Eric Wong wrote:
> eregontp / gmail.com wrote:
>> Would it be simpler to track a set of pids created by MJIT, ignore those in
>> waitpid() and synchronize around both creating GCC processes and when checking
>> the result of waitpid()?
> 
> Not possible, if there's a thread running in waitpid(-1, 0),
> it can steal the result of waitpid(mjit_used_by_pid).

Could it help at all to delegate to a single "JIT Controller" daemon?

E.g.

 - At startup, Ruby forks a single JIT Controller daemon,
   meant to persist for the duration of the Ruby process

 - JITC daemon removes itself from Ruby's process group

 - Ruby communicates with JITC daemon via pipes, and JITC
   in turn spawns and waits for JIT compiler processes


Just a thought.  (Apologies if not applicable; I'm not well-versed in
the current JIT architecture.)


Regards,

Bill





Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>