On 06/21/2016 05:51 PM, Joe Hillenbrand wrote: > I gathered the requested output. Unfortunately, it doesn't look very > helpful. Each command took about +95 minutes. Thanks, there are versions of backend libraries I need, so it is helpful. ALso it proves that for unlocking crypto works. >> cryptsetup benchmark -h sha256 --debug > > # cryptsetup 1.7.1 processing "cryptsetup benchmark -h sha256 --debug" > # Running command benchmark. > # Installing SIGINT/SIGTERM handler. > # Unblocking interruption on signal. > # Tests are approximate using memory only (no storage IO). > # Crypto backend (gcrypt 1.7.1) initialized in cryptsetup library version 1.7.1. > # Detected kernel Linux 4.6.2-1-ARCH x86_64. > PBKDF2-sha256 N/A > Command failed with code 22. It is failing the benchmark here, so there is probably another bug in benchmarking. What CPU is that? Is it some very fast machine? (paste lscpu output if possible). What is number of iterations for the used slot? - you masked that out in the first mail. I need to either reproduce myself or I have to add some more debugging output there. If I add some debugging code, can you compile it and run it on your hw? (Or I can provide debug binaries if you prefer that.) Thanks, Milan _______________________________________________ dm-crypt mailing list dm-crypt@xxxxxxxx http://www.saout.de/mailman/listinfo/dm-crypt