On Tue, Aug 30, 2022 at 7:44 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Buildfarm member mamba (NetBSD-current on prairiedog's former hardware)
> has failed repeatedly since I set it up. I have now run the cause of
> that to ground [1], and here's what's happening: if the postmaster
> receives a signal just before it first waits at the select() in
> ServerLoop, it can self-deadlock. During the postmaster's first use of
> select(), the dynamic loader needs to resolve the PLT branch table entry
> that the core executable uses to reach select() in libc.so, and it locks
> the loader's internal data structures while doing that. If we enter
> a signal handler while the lock is held, and the handler needs to do
> anything that also requires the lock, the postmaster is frozen.
. o O ( pselect() wouldn't have this problem, but it's slightly too
new for the back branches that didn't yet require SUSv3... drat )
> I'd originally intended to make this code "#ifdef __NetBSD__",
> but on looking into the FreeBSD sources I find much the same locking
> logic in their dynamic loader, and now I'm wondering if such behavior
> isn't pretty standard. The added calls should have negligible cost,
> so it doesn't seem unreasonable to do them everywhere.
FWIW I suspect FreeBSD can't break like this in a program linked with
libthr, because it has a scheme for deferring signals while the
runtime linker holds locks. _rtld_bind calls _thr_rtld_rlock_acquire,
which uses the THR_CRITICAL_ENTER mechanism to cause thr_sighandler to
defer until release. For a non-thread program, I'm not entirely sure,
but I don't think the fork() problem exists there. (Could be wrong,
based on a quick look.)
> (Of course, a much better answer is to get out of the business of
> doing nontrivial stuff in signal handlers. But even if we get that
> done soon, we'd surely not back-patch it.)
+1
