Generally it is not a problem at all!

I did have a serious problem however with a Philips RTC chip on an I2C bus that did not like having its transfer interrupted during an I2C transaction. The RTC would go into its RTC update cycle every so once in a while and drop the ability to respond with an ACK to the current transfer that had ben interrupted. I fixed that problem by doing two things. A) I used the RTC's block read and write modes as opposed to the single byte read and write modes. B) I disabled interruputs during the part of the software bit-banged data clocking through the ACK pulse. Interrupts turned out to be OK int he time between the byte+ACK transfers.

Michael Karas