public inbox for gcc-prs@sourceware.org
help / color / mirror / Atom feed
From: aaronw@net.com
To: gcc-gnats@gcc.gnu.org
Subject: target/8004: All C++ binaries crash in __register_frame_info_bases on Sparc Solaris 2.7
Date: Sun, 22 Sep 2002 10:56:00 -0000 [thread overview]
Message-ID: <20020922175041.5784.qmail@sources.redhat.com> (raw)
>Number: 8004
>Category: target
>Synopsis: All C++ binaries crash in __register_frame_info_bases on Sparc Solaris 2.7
>Confidential: no
>Severity: critical
>Priority: medium
>Responsible: unassigned
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Sun Sep 22 10:56:00 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator: Aaron Williams
>Release: 3.2
>Organization:
>Environment:
Sparc Sun Solaris 2.7 --disable-multilib (32-bit only)
>Description:
All binaries appear to crash in __register_frame_info_bases.
Here is the gdb info:
0xff2b5ffc <__register_frame_info_bases>: save %sp, -112, %sp
0xff2b6000 <__register_frame_info_bases+4>: sethi %hi(0), %o0
0xff2b6004 <__register_frame_info_bases+8>: sethi %hi(0x11400), %l7
0xff2b6008 <__register_frame_info_bases+12>:
call 0xff2b5e38 <base_of_encoded_value+176>
0xff2b600c <__register_frame_info_bases+16>:
add %l7, 0x1ac, %l7 ! 0x115ac <_start+32>
0xff2b6010 <__register_frame_info_bases+20>: or %o0, 0x84, %o0
0xff2b6014 <__register_frame_info_bases+24>: sethi %hi(0), %o1
0xff2b6018 <__register_frame_info_bases+28>: ld [ %l7 + %o0 ], %l1
0xff2b601c <__register_frame_info_bases+32>: or %o1, 0x88, %o1
0xff2b6020 <__register_frame_info_bases+36>: ld [ %l7 + %o1 ], %l0
0xff2b6024 <__register_frame_info_bases+40>: ld [ %l1 ], %o2
0xff2b6028 <__register_frame_info_bases+44>: mov -1, %o1
0xff2b602c <__register_frame_info_bases+48>: st %o1, [ %i1 ]
0xff2b6030 <__register_frame_info_bases+52>: sethi %hi(0x1fe00000), %o1
0xff2b6034 <__register_frame_info_bases+56>: mov %l0, %o0
0xff2b6038 <__register_frame_info_bases+60>: st %i2, [ %i1 + 4 ]
0xff2b603c <__register_frame_info_bases+64>: st %i3, [ %i1 + 8 ]
0xff2b6040 <__register_frame_info_bases+68>: st %i0, [ %i1 + 0xc ]
0xff2b6044 <__register_frame_info_bases+72>: cmp %o2, 0
0xff2b6048 <__register_frame_info_bases+76>:
be 0xff2b6058 <__register_frame_info_bases+92>
0xff2b604c <__register_frame_info_bases+80>: st %o1, [ %i1 + 0x10 ]
0xff2b6050 <__register_frame_info_bases+84>:
call 0xff2c7818 <_PROCEDURE_LINKAGE_TABLE_+432>
0xff2b6054 <__register_frame_info_bases+88>: nop
0xff2b6058 <__register_frame_info_bases+92>: sethi %hi(0), %o1
0xff2b605c <__register_frame_info_bases+96>: or %o1, 0x8c, %o1 ! 0x8c
0xff2b6060 <__register_frame_info_bases+100>: ld [ %l7 + %o1 ], %o2
0xff2b6064 <__register_frame_info_bases+104>: mov %l0, %o0
0xff2b6068 <__register_frame_info_bases+108>: ld [ %l1 ], %o3
0xff2b606c <__register_frame_info_bases+112>: ld [ %o2 ], %o1
0xff2b6070 <__register_frame_info_bases+116>: cmp %o3, 0
0xff2b6074 <__register_frame_info_bases+120>: st %o1, [ %i1 + 0x14 ]
0xff2b6078 <__register_frame_info_bases+124>:
be 0xff2b6088 <__register_frame_info_bases+140>
0xff2b607c <__register_frame_info_bases+128>: st %i1, [ %o2 ]
0xff2b6080 <__register_frame_info_bases+132>:
call 0xff2c7824 <_PROCEDURE_LINKAGE_TABLE_+444>
0xff2b6084 <__register_frame_info_bases+136>: nop
0xff2b6088 <__register_frame_info_bases+140>: ret
0xff2b608c <__register_frame_info_bases+144>: restore
End of assembler dump.
(gdb) info registers
g0 0x0 0
g1 0xff2b6090 -13934448
g2 0x0 0
g3 0x0 0
g4 0x0 0
g5 0x0 0
g6 0x0 0
g7 0x0 0
o0 0x84 132
o1 0xffffffff -1
o2 0xff215950 -14591664
o3 0x185 389
o4 0xff30b756 -13584554
o5 0xff376eb4 -13144396
sp 0xffbee8a0 4290701472
o7 0xff2b6008 -13934584
l0 0xff2c7980 -13862528
l1 0xff2c7978 -13862536
l2 0x0 0
l3 0x0 0
l4 0x0 0
l5 0x0 0
l6 0x0 0
l7 0xff2c75b4 -13863500
i0 0xff300000 -13631488
i1 0xff300000 -13631488
i2 0x0 0
i3 0x0 0
i4 0x0 0
i5 0x0 0
fp 0xffbee910 4290701584
i7 0xff2b60a0 -13934432
y 0x0 0
psr 0xfe400000 -29360128 icc:-Z--, pil:0, s:0, ps:0, et:0, cwp:0
wim 0x0 0
tbr 0x0 0
pc 0xff2b602c 4281032748
npc 0xff2b6030 -13934544
fpsr 0x0 0 rd:N, tem:0, ns:0, ver:0, ftt:0, qne:0, fcc:=, aexc:0, cexc:0
cpsr 0x0 0
#0 0xff2b602c in __register_frame_info_bases (begin=0xff300000,
ob=0xff300000, tbase=0x0, dbase=0x0) from /tools/kde/gnu/lib/libgcc_s.so.1
#1 0xff2b60a8 in __register_frame_info (begin=0xff300000, ob=0xff300000)
from /tools/kde/gnu/lib/libgcc_s.so.1
#2 0xff325b20 in frame_dummy () from /tools/gcc-3.2/lib/libstdc++.so.5
#3 0xff325a08 in _init () from /tools/gcc-3.2/lib/libstdc++.so.5
#4 0xff3bad04 in ?? ()
#5 0xff3ba990 in ?? ()
#6 0xff3c4900 in ?? ()
#7 0xff3b2940 in ?? ()
The line of code crashing appears to be
ob->pc_begin = (void *)-1;
in unwind-dw2-fde.c
Disassembly of the caller:
Dump of assembler code for function __register_frame_info:
0xff2b6090 <__register_frame_info>: save %sp, -112, %sp
0xff2b6094 <__register_frame_info+4>: mov %i0, %o0
0xff2b6098 <__register_frame_info+8>: mov %i1, %o1
0xff2b609c <__register_frame_info+12>: clr %o2
0xff2b60a0 <__register_frame_info+16>:
call 0xff2c7830 <_PROCEDURE_LINKAGE_TABLE_+456>
0xff2b60a4 <__register_frame_info+20>: clr %o3
0xff2b60a8 <__register_frame_info+24>: ret
0xff2b60ac <__register_frame_info+28>: restore
End of assembler dump.
As far as I can tell, the problem is caused by the following call:
0xff2b6008 <__register_frame_info_bases+12>:
call 0xff2b5e38 <base_of_encoded_value+176>
This seems to overwrite %o1 with 0 since %i1, where it was saved in the caller, looks valid to me.
Any help would be greatly appreciated, as I would like to move from gcc 2.95.3 to 3.2 on Solaris. So far, upgrading beyond 2.95.3 has been a nightmare.
>How-To-Repeat:
Compile C++ program, execute binary
>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted:
next reply other threads:[~2002-09-22 17:56 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-09-22 10:56 aaronw [this message]
2002-10-01 20:59 davem
2002-10-01 21:16 Aaron Williams
2002-10-01 22:00 davem
2002-10-01 22:06 David S. Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20020922175041.5784.qmail@sources.redhat.com \
--to=aaronw@net.com \
--cc=gcc-gnats@gcc.gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).