The function opencodes for_each_cpu() by using a plain for-loop. The
loop calls cpumask_weight() inside the conditional section. Because
cpumask_weight() is O(1), the overall complexity of the function is
O(node * node_cpus^2). Also, cpumask_nth() internally calls hweight(),
which, if not hardware accelerated, is slower than cpumask_next() in
for_each_cpu().

If switched to the dedicated for_each_cpu(), the rebalance_wq_table()
can drop calling cpumask_weight(), together with some housekeeping code.
This makes the overall complexity O(node * node_cpus), or simply speaking
O(nr_cpu_ids).

While there, fix opencoded for_each_possible_cpu() too.

Signed-off-by: Yury Norov <yury.norov@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

In preparation for making the kmalloc family of allocators type aware,
we need to make sure that the returned type from the allocation matches
the type of the variable being assigned. (Before, the allocator would
always return "void *", which can be implicitly cast to any pointer type.)

The assigned type is "struct idxd_wq **", but the returned type will be
"struct wq **". These are the same size allocation (pointer sized), but
the types don't match. Adjust the allocation type to match the assignment.

Signed-off-by: Kees Cook <kees@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Use the common reqsize field for acomp algorithms.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Rather than copying the request by hand, use the ACOMP_FBREQ_ON_STACK
helper to do it.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Remove the disable_async field left over after the NULL dst removal.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The req->base.data field is for the user and must not be touched by
the driver, unless you save it first.

The iaa driver doesn't seem to be using the req->base.data value
so just remove the assignment.

Fixes: 09646c98d0bf ("crypto: iaa - Add irq support for the crypto async interface")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Rather than passing around a CRC between the functions, embed it
into the acomp_request context.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Use ACOMP_REQUEST_ON_STACK instead of allocating legacy fallback
compression transform.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Remove the unused dst_null support.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Test the correct flags for the MAY_SLEEP bit.

Fixes: 2ec6761df889 ("crypto: iaa - Add support for deflate-iaa compression algorithm")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>