[Bug threads/20291] Set scheduler-locking on doesn't work correctly with pthread_create

["palves at redhat dot com" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=20291

--- Comment #10 from Pedro Alves <palves at redhat dot com> ---
Well, I actually said:

 "IF there's synchronization (mutex/futex/whatnot) between
 the parent/child threads inside glibc's pthread_create"

New emphasis on the IF.  I haven't looked at the code in that level of detail. 
I was merely pointing out a _potential_ problem with the use of schedlock in
the new glibc printers you're working on (as discussed in the libc-alpha@).

(It's still true that the child manages to run a bit with older gdbs.  But with
your observation, that ends up not making a practical difference for the issues
at hand.)

> What this means is, the child waits for the parent and not the other way around.

OK, then the concern is moot.  (unless glibc changes, of course.)

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