[Bug c++/15689] New: can't compute CFA for this frame

[Bug c++/15689] New: can't compute CFA for this frame

[sebastian.huber@embedded-brains.de]

http://sourceware.org/bugzilla/show_bug.cgi?id=15689

            Bug ID: 15689
           Summary: can't compute CFA for this frame
           Product: gdb
           Version: 7.6
            Status: NEW
          Severity: normal
          Priority: P2
         Component: c++
          Assignee: unassigned at sourceware dot org
          Reporter: sebastian.huber@embedded-brains.de

Created attachment 7097
  --> http://sourceware.org/bugzilla/attachment.cgi?id=7097&action=edit
ELF file

I use powerpc-rtems4.11-gcc (GCC) 4.8.2 20130627 (prerelease) (a similar
problem occurs also on SPARC and ARM) with -O2 -g.

(gdb) b Task1::screen1() 
Breakpoint 4 at 0x8f8: file
/home/sh/rtems-4.11/c/src/../../testsuites/libtests/rtems++/Task1.cc, line 93.
(gdb) r
Breakpoint 4, Task1::screen1 (this=this@entry=0x480c0 <task_1>) at
/home/sh/rtems-4.11/c/src/../../testsuites/libtests/rtems++/Task1.cc:93
93      {
(gdb) n
95        rtemsTask local_task_1 = *this;
(gdb) n
95        rtemsTask local_task_1 = *this;
(gdb) p local_task_1
can't compute CFA for this frame
(gdb)

In case I use -gdwarf-2, then I get "Could not find the frame base for".

You can run the example with the built in GDB PSIM simulator:

tar sim -f psim_tree.sh

psim_tree.sh:

#
#  Device Tree for PSIM
#
#  Automatically Generated -- DO NOT EDIT!!
#
/#address-cells 1
/openprom/init/register/pvr 0xfffe0000
/openprom/options/oea-memory-size 16777216
##### EEPROM @ 0x0c000000 for 512K
/eeprom@0x0c000000/reg 0x0c000000 0x80000
/eeprom@0x0c000000/nr-sectors 8
/eeprom@0x0c000000/sector-size 0x10000
/eeprom@0x0c000000/byte-write-delay 1000
/eeprom@0x0c000000/sector-start-delay 100
/eeprom@0x0c000000/erase-delay 1000
/eeprom@0x0c000000/manufacture-code 0x01
/eeprom@0x0c000000/device-code 0xa4

##### NVRAM/RTC NVRAM Portion is 0x0c080000 for 512K
##### NVRAM/RTC RTC   Portion is 0x0c100000 for 12
/nvram@0x0c080000/reg 0x0c080000 524300
/nvram@0x0c080000/timezone -3600

##### OPENPIC @ 0x0c130000 - 0x0c170000 (512K)
/opic@0x0c130000/reg              0x0c130000 0 0x0c130000 0x40000
/opic@0x0c130000/interrupt-ranges 0 0 0 16
/opic@0x0c130000/device_type      open-pic
## interupt out -> CPU's interrupt pin
/opic@0x0c130000 > intr0 int /cpus/cpu@0

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

[Bug c++/15689] can't compute CFA for this frame

[sebastian.huber@embedded-brains.de]

http://sourceware.org/bugzilla/show_bug.cgi?id=15689

Sebastian Huber <sebastian.huber@embedded-brains.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |INVALID

--- Comment #1 from Sebastian Huber <sebastian.huber@embedded-brains.de> ---
It seems that this is due to my linker command file.  I had to aggregate
several output sections due to a recent GNU LD change.  I reverted this and
tried this LD patch

http://sourceware.org/ml/binutils/2013-06/msg00246.html

and now the debugging works as usual.

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