I've compared charge injection runs 308553662 collected on October 12, 2010 and 306938342 collected on September 23, 2010. The modified version of flight software was uploaded between these two runs.
The following plot shows the difference between charge injection curves in these two runs for all CAL channels.
Four panels correspond to different energy ranges: LEX8 (top left), HEX8 (top right), LEX1 (bottom left), HEX1 (bottom right).
For LEX8 and HEX8 we see no difference except noise, for LEX1 and HEX1 ranges there is some very small systematic deviation ~0.5 ADC units on full dynamic range or ~0.01%.
Conclusion: the system update caused no change in charge injection calibration.