>Poll for TRAN state if the ioctl command will eventually return to TRAN > >The ioctl submitted command should not be considered completed until >the card has returned back to TRAN state. Waiting just for the card >to no longer signal busy is not enough as they might remain in a >non-busy PROG state for a while after the command. >Further commands requiring TRAN will fail then. >It should not be the responsibility of the user to check if their command >has completed until sending the next via ioctl, >instead the check should be made here. >So now, in doubt, wait for TRAN except for the few commands that will >never return to TRAN state. So apart from the fact that I missed a couple of non-TRAN returning MMC commands, which I will add in v2, are there any other thoughts about this patch? It would change the behavior of the ioctl interface, but I think it is the only way to prevent race conditions here. Best Regards, Christian Hyperstone GmbH | Line-Eid-Strasse 3 | 78467 Konstanz Managing Directors: Dr. Jan Peter Berns. Commercial register of local courts: Freiburg HRB381782
