As of the [pernosco session](https://pernos.co/debug/FrbZoY4merqIPdBpVH14vA/index.html#f{m[DAcH,Boc9_,t[3Q,+SU_,f{e[DAcH,BoKZ_,s{afISCIAAA,bAbo,uM+8b9g,oM/gdLQ___/) it looks like [we hit the JS memory limit](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#488) (where `gcMaxBytes_.value=4294967295`) and then [GCRuntime::attemptLastDitchGC](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#240-256) is what we try currently, and it seems to be [only ever called here](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#229) in `RetryTenuredAlloc`. Not sure if that would be the best moment, but it looks like a last possible attempt one could make easily? (In reply to Andrew McCreight [:mccr8] from comment #51) > Boris spent like 3 months investigating this crash with adding lots of instrumentation (starting with bug 1491313 and ending with removing it in bug 1491925) so I'm not sure how fruitful that will be. Well, that was pre-pernosco, AFAICT. The pernosco session might also give some hint on what is actually holding that JS memory, but I don't know how to inspect that easily.
Bug 1405521 Comment 53 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
As of the [pernosco session](https://pernos.co/debug/FrbZoY4merqIPdBpVH14vA/index.html#f{m[DAcH,Boc9_,t[3Q,+SU_,f{e[DAcH,BoKZ_,s{afISCIAAA,bAbo,uM+8b9g,oM/gdLQ___/) it looks like [we hit the JS memory limit](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#488) (where `gcMaxBytes_.value=4294967295`) and then [GCRuntime::attemptLastDitchGC](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#240-256) is what we try currently, and it seems to be [only ever called here](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#229) in `RetryTenuredAlloc`. Not sure if that would be the best moment, but it looks like a last possible attempt one could make easily? (In reply to Andrew McCreight [:mccr8] from comment #51) > Boris spent like 3 months investigating this crash with adding lots of instrumentation (starting with bug 1491313 and ending with removing it in bug 1491925) so I'm not sure how fruitful that will be. Well, that was pre-pernosco, AFAICT. The pernosco session might also give some hint on what is actually holding that JS memory, but I don't know how to inspect that easily (or how representative that would be for all the crashes).
As of the [pernosco session](https://pernos.co/debug/FrbZoY4merqIPdBpVH14vA/index.html#f{m[DAcH,Boc9_,t[3Q,+SU_,f{e[DAcH,BoKZ_,s{afISCIAAA,bAbo,uM+8b9g,oM/gdLQ___/) it looks like [we hit the JS memory limit](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#488) (where `gcMaxBytes_.value=4294967295`) and then [GCRuntime::attemptLastDitchGC](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#240-256) is what we try currently, and it seems to be [only ever called here](https://searchfox.org/firefox-main/rev/d0ff31da7cb418d2d86b0d83fecd7114395e5d46/js/src/gc/Allocator.cpp#229) in `RetryTenuredAlloc`. Not sure if that would be the best moment to CC, but it looks like a last possible attempt one could make easily? (In reply to Andrew McCreight [:mccr8] from comment #51) > Boris spent like 3 months investigating this crash with adding lots of instrumentation (starting with bug 1491313 and ending with removing it in bug 1491925) so I'm not sure how fruitful that will be. Well, that was pre-pernosco, AFAICT. The pernosco session might also give some hint on what is actually holding that JS memory, but I don't know how to inspect that easily (or how representative that would be for all the crashes).