Do the assumptions for linear regression apply to AR(p) models?

If we have a stationary time series and we want to model it as an AR(p) process, what conditions must hold besides the stationarity itself?

Are they the same a the assumptions for linear regression:

• Linear relationship (this one seems obvious)
• Gaussian errors
• No collinearity
• Homoscedasticity (I suppose this is covered by the stationarity requirement anyway?)
• Assumptions are used to ensure properties of estimators. Different assumptions yield different properties. Tell us what properties you are interested in, then we will be able tell you what assumptions are needed to ensure them. – Richard Hardy May 3 '19 at 8:56