Bug 12237 - set backtrace past-zero-pc
Summary: set backtrace past-zero-pc
Status: NEW
Alias: None
Product: gdb
Classification: Unclassified
Component: backtrace (show other bugs)
Version: HEAD
: P2 enhancement
Target Milestone: ---
Assignee: Not yet assigned to anyone
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-11-18 23:29 UTC by Jan Kratochvil
Modified: 2010-11-18 23:29 UTC (History)
0 users

See Also:
Host:
Target:
Build:
Last reconfirmed:

Attachments
Draft patch. (1.76 KB, application/octet-stream)
2010-11-18 23:29 UTC, Jan Kratochvil 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Jan Kratochvil 2010-11-18 23:29:58 UTC 
Created attachment 5129 [details]
Draft patch.

Sometimes corrupted stack stops stack unwinding.  User should have a way to request further unwinding so that the backtraces "synces" up again.

Fedora contained:
gdb-6.3-framepczero-20040927.patch
gdb-6.3-bt-past-zero-20051201.patch

but those work only in marginal cases, as shown on the attached testcase.

Still there is a problem:

(gdb) bt
#0  marker () at ./gdb.base/bt-past-zero-pc.c:23
#1  0x0000000000400573 in clear_stack (end=0x0) at ./gdb.base/bt-past-zero-pc.c:38
#2  0x0000000000000000 in ?? ()
#3  0x0000000000000000 in ?? ()
#4  0x0000000000000000 in ?? ()
#5  0x00007ffff7b4c730 in ?? () from /lib64/libc.so.6
#6  0x00007fffffffdd18 in ?? ()
#7  0x00007fffffffdd40 in ?? ()
#8  0x00000000004005a5 in stub () at ./gdb.base/bt-past-zero-pc.c:52
Backtrace stopped: previous frame inner to this frame (corrupt stack?)

Frames #2 upwards already have $rbp == 0.
DWARF frame #8 therefore cannot unwind futher.

There should be a main() frame #9.