Re: [PATCH glibc] nptl_db: different libpthread/ld.so load orders (bug 27744)

[Florian Weimer <fweimer@redhat.com>]

* Simon Marchi:

> On 2021-04-16 11:56 a.m., Florian Weimer wrote:
>> libthread_db is loaded once GDB encounters libpthread, and at this
>> point, ld.so may not have been loaded yet.

> Can this state (libpthread loaded, ld.so not loaded) really happen
> during the normal lifetime of a process?  My understanding is that this
> state happens when attaching only because GDB reads the shared libraries
> from the process in an undefined order, so libpthread may be discovered
> before ld.so.  So we present to libthread_db a state that doesn't really
> make sense.

I think it depends on the order of the link map list, I think, and there
ld.so ends up being sorted last.

If libpthread gets merged into libc for glibc 2.34, I will try to make
this work with old GDB by adding a fake libpthread.so.0 at the end of
the link map list, after ld.so.  But none of this code exists yet, so
it's impossible to know if it actually will work as expected.

Thanks,
Florian