when autocorrelated errors are correlated

[remuct]

Hello,
I am estimating a system of equations with correlated errors, where each error also exhibits autocorrelation. To address this, I specified a system in which each equation includes the component +[ar(1)=c(i)] to account for first-order autocorrelation. I then estimated the entire system using the SUR method in EViews.
EViews successfully performs this estimation, and I would like to understand what method is implemented for the estimation process regarding the two issues of autocorrelation and correlation. According to the tutorial, EViews uses maximum likelihood method to simultaneously estimate the autocorrelation coefficients and the structural coefficients. How is this procedure linked with the FGLS method typically involved in SUR? Does Eviews reverse the process and maximize the joint likelihood of the correlated errors, although my program ends with the instruction syst.sur ?
Thank you for your clarifications