https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70690
Bug ID: 70690
Summary: [6/7 Regression] r235002 miscompiles tcmalloc
Product: gcc
Version: 6.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: trippels at gcc dot gnu.org
CC: jason at gcc dot gnu.org
Target Milestone: ---
Created attachment 38286
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38286&action=edit
unreduced testcase
Starting with r235002 tcmalloc apparently gets miscompiled.
Chromium uses tcmalloc and crashes on startup:
Program received signal SIGSEGV, Segmentation fault.
0x00005555564f7dbf in tcmalloc::CentralFreeList::FetchFromSpans() ()
(gdb) bt
#0 0x00005555564f7dbf in tcmalloc::CentralFreeList::FetchFromSpans() ()
#1 0x00005555564f8429 in tcmalloc::CentralFreeList::FetchFromSpansSafe() ()
#2 0x00005555564f84c2 in tcmalloc::CentralFreeList::RemoveRange(void**,
void**, int) ()
#3 0x00005555564f3b94 in tcmalloc::ThreadCache::FetchFromCentralCache(unsigned
long, unsigned long) ()
#4 0x000055555b24873a in tc_malloc ()
#5 0x000055555650a2ba in malloc ()
#6 0x00007ffff606b07d in __fopen_internal (filename=0x8a049a4f740
"/proc/31709/stat", mode=0x55555b252e53 "rb", is32=1) at iofopen.c:69
#7 0x00005555564849c6 in base::OpenFile(base::FilePath const&, char const*) ()
#8 0x0000555556483787 in base::ReadFileToStringWithMaxSize(base::FilePath
const&, std::__cxx11::basic_string<char, std::char_traits<char>,
std::allocator<char> >*, unsigned long) ()
#9 0x00005555564dae45 in base::internal::ReadProcStats(int,
std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>
>*) ()
#10 0x00005555564dbb03 in base::internal::ReadProcStatsAndGetFieldAsInt64(int,
base::internal::ProcStatsFields) ()
#11 0x00005555564a2a72 in base::CurrentProcessInfo::CreationTime() ()
#12 0x000055555602a970 in ChromeMainDelegate::ChromeMainDelegate() ()
#13 0x00005555560299db in ChromeMain ()
I've narrowed the issue down to a single file (attached).
Assembly difference using:
g++ -c -fPIC static_vars.ii -O2 -S
markus@x4 Release % diff -u static_vars.s static_vars_bad.s
--- static_vars.s 2016-04-16 13:52:18.332989892 +0200
+++ static_vars_bad.s 2016-04-16 13:34:21.899894073 +0200
@@ -1118,6 +1118,31 @@
.cfi_endproc
.LFE1443:
.size _ZN8tcmalloc6Static14InitStaticVarsEv,
.-_ZN8tcmalloc6Static14InitStaticVarsEv
+ .section .text.startup,"ax",@progbits
+ .p2align 4,,15
+ .type _GLOBAL__sub_I_static_vars.cc, @function
+_GLOBAL__sub_I_static_vars.cc:
+.LFB1718:
+ .cfi_startproc
+ movq _ZN8tcmalloc6Static14central_cache_E@GOTPCREL(%rip), %rdx
+ xorl %eax, %eax
+ leaq 64448(%rdx), %rsi
+ .p2align 4,,10
+ .p2align 3
+.L68:
+ movq %rdx, %rdi
+ addq $1216, %rdx
+ movl $152, %ecx
+ cmpq %rdx, %rsi
+ rep stosq
+ jne .L68
+ rep ret
+ .cfi_endproc
+.LFE1718:
+ .size _GLOBAL__sub_I_static_vars.cc, .-_GLOBAL__sub_I_static_vars.cc
+ .section .init_array,"aw"
+ .align 8
+ .quad _GLOBAL__sub_I_static_vars.cc
.globl _ZN8tcmalloc6Static9pageheap_E
.bss
.align 8
@@ -1173,5 +1198,5 @@
.size _ZN8tcmalloc6Static14pageheap_lock_E, 4
_ZN8tcmalloc6Static14pageheap_lock_E:
.zero 4
- .ident "GCC: (GNU) 6.0.0 20160415 (experimental)"
+ .ident "GCC: (GNU) 6.0.1 20160416 (prerelease)"
.section .note.GNU-stack,"",@progbits
It is weekend, so I haven't looked deeper yet.
