[Bug libc/29635] Debian libc6: 2.35-1 breaks gdb on hppa

["danglin at gcc dot gnu.org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=29635

--- Comment #1 from John David Anglin <danglin at gcc dot gnu.org> ---
Updated backtrace:

(gdb) c
Continuing.
warning: Unable to find libthread_db matching inferior's thread library, thread
debugging will not be available.

Breakpoint 10, 0x006b5500 in ?? ()
(gdb) bt
#0  0x006b5500 in ?? ()
#1  0xf5498c00 in boost::basic_regex<char, boost::regex_traits<char,
boost::cpp_regex_traits<char> > >::do_assign (this=0xf710b6c8
<srchilite::datadir_exp>,
    p1=0xf70e376c
"[[:blank:]]*(datadir)[[:blank:]]*=[[:blank:]]*\"([^[:blank:]\\r]+)\"[[:blank:]\\r]*|([[:space:]]+)|([[:space:]]*#.*)",
p2=0xf70e37dc "",
    f=0) at ./boost/regex/v4/basic_regex.hpp:671
#2  0xf704c434 in boost::basic_regex<char, boost::regex_traits<char,
boost::cpp_regex_traits<char> > >::assign (f=0, p2=0xf70e37dc "",
    p1=0xf70e376c
"[[:blank:]]*(datadir)[[:blank:]]*=[[:blank:]]*\"([^[:blank:]\\r]+)\"[[:blank:]\\r]*|([[:space:]]+)|([[:space:]]*#.*)",
    this=0xf710b6c8 <srchilite::datadir_exp>)
    at /usr/include/boost/regex/v4/basic_regex.hpp:394
#3  boost::basic_regex<char, boost::regex_traits<char,
boost::cpp_regex_traits<char> > >::assign (f=0,
    p=0xf70e376c
"[[:blank:]]*(datadir)[[:blank:]]*=[[:blank:]]*\"([^[:blank:]\\r]+)\"[[:blank:]\\r]*|([[:space:]]+)|([[:space:]]*#.*)",
    this=0xf710b6c8 <srchilite::datadir_exp>)
    at /usr/include/boost/regex/v4/basic_regex.hpp:383
#4  boost::basic_regex<char, boost::regex_traits<char,
boost::cpp_regex_traits<char> > >::basic_regex (f=0,
    p=0xf70e376c
"[[:blank:]]*(datadir)[[:blank:]]*=[[:blank:]]*\"([^[:blank:]\\r]+)\"[[:blank:]\\r]*|([[:space:]]+)|([[:space:]]*#.*)",
    this=0xf710b6c8 <srchilite::datadir_exp>)
    at /usr/include/boost/regex/v4/basic_regex.hpp:352
#5  __static_initialization_and_destruction_0 (__initialize_p=1,
    __priority=65535) at settings.cpp:52
#6  _GLOBAL__sub_I_settings.cpp(void) () at settings.cpp:200
#7  0xf96f5e50 in call_init (env=0xf70e37dc, argv=0xf70e376c, argc=0,
    l=<optimized out>) at dl-init.c:70
#8  call_init (l=<optimized out>, argc=0, argv=0xf70e376c, env=0xf70e37dc)
    at dl-init.c:26
#9  0xf96f5f88 in _dl_init (main_map=0xf710b4c8, argc=-150063140, argv=0x0,
    env=0x0) at dl-init.c:117
#10 0xf970bf9c in _dl_start_user () from /lib/ld.so.1
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

The boost initialization has called _Znwj@@Base in gdb.

006b5500 <_Znwj@@Base>:
  6b5500:       6b c2 3f d9     stw rp,-14(sp)
  6b5504:       37 de 01 00     ldo 80(sp),sp
  6b5508:       6b c6 3f 19     stw r6,-74(sp)
  6b550c:       08 13 02 46     copy r19,r6
  6b5510:       6b c3 3f 31     stw r3,-68(sp)
  6b5514:       34 03 00 02     ldi 1,r3
  6b5518:       6b d3 3f c1     stw r19,-20(sp)
  6b551c:       08 7a 88 80     cmpclr,<< r26,r3,r0
  6b5520:       08 1a 02 43     copy r26,r3
  6b5524:       6b c5 3f 21     stw r5,-70(sp)
  6b5528:       6b c4 3f 29     stw r4,-6c(sp)
  6b552c:       e8 43 0d 98     b,l 6bbc00 <_ZdaPvj@@Base+0x64a4>,rp
  6b5530:       08 03 02 5a     copy r3,r26
  6b5534:       c8 bc 20 38     movb,= ret0,r5,6b5558 <_Znwj@@Base+0x58>
  6b5538:       08 06 02 53     copy r6,r19
  6b553c:       4b c2 3e d9     ldw -94(sp),rp
  6b5540:       4b c6 3f 19     ldw -74(sp),r6
  6b5544:       4b c5 3f 21     ldw -70(sp),r5
  6b5548:       4b c4 3f 29     ldw -6c(sp),r4
  6b554c:       4b c3 3f 31     ldw -68(sp),r3
  6b5550:       e8 40 c0 00     bv r0(rp)
  6b5554:       37 de 3f 01     ldo -80(sp),sp

The long branch stub for the call at 6b552c is:

Dump of assembler code from 0x6bbc00 to 0x6bbc18:
=> 0x006bbc00:  addil L%15800,dp,r1
   0x006bbc04:  ldo 4d0(r1),r22
   0x006bbc08:  ldw 0(r22),r21
   0x006bbc0c:  bv r0(r21)
   0x006bbc10:  ldw 4(r22),r19
   0x006bbc14:  addil L%15000,dp,r1
End of assembler dump.

At 0x006bbc0c, we have:

(gdb) p/x $r21
$17 = 0x81c0259
(gdb) x/2x $r22
0x146dd8:       0x081c0259      0xe85b1779

It seems elf_machine_runtime_setup() has not yet been called for the main
executable, so register dp and likely the PLT is not yet setup.

-- 
You are receiving this mail because:
You are on the CC list for the bug.