[Bug gdb/30872] Assertion `buffer.copy_insn_closure.get () != nullptr' failed

["luis.machado at arm dot com" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=30872

--- Comment #8 from Luis Machado <luis.machado at arm dot com> ---
Well, just should've been a bit obvious. I'm still checking this, but here's
some useful information.

The displaced stepping machinery in gdb, which is used to side-step breakpoints
without removing them, uses scratch space for executing the instructions out of
their original location.

That scratch location was picked a while ago to be the entry point of the
program, from the auxv's AT_ENTRY value. That's usually the first instruction
of _start.

Now, your example program starts from _start. So things get a bit confusing
because gdb is trying to execute and modify things using the same address.

As a temporary workaround, you can disable displaced stepping. It likely won't
make a difference for your case.

set displaced-stepping off

That should allow you to resume your debugging session.

I'll see what can be done about this corner case.

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