Attachment 4764 Details for Bug 10746 – Annotated debugging log based on Oyvind's executable

Annotated debugging log based on Oyvind's executable

gdb_log.txt (text/plain), 20.38 KB, created by Hans Boehm on 2003-09-15 23:59:27 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Hans Boehm

Created: 2003-09-15 23:59:27 UTC

Size: 20.38 KB

patch

obsolete

>first thread 3469000 fgfgff
>second thread 2076000 fbfd
>first thread 3470000 cccgecab
>second thread 2077000 fdcdd
>
>Program received signal SIGSEGV, Segmentation fault.
>0x00410a99 in java::lang::Object::getClass() ()
>(gdb) where
>#0  0x00410a99 in java::lang::Object::getClass() ()
>#1  0x0040145e in Test.replaceAll(java.lang.String, java.lang.String, java.lang.
>String) (in=0x1062000, a=0x1025fd8, b=0xb5efd8) at Test.java:12
>#2  0x00401aec in Test.stressGB(java.lang.String) (thread=0xb5b930)
>    at Test.java:64
>#3  0x0040173d in Test.main(java.lang.String[]) (args=0xb56fe0) at Test.java:40
>#4  0x004147d2 in gnu::gcj::runtime::FirstThread::run() ()
>#5  0x0041de67 in _Jv_ThreadRun(java::lang::Thread*) ()
>#6  0x0040381f in _Jv_RunMain(java::lang::Class*, char const*, int, char const**
>, bool) ()
>#7  0x004038ef in JvRunMain ()
>#8  0x004012e2 in main (argc=1, argv=0x3f24c8) at c:/temp/cc2Fbaaa.i:11
>(gdb) x/20i 0x00410a80
>...
>0x410a90 <_ZN4java4lang6Object8getClassEv>:     push   %ebp
>0x410a91 <_ZN4java4lang6Object8getClassEv+1>:   mov    %esp,%ebp
>0x410a93 <_ZN4java4lang6Object8getClassEv+3>:   mov    0x8(%ebp),%eax
>0x410a96 <_ZN4java4lang6Object8getClassEv+6>:   pop    %ebp
>0x410a97 <_ZN4java4lang6Object8getClassEv+7>:   mov    (%eax),%eax
>0x410a99 <_ZN4java4lang6Object8getClassEv+9>:   mov    (%eax),%eax
>0x410a9b <_ZN4java4lang6Object8getClassEv+11>:  ret
>0x410a9c <_ZN4java4lang6Object8getClassEv+12>:  lea    0x0(%esi,1),%esi
>0x410aa0 <_ZN4java4lang6Object8hashCodeEv>:     push   %ebp
>0x410aa1 <_ZN4java4lang6Object8hashCodeEv+1>:   mov    $0x3,%ecx
>0x410aa6 <_ZN4java4lang6Object8hashCodeEv+6>:   mov    %esp,%ebp
>0x410aa8 <_ZN4java4lang6Object8hashCodeEv+8>:   sub    $0xc,%esp
>0x410aab <_ZN4java4lang6Object8hashCodeEv+11>:  mov    %ebx,(%esp,1)
>(gdb) p/x $ebp
>$1 = 0x22fd08
>[The following is the wrong address for the argument.  But it's an example of a normal string rep]
>(gdb) x/wx 8+$ebp
>0x22fd10:       0x01024f90
>(gdb) x/8wx 0x01024f90
>0x1024f90:      0x005a6288      0x01024f90      0x00000014      0x00000000
>0x1024fa0:      0x00000000      0x00000065      0x005a6288      0x01024fa8
>(gdb) x/8wx 0x005a6288
>0x5a6288 <_ZTVN4java4lang6StringE+8>:   0x005a6300      0x00000008      0x00405650      0x00409d40
>0x5a6298 <_ZTVN4java4lang6StringE+24>:  0x0040ae20      0x0040c9f0      0x00410af0      0x005a8e50
>[And now back to more interesting things:]
>(gdb) p/x $eax
>$2 = 0x0
>(gdb) p GC_gc_no
>$3 = 15039
>(gdb) x/4wx $esp
>0x22fcec:       0x0040ae49      0x01025fd8      0x01021ff0      0x00000000
>(gdb) x/8wx  0x01025fd8
>0x1025fd8:      0x00000000      0x00000000      0x00000000      0x00000000
>0x1025fe8:      0x00000000      0x00000000      0x00000000      0x00000000
>(gdb) p GC_find_header(0x01025fd8)
>
>Program received signal SIGSEGV, Segmentation fault.
>0x00410a99 in java::lang::Object::getClass() () ...
>
>
>[New crash, same addresses!:]
>second thread 163000 cbcbgc
>first thread 331000 ddggffca
>second thread 164000 cegcc
>first thread 332000 fggadac
>second thread 165000 ggdgdfcce
>
>Program received signal SIGSEGV, Segmentation fault.
>0x00410a99 in java::lang::Object::getClass() ()
>(gdb) where 5
>#0  0x00410a99 in java::lang::Object::getClass() ()
>#1  0x0040145e in Test.replaceAll(java.lang.String, java.lang.String, java.lang.
>String) (in=0x106a000, a=0x1025fd8, b=0xb5efd8) at Test.java:12
>#2  0x00401aec in Test.stressGB(java.lang.String) (thread=0xb5b930)
>    at Test.java:64
>#3  0x0040173d in Test.main(java.lang.String[]) (args=0xb56fe0) at Test.java:40
>#4  0x004147d2 in gnu::gcj::runtime::FirstThread::run() ()
>(More stack frames follow...)
>(gdb) p/x $esp
>$1 = 0x22fcec
>(gdb) x/4wx $esp
>0x22fcec:       0x0040ae49      0x01025fd8      0x77f8f2eb      0x00000001
>(gdb) x/wx $esp-4
>0x22fce8:       0x0022fd08
>(gdb) x/20wx  0x01025fb0
>0x1025fb0:      0x00000000      0x00000000      0x00000000      0x00000000
>0x1025fc0:      0x00000000      0x00000000      0x005f92e0      0x00000010
>0x1025fd0:      0x00630062      0x00000000      0x00000000      0x0106ea00
>0x1025fe0:      0x00000008      0x00000001      0x00000000      0x00000000
>0x1025ff0:      0x00000000      0x00000000      0x00000000      0x00000000
>(gdb) x/8wx 0x01025fd8
>0x1025fd8:      0x00000000      0x0106ea00      0x00000008      0x00000001
>0x1025fe8:      0x00000000      0x00000000      0x00000000      0x00000000
>(gdb) p GC_gc_no
>$2 = 1212
>(gdb) p/x $eax
>$3 = 0x0
>(gdb) p &GC_arrays
>$4 = (<data variable, no debug info> *) 0x606a30
>(gdb) x/100wx $4 [with annotations]
>0x606a30 <GC_arrays>:   0x0009d000[heapsize]      0x00000000      0x00000000      0x01050000[last_heap_addr]
>0x606a40 <GC_arrays+16>:        0x01030000[prev_heap_addr]  	0x00063000[large_free]     0x00003000[large_allocd] 0x00003000
>0x606a50 <GC_arrays+32>:        0x086dc12d[allocd_before]      0x0000000c[words_allocd]      0x00000000[wasted] 0x00000000[finalized]
>0x606a60 <GC_arrays+48>:        0x00000000      0x00000000      0x00000000 0x01000000[scratch_end]
>0x606a70 <GC_arrays+64>:        0x01000000      0x0041be80      0x00000000 0x00000000
>0x606a80 <GC_arrays+80>:        0x00000000      0x00000000      0x00000000 0x00000000
>0x606a90 <GC_arrays+96>:        0x00000000      0x0041c790      0x00000000 0x00000000
>0x606aa0 <GC_arrays+112>:       0x00000000      0x00000000      0x00000000 0x00000000
>0x606ab0 <GC_arrays+128>:       0x00000000      0x00000000      0x00000000 0x00000000
>0x606ac0 <GC_arrays+144>:       0x00000000      0x00000000      0x00000000 0x00000000
>0x606ad0 <GC_arrays+160>:       0x00000000      0x00000000      0x00000000 0x00000000
>0x606ae0 <GC_arrays+176>:       0x00000000      0x00000000      0x00000000 0x00000000
>---Type <return> to continue, or q <return> to quit---q
>Quit
>[Looks like it occurred just after (12 bytes after!) GC]
>(gdb) x/100wx 0x1025000
>0x1025000:      0x005f92e0      0x00000010      0x00000062      0x00000000
>0x1025010:      0x00000000      0x00000000      0x00000000      0x00000000
>0x1025020:      0x00000000      0x00000000      0x005f92e0      0x00000010
>0x1025030:      0x00610061      0x00630061      0x00650061      0x00000061
>0x1025040:      0x00000000      0x00000000      0x00000000      0x00000000
>0x1025050:      0x005f92e0      0x00000010      0x00610061      0x00630061
>0x1025060:      0x00650061      0x00000000      0x00000000      0x00000000
>0x1025070:      0x00000000      0x00000000      0x005f92e0      0x00000010
>0x1025080:      0x00610061      0x00630061      0x00000061      0x00000000
>0x1025090:      0x00000000      0x00000000      0x00000000      0x00000000
>0x10250a0:      0x005f92e0      0x00000010      0x00610061      0x00630061
>0x10250b0:      0x00000000      0x00000000      0x00000000      0x00000000
>0x10250c0:      0x00000000      0x00000000      0x005f92e0      0x00000010
>0x10250d0:      0x00610061      0x00000061      0x00000000      0x00000000
>[Definitely looks like 10 word objects.  Unfortunately the debugger doesn't work well enough to get
>at block header.]
>(gdb) p 0xfd8/40
>$7 = 101
>(gdb) p/x $7*40
>$8 = 0xfc8
>(gdb)
>[0x1025fd8 points to 8 bytes past the start of an object!  Where did this pointer come from? ]
>
>
>[Third crash:]
>first thread 4109000 abeg
>second thread 2876000 gccbag
>first thread 4110000 ccggaeee
>first thread 4111000 ffebabfd
>second thread 2877000 dddddecbbde
>
>Program received signal SIGSEGV, Segmentation fault.
>0x00410a99 in java::lang::Object::getClass() ()
>(gdb) p GC_gc_no
>$1 = 17032
>(gdb) where 8
>#0  0x00410a99 in java::lang::Object::getClass() ()
>#1  0x0040145e in Test.replaceAll(java.lang.String, java.lang.String, java.lang.
>String) (in=0x1020000, a=0x1014fd8, b=0xb5efd8) at Test.java:12
>#2  0x00401aec in Test.stressGB(java.lang.String) (thread=0xb5b930)
>    at Test.java:64
>#3  0x0040173d in Test.main(java.lang.String[]) (args=0xb56fe0) at Test.java:40
>#4  0x004147d2 in gnu::gcj::runtime::FirstThread::run() ()
>#5  0x0041de67 in _Jv_ThreadRun(java::lang::Thread*) ()
>#6  0x0040381f in _Jv_RunMain(java::lang::Class*, char const*, int, char const**
>, bool) ()
>#7  0x004038ef in JvRunMain ()
>(More stack frames follow...)
>(gdb) x/4wx $esp
>0x22fcec:       0x0040ae49      0x01014fd8      0x77f8f2eb      0x00000001
>(gdb) x/i 0x0040ae49
>0x40ae49 <_ZN4java4lang6String6equalsEPNS0_6ObjectE+41>:
>    xor    %edx,%edx
>(gdb) info line *0x0040ae49
>No line number information available for address
>  0x40ae49 <_ZN4java4lang6String6equalsEPNS0_6ObjectE+41>
>(gdb) x/32wx 0x01014000
>0x1014000:      0x005f92e0      0x00000010      0x00000067      0x00000000
>0x1014010:      0x00000000      0x00000000      0x00000000      0x00000000
>0x1014020:      0x00000000      0x00000000      0x005f92e0      0x00000010
>0x1014030:      0x00650062      0x00670061      0x00610066      0x00000064
>0x1014040:      0x00000000      0x00000000      0x00000000      0x00000000
>0x1014050:      0x005f92e0      0x00000010      0x00650062      0x00670061
>0x1014060:      0x00610066      0x00000000      0x00000000      0x00000000
>0x1014070:      0x00000000      0x00000000      0x005f92e0      0x00000010
>(gdb) p 0xfd8/40
>$2 = 101
>(gdb) p 0x1014000+$2*40
>$3 = 16863176
>(gdb) x/10wx $3
>0x1014fc8:      0x005f92e0      0x00000010      0x00670061      0x00000000
>0x1014fd8:      0x00000000      0x01024a28      0x00000008      0x00000001
>0x1014fe8:      0x00000000      0x00000000
>[Again failed on "object" with trailing address bits fd8, though this time
>in a different page.  The actual 10 byte object start appears to be at
>fc8, and it looks like a valid character object.  It looks like the call
>stack is missing a frame, since %ebp has already been popped, and gdb
>can't compensate without debug info.]
>(gdb) x/16wx &GC_arrays
>0x606a30 <GC_arrays>:   0x0009d000[heap_size]      0x00000000      0x00000000      0x01050000[last_heap_addr]
>0x606a40 <GC_arrays+16>:        0x01030000      0x0001f000      0x00003000  0x00003000
>0x606a50 <GC_arrays+32>:        0x765f948b[allocd_before]      0x00014a52[words_allocd]      0x00000000  0x00000000
>0x606a60 <GC_arrays+48>:        0x00000000      0x00000000      0x00000000  0x01000000
>(gdb) up
>#1  0x0040145e in Test.replaceAll(java.lang.String, java.lang.String, java.lang.
>String) (in=@1020000, a=@1014fd8, b=@b5efd8) at Test.java:12
>12                              if (c.equals(a))
>[Note that the last two string addresses end in fd8.]
>(gdb) x/16wx 0xb5efd8
>0xb5efd8:       0x005a6288      0x00b5efd8      0x00000014      0x00000000
>0xb5efe8:       0x00000000      0x00000000      0x00000000      0x00000000
>0xb5eff8:       0x00000000      0x00000000      0x00000000      0x00000022
>0xb5f008:       0x00610061      0x00610061      0x00610061      0x00610061
>(gdb) x/32wx 0xb5e000
>0xb5e000:       0x005a6288      0x00b5e000      0x00000014      0x00000001
>0xb5e010:       0x00000000      0x00000062      0x005a6288      0x00b5e018
>0xb5e020:       0x00000014      0x00000001      0x00000000      0x00000062
>0xb5e030:       0x005a6288      0x00b5e030      0x00000014      0x00000001
>0xb5e040:       0x00000000      0x00000065      0x005a6288      0x00b5e048
>0xb5e050:       0x00000014      0x00000001      0x00000000      0x00000065
>0xb5e060:       0x005a6288      0x00b5e060      0x00000014      0x00000001
>0xb5e070:       0x00000000      0x00000062      0x005a6288      0x00b5e078
>(gdb) p 0xfd8/24
>$6 = 169
>(gdb) p 169*24
>$7 = 4056
>(gdb) p/x $7
>$8 = 0xfd8
>[It looks like b is a valid string, the last one in a block of objects 6 words in size.
>Clearly a is not.  But perhaps it was.  Could we be reclaiming entire pages with
>the last object still in use?]
>(gdb) x/30i GC_block_empty
>0x471030 <GC_block_empty>:      push   %ebp
>0x471031 <GC_block_empty+1>:    mov    %esp,%ebp
>0x471033 <GC_block_empty+3>:    mov    0x8(%ebp),%eax
>0x471036 <GC_block_empty+6>:    lea    0x18(%eax),%edx
>0x471039 <GC_block_empty+9>:    lea    0x98(%eax),%ecx
>0x47103f <GC_block_empty+15>:   jmp    0x47104a <GC_block_empty+26>
>0x471041 <GC_block_empty+17>:   mov    (%edx),%eax
>0x471043 <GC_block_empty+19>:   add    $0x4,%edx
>0x471046 <GC_block_empty+22>:   test   %eax,%eax
>0x471048 <GC_block_empty+24>:   jne    0x471055 <GC_block_empty+37>
>0x47104a <GC_block_empty+26>:   cmp    %ecx,%edx
>0x47104c <GC_block_empty+28>:   jb     0x471041 <GC_block_empty+17>
>0x47104e <GC_block_empty+30>:   pop    %ebp
>0x47104f <GC_block_empty+31>:   mov    $0x1,%eax
>0x471054 <GC_block_empty+36>:   ret
>0x471055 <GC_block_empty+37>:   xor    %eax,%eax
>0x471057 <GC_block_empty+39>:   pop    %ebp
>0x471058 <GC_block_empty+40>:   ret
>0x471059 <GC_block_empty+41>:   lea    0x0(%esi,1),%esi
>...
>[Looks correct.]
>(gdb) x/30i GC_find_header
>0x473240 <GC_find_header>:      push   %ebp
>0x473241 <GC_find_header+1>:    mov    %esp,%ebp
>0x473243 <GC_find_header+3>:    mov    0x8(%ebp),%eax
>0x473246 <GC_find_header+6>:    pop    %ebp
>0x473247 <GC_find_header+7>:    mov    %eax,%edx
>0x473249 <GC_find_header+9>:    shr    $0x16,%edx
>0x47324c <GC_find_header+12>:   mov    0x6145a4(,%edx,4),%edx  [Address of top_index]
>0x473253 <GC_find_header+19>:   shr    $0xc,%eax
>0x473256 <GC_find_header+22>:   and    $0x3ff,%eax
>0x47325b <GC_find_header+27>:   mov    (%edx,%eax,4),%eax
>0x47325e <GC_find_header+30>:   ret
>0x47325f <GC_find_header+31>:   nop
>[And now to do this manually for the block including 0x1014fd8:
>(gdb) p/x 0x1014fd8>>22
>$9 = 0x4		[Index in top_index]
>(gdb) p/x 0x6145a4+16
>$10 = 0x6145b4
>(gdb) x/wx $10
>0x6145b4 <GC_arrays+56196>:     0x00ff1a28 [Pointer to bottom_index]
>(gdb) p/x 0x00ff1a28+(4*0x14)
>$11 = 0xff1a78			[Address of pointer to header]
>(gdb) x/wx $11
>0xff1a78:       0x00ff4d40	[Pointer to header]
>(gdb) x/38wx  0x00ff4d40	[Header]
>0xff4d40:       0x0000000a[size]      0x00000000      0x00000000      0x00000000[descr]
>0xff4d50:       0x00b3bc40[map]      0x4288[last_reclaimed]0000[kind,flags]      0x00000000      0x00000000
>0xff4d60:       0x00000000      0x00000000      0x00000000      0x00000000
>0xff4d70:       0x00000000      0x00000000      0x00000000      0x00000000
>0xff4d80:       0x00000000      0x00000000      0x00000000      0x00000000
>0xff4d90:       0x00000000      0x00000000      0x00000000      0x00000000
>0xff4da0:       0x00000000      0x00000000      0x00000000      0x00000000
>0xff4db0:       0x00000000      0x00000000      0x00000000      0x00000000
>0xff4dc0:       0x00000000      0x00000000      0x00000000      0x00000000
>0xff4dd0:       0x00000000      0x00000000
>[A block of completely unmareked pointerfree objects of size 10, last reclaimed, and presumably
>determined to be empty this cycle.]
>[Could also  be caused by failure to mark just allocated object?]
>(gdb) x/100i GC_malloc_atomic
>0x432e90 <GC_malloc_atomic>:    push   %ebp
>0x432e91 <GC_malloc_atomic+1>:  mov    $0x800,%eax
>0x432e96 <GC_malloc_atomic+6>:  mov    %esp,%ebp
>0x432e98 <GC_malloc_atomic+8>:  sub    $0x18,%esp
>0x432e9b <GC_malloc_atomic+11>: mov    %ebx,0xfffffff4(%ebp)
>0x432e9e <GC_malloc_atomic+14>: mov    0x8(%ebp),%ebx
>0x432ea1 <GC_malloc_atomic+17>: mov    %esi,0xfffffff8(%ebp)
>0x432ea4 <GC_malloc_atomic+20>: mov    %edi,0xfffffffc(%ebp)
>0x432ea7 <GC_malloc_atomic+23>: mov    0x5b32f0,%edx		[GC_all_interior_pointers]
>0x432ead <GC_malloc_atomic+29>: sub    %edx,%eax
>0x432eaf <GC_malloc_atomic+31>: cmp    %eax,%ebx
>0x432eb1 <GC_malloc_atomic+33>: jb     0x432ed7 <GC_malloc_atomic+71>
>0x432eb3 <GC_malloc_atomic+35>: mov    %ebx,(%esp,1)
>0x432eb6 <GC_malloc_atomic+38>: movl   $0x0,0x4(%esp,1)
>0x432ebe <GC_malloc_atomic+46>: call   0x432d10 <GC_generic_malloc>
>0x432ec3 <GC_malloc_atomic+51>: mov    %eax,0x8(%ebp)
>0x432ec6 <GC_malloc_atomic+54>: mov    0xfffffff4(%ebp),%ebx
>0x432ec9 <GC_malloc_atomic+57>: mov    0xfffffff8(%ebp),%esi
>0x432ecc <GC_malloc_atomic+60>: mov    0xfffffffc(%ebp),%edi
>0x432ecf <GC_malloc_atomic+63>: mov    %ebp,%esp
>0x432ed1 <GC_malloc_atomic+65>: pop    %ebp
>0x432ed2 <GC_malloc_atomic+66>: jmp    0x464820 <GC_clear_stack>
>0x432ed7 <GC_malloc_atomic+71>: mov    0x608b84(,%ebx,4),%eax    [GC_size_map entry]
>0x432ede <GC_malloc_atomic+78>: mov    %eax,%edi
>0x432ee0 <GC_malloc_atomic+80>: shl    $0x2,%edi
>0x432ee3 <GC_malloc_atomic+83>: mov    %eax,0xfffffff0(%ebp)
>0x432ee6 <GC_malloc_atomic+86>: movl   $0x606a10,(%esp,1)
>0x432eed <GC_malloc_atomic+93>: call   0x57ee10 <EnterCriticalSection@4>
>0x432ef2 <GC_malloc_atomic+98>: mov    0x607378(%edi),%esi	[Load from free list]
>0x432ef8 <GC_malloc_atomic+104>:        sub    $0x4,%esp
>0x432efb <GC_malloc_atomic+107>:        test   %esi,%esi
>0x432efd <GC_malloc_atomic+109>:        je     0x432f30 <GC_malloc_atomic+160>
>0x432eff <GC_malloc_atomic+111>:        mov    (%esi),%eax
>0x432f01 <GC_malloc_atomic+113>:        mov    %eax,0x607378(%edi)
>0x432f07 <GC_malloc_atomic+119>:        mov    0xfffffff0(%ebp),%eax
>0x432f0a <GC_malloc_atomic+122>:        add    %eax,0x606a54
>0x432f10 <GC_malloc_atomic+128>:        movl   $0x606a10,(%esp,1)
>0x432f17 <GC_malloc_atomic+135>:
>    call   0x57ee40 <LeaveCriticalSection@4>
>0x432f1c <GC_malloc_atomic+140>:        sub    $0x4,%esp
>0x432f1f <GC_malloc_atomic+143>:        mov    %esi,%eax
>0x432f21 <GC_malloc_atomic+145>:        mov    0xfffffff4(%ebp),%ebx
>0x432f24 <GC_malloc_atomic+148>:        mov    0xfffffff8(%ebp),%esi
>0x432f27 <GC_malloc_atomic+151>:        mov    0xfffffffc(%ebp),%edi
>0x432f2a <GC_malloc_atomic+154>:        mov    %ebp,%esp
>0x432f2c <GC_malloc_atomic+156>:        pop    %ebp
>0x432f2d <GC_malloc_atomic+157>:        ret
>0x432f2e <GC_malloc_atomic+158>:        mov    %esi,%esi
>0x432f30 <GC_malloc_atomic+160>:        movl   $0x606a10,(%esp,1)
>0x432f37 <GC_malloc_atomic+167>:
>    call   0x57ee40 <LeaveCriticalSection@4>
>0x432f3c <GC_malloc_atomic+172>:        sub    $0x4,%esp
>0x432f3f <GC_malloc_atomic+175>:        mov    %ebx,(%esp,1)
>0x432f42 <GC_malloc_atomic+178>:        movl   $0x0,0x4(%esp,1)
>0x432f4a <GC_malloc_atomic+186>:        call   0x432d10 <GC_generic_malloc>
>0x432f4f <GC_malloc_atomic+191>:        mov    %eax,0x8(%ebp)
>0x432f52 <GC_malloc_atomic+194>:        mov    0xfffffff4(%ebp),%ebx
>0x432f55 <GC_malloc_atomic+197>:        mov    0xfffffff8(%ebp),%esi
>0x432f58 <GC_malloc_atomic+200>:        mov    0xfffffffc(%ebp),%edi
>0x432f5b <GC_malloc_atomic+203>:        mov    %ebp,%esp
>0x432f5d <GC_malloc_atomic+205>:        pop    %ebp
>0x432f5e <GC_malloc_atomic+206>:        jmp    0x464820 <GC_clear_stack>
>0x432f63 <GC_malloc_atomic+211>:        lea    0x0(%esi),%esi
>0x432f69 <GC_malloc_atomic+217>:        lea    0x0(%edi,1),%edi
>...
>[Atomic object free list:]
>(gdb) x/20wx 0x607378
>0x607378 <GC_arrays+2376>:      0x00000000      0x00000000      0x00000000	0x00000000
>0x607388 <GC_arrays+2392>:      0x00000000      0x00000000      0x00fdc2a0	0x00000000
>0x607398 <GC_arrays+2408>:      0x00000000      0x00000000      0x00fdd5c8	0x00000000
>0x6073a8 <GC_arrays+2424>:      0x00000000      0x00000000      0x00000000	0x00000000
>0x6073b8 <GC_arrays+2440>:      0x00000000      0x00000000      0x00000000	0x00000000
>(gdb) x/4wx 0x00fdc2a0
>0xfdc2a0:       0x00fdc288      0x00fde988      0x00000008      0x00000004
>(gdb) x/4wx 0x00fdd5c8
>0xfdd5c8:       0x00fdd5a0      0x00000000      0x005a4cd8      0x00000004
>(gdb) x/4wx 0x00fdc000
>0xfdc000:       0x00000000      0x00fde528      0x00000008      0x00000003
>(gdb) x/4wx 0x00fdd000
>0xfdd000:       0x00000000      0x00000000      0x01032618      0x00000000
>[Only size 6 and 10 entries are non-null.  They look as expected.]
>(gdb) x/20i GC_push_regs
>No symbol "GC_push_regs" in current context.
>(gdb) x/50i GC_generic_push_regs
>0x501690 <GC_generic_push_regs>:        push   %ebp
>0x501691 <GC_generic_push_regs+1>:      mov    %esp,%ebp
>0x501693 <GC_generic_push_regs+3>:      sub    $0xc,%esp
>0x501696 <GC_generic_push_regs+6>:      mov    %ebx,(%esp,1)
>0x501699 <GC_generic_push_regs+9>:      mov    (%esp,1),%ebx
>0x50169c <GC_generic_push_regs+12>:     mov    %esi,0x4(%esp,1)
>0x5016a0 <GC_generic_push_regs+16>:     mov    0x4(%esp,1),%esi
>0x5016a4 <GC_generic_push_regs+20>:     mov    %edi,0x8(%esp,1)
>0x5016a8 <GC_generic_push_regs+24>:     mov    0x8(%esp,1),%edi
>0x5016ac <GC_generic_push_regs+28>:     mov    %ebp,%esp
>0x5016ae <GC_generic_push_regs+30>:     pop    %ebp
>0x5016af <GC_generic_push_regs+31>:     jmp    0x4b9e50 <GC_push_current_stack>
>0x5016b4 <GC_generic_push_regs+36>:     nop
>0x5016b5 <GC_generic_push_regs+37>:     nop
>0x5016b6 <GC_generic_push_regs+38>:     nop
>0x5016b7 <GC_generic_push_regs+39>:     nop
>0x5016b8 <GC_generic_push_regs+40>:     nop
>...
>[This looks very fishy.  The registers are pushed on the stack, but they are popped
>off again before the tail call.  Oops.]

Actions: View

Attachments on bug 10746: 4528 | 4716 | 4732 | 4733 | 4764 | 4765 | 4783 | 4888