swrast: use better _swrast_compute_lambda() function

swrast: use better _swrast_compute_lambda() function

The MAX-based function can produce values that are non-monotonic for a span
which causes glitches in texture filtering.  The sqrt-based one avoids that.

This is perhaps slightly slower than before, but the difference
probably isn't noticable given we're doing software mipmap filtering.

Issue reported by Nir Radian <nirr@horizonsemi.com>