runtime: preemptive scheduling

[rsc]

What steps will reproduce the problem?
1. go func() { for { } }

What is the expected output? What do you see instead?
Shouldn't wedge the rest of the program but does.

[rsc]

Comment 1:

http://groups.google.com/group/golang-nuts/t/7a9535c4136d3e2 
http://groups.google.com/group/golang-nuts/t/b546803a0faf3ade

[rsc]

Comment 2:

Step 1 should read go func() { for { } }()

[rsc]

Comment 3:

Labels changed: added priority-later.

[rsc]

Comment 5:

Labels changed: added go1.1maybe.

[rsc]

Comment 6:

Labels changed: added size-xl.

[robpike]

Comment 7:

Labels changed: removed go1.1maybe.

[dvyukov]

Comment 8:

Issue #4711 has been merged into this issue.

[dvyukov]

Comment 9:

Issue #5324 has been merged into this issue.

[dvyukov]

Comment 10:

There is a partial solution we can do for Go1.2 with few casualties.
Namely, make gcwaiting variable per-M. Sysmon() thread detects when an M runs the same
goroutine for a long time and sets m->gcwaiting. When a thread notices m->gcwaiting, it
calls into scheduler. If global gcwaiting is not set, it just causes rescheduling. This
way worker threads be preempted on chan/map/malloc operations.

Owner changed to @dvyukov.

[dvyukov]

Comment 11:

Issue #5324 has been merged into this issue.

[dvyukov]

Comment 12:

The following change partially addresses the issue:
https://golang.org/cl/10264044
runtime: preempt goroutines for GC
The last patch for preemptive scheduler,
with this change stoptheworld issues preemption
requests every 100us.
Update issue #543.

[rsc]

Comment 13:

This issue was updated by revision b913cf8.

I believe the runtime is strong enough now to reenable
preemption during the function prologue.
Assuming this is or can be made stable, it will be in Go 1.2.
More aggressive preemption is not planned for Go 1.2.
R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/11433045

[rsc]

Comment 14:

Labels changed: added go1.3maybe.

[rsc]

Comment 15:

Labels changed: added release-none, removed go1.3maybe.

[rsc]

Comment 16:

Labels changed: added repo-main.

[AlekSi]

I think discussion moved to #10958.

[ianlancetaylor]

Thanks. I agree that everything remaining here is captured by #10958, so closing this one.