[Qemu-devel] [PATCH v2] Use the GDB JIT debugging interface

[Richard Henderson]

I was reminded about this from a comment on the list, where
a new developer had gotten confused by our backtrace.

Changes from V1 to V2:
  * Rebase to master.

  * Do not try to allocate the ELF stuff within the code_gen_buffer.
    Instead we can use a SHT_NOBITS trick to say that's out-of-line.

  * Use C structures (with alignment and packing annotations) to create
    the Dwarf2 unwind info.  Previously I had used inline assembly.


Example:

Breakpoint 1, helper_idivl_EAX (t0=512)
    at /home/rth/work/qemu/qemu/target-i386/op_helper.c:1882
1882    {
(gdb) where
#0  helper_idivl_EAX (t0=512)
    at /home/rth/work/qemu/qemu/target-i386/op_helper.c:1882
#1  0x00005555557c0fdf in static_code_gen_buffer ()
#2  0x000055555559714d in cpu_x86_exec (env=0x5555577fba30)
    at /home/rth/work/qemu/qemu/cpu-exec.c:564
#3  0x00005555555b3290 in cpu_loop (env=0x5555577fba30)
    at /home/rth/work/qemu/qemu/linux-user/main.c:317
#4  0x00005555555b40a4 in main (argc=<optimized out>, argv=<optimized out>, 
    envp=<optimized out>) at /home/rth/work/qemu/qemu/linux-user/main.c:3826


r~


Richard Henderson (1):
  tcg: Use the GDB JIT debugging interface.

 elf.h                 |    1 +
 exec.c                |    1 +
 tcg/i386/tcg-target.c |  114 ++++++++++++++++++++++++++--
 tcg/tcg.c             |  194 +++++++++++++++++++++++++++++++++++++++++++++++++
 tcg/tcg.h             |    2 +
 5 files changed, 303 insertions(+), 9 deletions(-)

-- 
1.7.7.6