Took an early Roth Distribution Feb 4 2021 for what was thought to be 2020 Overfunding, but was not. 1099-R Received 2021

We thought that we overfunded my wife's (age in her forties) 2020 Roth. Took an early distribution 2/4/2021 to remove $1289.75 (thought to be an overcontribution, but was not) + $95.80 earnings + 27.38 Fed W/holding + 13.69 Ore State W/H for a total of $1,426.62. (We initially didn't add her self-employment to her W-2 earnings and thought she had not made enough money for a full $6,000 Roth contribution.)  Then, we discovered that we could fully fund her Roth to the $6,000 limit and we added $1289.75 back into her Roth on 2/16/21. The distribution had already occurred, so Vanguard computed the earnings of 95.80 that they said would be taxable for 2020. I don't understand that totally since we received the distribution in 2021, but they were made in 2020. The sent us the 1099-R for the early distribution as a 2021 1099-R that has the Codes J,P. When I imported the 1099-R to my 2021 TurboTax, it advised me that it appears that I must amend the 2020 return because "Your 1099-R reports earnings on a contribution you made in 2020. This information needs to go on your 2020 return."

 

Q1: JP applies to and overcontribition, but since it really did not turn out to be, will we be penalized for it anyway?

 

Q2: The $95.80 earnings were made on that $1289.75 contribution, but weren't they legitimate earnings since they were actually earned on funds that were in not in excess of the allowed total contribution and that distributed money was reinvested back into the 2020 Roth 14 days after the errant distribution? I presume that we have to pay taxes on them because they were sent to us and could not be returned to the Roth. Do we indeed amend the 2020 in order to pay the taxes owed on those earnings?

 

Q3: The total distribution of 1426.62 is more than the $1289.75 "overcontribution" due to the earnings and tax withholding. Are we to pay taxes and/or early withdrawal penalties on the full $1426.62?

 

Sorry if this is asking similar Q's addressed last year that I found, but none of those involved a distribution for an overcontribution that was in fact, not actually an overcontribution.