On Tue, 2021-11-30 at 15:14 -0600, Benjamin Marzinski wrote: > > > > If you think we should use a different strategy, please explain. > > We *could* try to go on even after encountering broken designators, > > assuming the length field is correct if it's within the given > > limits, > > even if the rest is bogus. So baiscally instead of the break > > statement > > above, we'd continue the loop. > > > > Would you prefer that? > > Again looking at sg_inq, it will loop through all the descriptors, > trusting the length field, until it gets to the whole data length. It > prints a warning if it doesn't end up at exactly the data length, but > still exports all the IDs it finds. If an individual descriptor > doesn't > make sense, it gets skipped. That would be my preference. Ok, I'll try to fix it up. Martin -- dm-devel mailing list dm-devel@xxxxxxxxxx https://listman.redhat.com/mailman/listinfo/dm-devel
