Re: Deadlock when using SslFilter and ProxyFilter together
If you look at the stacktraces, I sent in my previous mails, you see, that the order, in which the filters are called, is right: When receiving: ... at org.apache.mina.filter.ssl.SslHandler.flushScheduledEvents(SslHandler.java:310) <--- SslFilter at org.apache.mina.filter.ssl.SslFilter.messageReceived(SslFilter.java:534) at org.apache.mina.core.filterchain.DefaultIoFilterChain.callNextMessageReceived(DefaultIoFilterChain.java:542) at org.apache.mina.core.filterchain.DefaultIoFilterChain.access$1300(DefaultIoFilterChain.java:48) at org.apache.mina.core.filterchain.DefaultIoFilterChain$EntryImpl$1.messageReceived(DefaultIoFilterChain.java:943) at org.apache.mina.proxy.filter.ProxyFilter.messageReceived(ProxyFilter.java:153) <--- ProxyFilter ... When sending: --- at org.apache.mina.proxy.filter.ProxyFilter.writeData(ProxyFilter.java:208) < ProxyFilter - waiting to lock <0x00076ef8b1d8> (a org.apache.mina.proxy.handlers.socks.Socks4LogicHandler) at org.apache.mina.proxy.filter.ProxyFilter.filterWrite(ProxyFilter.java:192) at org.apache.mina.core.filterchain.DefaultIoFilterChain.callPreviousFilterWrite(DefaultIoFilterChain.java:625) at org.apache.mina.core.filterchain.DefaultIoFilterChain.access$1500(DefaultIoFilterChain.java:48) at org.apache.mina.core.filterchain.DefaultIoFilterChain$EntryImpl$1.filterWrite(DefaultIoFilterChain.java:953) at org.apache.mina.filter.ssl.SslHandler.flushScheduledEvents(SslHandler.java:316) at org.apache.mina.filter.ssl.SslFilter.filterWrite(SslFilter.java:653) < SslFilter I am only writing the client part of an application. I communicate with a FIX/SSL server behind a SOCKS proxy, so I cannot change the order in which Ssl encryption and proxy framing must be done.
Re: Deadlock when using SslFilter and ProxyFilter together
Le 11/02/16 10:10, Norbert Irmer a écrit : > If you look at the stacktraces, I sent in my previous mails, you see, that > the order, in which the > filters are called, is right: > > When receiving: > > ... > at > org.apache.mina.filter.ssl.SslHandler.flushScheduledEvents(SslHandler.java:310) > <--- SslFilter > at org.apache.mina.filter.ssl.SslFilter.messageReceived(SslFilter.java:534) > at > org.apache.mina.core.filterchain.DefaultIoFilterChain.callNextMessageReceived(DefaultIoFilterChain.java:542) > at > org.apache.mina.core.filterchain.DefaultIoFilterChain.access$1300(DefaultIoFilterChain.java:48) > at > org.apache.mina.core.filterchain.DefaultIoFilterChain$EntryImpl$1.messageReceived(DefaultIoFilterChain.java:943) > at > org.apache.mina.proxy.filter.ProxyFilter.messageReceived(ProxyFilter.java:153) > > <--- ProxyFilter > ... > > > When sending: > > --- > at org.apache.mina.proxy.filter.ProxyFilter.writeData(ProxyFilter.java:208) >< ProxyFilter > - waiting to lock <0x00076ef8b1d8> (a > org.apache.mina.proxy.handlers.socks.Socks4LogicHandler) > at org.apache.mina.proxy.filter.ProxyFilter.filterWrite(ProxyFilter.java:192) > at > org.apache.mina.core.filterchain.DefaultIoFilterChain.callPreviousFilterWrite(DefaultIoFilterChain.java:625) > at > org.apache.mina.core.filterchain.DefaultIoFilterChain.access$1500(DefaultIoFilterChain.java:48) > at > org.apache.mina.core.filterchain.DefaultIoFilterChain$EntryImpl$1.filterWrite(DefaultIoFilterChain.java:953) > at > org.apache.mina.filter.ssl.SslHandler.flushScheduledEvents(SslHandler.java:316) > at org.apache.mina.filter.ssl.SslFilter.filterWrite(SslFilter.java:653) > < SslFilter > > > > I am only writing the client part of an application. I communicate with a > FIX/SSL server behind a SOCKS proxy, > so I cannot change the order in which Ssl encryption and proxy framing must > be done. Ok, I *think* that the problem is due to the fact that the flushScheduledEvents() method was not protecting writes *and* reads, so they can be executed at the same time by two threads. Then what happens is that both reach the ProxyFilter synchronized section, FixMessageProcessor takes the SslHandler lock FixMessageProcessor takes the ProxyFilter lock and at the same time : NioProcessor takes the ProxyFilter lock NioProcessor takes the SslHandler lock resulting in a Deadlock. NioProcessor >[ProxyFilter]--->[SslHandler]> IoHandler NioProcessor <[ProxyFilter]<---[SslHandler]< IoHandler The new version I have put on http://people.apache.org/~elecharny should solve the problem : instead of getting the lock whe we enter in the flushScheduledEvents() event, we exit immediately, doing nothing but incrementing the number of scheduled-events, instead of getting the lock. That means the thread which actually hold this lock will process all the enqueued messages itself, until the counter get down to 0. Can you give this version a try ?
Re: Deadlock when using SslFilter and ProxyFilter together
> The new version I have put on http://people.apache.org/~elecharny should > solve the problem : instead of getting the lock whe we enter in the > flushScheduledEvents() event, we exit immediately, doing nothing but > incrementing the number of scheduled-events, instead of getting the > lock. That means the thread which actually hold this lock will process > all the enqueued messages itself, until the counter get down to 0. > >Can you give this version a try ? Looks very good, I am running my application now for 20 minutes or so, and still have no deadlock. I am using now your snapshot versions without any changes. This is very nice, now it's possible to send a pull request for proxy support to the QuickFixJ people the next time you do a Mina release. Thanks a lot, Norbert
Re: Deadlock when using SslFilter and ProxyFilter together
Le 11/02/16 17:38, Norbert Irmer a écrit : > >> The new version I have put on http://people.apache.org/~elecharny should >> solve the problem : instead of getting the lock whe we enter in the >> flushScheduledEvents() event, we exit immediately, doing nothing but >> incrementing the number of scheduled-events, instead of getting the >> lock. That means the thread which actually hold this lock will process >> all the enqueued messages itself, until the counter get down to 0. >> >> Can you give this version a try ? > > Looks very good, I am running my application now for 20 minutes or so, and > still > have no deadlock. > > I am using now your snapshot versions without any changes. This is very nice, > now it's possible to send a pull request for proxy support to the QuickFixJ > people > the next time you do a Mina release. > > Thanks a lot, > Norbert > > Glad that seems to be fixed ! Norbert, I have to tell you that I really appreciated your patience, and I'm really sorry to have tried to deflect the prb for a couple of days. I really thought it was something else, and it took me some time (and the price was sleep deprivation :/) to realize that there were a deadlock. Worst part : this deadlock was already detected months ago by Terence Marks (https://issues.apache.org/jira/browse/DIRMINA-1019), a patch was proposed, and I fucked it up thinking it was overdoing. My bad... I'm going to cut a release asap. Expect it to be ready by the beginning of next week. thanks a lot !
ApacheCon NA 2016 - Important Dates!!!
Hello everyone! I hope this email finds you well. I hope everyone is as excited about ApacheCon as I am! I'd like to remind you all of a couple of important dates, as well as ask for your assistance in spreading the word! Please use your social media platform(s) to get the word out! The more visibility, the better ApacheCon will be for all!! :) CFP Close: February 12, 2016CFP Notifications: February 29, 2016Schedule Announced: March 3, 2016 To submit a talk, please visit: http://events.linuxfoundation.org/events/apache-big-data-north-america/program/cfp Link to the main site can be found here: http://events.linuxfoundation.org/events/apache-big-data-north-america Apache: Big Data North America 2016 Registration Fees: Attendee Registration Fee: US$599 through March 6, US$799 through April 10, US$999 thereafterCommitter Registration Fee: US$275 through April 10, US$375 thereafterStudent Registration Fee: US$275 through April 10, $375 thereafter Planning to attend ApacheCon North America 2016 May 11 - 13, 2016? There is an add-on option on the registration form to join the conference for a discounted fee of US$399, available only to Apache: Big Data North America attendees. So, please tweet away!! I look forward to seeing you in Vancouver! Have a groovy day!! ~Melissaon behalf of the ApacheCon Team
ApacheCon NA 2016 - Important Dates!!!
Hello everyone! I hope this email finds you well. I hope everyone is as excited about ApacheCon as I am! I'd like to remind you all of a couple of important dates, as well as ask for your assistance in spreading the word! Please use your social media platform(s) to get the word out! The more visibility, the better ApacheCon will be for all!! :) CFP Close: February 12, 2016CFP Notifications: February 29, 2016Schedule Announced: March 3, 2016 To submit a talk, please visit: http://events.linuxfoundation.org/events/apache-big-data-north-america/program/cfp Link to the main site can be found here: http://events.linuxfoundation.org/events/apache-big-data-north-america Apache: Big Data North America 2016 Registration Fees: Attendee Registration Fee: US$599 through March 6, US$799 through April 10, US$999 thereafterCommitter Registration Fee: US$275 through April 10, US$375 thereafterStudent Registration Fee: US$275 through April 10, $375 thereafter Planning to attend ApacheCon North America 2016 May 11 - 13, 2016? There is an add-on option on the registration form to join the conference for a discounted fee of US$399, available only to Apache: Big Data North America attendees. So, please tweet away!! I look forward to seeing you in Vancouver! Have a groovy day!! ~Melissaon behalf of the ApacheCon Team
