It is, but it upsets editors and reviewers, who already invested time in the previous version, wonder why you did not do true diligence before submitting, and who are now wondering whether what you send is the final word.
That being said, alternative plans of actions that might be preferable are:
- retract the paper because you realized it is not done yet, apologizing profusely about wasting people's time.
- wait for the reviewer comments and then update, if the changes are small. (You also need to point this out in your answer to the reviewers.)*
- put the new results into a new paper if they are really significant.
Added: If in option 2, you change your results, then it needs to be reviewed again for the changes. If you do not alert to this, then you could be engaging in academic misconduct. Thanks to semmyk-research for underlying the importance of this parenthetical remark.