Bug 54515 - cc1plus sigsegv -O2 anonymous namespace
Summary: cc1plus sigsegv -O2 anonymous namespace
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: middle-end (show other bugs)
Version: 4.6.3
: P3 normal
Target Milestone: ---
Assignee: Richard Biener
URL:
Keywords: ice-on-valid-code
Depends on:
Blocks:
 
Reported: 2012-09-07 10:02 UTC by Werner Henze
Modified: 2012-09-11 08:33 UTC (History)
1 user (show)

See Also:
Host:
Target:
Build:
Known to work: 4.6.4, 4.7.2, 4.8.0
Known to fail: 4.6.3, 4.7.1
Last reconfirmed: 2012-09-07 00:00:00


Attachments
preprocessed file as written by the sigsegv handler (244.26 KB, application/x-gzip)
2012-09-07 10:53 UTC, Werner Henze
Details
preprocessed file as written by the sigsegv handler (201.87 KB, application/x-gzip)
2012-09-07 11:49 UTC, Werner Henze
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Werner Henze 2012-09-07 10:02:02 UTC
cc1plus crashes with a sigsegv.
The crash does occur with -O2, not with -Os or -O1.
The crash does not occur if I do not use an anonymous namespace for the functions ReadPIB7400, ReadPIB7420, AddAdapterAndReferencedAdapters and GenerateTrafficForPhyRate (by commenting out the lines #73452, #73458, #74520 and #74616).
Comment 1 Richard Biener 2012-09-07 10:24:50 UTC
Attachment missing.
Comment 2 Werner Henze 2012-09-07 10:53:33 UTC
Created attachment 28144 [details]
preprocessed file as written by the sigsegv handler
Comment 3 Werner Henze 2012-09-07 11:49:43 UTC
Created attachment 28145 [details]
preprocessed file as written by the sigsegv handler

Just saw a second sigsegv. This might be related to the bug (including same project files), but it might also be another bug.
This compilation succeeds with -O0 and -O1, it crashes with -O3, -O2 and -Os.
In this case there is no anonymous namespace involved.
Comment 4 Markus Trippelsdorf 2012-09-07 11:56:34 UTC
gcc-4.7 and 4.8 are also affected.

markus@x4 tmp % cat test.ii
template < typename T > T h2le (T)
{
    T a;
    unsigned short &b = a;
    short c = 0;
    unsigned char (&d)[2] = reinterpret_cast < unsigned char (&)[2] > (c);
    unsigned char (&e)[2] = reinterpret_cast < unsigned char (&)[2] > (b);
    e[0] = d[0];
    return a;
}

void
bar ()
{
    h2le ((unsigned short) 0);
}

markus@x4 tmp % gdb /var/tmp/foo/usr/x86_64-pc-linux-gnu/gcc-bin/4.8.0/g++
Reading symbols from /var/tmp/foo/usr/x86_64-pc-linux-gnu/gcc-bin/4.8.0/g++...done.
(gdb) run -c -O2 -std=c++0x test.ii
Starting program: /var/tmp/foo/usr/x86_64-pc-linux-gnu/gcc-bin/4.8.0/g++ -c -O2 -std=c++0x test.ii
warning: no loadable sections found in added symbol-file system-supplied DSO at 0x7ffff7ffa000
[New process 18282]
process 18282 is executing new program: /var/tmp/foo/usr/libexec/gcc/x86_64-pc-linux-gnu/4.8.0/cc1plus

Program received signal SIGSEGV, Segmentation fault.
[Switching to process 18282]
0x00000000008b5803 in disqualify_base_of_expr(tree_node*, char const*) () at ../../gcc/gcc/tree-sra.c:987
987     ../../gcc/gcc/tree-sra.c: No such file or directory.
(gdb) bt
#0  0x00000000008b5803 in disqualify_base_of_expr(tree_node*, char const*) () at ../../gcc/gcc/tree-sra.c:987
#1  0x00000000008b6eb5 in build_access_from_expr_1(tree_node*, gimple_statement_d*, bool) () at ../../gcc/gcc/tree-sra.c:1025
#2  0x00000000008b74ab in scan_function() () at ../../gcc/gcc/tree-sra.c:1119
#3  0x00000000008bb622 in ipa_early_sra() () at ../../gcc/gcc/tree-sra.c:4814
#4  0x00000000007ca012 in execute_one_pass(opt_pass*) () at ../../gcc/gcc/passes.c:2158
#5  0x00000000007ca385 in execute_pass_list(opt_pass*) () at ../../gcc/gcc/passes.c:2213
#6  0x00000000007ca397 in execute_pass_list(opt_pass*) () at ../../gcc/gcc/passes.c:2214
#7  0x00000000007c95fd in do_per_function_toporder(void (*)(void*), void*) () at ../../gcc/gcc/passes.c:1702
#8  0x00000000007ca7ec in execute_ipa_pass_list(opt_pass*) () at ../../gcc/gcc/passes.c:2527
#9  0x000000000064f758 in compile() ()
#10 0x000000000064fb15 in finalize_compilation_unit() ()
#11 0x0000000000528a57 in cp_write_global_declarations() () at ../../gcc/gcc/cp/decl2.c:4024
#12 0x000000000085a565 in compile_file() ()
#13 0x000000000085bd5a in toplev_main(int, char**) ()
#14 0x00007ffff74ef755 in __libc_start_main () from /lib/libc.so.6
#15 0x00000000004c8249 in _start ()
Comment 5 Markus Trippelsdorf 2012-09-07 11:59:45 UTC
(In reply to comment #3)
> Created attachment 28145 [details]
> preprocessed file as written by the sigsegv handler
> 
> Just saw a second sigsegv. This might be related to the bug (including same
> project files), but it might also be another bug.
> This compilation succeeds with -O0 and -O1, it crashes with -O3, -O2 and -Os.
> In this case there is no anonymous namespace involved.

It's the same issue.
Comment 6 Markus Trippelsdorf 2012-09-07 15:23:04 UTC
The following patch fixes the issue for me:

diff --git a/gcc/tree-sra.c b/gcc/tree-sra.c
index aafaa15..2bb92e9 100644
--- a/gcc/tree-sra.c
+++ b/gcc/tree-sra.c
@@ -984,7 +984,8 @@ static void
 disqualify_base_of_expr (tree t, const char *reason)
 {
   t = get_base_address (t);
-  if (sra_mode == SRA_MODE_EARLY_IPA
+  if (t
+      && sra_mode == SRA_MODE_EARLY_IPA
       && TREE_CODE (t) == MEM_REF)
     t = get_ssa_base_param (TREE_OPERAND (t, 0));
Comment 7 Richard Biener 2012-09-10 09:48:32 UTC
Author: rguenth
Date: Mon Sep 10 09:48:28 2012
New Revision: 191130

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=191130
Log:
2012-09-07  Markus Trippelsdorf  <markus@trippelsdorf.de>

	PR middle-end/54515
	* tree-sra.c (disqualify_base_of_expr): Check for possible
	NULL_TREE returned by get_base_address()

	* g++.dg/tree-ssa/pr54515.C: new testcase

Added:
    branches/gcc-4_7-branch/gcc/testsuite/g++.dg/tree-ssa/pr54515.C
Modified:
    branches/gcc-4_7-branch/gcc/ChangeLog
    branches/gcc-4_7-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_7-branch/gcc/tree-sra.c
Comment 8 Richard Biener 2012-09-10 09:50:07 UTC
Author: rguenth
Date: Mon Sep 10 09:50:02 2012
New Revision: 191131

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=191131
Log:
2012-09-10  Markus Trippelsdorf  <markus@trippelsdorf.de>

	PR middle-end/54515
	* tree-sra.c (disqualify_base_of_expr): Check for possible
	NULL_TREE returned by get_base_address()

	* g++.dg/tree-ssa/pr54515.C: new testcase

Added:
    branches/gcc-4_6-branch/gcc/testsuite/g++.dg/tree-ssa/pr54515.C
Modified:
    branches/gcc-4_6-branch/gcc/ChangeLog
    branches/gcc-4_6-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_6-branch/gcc/tree-sra.c
Comment 9 Richard Biener 2012-09-10 11:33:03 UTC
Testing sth different for trunk.
Comment 10 Richard Biener 2012-09-11 08:32:43 UTC
Author: rguenth
Date: Tue Sep 11 08:32:29 2012
New Revision: 191174

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=191174
Log:
2012-09-11  Richard Guenther  <rguenther@suse.de>

	PR middle-end/54515
	* gimple.c (get_base_address): Do not return NULL_TREE apart
	from for WITH_SIZE_EXPR.
	* gimple-fold.c (canonicalize_constructor_val): Do not call
	get_base_address when not necessary.

	* g++.dg/tree-ssa/pr54515.C: New testcase.

Added:
    trunk/gcc/testsuite/g++.dg/tree-ssa/pr54515.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/gimple-fold.c
    trunk/gcc/gimple.c
    trunk/gcc/testsuite/ChangeLog
Comment 11 Richard Biener 2012-09-11 08:33:36 UTC
Fixed.