3) change get_unnamed_section to have const char * as
last argument instead of const void *, change
unnamed_section::data also to const char * and update
everything related to that
4) maybe add a host hook whether it is ok to support binaries
changing addresses (the only thing I'm worried is if
some host that uses function descriptors allocates them
dynamically instead of having them somewhere in the
executable)
5) maybe add a gengtype warning if it sees in GTY tracked
structure a function pointer without that new callback
option
Here is 1), 2), 3) implemented.
Note, on stdc++.h.gch/O2g.gch there are just those 10 relocations without
the second patch, with it a few more, but nothing huge. And for non-PIEs
there isn't really any extra work on the load side except freading two scalar
values and fseek.
2021-12-03 Jakub Jelinek <jakub@redhat.com>
PR pch/71934
gcc/
* ggc.h (gt_pch_note_callback): Declare.
* gengtype.h (enum typekind): Add TYPE_CALLBACK.
(callback_type): Declare.
* gengtype.c (dbgprint_count_type_at): Handle TYPE_CALLBACK.
(callback_type): New variable.
(process_gc_options): Add CALLBACK argument, handle callback
option.
(set_gc_used_type): Adjust process_gc_options caller, if callback,
set type to &callback_type.
(output_mangled_typename): Handle TYPE_CALLBACK.
(walk_type): Likewise. Handle callback option.
(write_types_process_field): Handle TYPE_CALLBACK.
(write_types_local_user_process_field): Likewise.
(write_types_local_process_field): Likewise.
(write_root): Likewise.
(dump_typekind): Likewise.
(dump_type): Likewise.
* gengtype-state.c (type_lineloc): Handle TYPE_CALLBACK.
(state_writer::write_state_callback_type): New method.
(state_writer::write_state_type): Handle TYPE_CALLBACK.
(read_state_callback_type): New function.
(read_state_type): Handle TYPE_CALLBACK.
* ggc-common.c (callback_vec): New variable.
(gt_pch_note_callback): New function.
(gt_pch_save): Stream out gt_pch_save function address and relocation
table.
(gt_pch_restore): Stream in saved gt_pch_save function address and
relocation table and apply relocations if needed.
* doc/gty.texi (callback): Document new GTY option.
* varasm.c (get_unnamed_section): Change callback argument's type and
last argument's type from const void * to const char *.
(output_section_asm_op): Change argument's type from const void *
to const char *, remove unnecessary cast.
* tree-core.h (struct tree_translation_unit_decl): Drop GTY((skip))
from language member.
* output.h (unnamed_section_callback): Change argument type from
const void * to const char *.
(struct unnamed_section): Use GTY((callback)) instead of GTY((skip))
for callback member. Change data member type from const void *
to const char *.
(struct noswitch_section): Use GTY((callback)) instead of GTY((skip))
for callback member.
(get_unnamed_section): Change callback argument's type and
last argument's type from const void * to const char *.
(output_section_asm_op): Change argument's type from const void *
to const char *.
* config/avr/avr.c (avr_output_progmem_section_asm_op): Likewise.
Remove unneeded cast.
* config/darwin.c (output_objc_section_asm_op): Change argument's type
from const void * to const char *.
* config/pa/pa.c (som_output_text_section_asm_op): Likewise.
(som_output_comdat_data_section_asm_op): Likewise.
* config/rs6000/rs6000.c (rs6000_elf_output_toc_section_asm_op):
Likewise.
(rs6000_xcoff_output_readonly_section_asm_op): Likewise. Instead
of dereferencing directive hardcode variable names and decide based on
whether directive is NULL or not.
(rs6000_xcoff_output_readwrite_section_asm_op): Change argument's type
from const void * to const char *.
(rs6000_xcoff_output_tls_section_asm_op): Likewise. Instead
of dereferencing directive hardcode variable names and decide based on
whether directive is NULL or not.
(rs6000_xcoff_output_toc_section_asm_op): Change argument's type
from const void * to const char *.
(rs6000_xcoff_asm_init_sections): Adjust get_unnamed_section callers.
gcc/c-family/
* c-pch.c (struct c_pch_validity): Remove pch_init member.
(pch_init): Don't initialize v.pch_init.
(c_common_valid_pch): Don't warn and punt if .text addresses change.
libcpp/
* include/line-map.h (class line_maps): Add GTY((callback)) to
reallocator and round_alloc_size members.
commit 4dc6d19222581c77a174d44d97507d234fb7e39b
Author: Jakub Jelinek <jakub@redhat.com>
Date: Mon Dec 6 11:18:58 2021 +0100
avr: Fix AVR build [PR71934]
On Mon, Dec 06, 2021 at 11:00:30AM +0100, Martin Liška wrote:
> Jakub, I think the patch broke avr-linux target:
>
> g++ -fno-PIE -c -g -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE -fno-exceptions -fno-rtti -fasynchronous-unwind-tables -W -Wall -Wno-narrowing -Wwrite-strings -Wcast-qual -Wno-erro
> /home/marxin/Programming/gcc/gcc/config/avr/avr.c: In function ‘void avr_output_data_section_asm_op(const void*)’:
> /home/marxin/Programming/gcc/gcc/config/avr/avr.c:10097:26: error: invalid conversion from ‘const void*’ to ‘const char*’ [-fpermissive]
This patch fixes that.
2021-12-06 Jakub Jelinek <jakub@redhat.com>
PR pch/71934
* config/avr/avr.c (avr_output_data_section_asm_op,
avr_output_bss_section_asm_op): Change argument type from const void *
to const char *.
diff --git a/gcc/c-family/c-pch.c b/gcc/c-family/c-pch.c
index 5da60423354..2cafa1387bb 100644
--- a/gcc/c-family/c-pch.c
+++ b/gcc/c-family/c-pch.c
@@ -58,7 +58,6 @@ struct c_pch_validity
{
unsigned char debug_info_type;
signed char match[MATCH_SIZE];
- void (*pch_init) (void);
size_t target_data_length;
};
@@ -123,7 +122,6 @@ pch_init (void)
gcc_assert (v.match[i] == *pch_matching[i].flag_var);
}
}
- v.pch_init = &pch_init;
target_validity = targetm.get_pch_validity (&v.target_data_length);
if (fwrite (partial_pch, IDENT_LENGTH, 1, f) != 1
@@ -287,20 +285,6 @@ c_common_valid_pch (cpp_reader *pfile, c
}
}
- /* If the text segment was not loaded at the same address as it was
- when the PCH file was created, function pointers loaded from the
- PCH will not be valid. We could in theory remap all the function
- pointers, but no support for that exists at present.
- Since we have the same executable, it should only be necessary to
- check one function. */
- if (v.pch_init != &pch_init)
- {
- if (cpp_get_options (pfile)->warn_invalid_pch)
- cpp_error (pfile, CPP_DL_WARNING,
- "%s: had text segment at different address", name);
- return 2;
- }
-
/* Check the target-specific validity data. */
{
void *this_file_data = xmalloc (v.target_data_length);
diff --git a/gcc/config/avr/avr.c b/gcc/config/avr/avr.c
index 200701a583c..6ba038881d6 100644
--- a/gcc/config/avr/avr.c
+++ b/gcc/config/avr/avr.c
@@ -10114,10 +10114,9 @@ avr_output_bss_section_asm_op (const void *data)
/* Unnamed section callback for progmem*.data sections. */
static void
-avr_output_progmem_section_asm_op (const void *data)
+avr_output_progmem_section_asm_op (const char *data)
{
- fprintf (asm_out_file, "\t.section\t%s,\"a\",@progbits\n",
- (const char*) data);
+ fprintf (asm_out_file, "\t.section\t%s,\"a\",@progbits\n", data);
}
diff --git a/gcc/config/darwin.c b/gcc/config/darwin.c
index c5ba7927ce1..8ad5b26c980 100644
--- a/gcc/config/darwin.c
+++ b/gcc/config/darwin.c
@@ -134,7 +134,7 @@ int emit_aligned_common = false;
DIRECTIVE is as for output_section_asm_op. */
static void
-output_objc_section_asm_op (const void *directive)
+output_objc_section_asm_op (const char *directive)
{
static bool been_here = false;
diff --git a/gcc/config/pa/pa.c b/gcc/config/pa/pa.c
index f22d25a4066..2b10ef34061 100644
--- a/gcc/config/pa/pa.c
+++ b/gcc/config/pa/pa.c
@@ -10009,7 +10009,7 @@ pa_arg_partial_bytes (cumulative_args_t cum_v, const function_arg_info &arg)
to the default text subspace. */
static void
-som_output_text_section_asm_op (const void *data ATTRIBUTE_UNUSED)
+som_output_text_section_asm_op (const char *data ATTRIBUTE_UNUSED)
{
gcc_assert (TARGET_SOM);
if (TARGET_GAS)
@@ -10053,7 +10053,7 @@ som_output_text_section_asm_op (const void *data ATTRIBUTE_UNUSED)
sections. This function is only used with SOM. */
static void
-som_output_comdat_data_section_asm_op (const void *data)
+som_output_comdat_data_section_asm_op (const char *data)
{
in_section = NULL;
output_section_asm_op (data);
diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c
index 945157b1c1a..34089743759 100644
--- a/gcc/config/rs6000/rs6000.c
+++ b/gcc/config/rs6000/rs6000.c
@@ -20599,7 +20599,7 @@ rs6000_ms_bitfield_layout_p (const_tree record_type)
/* A get_unnamed_section callback, used for switching to toc_section. */
static void
-rs6000_elf_output_toc_section_asm_op (const void *data ATTRIBUTE_UNUSED)
+rs6000_elf_output_toc_section_asm_op (const char *data ATTRIBUTE_UNUSED)
{
if ((DEFAULT_ABI == ABI_AIX || DEFAULT_ABI == ABI_ELFv2)
&& TARGET_MINIMAL_TOC)
@@ -21303,35 +21303,39 @@ rs6000_xcoff_asm_globalize_label (FILE *stream, const char *name)
points to the section string variable. */
static void
-rs6000_xcoff_output_readonly_section_asm_op (const void *directive)
+rs6000_xcoff_output_readonly_section_asm_op (const char *directive)
{
fprintf (asm_out_file, "\t.csect %s[RO],%s\n",
- *(const char *const *) directive,
+ directive
+ ? xcoff_private_rodata_section_name
+ : xcoff_read_only_section_name,
XCOFF_CSECT_DEFAULT_ALIGNMENT_STR);
}
/* Likewise for read-write sections. */
static void
-rs6000_xcoff_output_readwrite_section_asm_op (const void *directive)
+rs6000_xcoff_output_readwrite_section_asm_op (const char *)
{
fprintf (asm_out_file, "\t.csect %s[RW],%s\n",
- *(const char *const *) directive,
+ xcoff_private_data_section_name,
XCOFF_CSECT_DEFAULT_ALIGNMENT_STR);
}
static void
-rs6000_xcoff_output_tls_section_asm_op (const void *directive)
+rs6000_xcoff_output_tls_section_asm_op (const char *directive)
{
fprintf (asm_out_file, "\t.csect %s[TL],%s\n",
- *(const char *const *) directive,
+ directive
+ ? xcoff_private_data_section_name
+ : xcoff_tls_data_section_name,
XCOFF_CSECT_DEFAULT_ALIGNMENT_STR);
}
/* A get_unnamed_section callback, used for switching to toc_section. */
static void
-rs6000_xcoff_output_toc_section_asm_op (const void *data ATTRIBUTE_UNUSED)
+rs6000_xcoff_output_toc_section_asm_op (const char *data ATTRIBUTE_UNUSED)
{
if (TARGET_MINIMAL_TOC)
{
@@ -21358,26 +21362,26 @@ rs6000_xcoff_asm_init_sections (void)
{
read_only_data_section
= get_unnamed_section (0, rs6000_xcoff_output_readonly_section_asm_op,
- &xcoff_read_only_section_name);
+ NULL);
private_data_section
= get_unnamed_section (SECTION_WRITE,
rs6000_xcoff_output_readwrite_section_asm_op,
- &xcoff_private_data_section_name);
+ NULL);
read_only_private_data_section
= get_unnamed_section (0, rs6000_xcoff_output_readonly_section_asm_op,
- &xcoff_private_rodata_section_name);
+ "");
tls_data_section
= get_unnamed_section (SECTION_TLS,
rs6000_xcoff_output_tls_section_asm_op,
- &xcoff_tls_data_section_name);
+ NULL);
tls_private_data_section
= get_unnamed_section (SECTION_TLS,
rs6000_xcoff_output_tls_section_asm_op,
- &xcoff_private_data_section_name);
+ "");
toc_section
= get_unnamed_section (0, rs6000_xcoff_output_toc_section_asm_op, NULL);
diff --git a/gcc/doc/gty.texi b/gcc/doc/gty.texi
index b996ff2c44e..ca2c8404894 100644
--- a/gcc/doc/gty.texi
+++ b/gcc/doc/gty.texi
@@ -205,6 +205,15 @@ If @code{skip} is applied to a field, the type machinery will ignore it.
This is somewhat dangerous; the only safe use is in a union when one
field really isn't ever used.
+@findex callback
+@item callback
+
+@code{callback} should be applied to fields with pointer to function type
+and causes the field to be ignored similarly to @code{skip}, except when
+writing PCH and the field is non-NULL it will remember the field's address
+for relocation purposes if the process writing PCH has different load base
+from a process reading PCH.
+
@findex for_user
@item for_user
diff --git a/gcc/gengtype-state.c b/gcc/gengtype-state.c
index ac9d536963f..36a96e84574 100644
--- a/gcc/gengtype-state.c
+++ b/gcc/gengtype-state.c
@@ -57,6 +57,7 @@ type_lineloc (const_type_p ty)
case TYPE_STRING:
case TYPE_POINTER:
case TYPE_ARRAY:
+ case TYPE_CALLBACK:
return NULL;
default:
gcc_unreachable ();
@@ -171,6 +172,7 @@ private:
void write_state_version (const char *version);
void write_state_scalar_type (type_p current);
void write_state_string_type (type_p current);
+ void write_state_callback_type (type_p current);
void write_state_undefined_type (type_p current);
void write_state_struct_union_type (type_p current, const char *kindstr);
void write_state_struct_type (type_p current);
@@ -898,6 +900,20 @@ state_writer::write_state_string_type (type_p current)
fatal ("Unexpected type in write_state_string_type");
}
+/* Write the callback type. There is only one such thing! */
+void
+state_writer::write_state_callback_type (type_p current)
+{
+ if (current == &callback_type)
+ {
+ write_any_indent (0);
+ fprintf (state_file, "callback ");
+ write_state_common_type_content (current);
+ }
+ else
+ fatal ("Unexpected type in write_state_callback_type");
+}
+
/* Write an undefined type. */
void
state_writer::write_state_undefined_type (type_p current)
@@ -1143,6 +1159,9 @@ state_writer::write_state_type (type_p current)
case TYPE_STRING:
write_state_string_type (current);
break;
+ case TYPE_CALLBACK:
+ write_state_callback_type (current);
+ break;
}
}
@@ -1477,6 +1496,14 @@ read_state_string_type (type_p *type)
read_state_common_type_content (*type);
}
+/* Read the callback_type. */
+static void
+read_state_callback_type (type_p *type)
+{
+ *type = &callback_type;
+ read_state_common_type_content (*type);
+}
+
/* Read a lang_bitmap representing a set of GCC front-end languages. */
static void
@@ -1834,6 +1861,11 @@ read_state_type (type_p *current)
next_state_tokens (1);
read_state_string_type (current);
}
+ else if (state_token_is_name (t0, "callback"))
+ {
+ next_state_tokens (1);
+ read_state_callback_type (current);
+ }
else if (state_token_is_name (t0, "undefined"))
{
*current = XCNEW (struct type);
diff --git a/gcc/gengtype.c b/gcc/gengtype.c
index a77cfd92bfa..b9daaa43689 100644
--- a/gcc/gengtype.c
+++ b/gcc/gengtype.c
@@ -172,6 +172,7 @@ dbgprint_count_type_at (const char *fil, int lin, const char *msg, type_p t)
int nb_struct = 0, nb_union = 0, nb_array = 0, nb_pointer = 0;
int nb_lang_struct = 0;
int nb_user_struct = 0, nb_undefined = 0;
+ int nb_callback = 0;
type_p p = NULL;
for (p = t; p; p = p->next)
{
@@ -202,6 +203,9 @@ dbgprint_count_type_at (const char *fil, int lin, const char *msg, type_p t)
case TYPE_ARRAY:
nb_array++;
break;
+ case TYPE_CALLBACK:
+ nb_callback++;
+ break;
case TYPE_LANG_STRUCT:
nb_lang_struct++;
break;
@@ -217,6 +221,8 @@ dbgprint_count_type_at (const char *fil, int lin, const char *msg, type_p t)
fprintf (stderr, "@@%%@@ %d structs, %d unions\n", nb_struct, nb_union);
if (nb_pointer > 0 || nb_array > 0)
fprintf (stderr, "@@%%@@ %d pointers, %d arrays\n", nb_pointer, nb_array);
+ if (nb_callback > 0)
+ fprintf (stderr, "@@%%@@ %d callbacks\n", nb_callback);
if (nb_lang_struct > 0)
fprintf (stderr, "@@%%@@ %d lang_structs\n", nb_lang_struct);
if (nb_user_struct > 0)
@@ -495,6 +501,10 @@ struct type scalar_char = {
TYPE_SCALAR, 0, 0, 0, GC_USED, {0}
};
+struct type callback_type = {
+ TYPE_CALLBACK, 0, 0, 0, GC_USED, {0}
+};
+
/* Lists of various things. */
pair_p typedefs = NULL;
@@ -1464,7 +1474,7 @@ static void set_gc_used (pair_p);
static void
process_gc_options (options_p opt, enum gc_used_enum level, int *maybe_undef,
- int *length, int *skip, type_p *nested_ptr)
+ int *length, int *skip, int *callback, type_p *nested_ptr)
{
options_p o;
for (o = opt; o; o = o->next)
@@ -1478,6 +1488,8 @@ process_gc_options (options_p opt, enum gc_used_enum level, int *maybe_undef,
*length = 1;
else if (strcmp (o->name, "skip") == 0)
*skip = 1;
+ else if (strcmp (o->name, "callback") == 0)
+ *callback = 1;
else if (strcmp (o->name, "nested_ptr") == 0
&& o->kind == OPTION_NESTED)
*nested_ptr = ((const struct nested_ptr_data *) o->info.nested)->type;
@@ -1526,7 +1538,7 @@ set_gc_used_type (type_p t, enum gc_used_enum level,
type_p dummy2;
bool allow_undefined_field_types = (t->kind == TYPE_USER_STRUCT);
- process_gc_options (t->u.s.opt, level, &dummy, &dummy, &dummy,
+ process_gc_options (t->u.s.opt, level, &dummy, &dummy, &dummy, &dummy,
&dummy2);
if (t->u.s.base_class)
@@ -1542,9 +1554,10 @@ set_gc_used_type (type_p t, enum gc_used_enum level,
int maybe_undef = 0;
int length = 0;
int skip = 0;
+ int callback = 0;
type_p nested_ptr = NULL;
process_gc_options (f->opt, level, &maybe_undef, &length, &skip,
- &nested_ptr);
+ &callback, &nested_ptr);
if (nested_ptr && f->type->kind == TYPE_POINTER)
set_gc_used_type (nested_ptr, GC_POINTED_TO);
@@ -1554,6 +1567,8 @@ set_gc_used_type (type_p t, enum gc_used_enum level,
set_gc_used_type (f->type->u.p, GC_MAYBE_POINTED_TO);
else if (skip)
; /* target type is not used through this field */
+ else if (callback)
+ f->type = &callback_type;
else
set_gc_used_type (f->type, GC_USED, allow_undefined_field_types);
}
@@ -2519,6 +2534,7 @@ output_mangled_typename (outf_p of, const_type_p t)
{
case TYPE_NONE:
case TYPE_UNDEFINED:
+ case TYPE_CALLBACK:
gcc_unreachable ();
break;
case TYPE_POINTER:
@@ -2719,6 +2735,8 @@ walk_type (type_p t, struct walk_type_data *d)
;
else if (strcmp (oo->name, "for_user") == 0)
;
+ else if (strcmp (oo->name, "callback") == 0)
+ ;
else
error_at_line (d->line, "unknown option `%s'\n", oo->name);
@@ -2744,6 +2762,7 @@ walk_type (type_p t, struct walk_type_data *d)
{
case TYPE_SCALAR:
case TYPE_STRING:
+ case TYPE_CALLBACK:
d->process_field (t, d);
break;
@@ -3275,6 +3294,7 @@ write_types_process_field (type_p f, const struct walk_type_data *d)
break;
case TYPE_SCALAR:
+ case TYPE_CALLBACK:
break;
case TYPE_ARRAY:
@@ -3820,6 +3840,7 @@ write_types_local_user_process_field (type_p f, const struct walk_type_data *d)
break;
case TYPE_SCALAR:
+ case TYPE_CALLBACK:
break;
case TYPE_ARRAY:
@@ -3906,6 +3927,13 @@ write_types_local_process_field (type_p f, const struct walk_type_data *d)
case TYPE_SCALAR:
break;
+ case TYPE_CALLBACK:
+ oprintf (d->of, "%*sif ((void *)(%s) == this_obj)\n", d->indent, "",
+ d->prev_val[3]);
+ oprintf (d->of, "%*s gt_pch_note_callback (&(%s), this_obj);\n",
+ d->indent, "", d->val);
+ break;
+
case TYPE_ARRAY:
case TYPE_NONE:
case TYPE_UNDEFINED:
@@ -4434,6 +4462,7 @@ write_root (outf_p f, pair_p v, type_p type, const char *name, int has_length,
case TYPE_UNDEFINED:
case TYPE_UNION:
case TYPE_LANG_STRUCT:
+ case TYPE_CALLBACK:
error_at_line (line, "global `%s' is unimplemented type", name);
}
}
@@ -4728,6 +4757,9 @@ dump_typekind (int indent, enum typekind kind)
case TYPE_ARRAY:
printf ("TYPE_ARRAY");
break;
+ case TYPE_CALLBACK:
+ printf ("TYPE_CALLBACK");
+ break;
case TYPE_LANG_STRUCT:
printf ("TYPE_LANG_STRUCT");
break;
@@ -4894,6 +4926,7 @@ dump_type (int indent, type_p t)
t->u.scalar_is_char ? "true" : "false");
break;
case TYPE_STRING:
+ case TYPE_CALLBACK:
break;
case TYPE_STRUCT:
case TYPE_UNION:
diff --git a/gcc/gengtype.h b/gcc/gengtype.h
index 8a7a54957ea..8fa7064ca85 100644
--- a/gcc/gengtype.h
+++ b/gcc/gengtype.h
@@ -154,6 +154,9 @@ enum typekind {
TYPE_UNION, /* Type for GTY-ed discriminated unions. */
TYPE_POINTER, /* Pointer type to GTY-ed type. */
TYPE_ARRAY, /* Array of GTY-ed types. */
+ TYPE_CALLBACK, /* A function pointer that needs relocation if
+ the executable has been loaded at a different
+ address. */
TYPE_LANG_STRUCT, /* GCC front-end language specific structs.
Various languages may have homonymous but
different structs. */
@@ -331,6 +334,9 @@ extern struct type string_type;
extern struct type scalar_nonchar;
extern struct type scalar_char;
+/* The one and only TYPE_CALLBACK. */
+extern struct type callback_type;
+
/* Test if a type is a union, either a plain one or a language
specific one. */
#define UNION_P(x) \
diff --git a/gcc/ggc-common.c b/gcc/ggc-common.c
index b6abed1d9a2..7c998e95473 100644
--- a/gcc/ggc-common.c
+++ b/gcc/ggc-common.c
@@ -256,6 +256,7 @@ saving_hasher::equal (const ptr_data *p1
}
static hash_table<saving_hasher> *saving_htab;
+static vec<void *> callback_vec;
/* Register an object in the hash table. */
@@ -288,6 +289,23 @@ gt_pch_note_object (void *obj, void *not
return 1;
}
+/* Register address of a callback pointer. */
+void
+gt_pch_note_callback (void *obj, void *base)
+{
+ void *ptr;
+ memcpy (&ptr, obj, sizeof (void *));
+ if (ptr != NULL)
+ {
+ struct ptr_data *data
+ = (struct ptr_data *)
+ saving_htab->find_with_hash (base, POINTER_HASH (base));
+ gcc_assert (data);
+ callback_vec.safe_push ((char *) data->new_addr
+ + ((char *) obj - (char *) base));
+ }
+}
+
/* Register an object in the hash table. */
void
@@ -582,10 +600,20 @@ gt_pch_save (FILE *f)
ggc_pch_finish (state.d, state.f);
gt_pch_fixup_stringpool ();
+ unsigned num_callbacks = callback_vec.length ();
+ void (*pch_save) (FILE *) = &gt_pch_save;
+ if (fwrite (&pch_save, sizeof (pch_save), 1, f) != 1
+ || fwrite (&num_callbacks, sizeof (num_callbacks), 1, f) != 1
+ || (num_callbacks
+ && fwrite (callback_vec.address (), sizeof (void *), num_callbacks,
+ f) != num_callbacks))
+ fatal_error (input_location, "cannot write PCH file: %m");
+
XDELETE (state.ptrs);
XDELETE (this_object);
delete saving_htab;
saving_htab = NULL;
+ callback_vec.release ();
}
/* Read the state of the compiler back in from F. */
@@ -639,6 +667,30 @@ gt_pch_restore (FILE *f)
ggc_pch_read (f, mmi.preferred_base);
gt_pch_restore_stringpool ();
+
+ void (*pch_save) (FILE *);
+ unsigned num_callbacks;
+ if (fread (&pch_save, sizeof (pch_save), 1, f) != 1
+ || fread (&num_callbacks, sizeof (num_callbacks), 1, f) != 1)
+ fatal_error (input_location, "cannot read PCH file: %m");
+ if (pch_save != &gt_pch_save)
+ {
+ uintptr_t bias = (uintptr_t) &gt_pch_save - (uintptr_t) pch_save;
+ void **ptrs = XNEWVEC (void *, num_callbacks);
+ unsigned i;
+
+ if (fread (ptrs, sizeof (void *), num_callbacks, f) != num_callbacks)
+ fatal_error (input_location, "cannot read PCH file: %m");
+ for (i = 0; i < num_callbacks; ++i)
+ {
+ memcpy (&pch_save, ptrs[i], sizeof (pch_save));
+ pch_save = (void (*) (FILE *)) ((uintptr_t) pch_save + bias);
+ memcpy (ptrs[i], &pch_save, sizeof (pch_save));
+ }
+ XDELETE (ptrs);
+ }
+ else if (fseek (f, num_callbacks * sizeof (void *), SEEK_CUR) != 0)
+ fatal_error (input_location, "cannot read PCH file: %m");
}
/* Default version of HOST_HOOKS_GT_PCH_GET_ADDRESS when mmap is not present.
diff --git a/gcc/ggc.h b/gcc/ggc.h
index 5e921d957fd..c005f7e0412 100644
--- a/gcc/ggc.h
+++ b/gcc/ggc.h
@@ -46,6 +46,10 @@ typedef void (*gt_handle_reorder) (void *, void *, gt_pointer_operator,
/* Used by the gt_pch_n_* routines. Register an object in the hash table. */
extern int gt_pch_note_object (void *, void *, gt_note_pointers);
+/* Used by the gt_pch_p_* routines. Register address of a callback
+ pointer. */
+extern void gt_pch_note_callback (void *, void *);
+
/* Used by the gt_pch_n_* routines. Register that an object has a reorder
function. */
extern void gt_pch_note_reorder (void *, void *, gt_handle_reorder);
diff --git a/gcc/output.h b/gcc/output.h
index 8f6f15308f4..4a23795bf7e 100644
--- a/gcc/output.h
+++ b/gcc/output.h
@@ -456,7 +456,7 @@ struct GTY(()) named_section {
/* A callback that writes the assembly code for switching to an unnamed
section. The argument provides callback-specific data. */
-typedef void (*unnamed_section_callback) (const void *);
+typedef void (*unnamed_section_callback) (const char *);
/* Information about a SECTION_UNNAMED section. */
struct GTY(()) unnamed_section {
@@ -464,8 +464,8 @@ struct GTY(()) unnamed_section {
/* The callback used to switch to the section, and the data that
should be passed to the callback. */
- unnamed_section_callback GTY ((skip)) callback;
- const void *GTY ((skip)) data;
+ unnamed_section_callback GTY ((callback)) callback;
+ const char *data;
/* The next entry in the chain of unnamed sections. */
section *next;
@@ -489,7 +489,7 @@ struct GTY(()) noswitch_section {
struct section_common common;
/* The callback used to assemble decls in this section. */
- noswitch_section_callback GTY ((skip)) callback;
+ noswitch_section_callback GTY ((callback)) callback;
};
/* Information about a section, which may be named or unnamed. */
@@ -524,8 +524,8 @@ extern GTY(()) section *bss_noswitch_sec
extern GTY(()) section *in_section;
extern GTY(()) bool in_cold_section_p;
-extern section *get_unnamed_section (unsigned int, void (*) (const void *),
- const void *);
+extern section *get_unnamed_section (unsigned int, void (*) (const char *),
+ const char *);
extern section *get_section (const char *, unsigned int, tree);
extern section *get_named_section (tree, const char *, int);
extern section *get_variable_section (tree, bool);
@@ -546,7 +546,7 @@ extern section *get_cdtor_priority_secti
extern bool unlikely_text_section_p (section *);
extern void switch_to_section (section *);
-extern void output_section_asm_op (const void *);
+extern void output_section_asm_op (const char *);
extern void record_tm_clone_pair (tree, tree);
extern void finish_tm_clone_pairs (void);
diff --git a/gcc/tree-core.h b/gcc/tree-core.h
index 8ab119dc9a2..91ae5237d7e 100644
--- a/gcc/tree-core.h
+++ b/gcc/tree-core.h
@@ -1961,7 +1961,7 @@ struct GTY(()) tree_function_decl {
struct GTY(()) tree_translation_unit_decl {
struct tree_decl_common common;
/* Source language of this translation unit. Used for DWARF output. */
- const char * GTY((skip(""))) language;
+ const char *language;
/* TODO: Non-optimization used to build this translation unit. */
/* TODO: Root of a partial DWARF tree for global types and decls. */
};
diff --git a/gcc/varasm.c b/gcc/varasm.c
index 9315e2c6936..aff93ca5de9 100644
--- a/gcc/varasm.c
+++ b/gcc/varasm.c
@@ -250,8 +250,8 @@ object_block_hasher::hash (object_block *old)
/* Return a new unnamed section with the given fields. */
section *
-get_unnamed_section (unsigned int flags, void (*callback) (const void *),
- const void *data)
+get_unnamed_section (unsigned int flags, void (*callback) (const char *),
+ const char *data)
{
section *sect;
@@ -7778,9 +7778,9 @@ file_end_indicate_split_stack (void)
a get_unnamed_section callback. */
void
-output_section_asm_op (const void *directive)
+output_section_asm_op (const char *directive)
{
- fprintf (asm_out_file, "%s\n", (const char *) directive);
+ fprintf (asm_out_file, "%s\n", directive);
}
/* Emit assembly code to switch to section NEW_SECTION. Do nothing if
diff --git a/libcpp/include/line-map.h b/libcpp/include/line-map.h
index 8b5e2f82982..bc40e333579 100644
--- a/libcpp/include/line-map.h
+++ b/libcpp/include/line-map.h
@@ -758,11 +758,11 @@ struct GTY(()) line_maps {
/* If non-null, the allocator to use when resizing 'maps'. If null,
xrealloc is used. */
- line_map_realloc reallocator;
+ line_map_realloc GTY((callback)) reallocator;
/* The allocators' function used to know the actual size it
allocated, for a certain allocation size requested. */
- line_map_round_alloc_size_func round_alloc_size;
+ line_map_round_alloc_size_func GTY((callback)) round_alloc_size;
struct location_adhoc_data_map location_adhoc_data_map;
diff --git a/gcc/config/avr/avr.c b/gcc/config/avr/avr.c
index 6ba038881d6..1c2f7d564e7 100644
--- a/gcc/config/avr/avr.c
+++ b/gcc/config/avr/avr.c
@@ -10089,7 +10089,7 @@ avr_asm_asm_output_aligned_bss (FILE *file, tree decl, const char *name,
to track need of __do_copy_data. */
static void
-avr_output_data_section_asm_op (const void *data)
+avr_output_data_section_asm_op (const char *data)
{
avr_need_copy_data_p = true;
@@ -10102,7 +10102,7 @@ avr_output_data_section_asm_op (const void *data)
to track need of __do_clear_bss. */
static void
-avr_output_bss_section_asm_op (const void *data)
+avr_output_bss_section_asm_op (const char *data)
{
avr_need_clear_bss_p = true;

254
SOURCES/gcc8-pr96796.patch Normal file
View File

@ -0,0 +1,254 @@
commit 6001db79c477b03eacc7e7049560921fb54b7845
Author: Richard Sandiford <richard.sandiford@arm.com>
Date: Mon Sep 7 20:15:36 2020 +0100
lra: Avoid cycling on certain subreg reloads [PR96796]
This PR is about LRA cycling for a reload of the form:
----------------------------------------------------------------------------
Changing pseudo 196 in operand 1 of insn 103 on equiv [r105:DI*0x8+r140:DI]
Creating newreg=287, assigning class ALL_REGS to slow/invalid mem r287
Creating newreg=288, assigning class ALL_REGS to slow/invalid mem r288
103: r203:SI=r288:SI<<0x1+r196:DI#0
REG_DEAD r196:DI
Inserting slow/invalid mem reload before:
316: r287:DI=[r105:DI*0x8+r140:DI]
317: r288:SI=r287:DI#0
----------------------------------------------------------------------------
The problem is with r287. We rightly give it a broad starting class of
POINTER_AND_FP_REGS (reduced from ALL_REGS by preferred_reload_class).
However, we never make forward progress towards narrowing it down to
a specific choice of class (POINTER_REGS or FP_REGS).
I think in practice we rely on two things to narrow a reload pseudo's
class down to a specific choice:
(1) a restricted class is specified when the pseudo is created
This happens for input address reloads, where the class is taken
from the target's chosen base register class. It also happens
for simple REG reloads, where the class is taken from the chosen
alternative's constraints.
(2) uses of the reload pseudo as a direct input operand
In this case get_reload_reg tries to reuse the existing register
and narrow its class, instead of creating a new reload pseudo.
However, neither occurs here. As described above, r287 rightly
starts out with a wide choice of class, ultimately derived from
ALL_REGS, so we don't get (1). And as the comments in the PR
explain, r287 is never used as an input reload, only the subreg is,
so we don't get (2):
----------------------------------------------------------------------------
Choosing alt 13 in insn 317: (0) r (1) w {*movsi_aarch64}
Creating newreg=291, assigning class FP_REGS to r291
317: r288:SI=r291:SI
Inserting insn reload before:
320: r291:SI=r287:DI#0
----------------------------------------------------------------------------
IMO, in this case we should rely on the reload of r316 to narrow
down the class of r278. Currently we do:
----------------------------------------------------------------------------
Choosing alt 7 in insn 316: (0) r (1) m {*movdi_aarch64}
Creating newreg=289 from oldreg=287, assigning class GENERAL_REGS to r289
316: r289:DI=[r105:DI*0x8+r140:DI]
Inserting insn reload after:
318: r287:DI=r289:DI
---------------------------------------------------
i.e. we create a new pseudo register r289 and give *that* pseudo
GENERAL_REGS instead. This is because get_reload_reg only narrows
down the existing class for OP_IN and OP_INOUT, not OP_OUT.
But if we have a reload pseudo in a reload instruction and have chosen
a specific class for the reload pseudo, I think we should simply install
it for OP_OUT reloads too, if the class is a subset of the existing class.
We will need to pick such a register whatever happens (for r289 in the
example above). And as explained in the PR, doing this actually avoids
an unnecessary move via the FP registers too.
The patch is quite aggressive in that it does this for all reload
pseudos in all reload instructions. I wondered about reusing the
condition for a reload move in in_class_p:
INSN_UID (curr_insn) >= new_insn_uid_start
&& curr_insn_set != NULL
&& ((OBJECT_P (SET_SRC (curr_insn_set))
&& ! CONSTANT_P (SET_SRC (curr_insn_set)))
|| (GET_CODE (SET_SRC (curr_insn_set)) == SUBREG
&& OBJECT_P (SUBREG_REG (SET_SRC (curr_insn_set)))
&& ! CONSTANT_P (SUBREG_REG (SET_SRC (curr_insn_set)))))))
but I can't really justify that on first principles. I think we
should apply the rule consistently until we have a specific reason
for doing otherwise.
gcc/
PR rtl-optimization/96796
* lra-constraints.c (in_class_p): Add a default-false
allow_all_reload_class_changes_p parameter. Do not treat
reload moves specially when the parameter is true.
(get_reload_reg): Try to narrow the class of an existing OP_OUT
reload if we're reloading a reload pseudo in a reload instruction.
gcc/testsuite/
PR rtl-optimization/96796
* gcc.c-torture/compile/pr96796.c: New test.
diff --git a/gcc/lra-constraints.c b/gcc/lra-constraints.c
index 580da9c3ed6..161b721efb1 100644
--- a/gcc/lra-constraints.c
+++ b/gcc/lra-constraints.c
@@ -236,12 +236,17 @@ get_reg_class (int regno)
CL. Use elimination first if REG is a hard register. If REG is a
reload pseudo created by this constraints pass, assume that it will
be allocated a hard register from its allocno class, but allow that
- class to be narrowed to CL if it is currently a superset of CL.
+ class to be narrowed to CL if it is currently a superset of CL and
+ if either:
+
+ - ALLOW_ALL_RELOAD_CLASS_CHANGES_P is true or
+ - the instruction we're processing is not a reload move.
If NEW_CLASS is nonnull, set *NEW_CLASS to the new allocno class of
REGNO (reg), or NO_REGS if no change in its class was needed. */
static bool
-in_class_p (rtx reg, enum reg_class cl, enum reg_class *new_class)
+in_class_p (rtx reg, enum reg_class cl, enum reg_class *new_class,
+ bool allow_all_reload_class_changes_p = false)
{
enum reg_class rclass, common_class;
machine_mode reg_mode;
@@ -266,7 +271,8 @@ in_class_p (rtx reg, enum reg_class cl, enum reg_class *new_class)
typically moves that have many alternatives, and restricting
reload pseudos for one alternative may lead to situations
where other reload pseudos are no longer allocatable. */
- || (INSN_UID (curr_insn) >= new_insn_uid_start
+ || (!allow_all_reload_class_changes_p
+ && INSN_UID (curr_insn) >= new_insn_uid_start
&& curr_insn_set != NULL
&& ((OBJECT_P (SET_SRC (curr_insn_set))
&& ! CONSTANT_P (SET_SRC (curr_insn_set)))
@@ -551,13 +557,12 @@ init_curr_insn_input_reloads (void)
curr_insn_input_reloads_num = 0;
}
-/* Create a new pseudo using MODE, RCLASS, ORIGINAL or reuse already
- created input reload pseudo (only if TYPE is not OP_OUT). Don't
- reuse pseudo if IN_SUBREG_P is true and the reused pseudo should be
- wrapped up in SUBREG. The result pseudo is returned through
- RESULT_REG. Return TRUE if we created a new pseudo, FALSE if we
- reused the already created input reload pseudo. Use TITLE to
- describe new registers for debug purposes. */
+/* Create a new pseudo using MODE, RCLASS, ORIGINAL or reuse an existing
+ reload pseudo. Don't reuse an existing reload pseudo if IN_SUBREG_P
+ is true and the reused pseudo should be wrapped up in a SUBREG.
+ The result pseudo is returned through RESULT_REG. Return TRUE if we
+ created a new pseudo, FALSE if we reused an existing reload pseudo.
+ Use TITLE to describe new registers for debug purposes. */
static bool
get_reload_reg (enum op_type type, machine_mode mode, rtx original,
enum reg_class rclass, bool in_subreg_p,
@@ -616,6 +621,35 @@ get_reload_reg (enum op_type type, machine_mode mode, rtx original,
if (type == OP_OUT)
{
+ /* Output reload registers tend to start out with a conservative
+ choice of register class. Usually this is ALL_REGS, although
+ a target might narrow it (for performance reasons) through
+ targetm.preferred_reload_class. It's therefore quite common
+ for a reload instruction to require a more restrictive class
+ than the class that was originally assigned to the reload register.
+
+ In these situations, it's more efficient to refine the choice
+ of register class rather than create a second reload register.
+ This also helps to avoid cycling for registers that are only
+ used by reload instructions. */
+ if (REG_P (original)
+ && (int) REGNO (original) >= new_regno_start
+ && INSN_UID (curr_insn) >= new_insn_uid_start
+ && in_class_p (original, rclass, &new_class, true))
+ {
+ unsigned int regno = REGNO (original);
+ if (lra_dump_file != NULL)
+ {
+ fprintf (lra_dump_file, " Reuse r%d for output ", regno);
+ dump_value_slim (lra_dump_file, original, 1);
+ }
+ if (new_class != lra_get_allocno_class (regno))
+ lra_change_class (regno, new_class, ", change to", false);
+ if (lra_dump_file != NULL)
+ fprintf (lra_dump_file, "\n");
+ *result_reg = original;
+ return false;
+ }
*result_reg
= lra_create_new_reg_with_unique_value (mode, original, rclass, title);
return true;
diff --git a/gcc/testsuite/gcc.c-torture/compile/pr96796.c b/gcc/testsuite/gcc.c-torture/compile/pr96796.c
new file mode 100644
index 00000000000..8808e62fe77
--- /dev/null
+++ b/gcc/testsuite/gcc.c-torture/compile/pr96796.c
@@ -0,0 +1,55 @@
+/* { dg-additional-options "-fcommon" } */
+
+struct S0 {
+ signed f0 : 8;
+ unsigned f1;
+ unsigned f4;
+};
+struct S1 {
+ long f3;
+ char f4;
+} g_3_4;
+
+int g_5, func_1_l_32, func_50___trans_tmp_31;
+static struct S0 g_144, g_834, g_1255, g_1261;
+
+int g_273[120] = {};
+int *g_555;
+char **g_979;
+static int g_1092_0;
+static int g_1193;
+int safe_mul_func_int16_t_s_s(int si1, int si2) { return si1 * si2; }
+static struct S0 *func_50();
+int func_1() { func_50(g_3_4, g_5, func_1_l_32, 8, 3); }
+void safe_div_func_int64_t_s_s(int *);
+void safe_mod_func_uint32_t_u_u(struct S0);
+struct S0 *func_50(int p_51, struct S0 p_52, struct S1 p_53, int p_54,
+ int p_55) {
+ int __trans_tmp_30;
+ char __trans_tmp_22;
+ short __trans_tmp_19;
+ long l_985_1;
+ long l_1191[8];
+ safe_div_func_int64_t_s_s(g_273);
+ __builtin_printf((char*)g_1261.f4);
+ safe_mod_func_uint32_t_u_u(g_834);
+ g_144.f0 += 1;
+ for (;;) {
+ struct S1 l_1350 = {&l_1350};
+ for (; p_53.f3; p_53.f3 -= 1)
+ for (; g_1193 <= 2; g_1193 += 1) {
+ __trans_tmp_19 = safe_mul_func_int16_t_s_s(l_1191[l_985_1 + p_53.f3],
+ p_55 % (**g_979 = 10));
+ __trans_tmp_22 = g_1255.f1 * p_53.f4;
+ __trans_tmp_30 = __trans_tmp_19 + __trans_tmp_22;
+ if (__trans_tmp_30)
+ g_1261.f0 = p_51;
+ else {
+ g_1255.f0 = p_53.f3;
+ int *l_1422 = g_834.f0 = g_144.f4 != (*l_1422)++ > 0 < 0 ^ 51;
+ g_555 = ~0;
+ g_1092_0 |= func_50___trans_tmp_31;
+ }
+ }
+ }
+}

101
SOURCES/gcc8-rh2028609.patch Normal file
View File

@ -0,0 +1,101 @@
The cprop_hardreg pass is built around the assumption that accessing a
register in a narrower mode is the same as accessing the lowpart of
the register. This unfortunately is not true for vector registers on
IBM Z. This caused a miscompile of LLVM with GCC 8.5. The problem
could not be reproduced with upstream GCC unfortunately but we have to
assume that it is latent there. The right fix would require
substantial changes to the cprop pass and is certainly something we
would want for our platform. But since this would not be acceptable
for older GCCs I'll go with what Vladimir proposed in the RedHat BZ
and introduce a hopefully temporary and undocumented target hook to
disable that specific transformation in regcprop.c.
--- a/gcc/config/s390/s390.c
+++ b/gcc/config/s390/s390.c
@@ -10488,6 +10488,18 @@ s390_hard_regno_mode_ok (unsigned int regno, machine_mode mode)
return false;
}
+/* Implement TARGET_NARROW_MODE_REFERS_LOW_PART_P. */
+
+static bool
+s390_narrow_mode_refers_low_part_p (unsigned int regno)
+{
+ if (reg_classes_intersect_p (VEC_REGS, REGNO_REG_CLASS (regno)))
+ return false;
+
+ return true;
+}
+
+
/* Implement TARGET_MODES_TIEABLE_P. */
static bool
@@ -16956,6 +16968,9 @@ s390_case_values_threshold (void)
#undef TARGET_CASE_VALUES_THRESHOLD
#define TARGET_CASE_VALUES_THRESHOLD s390_case_values_threshold
+#undef TARGET_NARROW_MODE_REFERS_LOW_PART_P
+#define TARGET_NARROW_MODE_REFERS_LOW_PART_P s390_narrow_mode_refers_low_part_p
+
struct gcc_target targetm = TARGET_INITIALIZER;
#include "gt-s390.h"
--- a/gcc/regcprop.c
+++ b/gcc/regcprop.c
@@ -426,7 +426,8 @@ maybe_mode_change (machine_mode orig_mode, machine_mode copy_mode,
if (orig_mode == new_mode)
return gen_raw_REG (new_mode, regno);
- else if (mode_change_ok (orig_mode, new_mode, regno))
+ else if (mode_change_ok (orig_mode, new_mode, regno)
+ && targetm.narrow_mode_refers_low_part_p (copy_regno))
{
int copy_nregs = hard_regno_nregs (copy_regno, copy_mode);
int use_nregs = hard_regno_nregs (copy_regno, new_mode);
--- a/gcc/target.def
+++ b/gcc/target.def
@@ -5446,6 +5446,16 @@ value that the middle-end intended.",
bool, (machine_mode from, machine_mode to, reg_class_t rclass),
hook_bool_mode_mode_reg_class_t_true)
+/* This hook is used to work around a problem in regcprop. Hardcoded
+assumptions currently prevent it from working correctly for targets
+where the low part of a multi-word register doesn't align to accessing
+the register with a narrower mode. */
+DEFHOOK_UNDOC
+(narrow_mode_refers_low_part_p,
+"",
+bool, (unsigned int regno),
+hook_bool_uint_true)
+
/* Change pseudo allocno class calculated by IRA. */
DEFHOOK
(ira_change_pseudo_allocno_class,
--- a/gcc/hooks.h
+++ b/gcc/hooks.h
@@ -86,6 +86,7 @@ extern void hook_void_tree (tree);
extern void hook_void_tree_treeptr (tree, tree *);
extern void hook_void_int_int (int, int);
extern void hook_void_gcc_optionsp (struct gcc_options *);
+extern bool hook_bool_uint_true (unsigned int);
extern bool hook_bool_uint_uintp_false (unsigned int, unsigned int *);
extern int hook_int_uint_mode_1 (unsigned int, machine_mode);
--- a/gcc/hooks.c
+++ b/gcc/hooks.c
@@ -498,6 +498,14 @@ hook_void_gcc_optionsp (struct gcc_optio
{
}
+/* Generic hook that takes an unsigned int and returns true. */
+
+bool
+hook_bool_uint_true (unsigned int)
+{
+ return true;
+}
+
/* Generic hook that takes an unsigned int, an unsigned int pointer and
returns false. */

36
SPECS/gcc.spec
View File

@ -4,7 +4,7 @@
%global gcc_major 8
# Note, gcc_release must be integer, if you want to add suffixes to
# %%{release}, append them after %%{gcc_release} on Release: line.
%global gcc_release 4
%global gcc_release 10
%global nvptx_tools_gitrev c28050f60193b3b95a18866a96f03334e874e78f
%global nvptx_newlib_gitrev aadc8eb0ec43b7cd0dd2dfb484bae63c8b05ef24
%global _unpackaged_files_terminate_build 0
@ -279,7 +279,11 @@ Patch18: gcc8-remove-old-demangle.patch
Patch19: gcc8-rh1960701.patch
Patch20: gcc8-pr100797.patch
Patch21: gcc8-rh1981822.patch
Patch22: gcc8-add-Wbidirectional.patch
Patch22: gcc8-Wbidi-chars.patch
Patch23: gcc8-pr96796.patch
Patch24: gcc8-pch-tweaks.patch
Patch25: gcc8-aarch64-mtune-neoverse-512tvb.patch
Patch26: gcc8-rh2028609.patch
Patch30: gcc8-rh1668903-1.patch
Patch31: gcc8-rh1668903-2.patch
@ -861,6 +865,10 @@ to NVidia PTX capable devices if available.
%patch20 -p0 -b .pr100797~
%patch21 -p0 -b .rh1981822~
%patch22 -p1 -b .bidi~
%patch23 -p1 -b .pr96796~
%patch24 -p1 -b .pch-tweaks~
%patch25 -p1 -b .neoverse~
%patch26 -p1 -b .rh2028609~
%patch30 -p0 -b .rh1668903-1~
%patch31 -p0 -b .rh1668903-2~
@ -1359,7 +1367,7 @@ mkdir -p %{buildroot}/%{_lib}
mv -f %{buildroot}%{_prefix}/%{_lib}/libgcc_s.so.1 %{buildroot}/%{_lib}/libgcc_s-%{gcc_major}-%{DATE}.so.1
chmod 755 %{buildroot}/%{_lib}/libgcc_s-%{gcc_major}-%{DATE}.so.1
ln -sf libgcc_s-%{gcc_major}-%{DATE}.so.1 %{buildroot}/%{_lib}/libgcc_s.so.1
%ifarch %{ix86} x86_64 ppc ppc64 ppc64p7 ppc64le %{arm}
%ifarch %{ix86} x86_64 ppc ppc64 ppc64p7 ppc64le %{arm} aarch64
rm -f $FULLPATH/libgcc_s.so
echo '/* GNU ld script
Use the shared library, but some functions are only in
@ -3177,8 +3185,26 @@ fi
%endif
%changelog
* Fri Oct 29 2021 Marek Polacek <polacek@redhat.com> 8.5.0-4
- add -Wbidirectional patch (#2008391)
* Thu Jan 27 2022 Marek Polacek <polacek@redhat.com> 8.5.0-10
- fix typo in the cprop_hardreg patch (#2028609)
* Mon Jan 24 2022 Marek Polacek <polacek@redhat.com> 8.5.0-9
- apply cprop_hardreg fix for narrow mode != lowpart targets (#2028609)
* Mon Jan 24 2022 Marek Polacek <polacek@redhat.com> 8.5.0-8
- aarch64: Add -mtune=neoverse-512tvb (#1845932)
* Fri Dec 10 2021 Marek Polacek <polacek@redhat.com> 8.5.0-7
- backport PCH tweaks (#2030878)
* Fri Dec 3 2021 Marek Polacek <polacek@redhat.com> 8.5.0-6
- avoid cycling on certain subreg reloads (PR rtl-optimization/96796, #2028798)
* Tue Nov 30 2021 Marek Polacek <polacek@redhat.com> 8.5.0-5
- when linking against libgcc_s, link libgcc.a too (#2022588)
* Thu Nov 18 2021 Marek Polacek <polacek@redhat.com> 8.5.0-4
- add -Wbidi-chars patch (#2008392)
* Tue Jul 13 2021 Marek Polacek <polacek@redhat.com> 8.5.0-3
- fix mangling of lambdas in default args (PR c++/91241, #1981822)