What seems to be more interesting is xsaveopt, which I think could
make a difference. That would, however, need xsave are to be at
constant address. I've been thinking about storing it next to TEB, but
we can't do that as long as winsock is called on signal stack, so I
left experimenting with it for the future.

xsavec also performs an optimization (doesn't save the xstate in initial
state),

As I said, xsave already does that optimization, there is no need for xsavec for that.

 and I think it should not depend on the save address

Constant save address is for a different optimization: xsaveopt does not need to store parts that didn't change since the last xrstor. It means that it would save a lot of sores in applications rarely using AVX that xsave/xsavec can't optimize.

Jacek