On Tue, 2018-06-26 at 21:15 +0000, Winkler, Tomas wrote: > > Right now if I really put head into this I can understand the logic but it > > is a > > complete mess. > > I think what is the mess is that we have a recursive call to tpm_transmit > topped with retries. All other mess is just the result of that. > > > I will forgot the dependencies between flags within few > > weeks. > > Hope the reasons are well documented both in code and the commit message, if > not let's address that. We really cannot depend on one's memory. > It's not like I'm not striving for simplest possible code. > > > A fixed requirement (so that you know) is that they must be > > independent. > > The flags (hope this what you referring here to) are not independent and > weren't before, (RAW cannot be called alone as you will have double locking) > putting them under one name just should make it clear. > I beg you to go over the code one more time, don't get stuck with flags > names, maybe you even discover some real issue. > > Thanks > Tomas You should then find a solution where you can remove TPM_TRANSMIT_RAW completely and make it as a separate commit, not part of the bug fix. This is not in a shape that I would dare to put this in a pull request. /Jarkko
