Re: [PATCH v2] gdb: Modify until_break_command to act correctly on SIGTRAMP_FRAME

[Tomas Vanek <vanekt@fbl.cz>]

On 10/01/2023 16:31, Simon Marchi wrote:
>
> On 10/21/22 07:58, Tomas Vanek wrote:
>> This patch partially depends on
>> gdb/arm: Terminate frame unwinding in M-profile lockup state
>> (without it lockup state is unwound as if it were a normal
>> stack frame).
>>
>> The commands 'advance' and 'until' try to set a breakpoint
>> on the bogus return address derived from Arm M-profile magic
>> address (actually EXC_RETURN or a PC value indicating lockup).
>>
>> The offending breakpoint should be set at the return address in
>> the caller. The magic value 0xffffffff in LR indicates
>> there is no caller (return to this address would lock up the CPU).
>>
>> Similar behaviour of 'advance' and 'until' is observed in
>> an exception handler routine. In this case LR contains e.g.
>> 0xfffffff1 (EXC_RETURN) and GDB tries to set a breakpoint at
>> 0xfffffff0. It should use a return value stacked by the exception
>> instead.
>>
>> Testbench setup:
>> STM32G474, a Cortex-M4 device. Any Cortex-M device can be used.
>> A test application (an ordinary blink) with a standard startup
>> is loaded to the device flash.
>>
>> Steps to reproduce the problem:
>>
>> start GDB server
>>   $ openocd -f interface/cmsis-dap.cfg -f target/stm32g4x.cfg
>>
>> start GDB in second terminal
>>   $ arm-none-eabi-gdb blink.elf
>>
>>   (gdb) target extended-remote localhost:3333
>>
>> Reset the device and halt it:
>>   (gdb) monitor reset halt
>>   target halted due to debug-request, current mode: Thread
>>   xPSR: 0x01000000 pc: 0x08000e14 msp: 0x20020000
>>
>> Step by one instruction to re-read GDB register cache:
>>   (gdb) stepi
>>
>> Check registers, LR should be 0xffffffff after reset:
>>   (gdb) info registers
>>   ...
>>   sp             0x20020000          0x20020000
>>   lr             0xffffffff          -1
>>   pc             0x8000e16           0x8000e16
>>   xPSR           0x1000000           16777216
>>   ...
>>
>>   (gdb) set debug remote
>>
>> Issue 'advance' command:
>>   (gdb) advance main
>>   [remote] Sending packet: $mfffffffe,2#fa
>>   [remote] Packet received: 0000
>>   [remote] Sending packet: $mfffffffe,2#fa
>>   [remote] Packet received: 0000
>>   [remote] Sending packet: $m8000526,2#30
>>   [remote] Packet received: 2046
>>   [remote] Sending packet: $Z1,8000526,2#7a
>>   [remote] Packet received: OK
>>   [remote] packet_ok: Packet Z1 (hardware-breakpoint) is supported
>>   [remote] Sending packet: $Z0,fffffffe,2#43
>>   [remote] Packet received: E0E
>>   [remote] packet_ok: Packet Z0 (software-breakpoint) is supported
>>   Warning:
>>   Cannot insert breakpoint 0.
>>   Cannot access memory at address 0xfffffffe
>>
>>   Command aborted.
>>   (gdb)
>>
>> Relevant messages from OpenOCD:
>>   Error: Failed to read memory at 0xfffff000
>>   Error: can't add breakpoint: unknown reason
>>
>> This patch adds skipping over frames that are not suitable for
>> guarding with a breakpoint inspired by 'finish' command.
>> If no suitable frame is found, a momentary breakpoint is not set.
>>
>> v2: Comment fixes, bug reference.
>>
>> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28683
>> Signed-off-by: Tomas Vanek <vanekt@fbl.cz>
> Hi Tomas,
>
> In order to better understand if this is the right fix, can you describe
> what the different frames are in this situation, starting with the
> current frame?  I'd like to see what the frame ids and frame types are.
>
> Is there any chance this could be reproduced using the GDB simulator (or maybe
> qemu), so I could tinker with it?
>
> Simon

Hi Simon,

I'm not familiar with GDB sim nor qemu and have no idea if their 
Cortex-M profile
implementation is precise enough to get the same behaviour. I'll give it 
a try...

Please see gdb debug output, I marked compute_frame_id results. This is 
a real Cortex-M device (this time STM32H7A3) right after reset,
OpenOCD used as gdb server:

(gdb) maintenance flush register-cache
Register cache flushed.
(gdb) set debug frame 1
(gdb) bt -past-entry
[frame] get_prev_frame_always_1: enter
   [frame] get_prev_frame_always_1: this_frame=-1
   [frame] get_prev_frame_always_1:   -> 
{level=0,type=<unknown>,unwinder=<unknown>,pc=0x800609c,id=<not 
computed>,func=<unknown>} // cached
[frame] get_prev_frame_always_1: exit
[frame] get_prev_frame_always_1: enter
   [frame] get_prev_frame_always_1: this_frame=-1
   [frame] get_prev_frame_always_1:   -> 
{level=0,type=<unknown>,unwinder=<unknown>,pc=0x800609c,id=<not 
computed>,func=<unknown>} // cached
[frame] get_prev_frame_always_1: exit
[frame] compute_frame_id: enter
   [frame] compute_frame_id: fi=0
   [frame] frame_unwind_find_by_frame: enter
     [frame] frame_unwind_find_by_frame: this_frame=0
     [frame] frame_unwind_try_unwinder: trying unwinder "dummy"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "dwarf2 tailcall"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "inline"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "jit"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "arm m exception 
lockup sec_fnc"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "arm stub"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "dwarf2"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "dwarf2 signal"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "arm exidx"
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "arm epilogue"
     [frame] frame_unwind_register_value: enter
       [frame] frame_unwind_register_value: frame=-1, regnum=25(xPSR)
       [frame] frame_unwind_register_value:   -> register=25 
bytes=[00000001]
     [frame] frame_unwind_register_value: exit
     [frame] frame_unwind_try_unwinder: no
     [frame] frame_unwind_try_unwinder: trying unwinder "arm prologue"
     [frame] frame_unwind_try_unwinder: yes
   [frame] frame_unwind_find_by_frame: exit
   [frame] frame_unwind_register_value: enter
     [frame] frame_unwind_register_value: frame=-1, regnum=13(sp)
     [frame] frame_unwind_register_value:   -> register=13 bytes=[00000220]
   [frame] frame_unwind_register_value: exit
   [frame] frame_unwind_register_value: enter
     [frame] frame_unwind_register_value: frame=-1, regnum=91(msp)
     [frame] frame_unwind_register_value:   -> register=91 bytes=[00000220]
   [frame] frame_unwind_register_value: exit
   [frame] frame_unwind_register_value: enter
     [frame] frame_unwind_register_value: frame=-1, regnum=92(psp)
     [frame] frame_unwind_register_value:   -> register=92 bytes=[00000000]
   [frame] frame_unwind_register_value: exit
   [frame] frame_unwind_register_value: enter
     [frame] frame_unwind_register_value: frame=-1, regnum=25(xPSR)
     [frame] frame_unwind_register_value:   -> register=25 bytes=[00000001]
   [frame] frame_unwind_register_value: exit
   [frame] frame_unwind_register_value: enter
     [frame] frame_unwind_register_value: frame=-1, regnum=13(sp)
     [frame] frame_unwind_register_value:   -> register=13 bytes=[00000220]
   [frame] frame_unwind_register_value: exit
   [frame] get_frame_func_if_available: this_frame=0 -> 0x800609c
   [frame] frame_id_p: 
l={stack=0x20020000,code=0x000000000800609c,!special} -> 1
   [frame] compute_frame_id:   -> 
{stack=0x20020000,code=0x000000000800609c,!special} <------- *current frame*
[frame] compute_frame_id: exit
#0  Reset_Handler ([frame] frame_id_p: l={!stack,!code,!special} -> 0
) at startup_stm32h7a3xx.s:62
[frame] operator==: 
l={stack=0x20020000,code=0x000000000800609c,!special}, 
r={!stack,!code,!special} -> 0
[frame] get_prev_frame: enter
   [frame] get_prev_frame_always_1: enter
     [frame] get_prev_frame_always_1: this_frame=0
     [frame] get_prev_frame_raw:   -> 
{level=1,type=<unknown>,unwinder=<unknown>,pc=<unknown>,id=<not 
computed>,func=<unknown>}
     [frame] compute_frame_id: enter
       [frame] compute_frame_id: fi=1
       [frame] frame_unwind_find_by_frame: enter
         [frame] frame_unwind_find_by_frame: this_frame=1
         [frame] frame_unwind_arch: next_frame=0 -> armv7e-m
         [frame] frame_unwind_try_unwinder: trying unwinder "dummy"
         [frame] frame_unwind_try_unwinder: no
         [frame] frame_unwind_try_unwinder: trying unwinder "dwarf2 
tailcall"
         [frame] frame_unwind_try_unwinder: no
         [frame] frame_unwind_try_unwinder: trying unwinder "inline"
         [frame] frame_unwind_register_value: enter
           [frame] frame_unwind_register_value: frame=0, regnum=15(pc)
           [frame] frame_unwind_register_value: enter
             [frame] frame_unwind_register_value: frame=0, regnum=14(lr)
             [frame] frame_id_p: 
l={stack=<sentinel>,!code,special=0x0000000000000000} -> 1
             [frame] frame_id_p: 
l={stack=<sentinel>,!code,special=0x0000000000000000} -> 1
             [frame] operator==: 
l={stack=<sentinel>,!code,special=0x0000000000000000}, 
r={stack=<sentinel>,!code,special=0x0000000000000000} -> 1
             [frame] frame_unwind_register_value: enter
               [frame] frame_unwind_register_value: frame=-1, regnum=14(lr)
               [frame] frame_unwind_register_value:   -> register=14 
bytes=[ffffffff]
             [frame] frame_unwind_register_value: exit
             [frame] frame_id_p: 
l={stack=<sentinel>,!code,special=0x0000000000000000} -> 1
             [frame] operator==: 
l={stack=<sentinel>,!code,special=0x0000000000000000}, 
r={stack=<sentinel>,!code,special=0x0000000000000000} -> 1
             [frame] get_prev_frame_always_1: enter
               [frame] get_prev_frame_always_1: this_frame=-1
               [frame] get_prev_frame_always_1:   -> 
{level=0,type=NORMAL_FRAME,unwinder="arm 
prologue",pc=0x800609c,id={stack=0x20020000,code=0x000000000800609c,!special},func=0x800609c} 
//
  cached
             [frame] get_prev_frame_always_1: exit
             [frame] value_fetch_lazy_register: (frame=0, regnum=14(lr), 
...) -> register=14 bytes=[ffffffff]
             [frame] frame_unwind_register_value:   -> register=14 
bytes=[ffffffff]
           [frame] frame_unwind_register_value: exit
           [frame] frame_unwind_register_value:   -> computed 
bytes=[ffffffff]
         [frame] frame_unwind_register_value: exit
         [frame] frame_unwind_pc: this_frame=0 -> 0xffffffff
         [frame] frame_unwind_try_unwinder: no
         [frame] frame_unwind_try_unwinder: trying unwinder "jit"
         [frame] frame_unwind_try_unwinder: no
         [frame] frame_unwind_try_unwinder: trying unwinder "arm m 
exception lockup sec_fnc"
         [frame] frame_unwind_try_unwinder: yes
       [frame] frame_unwind_find_by_frame: exit
       [frame] frame_unwind_register_value: enter
         [frame] frame_unwind_register_value: frame=0, regnum=13(sp)
         [frame] frame_unwind_register_value:   -> computed bytes=[00000220]
       [frame] frame_unwind_register_value: exit
       [frame] frame_unwind_register_value: enter
         [frame] frame_unwind_register_value: frame=0, regnum=91(msp)
         [frame] frame_unwind_register_value:   -> computed bytes=[00000220]
       [frame] frame_unwind_register_value: exit
       [frame] frame_unwind_register_value: enter
         [frame] frame_unwind_register_value: frame=0, regnum=92(psp)
         [frame] frame_unwind_register_value:   -> computed bytes=[00000000]
       [frame] frame_unwind_register_value: exit
       [frame] frame_id_p: 
l={stack=0x0,code=0x00000000ffffffff,!special} -> 1
       [frame] compute_frame_id:   -> 
{stack=0x0,code=0x00000000ffffffff,!special}        <----- *magic value 
0xffffffff in LR indicates there is no caller*
     [frame] compute_frame_id: exit
   [frame] get_prev_frame_always_1: exit
[frame] get_prev_frame: exit
#1  <signal handler called>
[frame] operator==: l={stack=0x0,code=0x00000000ffffffff,!special}, 
r={!stack,!code,!special} -> 0
[frame] get_prev_frame: enter
   [frame] get_prev_frame_always_1: enter
     [frame] get_prev_frame_always_1: this_frame=1
     [frame] get_prev_frame_always_1:   -> nullptr // UNWIND_OUTERMOST
   [frame] get_prev_frame_always_1: exit
[frame] get_prev_frame: exit
[frame] get_prev_frame_always_1: enter
   [frame] get_prev_frame_always_1: this_frame=1
   [frame] get_prev_frame_always_1:   -> nullptr // UNWIND_OUTERMOST // 
cached
[frame] get_prev_frame_always_1: exit
[frame] get_prev_frame_always_1: enter
   [frame] get_prev_frame_always_1: this_frame=-1
   [frame] get_prev_frame_always_1:   -> 
{level=0,type=NORMAL_FRAME,unwinder="arm 
prologue",pc=0x800609c,id={stack=0x20020000,code=0x000000000800609c,!special},func=0x800609c} 
// cached
[frame] get_prev_frame_always_1: exit


The situation when stopped in an ISR is similar, just the magic value 
differs e.g. 0xfffffff1
and other frames follows.