Spec: 4.4.4 and 4.5: incorrect uses of `axpy`

[tgeoghegan]

2.1 defines: axpy(n, Y, A, X): sets Y[i] += A*X[i] for 0 <= i < n and clearly A is a scalar. In 4.4.4, dot_proof includes:

axpy(DBLOCK, y[0..DBLOCK], Af[0..DBLOCK], T[i + IW][0...DBLOCK])

Af[0..DBLOCK] is clearly an array. There's a similar problem in the loop of quadratic_proof, and 4.5's quadratic_check.

What's missing here is a definition vaxpy(n, Y, A, X) where A is an array or vector, matching the one in lib/algabra/blas.h.