[Cs-dev] Jack Crash

[Cs-dev] Jack Crash

Date	2014-02-20 15:40
From	Steven Yi
Subject	[Cs-dev] Jack Crash
	Hi All, Menno reported to me a crash from Blue when a Csound run completed. The log reported the crash originated from Csound's librtjack's processCallback function. The log didn't have much more information than that. My one thought from a quick look at the rtjack code is that it might be possible that the void* arg passed into the function might be NULL, maybe if csound deletes and cleans up just before an jack callback gets called. My guess is that we might need a NULL check at the top of the function and if NULL just return 0 (I don't know what return values are usable there). Does anyone have any thoughts on if the above analysis is plausible, and whether they might see some other problems? Thanks! steven ------------------------------------------------------------------------------ Managing the Performance of Cloud-Based Applications Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. Read the Whitepaper. http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net

Date	2014-02-20 15:59
From	Victor Lazzarini
Subject	Re: [Cs-dev] Jack Crash
	I see this csound->DestroyGlobalVariable(csound, "_rtjackGlobals"); if (p.client != (jack_client_t) NULL) { / deactivate client / //if (p.jackState != 2) { //if (p.jackState == 0) // csound->Sleep((size_t) // ((int) ((double) (p.bufSize p.nBuffers) // * 1000.0 / (double) p.sampleRate + 0.999))); jack_deactivate(p.client); //} csound->Sleep((size_t) 50); surely the DestroyGlobalVariable() needs to happen after deactivation. Maybe this line should be moved to after jack_deactivate(). Victor On 20 Feb 2014, at 15:40, Steven Yi wrote: > Hi All, > > Menno reported to me a crash from Blue when a Csound run completed. > The log reported the crash originated from Csound's librtjack's > processCallback function. The log didn't have much more information > than that. My one thought from a quick look at the rtjack code is > that it might be possible that the void* arg passed into the function > might be NULL, maybe if csound deletes and cleans up just before an > jack callback gets called. My guess is that we might need a NULL > check at the top of the function and if NULL just return 0 (I don't > know what return values are usable there). > > Does anyone have any thoughts on if the above analysis is plausible, > and whether they might see some other problems? > > Thanks! > steven > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel ------------------------------------------------------------------------------ Managing the Performance of Cloud-Based Applications Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. Read the Whitepaper. http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net

Date	2014-02-20 16:12
From	Steven Yi
Subject	Re: [Cs-dev] Jack Crash
	That seems reasonable. Could you commit a change and we can ask Menno to test? On Thu, Feb 20, 2014 at 10:59 AM, Victor Lazzarini wrote: > I see this > > csound->DestroyGlobalVariable(csound, "_rtjackGlobals"); > if (p.client != (jack_client_t) NULL) { > / deactivate client / > //if (p.jackState != 2) { > //if (p.jackState == 0) > // csound->Sleep((size_t) > // ((int) ((double) (p.bufSize p.nBuffers) > // * 1000.0 / (double) p.sampleRate + 0.999))); > jack_deactivate(p.client); > //} > csound->Sleep((size_t) 50); > > surely the DestroyGlobalVariable() needs to happen after deactivation. > Maybe this line should be moved to after jack_deactivate(). > > Victor > On 20 Feb 2014, at 15:40, Steven Yi wrote: > >> Hi All, >> >> Menno reported to me a crash from Blue when a Csound run completed. >> The log reported the crash originated from Csound's librtjack's >> processCallback function. The log didn't have much more information >> than that. My one thought from a quick look at the rtjack code is >> that it might be possible that the void* arg passed into the function >> might be NULL, maybe if csound deletes and cleans up just before an >> jack callback gets called. My guess is that we might need a NULL >> check at the top of the function and if NULL just return 0 (I don't >> know what return values are usable there). >> >> Does anyone have any thoughts on if the above analysis is plausible, >> and whether they might see some other problems? >> >> Thanks! >> steven >> >> ------------------------------------------------------------------------------ >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >> Read the Whitepaper. >> http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk >> _______________________________________________ >> Csound-devel mailing list >> Csound-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/csound-devel > > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel ------------------------------------------------------------------------------ Managing the Performance of Cloud-Based Applications Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. Read the Whitepaper. http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net

Date	2014-02-20 17:02
From	Victor Lazzarini
Subject	Re: [Cs-dev] Jack Crash
	done. On 20 Feb 2014, at 16:12, Steven Yi wrote: > That seems reasonable. Could you commit a change and we can ask Menno to test? > > On Thu, Feb 20, 2014 at 10:59 AM, Victor Lazzarini > wrote: >> I see this >> >> csound->DestroyGlobalVariable(csound, "_rtjackGlobals"); >> if (p.client != (jack_client_t) NULL) { >> / deactivate client / >> //if (p.jackState != 2) { >> //if (p.jackState == 0) >> // csound->Sleep((size_t) >> // ((int) ((double) (p.bufSize p.nBuffers) >> // * 1000.0 / (double) p.sampleRate + 0.999))); >> jack_deactivate(p.client); >> //} >> csound->Sleep((size_t) 50); >> >> surely the DestroyGlobalVariable() needs to happen after deactivation. >> Maybe this line should be moved to after jack_deactivate(). >> >> Victor >> On 20 Feb 2014, at 15:40, Steven Yi wrote: >> >>> Hi All, >>> >>> Menno reported to me a crash from Blue when a Csound run completed. >>> The log reported the crash originated from Csound's librtjack's >>> processCallback function. The log didn't have much more information >>> than that. My one thought from a quick look at the rtjack code is >>> that it might be possible that the void* arg passed into the function >>> might be NULL, maybe if csound deletes and cleans up just before an >>> jack callback gets called. My guess is that we might need a NULL >>> check at the top of the function and if NULL just return 0 (I don't >>> know what return values are usable there). >>> >>> Does anyone have any thoughts on if the above analysis is plausible, >>> and whether they might see some other problems? >>> >>> Thanks! >>> steven >>> >>> ------------------------------------------------------------------------------ >>> Managing the Performance of Cloud-Based Applications >>> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >>> Read the Whitepaper. >>> http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Csound-devel mailing list >>> Csound-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/csound-devel >> >> >> ------------------------------------------------------------------------------ >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >> Read the Whitepaper. >> http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk >> _______________________________________________ >> Csound-devel mailing list >> Csound-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/csound-devel > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel ------------------------------------------------------------------------------ Managing the Performance of Cloud-Based Applications Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. Read the Whitepaper. http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net

Date	2014-02-20 17:04
From	Steven Yi
Subject	Re: [Cs-dev] Jack Crash
	Thanks! I've asked Menno to give it a try. On Thu, Feb 20, 2014 at 12:02 PM, Victor Lazzarini wrote: > done. > On 20 Feb 2014, at 16:12, Steven Yi wrote: > >> That seems reasonable. Could you commit a change and we can ask Menno to test? >> >> On Thu, Feb 20, 2014 at 10:59 AM, Victor Lazzarini >> wrote: >>> I see this >>> >>> csound->DestroyGlobalVariable(csound, "_rtjackGlobals"); >>> if (p.client != (jack_client_t) NULL) { >>> / deactivate client / >>> //if (p.jackState != 2) { >>> //if (p.jackState == 0) >>> // csound->Sleep((size_t) >>> // ((int) ((double) (p.bufSize p.nBuffers) >>> // * 1000.0 / (double) p.sampleRate + 0.999))); >>> jack_deactivate(p.client); >>> //} >>> csound->Sleep((size_t) 50); >>> >>> surely the DestroyGlobalVariable() needs to happen after deactivation. >>> Maybe this line should be moved to after jack_deactivate(). >>> >>> Victor >>> On 20 Feb 2014, at 15:40, Steven Yi wrote: >>> >>>> Hi All, >>>> >>>> Menno reported to me a crash from Blue when a Csound run completed. >>>> The log reported the crash originated from Csound's librtjack's >>>> processCallback function. The log didn't have much more information >>>> than that. My one thought from a quick look at the rtjack code is >>>> that it might be possible that the void* arg passed into the function >>>> might be NULL, maybe if csound deletes and cleans up just before an >>>> jack callback gets called. My guess is that we might need a NULL >>>> check at the top of the function and if NULL just return 0 (I don't >>>> know what return values are usable there). >>>> >>>> Does anyone have any thoughts on if the above analysis is plausible, >>>> and whether they might see some other problems? >>>> >>>> Thanks! >>>> steven >>>> >>>> ------------------------------------------------------------------------------ >>>> Managing the Performance of Cloud-Based Applications >>>> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >>>> Read the Whitepaper. >>>> http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> Csound-devel mailing list >>>> Csound-devel@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/csound-devel >>> >>> >>> ------------------------------------------------------------------------------ >>> Managing the Performance of Cloud-Based Applications >>> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >>> Read the Whitepaper. >>> http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Csound-devel mailing list >>> Csound-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/csound-devel >> >> ------------------------------------------------------------------------------ >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >> Read the Whitepaper. >> http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk >> _______________________________________________ >> Csound-devel mailing list >> Csound-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/csound-devel > > > ------------------------------------------------------------------------------ > Managing the Performance of Cloud-Based Applications > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. > Read the Whitepaper. > http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk > _______________________________________________ > Csound-devel mailing list > Csound-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/csound-devel ------------------------------------------------------------------------------ Managing the Performance of Cloud-Based Applications Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. Read the Whitepaper. http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk _______________________________________________ Csound-devel mailing list Csound-devel@lists.sourceforge.net