b4dda138 imp Jan. 4, 2022, 11:01 p.m.
So #ifdef the clk.h include aarch64. Otherwise the right kernel options
aren't always present.

Sponsored by:		Netflix
c47e2e51 imp Jan. 4, 2022, 11:01 p.m.
clknode_if.h is needed for this device on aarch64. Create it always.

Sponsored by:		Netflix
44796b7e imp Jan. 4, 2022, 11:01 p.m.
The saf1761 OTG support was only for mips targets (BERI?). Retire it.

Sponsored by:		Netflix
Reviewed by:		brooks
Differential Revision:	https://reviews.freebsd.org/D33706
822fa675 jhb Jan. 4, 2022, 10:46 p.m.
Fixes:		74d3f1b63dbe OCF: Add crypto_clonereq().
8903d8e3 jhb Jan. 4, 2022, 10:37 p.m.
This matches icl_conn_task_setup() which passes the PDU and avoids the
need for a layering violation in cxgbei to fetch the request PDU from
the ctl_io.

Reviewed by:	mav
Sponsored by:	Chelsio Communications
Differential Revision:	https://reviews.freebsd.org/D33746
e43cf698 jhb Jan. 4, 2022, 10:22 p.m.
Previously the driver duplicated code from cryptosoft.c to handle
certain edge case AES-CCM and AES-GCM requests.  However, this
approach has a few downsides:

1) It only uses "plain" software and not accelerated software since it
   uses enc_xform directly.

2) It performs the operation synchronously even though the caller
   believes it is invoking an async driver.  This was fine for the
   original use case of requests with only AAD and no payload that
   execute quickly, but is a bit more disingenuous for large requests
   which fall back due to exceeding the size of a firmware work
   request (e.g. due to large scatter/gather lists).

3) It has required several updates since ccr(4) was added to the tree.

Instead, allocate a software session for AES-CCM and AES-GCM sessions
and dispatch a cloned request asynchronusly to the software session.

Reviewed by:	markj
Sponsored by:	Chelsio Communications
Differential Revision:	https://reviews.freebsd.org/D33608
74d3f1b6 jhb Jan. 4, 2022, 10:22 p.m.
This function clones an existing crypto request, but associates the
new request with a specified session.  The intended use case is for
drivers to be able to fall back to software by cloning a request and
dispatch it to an internally allocated software session.

Reviewed by:	markj
Sponsored by:	Chelsio Communications
Differential Revision:	https://reviews.freebsd.org/D33607
753c8513 jhb Jan. 4, 2022, 9:37 p.m.
Obtained from:	CheriBSD
f3d7ace4 imp Jan. 4, 2022, 9:12 p.m.
It turns out that we still need xlocal.h protection for when we're
cross building on Linux. Linux doesn't have this file, but os/x
does. Before, we'd assume we didn't have it, like old FreeBSD, when
cross compiling. After the latest update, all that code was removed so
cross compiling needs to be handled separaetly. Do so by defining
HAVE_XLOCALE_H only when we're not building on linux. This allows us
to build the mkmagic build tool needed to build file(1).

Sponsored by:		Netflix
Reviewed by:		jrtc27
Differential Revision:	https://reviews.freebsd.org/D33741
d373e61e imp Jan. 4, 2022, 8:57 p.m.
Sort arm alphabetically.

Reviewed by:		emaste
Differential Revision:	https://reviews.freebsd.org/D33744

Sponsored by:		Netflix
3a9688f8 jhibbits Jan. 4, 2022, 3:24 p.m.
The original logic was to check if there's no filter and the address is
misaligned relative to the requirements.  The refactoring in
c606ab59e7f9423f7027320e9a4514c7db39658d missed this, and instead caused
it to return failure if the address *is* properly aligned.
ad07e93f cy Jan. 4, 2022, 3:11 p.m.
Reported by:	jrtc27
Fixes:		2582ae5740181e0d2bab10003d66ae91c9b56329
MFC after:	1 month
ed166a01 markj Jan. 4, 2022, 2:46 p.m.
The error message is useful for diagnosing mvneta_attach() failures.

MFC after:	1 week
e2576591 markj Jan. 4, 2022, 2:46 p.m.
Otherwise MSG_CMSG_CLOEXEC is not decoded.

MFC after:	1 week
Sponsored by:	The FreeBSD Foundation
f7b04c53 trasz Jan. 4, 2022, 2:39 p.m.
No functional changes (except for the uprintf).

Discussed With:	kib
Sponsored By:	EPSRC