aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/OpenSim (unfollow)
Commit message (Collapse)AuthorFilesLines
2013-07-21Delay the enqueueing of non-longpoll requests for 100ms. No need to have ↵Diva Canto1-1/+11
these requests actively on the processing queue if it seems they're not ready.
2013-07-21Do some simple queue empty checks in the main outgoing udp loop instead of ↵Justin Clark-Casey (justincc)5-22/+92
always performing these on a separate fired thread. This appears to improve cpu usage since launching a new thread is more expensive than performing a small amount of inline logic. However, needs testing at scale.
2013-07-19Fix HGTravelStore.migrations in SQLite (mantis #6709)Diva Canto1-10/+10
2013-07-18BulletSim: add position resetting for stationary avatars so they don'tRobert Adams2-5/+20
move around when standing on a stationary object. Create proper linkage between BSCharacter and its actor by generating a UpdatedProperties event the same way BSPrim does.
2013-07-18Enable storing of environment settings in NullSimulationDataDan Lake1-3/+17
2013-07-18Changed the timoeut of EQ 502s (no events) to 50 secs. The viewer post ↵Diva Canto2-3/+10
requests timeout in 60 secs. There's plenty of room for improvement in handling the EQs. Some other time...
2013-07-18try Hacking in an AutoResetEvent to control the outgoing UDP loop instead of ↵Justin Clark-Casey (justincc)2-2/+28
a continuous loop with sleeps. Does appear to have a cpu impact but may need further tweaking
2013-07-18Revert "Revert "I confuse myself. Let's try this variable name instead.""Diva Canto1-3/+3
This reverts commit 52dc7b2a96a28798d55d07d79d003ce5e3d35216.
2013-07-18Revert "Revert "Do the same trick that dahlia did for Dequeue(timeout)""Diva Canto1-2/+3
This reverts commit 5495df74436d6c0039a1500d979a964b003abfdf.
2013-07-18Revert "Revert "Putting the requests back in the queue while testing for ↵Diva Canto1-3/+7
count >0 is not the smartest move..."" This reverts commit 71278919575b0e0222cdbe3c0cefa5919f9a75bc.
2013-07-18Revert "Revert "Didn't mean to commit this change in BlockingQueue.cs""Diva Canto1-1/+1
This reverts commit fda91d93dad1fa6f901e8db5829aa8b70477c97e.
2013-07-18Reverting the reverts I did yesterday. cpu-branch has now beenDiva Canto4-52/+41
successfully tested, and I'm merging back those changes, which proved to be good. Revert "Revert "Cleared up much confusion in PollServiceRequestManager. Here's the history:"" This reverts commit fa2370b32ee57a07f27501152c3c705a883b13d8.
2013-07-18Added MinPoolThreads to ini [Startup] section to control SmartThreadPool.Dan Lake2-4/+8
2013-07-18Fix what apepars to be a bug in DoubleQueue<T>.Enqueue(Queue<T> q, T data) ↵Justin Clark-Casey (justincc)1-1/+1
where the q parmater is ignored and everyghig is always placed on m_lowQueue. No actual impact presently since nothing ends up calling EnqueueHigh()
2013-07-17Revert "Cleared up much confusion in PollServiceRequestManager. Here's the ↵Diva Canto4-41/+52
history:" This reverts commit e46459ef21e1ee5ceaeca70365a7c881d33b09ce.
2013-07-17command line kick user now uses exact name match instead of substring search ↵Dan Lake1-2/+3
to avoid kicking the wrong user or multiple wrong users.
2013-07-17Revert "Didn't mean to commit this change in BlockingQueue.cs"Diva Canto1-1/+1
This reverts commit 0f5b616fb0ebf9207b3cc81771622ed1290ea7d6.
2013-07-17Revert "Putting the requests back in the queue while testing for count >0 is ↵Diva Canto1-7/+3
not the smartest move..." This reverts commit f4317dc26d670c853d0ea64b401b00f718f09474.
2013-07-17Revert "Do the same trick that dahlia did for Dequeue(timeout)"Diva Canto1-3/+2
This reverts commit af792bc7f2504e9ccf1c8ae7568919785dc397c9.
2013-07-17Revert "I confuse myself. Let's try this variable name instead."Diva Canto1-3/+3
This reverts commit 1d3deda10cf85abd68a5f904d6698ae597a67cc0.
2013-07-17Revert "Now trying DoubleQueue instead of BlockingQueue for the ↵Diva Canto1-2/+2
PollServiceRequestManager." This reverts commit 5f95f4d78e8c7d17b8ba866907156fe6d4444c04.
2013-07-17Revert "This is a completely unreasonable thing to do, effectively defying ↵Diva Canto1-52/+40
the purpose of BlockingQueues. Trying this, to see the effect on CPU." This reverts commit 5232ab0496eb4fe6903a0fd328974ac69df29ad8.
2013-07-17This is a completely unreasonable thing to do, effectively defying the ↵Diva Canto1-40/+52
purpose of BlockingQueues. Trying this, to see the effect on CPU.
2013-07-17Now trying DoubleQueue instead of BlockingQueue for the ↵Diva Canto1-2/+2
PollServiceRequestManager.
2013-07-17I confuse myself. Let's try this variable name instead.Diva Canto1-3/+3
2013-07-17Do the same trick that dahlia did for Dequeue(timeout)Diva Canto1-2/+3
2013-07-17Putting the requests back in the queue while testing for count >0 is not the ↵Diva Canto1-3/+7
smartest move...
2013-07-17Didn't mean to commit this change in BlockingQueue.csDiva Canto1-1/+1
2013-07-17Cleared up much confusion in PollServiceRequestManager. Here's the history:Diva Canto4-52/+41
When Melanie added the web fetch inventory throttle to core, she made the long poll requests (EQs) effectively be handled on an active loop. All those requests, if they existed, were being constantly dequeued, checked for events (which most often they didn't have), and requeued again. This was an active loop thread on a 100ms cycle! This fixes the issue. Now the inventory requests, if they aren't ready to be served, are placed directly back in the queue, but the long poll requests aren't placed there until there are events ready to be sent or timeout has been reached. This puts the LongPollServiceWatcherThread back to 1sec cycle, as it was before.
2013-07-17BulletSim: fix small bug where everything looked like it was collidingRobert Adams2-6/+10
before the first simulator step.
2013-07-16Removed the MapItems thread. Redirected the map items requests to the ↵Diva Canto2-9/+45
services throttle thread. Didn't change anything in how that processor is implemented, for better or for worse.
2013-07-16Improvements to the ServiceThrottleModule: added a category and an itemid to ↵Diva Canto3-11/+36
the interface, so that duplicate requests aren't enqueued more than once.
2013-07-16UserManagementModule: in the continuation, call the method that also looks ↵Diva Canto1-2/+2
up the cache, because the resource may be here in the meantime
2013-07-16Added config var to fiddle with the Interval for the service throttle threadDiva Canto1-1/+4
2013-07-16Attachments module only registers when enabled. This enables alternative ↵Dan Lake3-6/+12
attachments module implementations. All calls to Scene.AttachmentsModule are checking for null. Ideally, if we support disabling attachments then we need a null attachments module to register with the scene.
2013-07-16Deleted a couple of verbose messagesDiva Canto2-4/+2
2013-07-16Added IServiceThrottleModule.csDiva Canto1-0/+11
2013-07-16Changed the name to ServiceThrottle/ServiceThrottleModule in order to ↵Diva Canto1-0/+0
reflect its more generic nature.
2013-07-16Eliminated the UserManagement/UserManagementModule throttle thread. Made the ↵Diva Canto2-189/+148
other one generic, taking any continuation.
2013-07-16minor: Add warning method doc about possibly inconsistent results returned ↵Justin Clark-Casey (justincc)1-1/+19
from BlockingQueue.Contains(), Count() and GetQueueArray()
2013-07-16Revert "Revert "MSDN documentation is unclear about whether exiting a lock() ↵Justin Clark-Casey (justincc)1-5/+8
block will trigger a Monitor.Wait() to exit, so avoid some locks that don't actually affect the state of the internal queues in the BlockingQueue class."" This reverts commit 21a09ad3ad42b24bce4fc04c6bcd6f7d9a80af08. After more analysis and discussion, it is apparant that the Count(), Contains() and GetQueueArray() cannot be made thread-safe anyway without external locking And this change appears to have a positive impact on performance. I still believe that Monitor.Exit() will not release any thread for Monitor.Wait(), as per http://msdn.microsoft.com/en-gb/library/vstudio/system.threading.monitor.exit%28v=vs.100%29.aspx so this should in theory make no difference, though mono implementation issues could possibly be coming into play.
2013-07-16Revert "MSDN documentation is unclear about whether exiting a lock() block ↵Justin Clark-Casey (justincc)1-8/+5
will trigger a Monitor.Wait() to exit, so avoid some locks that don't actually affect the state of the internal queues in the BlockingQueue class." This reverts commit 42e2a0d66eaa7e322bce817e9e2cc9a288de167b Reverting because unfortunately this introduces race conditions because Contains(), Count() and GetQueueArray() may now end up returning the wrong result if another thread performs a simultaneous update on m_queue. Code such as PollServiceRequestManager.Stop() relies on the count being correct otherwise a request may be lost. Also, though some of the internal queue methods do not affect state, they are not thread-safe and could return the wrong result generating the same problem lock() generates Monitor.Enter() and Monitor.Exit() under the covers. Monitor.Exit() does not cause Monitor.Wait() to exist, only Pulse() and PulseAll() will do this Reverted with agreement.
2013-07-16In the pursuit of using less CPU: now trying to avoid blocking queues ↵Diva Canto2-15/+120
altogether. Instead, this uses a timer. No sure if it's better or worse, but worth the try.
2013-07-16Actually use DoubleQueue in UserManagement/UserManagementModuleDiva Canto1-1/+1
2013-07-16High CPU hunt: try a different blocking queue, DoubleQueueDiva Canto2-1/+3
2013-07-16revert last commit which seems to conflict with DoubleQueue internals. The ↵dahlia1-10/+5
random crash might be in DoubleQueue instead. See http://pastebin.com/XhNBNqsc
2013-07-16add locking to internal queue in WebFetchInvDescModule; lack of which caused ↵dahlia1-5/+10
a random crash in a load test yesterday
2013-07-16MSDN documentation is unclear about whether exiting a lock() block will ↵dahlia1-5/+8
trigger a Monitor.Wait() to exit, so avoid some locks that don't actually affect the state of the internal queues in the BlockingQueue class.
2013-07-16Simplify EventQueue cap setup so that it is also stat monitored.Justin Clark-Casey (justincc)1-45/+29
Curiously, the number of requests received is always one greater than that shown as handled - needs investigation
2013-07-15Add request received/handling stats for caps which are served by http poll ↵Justin Clark-Casey (justincc)11-126/+267
handlers. This adds explicit cap poll handler supporting to the Caps classes rather than relying on callers to do the complicated coding. Other refactoring was required to get logic into the right places to support this.
generated by cgit v1.1 at 2024-12-27 18:37:21 +0000