I didn't realize that your situation was identical in this respect to that of hl438. If you made an HSA contribution for 2021 or 2022 it seems that the HSA custodian thought that you were correcting an excess contribution for one of those years. If you made no HSA contribution for 2021 or 2022, the HSA custodian should have rejected what it thought was any attempt to obtain in 2022 any return of contribution that would be reportable with code 2. It seems that it is not unusual for HSA custodians to not any responsibility for making this sort of error.
Unfortunately, for dealing with erroneous Forms 1099-SA the IRS provides no form similar to the Form 4852 that would be used to provide the correct information and explanation. to get the numbers to appear on Forms 8889 and 5329 as I described you would have to enter this into TurboTax as a code-1 Form 1099-R, then provide an explanation statement, otherwise the IRS might mistakenly think that you over-reported your tax liability. I don't think that there is any way to include such a statement in an e-filed tax return using TurboTax, so I think the statement needs to be prepared outside of TurboTax and included with your mailed tax return. The code-1 Form 1099-SA that you enter should have the same gross amount in box 1 but nothing in box 2; box 2 only has meaning on a code-2 Form 1099-SA. Hopefully the gross amount distributed was not less than the amount of the excess contribution, otherwise some excess remains in the HSA. If more than the amount of the excess was distributed, you could potentially apply that to qualified medical expenses to avoid the tax and penalty on that portion.
Because both HSA custodians and IRA custodians are often inept at making the correct type of distribution to remove an excess after the due date of the tax return, it's best to tell the IRA custodian nothing about the distribution being to correct an excess and just ask for a regular distribution, but it's too late for that in this case. I only mention this so that it might help someone else avoid getting the wrong type of distribution.
