Hello.
One can see use-after-scope error in boostrap-asan:
Executing on host: /home/marxin/gcc/objdir2/gcc/xgcc
-B/home/marxin/gcc/objdir2/gcc/ -fno-diagnostics-show-caret
-fdiagnostics-color=never -w -O3 -g -dumpbase dump1/dump-noaddr.c -DMASK=1
-x c --param ggc-min-heapsize=1 -fdump-ipa-all -fdump-rtl-all -fdump-tree-all -
fdump-noaddr -c -o /home/marxin/gcc/objdir2/gcc/testsuite/gcc27/dump-noaddr.o
/home/marxin/gcc/gcc/testsuite/gcc.c-torture/unsorted/dump-noaddr.c (timeout
= 300)
spawn /home/marxin/gcc/objdir2/gcc/xgcc -B/home/marxin/gcc/objdir2/gcc/
-fno-diagnostics-show-caret -fdiagnostics-color=never -w -O3 -g -dumpbase
dump1/dump-noaddr.c -DMASK=1 -x c --param ggc-min-heapsize=1 -fdump-ipa-all
-fdump-rtl-all -fdump-tree-all -fdump-noaddr -c -o
/home/marxin/gcc/objdir2/gcc/testsuite/gcc27/dump-noaddr.o
/home/marxin/gcc/gcc/testsuite/gcc.c-torture/unsorted/dump-noaddr.c
=================================================================
==88887==ERROR: AddressSanitizer: stack-use-after-scope on address
0x7fffffff9890 at pc 0x0000009d9361 bp 0x7fffffff9280 sp 0x7fffffff8a30
READ of size 4 at 0x7fffffff9890 thread T0
#0 0x9d9360 in __interceptor_strlen
../../../../libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:225
#1 0x3189f03 in pp_string(pretty_printer*, char const*)
../../gcc/pretty-print.c:990
#2 0x318cb5a in pp_format(pretty_printer*, text_info*)
../../gcc/pretty-print.c:599
#3 0x318ecfe in pp_printf(pretty_printer*, char const*, ...)
../../gcc/pretty-print.c:937
#4 0x17ac387 in print_insn(pretty_printer*, rtx_insn const*, int)
../../gcc/print-rtl.c:1816
#5 0x17ac837 in print_insn_with_notes ../../gcc/print-rtl.c:1897
#6 0x17b5c73 in dump_insn_slim(_IO_FILE*, rtx_insn const*)
../../gcc/print-rtl.c:1934
#7 0x2e61d85 in combine_instructions ../../gcc/combine.c:1218
#8 0x2e61d85 in rest_of_handle_combine ../../gcc/combine.c:14784
#9 0x2e61d85 in execute ../../gcc/combine.c:14829
#10 0x173d971 in execute_one_pass(opt_pass*) ../../gcc/passes.c:2495
#11 0x173f126 in execute_pass_list_1 ../../gcc/passes.c:2584
#12 0x173f150 in execute_pass_list_1 ../../gcc/passes.c:2585
#13 0x173f1af in execute_pass_list(function*, opt_pass*)
../../gcc/passes.c:2595
#14 0xeb7957 in cgraph_node::expand() ../../gcc/cgraphunit.c:2115
#15 0xeba71e in expand_all_functions ../../gcc/cgraphunit.c:2251
#16 0xeba71e in symbol_table::compile() ../../gcc/cgraphunit.c:2599
#17 0xec0e40 in symbol_table::compile() ../../gcc/cgraphunit.c:2695
#18 0xec0e40 in symbol_table::finalize_compilation_unit()
../../gcc/cgraphunit.c:2692
#19 0x19e3a8b in compile_file ../../gcc/toplev.c:481
#20 0x9a2a3f in do_compile ../../gcc/toplev.c:2037
#21 0x9a2a3f in toplev::main(int, char**) ../../gcc/toplev.c:2172
#22 0x9acd24 in main ../../gcc/main.c:39
#23 0x7ffff6a396e4 in __libc_start_main (/lib64/libc.so.6+0x206e4)
#24 0x9adf28 in _start (/home/marxin/gcc/objdir2/gcc/cc1+0x9adf28)
Address 0x7fffffff9890 is located in stack of thread T0 at offset 96 in frame
#0 0x17abfdf in print_insn(pretty_printer*, rtx_insn const*, int)
../../gcc/print-rtl.c:1777
This frame has 2 object(s):
[32, 64) 'uid_prefix'
[96, 128) 'idbuf' <== Memory access at offset 96 is inside this variable
...
This is fix of that. It's quite clear.
Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
Ready to be installed?
Martin
gcc/ChangeLog:
2017-10-10 Martin Liska <[email protected]>
* print-rtl.c (print_insn): Move declaration of idbuf
to same scope as name.
---
gcc/print-rtl.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gcc/print-rtl.c b/gcc/print-rtl.c
index 79ec463df45..28d99862cad 100644
--- a/gcc/print-rtl.c
+++ b/gcc/print-rtl.c
@@ -1792,11 +1792,11 @@ print_insn (pretty_printer *pp, const rtx_insn *x, int verbose)
case DEBUG_INSN:
{
const char *name = "?";
+ char idbuf[32];
if (DECL_P (INSN_VAR_LOCATION_DECL (x)))
{
tree id = DECL_NAME (INSN_VAR_LOCATION_DECL (x));
- char idbuf[32];
if (id)
name = IDENTIFIER_POINTER (id);
else if (TREE_CODE (INSN_VAR_LOCATION_DECL (x))
