Brian Anderson
dc0b9f44e2
rt: Change the rust_port refcounting scheme to avoid races
...
Hopefully...
2012-03-06 17:14:40 -08:00
Brian Anderson
1347d04bb0
rt: Properly block tasks while waiting for port detach
2012-03-05 19:39:56 -08:00
Brian Anderson
4c4a2320eb
rt: Make rust_port_detach less contentious
...
It still utterly dominates some benchmarks with busy waiting, but at least it
doesn't create lock contention while doing so.
2012-03-05 19:39:56 -08:00
Brian Anderson
e08f46db68
rt: Move receive code into rust_port
2012-03-05 19:39:56 -08:00
Brian Anderson
b2cfb7ef82
rt: Add rust_port_select function
2012-02-16 11:12:22 -08:00
Brian Anderson
342dc21d2c
Disconnect ports before draining them. Issue #1155
2011-11-16 17:22:14 -08:00
Brian Anderson
138d9ca5d5
Drop enqueued elements when a port is destructed. Closes #1155
2011-11-13 15:43:58 -08:00
Brian Anderson
a6f9f73f06
rt: Make rust_port's ref counting non-atomic
...
Now that the task lock must be held there's no need for atomicity
2011-11-11 16:12:48 -08:00
Brian Anderson
5d1e321ecb
rt: Remove rust_chan
2011-11-11 12:11:21 -08:00
Brian Anderson
793da65a8e
rt: Move rust_chan::send to rust_port::send
2011-11-10 17:53:19 -08:00
Brian Anderson
371574f64b
rt: Remove rust_token. Unused
2011-11-10 17:35:21 -08:00
Brian Anderson
e4f980810b
rt: Simplify channel-port association
...
It turns out that there's only ever a single channel per port these days,
and it always has the same lifetime as the port, so we don't need a list or a
complex association protocol.
2011-11-10 17:26:22 -08:00
Eric Holk
39b16077bb
Port ID-based channels.
2011-08-15 09:26:51 -07:00
Eric Holk
04af99ecb0
First step towards port handles.
2011-08-15 09:26:51 -07:00
Eric Holk
b62e80c1f0
Atomic ref counting for chans.
2011-08-05 15:27:28 -07:00
Eric Holk
d1dbb99984
Removing proxies and message queues.
2011-07-29 18:54:59 -07:00
Graydon Hoare
49116adbe9
Fix compile-command lines in rt.
2011-07-13 13:51:20 -07:00
Eric Holk
8acadb17c2
Work on debugging race conditions.
...
Ports and channels have been moved to the kernel pool, since they've
been known to outlive their associated task. This probably isn't the
right thing to do, the life cycle needs fixed instead.
Some refactorying in memory_region.cpp. Added a helper function to
increment and decrement the allocation counter. This makes it easier
to switch between atomic and non-atomic increments. Using atomic
increments for now, although this still does not fix the problem.
2011-07-07 18:22:27 -07:00
Eric Holk
681c063ec0
Conservatively serialize nearly all upcalls. Successfuly ran make check with RUST_THREADS=8, so we're probably fairly safe now. In the future we can relax the synchronization to get better performance.
2011-06-27 09:58:39 -07:00
Eric Holk
4d99bf9af2
Added some locking to ports to prevent the case where two threads simultaneously wake up a task blocked on a certain port.
2011-06-27 09:58:39 -07:00
Michael Bebenita
97d6342bf0
Synthesize a flush_chan upcall right before a channel's ref_count drops to zero. This should only happen in the Rust code and not in the drop glue, or on the unwind path. This change allows the task owning the channel to block on a flush and delete its own channel. This change also cleans up some code around rust_port and rust_chan.
2010-08-09 08:15:34 -07:00
Michael Bebenita
30734a9a7b
Some pretty printing in the runtime.
2010-08-09 07:52:07 -07:00
Michael Bebenita
4246d567b7
Move ports out into their own file, add data_message and make communication system use it (and proxies) instead of existing token scheme.
2010-07-28 20:30:29 -07:00
