A64FX: Add support for SVE to SGEMV/DGEMV kernels.

Add missing parentheses

Fix axpby_rvv kernels for cases where inc_y = 0

Improve the error message for (p)thread creation failure

The following openblas_utest tests fail when the RISCV64_ZVL128B is
enabled.

TEST 89/103 axpby:zaxpby_inc_0 [FAIL]
TEST 92/103 axpby:caxpby_inc_0 [FAIL]
TEST 95/103 axpby:daxpby_inc_0 [FAIL]
TEST 98/103 axpby:saxpby_inc_0 [FAIL]

The issue is that the vectorized kernels do not work when inc_y == 0.
This patch updates the kernels to fall back to the scalar algorithms
when inc_y == 0, fixing the failing tests.

Signed-off-by: Mark Ryan <markdryan@rivosinc.com>

Suffix BUFFER_SIZEs on POWER as UL to prevent int overflow in computations

Fix NAN handling in ARM/generic SCAL; have AzureCI Windows show errors on failure

Vectorize SBGEMM incopy for Power10 - 4x faster.

Fix build on FreeBSD/powerpc64*

Rewrite "Install OpenBLAS" docs page

Fix core assignment in cpu detection for Intel family 15

Add support for RISCV64_GENERIC in cmake

openblas: fix `BUFFERSIZE` value

Add another CPUID for Intel Meteor Lake

fix CI job to deploy docs, and make it run on pull requests too

AzureCI: Update OSX jobs to use the macos-12 image

Improve documention content, formatting, and html theme

ignore the gesdd failure on codspeed

In https://github.com/OpenMathLib/OpenBLAS/issues/4776
we're hitting
** On entry to SLASCL parameter number  4 had an illegal value

on codspeed, but not outside (either locally or on github runners)

Guard against invalid thread_status.queue

BENCH: sync codspeed-benchmarks with BLAS-benchmarks

Correctly detect ARM Neoverse V2 CPUs.