Closed jafarsarif10 closed 1 year ago
Hey @jafarsarif10, thanks for reaching out.
Thanks, @ksteuck for responding to my query. I have recently started experimenting with Silifuzz, but I encountered an error because the args for the orchestrator have changed. Here's the error message I received:
./orchestrator/silifuzz_orchestrator_main --duration=30s --runner=./runner/reading_runner_main_nolibc --shard_list_file="/tmp/wd/runnable-corpus.00001" E0314 07:14:12.904824 2409483 byte_io.cc:37] Error reading from fd 3: errno=21 Segmentation fault
It's worth noting that I'm able to execute the same runnable corpus without any errors using the runner binary. Could you please help me identify what I'm doing wrong? Also, if possible, could you update the wiki with this information?
Hey @jafarsarif10 ,
the --shard_list_file
flag expects a file name containing a list of shard paths separated by line breaks. E.g. in your case the following should work:
./orchestrator/silifuzz_orchestrator_main --duration=30s --runner=./runner/reading_runner_main_nolibc --shard_list_file=shards.txt
where shards.txt contents is
/tmp/wd/runnable-corpus.00001
Hi, We are testing silifuzz to detect defects/bugs on the CPU. I am writing to ask few questions, which I have listed down below: 1) The Corpus provided https://github.com/google/silifuzz#using-pre-generated-corpus, what type of instructions are included in the corpus. 2) Can you update the WIKI for the latest code change. 3) In this commit d9056056306b9871798fc8f318e30f3cc1d3fe4b, supported platform included intel-sapphire-rapids, does unicorn also support saphire-rapids new instructions.