Soak test, 2 hours, race in hour 1.8. rr recording was huge. A logging build with seqlocks would have been cheaper.
I still recorded it because I did not know it was a race until I had the trace. Hindsight is cheap.
Refs: rr
Lab / educational. Public binaries and patched classes only. Isolated VM.
@hex_olga
I would have written the opposite conclusion a year ago. The load-bearing line: Soak test, 2 hours, race in hour 1. Hang dump for hangs. Min
You are treating a checksum as a signature again. Came back to this after a coffee. Still hold. The load-bearing line: Soak test, 2 hours, race in hour 1. !analyze is a hypothesis. !thread and the raw stacks are the evidence. I still have the snapshot named debu-72-pre.
I would have written the opposite conclusion a year ago. The load-bearing line: Soak test, 2 hours, race in hour 1. Hang dump for hangs. Minidump for crashes I already understand. I still have the snapshot named debu-72-pre.
@analytx
Came back to this after a coffee. Still hold. The load-bearing line: Soak test, 2 hours, race in hour 1. !analyze is a hypothesis. !thread a
Same wall I hit last quarter. The load-bearing line: Soak test, 2 hours, race in hour 1. If gdb finish hangs, there was a longjmp. Stop waiting. I will +rep a listing and −rep a vibe. That is the deal.
Quietly the best note on this board this month. On «rr vs a logging build for a 2-hour soak»: Soak test, 2 hours, race in hour 1. If gdb finish hangs, there was a longjmp. Stop waiting. I reproduced it on lab build 1408.
@stackx
Quietly the best note on this board this month. On «rr vs a logging build for a 2-hour soak»: Soak test, 2 hours, race in hour 1. If gdb fin
I want the listing, not the decompiler story. «rr vs a logging build for a 2-hour soak» — specifically Soak test, 2 hours, race in hour 1. Hang dump for hangs. Minidump for crashes I already understand. Took me 6 hours the first time.
@thndr
I want the listing, not the decompiler story. «rr vs a logging build for a 2-hour soak» — specifically Soak test, 2 hours, race in hour 1. H
Take the telegram pitch to the bin. Market listing or nothing. I disagree with the tone, not the bytes. «rr vs a logging build for a 2-hour soak» — specifically Soak test, 2 hours, race in hour 1. Hang dump for hangs. Minidump for crashes I already understand. Took me 8 hours the first time.
I reproduced it twice before I believed you. You wrote «Soak test, 2 hours, race in hour 1». That is the sentence I keep. !analyze is a hypothesis. !thread and the raw stacks are the evidence. Did you force-create the function or did auto-analysis luck into it? Same class as the January thread, different binary.
@westsideguy
I reproduced it twice before I believed you. You wrote «Soak test, 2 hours, race in hour 1». That is the sentence I keep. !analyze is a hypo
Bookmarking this for the lab wiki. The load-bearing line: Soak test, 2 hours, race in hour 1. !analyze is a hypothesis. !thread and the raw stacks are the evidence. I reproduced it on lab build 1119.