Lucas Manning
Lucas Manning
I tried my own hand at getting more speed with TPACKET_V3 and ran into a similar latency/throughput tradeoff issue. Even with the kernel fix @hbhasker mentioned packet_mmap v1/2 is still...
Earlier this year I ran experiments across lots of different frame/block sizes with tPACKET v3. The conclusion I came to was that PACKET_MMAP trades off raw packet throughput for CPU...
Yeah could be. I definitely did less testing on V2 since I thought v3 would be the answer. FWIW when I tried to increase the number of frames (in v2)...
I believe mountpoint-s3 is a FUSE, and gVisor's utimensat implementation calls vfs.SetStatAt to update timestamps. It could be an issue with our FUSE implementation.
Looks like a use-after-free to me. This could be coming from inside netstack or from your application which uses reference counted parts of the netstack API. Looking through the netstack...
Also, how frequently is this crash happening? Is it occasional or every time you try to run this code?
Thanks for the extra info. I wasn't able to determine the root cause after looking through both the gVisor UDP code and your code for some time. I will be...
@ignoramous #10958 may fix this issue. Let me know if you see any improvement once it's merged.
I think it was a race condition, not a leak. It should be fixed by #10958 so no need to create a new issue.
Kevin's hypothesis turned out to be incorrect. There was not any leak in Forwarder.HandlePacket. The clone() is already decref()'d properly in all the places it needs to be. That's what...