Re: [PROTOTYPE] Make "info breakpoints" show breakpoint's specs (Re: [PATCH] gdb: Don't skip prologue for explicit line breakpoints in assembler)

[Pedro Alves <palves@redhat.com>]

On 6/24/19 8:54 PM, Pedro Alves wrote:
> I did notice that the patch currently crashes with catchpoints that
> are implemented via breakpoints, for example, and possibly
> other types:
> 
>  (top-gdb) catch throw 
>  Catchpoint 5 (throw)
>  (top-gdb) info breakpoints 
>  5       breakpoint     keep y   ../../gdb/breakpoint.c:6387: internal-error: void print_one_breakpoint_location(breakpoint*, bp_location*, int, bp_location**, int): Assertion `b->loc == NULL || b->loc->next == NULL' failed.
>  A problem internal to GDB has been detected,
>  further debugging may prove unreliable.
>  Quit this debugging session? (y or n) 
> 
> whoops.  :-)  I did say this was a prototype, right?  :-)

Actually, turns out this is a preexisting problem.

GDB is not expecting to find more than one location for a catchpoint
backed by breakpoints.  But it is now finding two locations.
The breakpoint is set at "__cxa_begin_catch", and if we do that manually,
we see what are those locations:

(top-gdb) b __cxa_begin_catch
Breakpoint 5 at 0xb126c0 (2 locations)
(top-gdb) info breakpoints 
Num     Type           Disp Enb Address            What
1       breakpoint     keep y   <MULTIPLE>         
1.1                         y   0x0000000000b126c0 <__cxa_begin_catch>
1.2                         y   0x00007ffff2f4ddb0 in __cxxabiv1::__cxa_begin_catch(void*) at ../../../../libstdc++-v3/libsupc++/eh_catch.cc:41

Looks like a regression caused by the -qualified stuff.
I guess we want a breakpoint at <__cxa_begin_catch> only.
Should be easy to fix.

But I also suggest that "info breakpoints" shouldn't be showing any
address for "catch catch" breakpoints (and catch throw, etc.).
It's an implementation detail that these catchpoints are implemented
as magic breakpoints.

Thanks,
Pedro Alves