I should have said 'the magnitude of the material copied' since I'm not trying to make a judgement one way or other whether this rises to a copyright violation, which is a legal term of art no body on this review is licensed to proffer.

I'd not make this change. The vast majority of the copyright protectable parts of this file was developed by Andrew.

The more standard wording in the tree is 'derived from' which is why I suggested that.

The wording "and subject to" concerns me. I don't think it means anything legally here and I'd rather not imply they are 'subject to' a copyright without there being an attendant license statement.
A quick grep shows that we have about two dozen instances that use 'derived from' and none with the 'copied and modified' in the tree today.
There's also a smattering of other phrases like 'inspired by' and the like that I could quickly locate.
In code we've removed there was also "This software is a derivative work of if_ed.c version 1.56 by David Greenman" followed by his copyright for a similar copy but replace that's happened here.

The original license in aesni_wrap.c just says that the copyright needed to be preserved along with the disclaimer. Since this file is licensed under the same license, both those terms are satisfied by my suggestion. We should, imho, use the wording we use elsewhere here because the more variations we have in wording means more work for people doing license audits.

please include links to these examples of derived so I can compare wordings. I attempted to find derived from, but did not see any that resembled this wording.

The reason I included the subject to is that just because they are derived works does NOT remove their original copyright.

The originally the copyright was not preserved, which started this whole problem. If it had been retained like it was required, we wouldn't be here today.

Neither of us are lawyers, maybe you should get a lawyer to chime in on this, but IMO, the proposed derived from wording implied that the copyright does not apply any longer and I disagree w/ that.

I used the following grep to find it:

% grep -r 'derived from' sys | grep ":.*\.[ch]"
sys/arm64/broadcom/genet/if_genetreg.h:/* derived from NetBSD's bcmgenetreg.h */
sys/cam/nvme/nvme_xpt.c: * derived from ata_xpt.c: Copyright (c) 2009 Alexander Motin <mav@FreeBSD.org>
sys/contrib/device-tree/include/dt-bindings/reset/altr,rst-mgr-s10.h: * derived from Steffen Trumtrar's "altr,rst-mgr-a10.h"
sys/contrib/openzfs/module/os/freebsd/zfs/crypto_os.c: * Portions of this file are derived from sys/geom/eli/g_eli_hmac.c
sys/contrib/openzfs/module/icp/algs/aes/aes_impl_generic.c: * This file is derived from the file  rijndael-alg-fst.c  taken from the
sys/xen/interface/io/tpmif.h: * This code has been derived from tools/libxc/xen/io/netif.h
sys/dev/exca/excavar.h: * This software may be derived from NetBSD i82365.c and other files with
sys/dev/exca/excareg.h: * This software may be derived from NetBSD i82365.c and other files with
sys/dev/exca/exca.c: * This software may be derived from NetBSD i82365.c and other files with
sys/dev/sound/pci/via8233.c: * Portions of this code derived from via82c686.c:
sys/dev/sound/usb/uaudio.c:/* The following table is derived from Linux's quirks-table.h */
sys/dev/pcf/pcf_isa.c: * derived from sys/i386/isa/pcf.c which is:
sys/dev/usb/misc/udbp.h: * This file was derived from src/sys/netgraph/ng_sample.h, revision 1.1
sys/dev/bhnd/cores/pmu/bhnd_pmureg.h: * This file is derived from the sbchipc.h header contributed by Broadcom
sys/dev/bhnd/cores/pmu/bhnd_pmu_private.h: * This file is derived from the hndpmu.h header contributed by Broadcom
sys/dev/bhnd/cores/pmu/bhnd_pmu_subr.c: * This file is derived from the hndpmu.c source contributed by Broadcom
sys/dev/bhnd/cores/pcie2/bhnd_pcie2_reg.h: * This file is derived from the pcie_core.h and pcie2_core.h headers
sys/dev/bhnd/cores/pci/bhnd_pcireg.h: * This file is derived from the hndsoc.h, pci_core.h, and pcie_core.h headers
sys/dev/bhnd/cores/chipc/pwrctl/bhnd_pwrctl_subr.c: * This file is derived from the siutils.c source distributed with the
sys/dev/bhnd/cores/chipc/pwrctl/bhnd_pwrctl.c: * Portions of this file were derived from the siutils.c source distributed with
sys/dev/bhnd/cores/chipc/chipcreg.h: * This file is derived from the sbchipc.h header contributed by Broadcom
sys/dev/bhnd/bhndb/bhndb_pcireg.h: * Portions of this file were derived from the bcmdevs.h header contributed by
sys/dev/bhnd/bcma/bcma_eromreg.h: * Portions of this file were derived from the aidmp.h header
sys/dev/bhnd/bcma/bcma_dmp.h: * Portions of this file were derived from the aidmp.h header
sys/dev/bhnd/bhnd_ids.h: * This file is derived from the bcmdevs.h header contributed by Broadcom
sys/dev/bhnd/bhndreg.h: * Portions of this file were derived from the sbchipc.h header contributed by
sys/dev/bhnd/siba/sibareg.h: * This file was derived from the sbconfig.h header distributed with
sys/mips/broadcom/bcm_bmips_exts.h: * This file is derived from the sbmips32.h header distributed
sys/mips/include/elf.h:/* FreeBSD specific bits - derived from FreeBSD specific files and changes to old elf.h */
sys/netipsec/keysock.c:/* This code has derived from sys/net/rtsock.c on FreeBSD2.2.5 */

So ~30 files that use some variation of this language. There's a few old drivers that we removed that used 'derived from' or 'derivative work' both of which are standard terms in this area. "subject to" is something I've not seen in the last 20 years of consulting on licenses and such. So while I'm not a lawyer, I think that experience is at least suggestive that we should avoid it. Neither of us being lawyers: you are right. However, your license didn't specify a wording, so using a similar wording to what we use elsewhere would make it easier for people. For these two reasons, I'd strongly suggest we use 'derived from' since in my opinion it matches the prior uses of this wording for what happened here.

The originally the copyright was not preserved, which started this whole problem. If it had been retained like it was required, we wouldn't be here today.

Note that it is restored to this file in the lines below. There was no specification as to where and how it must be preserved, and I maintain that these lines below restore that in a way that's compliant with the license.