jsg [Mon, 26 Apr 2021 04:28:07 +0000 (04:28 +0000)]
Remove some commented out entries to make this easier to read
and enable openprom.
ok visa@
mortimer [Sun, 25 Apr 2021 16:36:56 +0000 (16:36 +0000)]
Bump pbuild stacksize to 8M.
Addresses a stack exhaustion issue with llvm11 and a small number of ports.
ok kettenis@
stsp [Sun, 25 Apr 2021 15:59:56 +0000 (15:59 +0000)]
Asus USB-N10 should work with urtwn(4). Patch by Jan Vlach.
stsp [Sun, 25 Apr 2021 15:59:19 +0000 (15:59 +0000)]
Make urtwn(4) attach to 'ASUS USB-N10 Nano' devices.
Patch by Jan Vlach.
stsp [Sun, 25 Apr 2021 15:58:14 +0000 (15:58 +0000)]
regen
stsp [Sun, 25 Apr 2021 15:58:01 +0000 (15:58 +0000)]
Add USB product ID for ASUS RTL8188EUS. Patch by Jan Vlach.
stsp [Sun, 25 Apr 2021 15:32:21 +0000 (15:32 +0000)]
Implement support for Rx aggregation offload in iwm(9) and iwx(4), and
re-enable de-aggregation of A-MSDUs in net80211 for all drivers capable
of 11n mode. This can provide improved Rx performance if the access point
supports transmission of A-MSDUs nested in A-MDPUs.
iwm(9) 9k and iwx(4) devices de-aggregate A-MSDUs in hardware.
Neither our drivers nor the net80211 stack were prepared to handle this.
Add two Rx-info flags which drivers can use to avoid having subframes which
arrived in the same A-MSDU rejected as duplicates in the net80211 input layer:
IEEE80211_RXI_HWDEC_SAME_PN allows the same CCMP packet number for a series
of subsequent frames. IEEE80211_RXI_SAME_SEQ allows the same 802.11 frame
header sequence number for a series of subsequent of frames.
Handle A-MPDU reordering on iwm 9k and iwx devices, based on code from iwlwifi.
Rx block ack window information is provided by firmware. So far this info was
ignored by drivers and reordering of A-MPDU subframes happened twice: Once in
firmware, and again in net80211.
Tested:
iwm 7260: bcallah, dv
iwm 7265: mpi, trondd, Matthias Schmidt
iwm 8260: bket, Marcus MERIGHI
iwm 8265: stsp, tracey, Uwe Werler
iwm 9260: phessler, matthieu
iwm 9560: stsp, Uwe Werler
iwx ax200: jmc, stsp
iwx ax201: stsp
jsing [Sun, 25 Apr 2021 13:15:22 +0000 (13:15 +0000)]
Clean up derivation of finished/peer finished.
Make this process more readable by having specific client/server functions,
calling the correct one based on s->server. This allows to remove various
SSL_ST_ACCEPT/SSL_ST_CONNECT checks, along with duplicate code.
ok inoguchi@ tb@
jsg [Sun, 25 Apr 2021 02:48:00 +0000 (02:48 +0000)]
cleanup riscv64 config glue
ok kettenis@ visa@
mvs [Sun, 25 Apr 2021 00:00:34 +0000 (00:00 +0000)]
Implement per-socket `so_lock' rwlock(9) and use it to protect routing
(PF_ROUTE) sockets. This can be done because we have no cases where one
thread should lock two sockets simultaneously.
Also this time `so_lock 'used for routing sockets only but in the future
it will be used to other socket types too.
ok bluhm@
kettenis [Sat, 24 Apr 2021 21:37:41 +0000 (21:37 +0000)]
Sync with arm64 version fixing bugs in some of the more obscure bus_space(9)
functions.
ok patrick@
deraadt [Sat, 24 Apr 2021 18:28:34 +0000 (18:28 +0000)]
sync
tb [Sat, 24 Apr 2021 18:16:39 +0000 (18:16 +0000)]
Do not leave errors on the error stack on PKCS12_parse() success.
Fix is the same as OpenSSL commit
ffbf304d4832bd51bb0618f8ca5b7c26647ee664
Found by Alex Gaynor with a new pyca cryptography regress test.
ok inoguchi
tb [Sat, 24 Apr 2021 18:10:12 +0000 (18:10 +0000)]
Enable the new verifier again so hopefully the remaining kinks get ironed
out in this release cycles.
discussed with deraadt and jsing
deraadt [Sat, 24 Apr 2021 17:04:24 +0000 (17:04 +0000)]
remove bgpd warning message exposed by "bgpctl show rib peer out" for
non-existant peer; the bgpctl command exposes the error itself
ok claudio
kettenis [Sat, 24 Apr 2021 10:33:09 +0000 (10:33 +0000)]
The variation in access size requirements and spacing of registers of
NS16550 clones is a bit of a mess. There are FDT properties that describe
these but these proprties are optional for some variants and the default
values depend on the compatible string being used. Most variants that have
shown up in the ARM world use 32-bit access and spacing. But true NS16550A
implementations of course use 8-bit access and spacing. So look at the
compatible strings to set up the appropriate defaults. Makes the "true"
NS16550A implemented by QEMU work for RISCV targets.
ok jsg@, drahn@
mpi [Sat, 24 Apr 2021 10:15:14 +0000 (10:15 +0000)]
Remove unused function pointers from `todr_handle'.
from miod@
mpi [Sat, 24 Apr 2021 09:44:44 +0000 (09:44 +0000)]
Convert allocations to km_alloc(9).
Also document an existing bug where the mapping for `pm_pdir_intel' is
lost when PAE is enabled.
With inputs from hshoexer@, tested by bluhm@, ok mlarkin@
jmatthew [Sat, 24 Apr 2021 09:37:46 +0000 (09:37 +0000)]
If MSI-X vectors are available, create multiple tx/rx queues, configure
RSS, and use the RSS hash value as a flow ID for received packets.
ok dlg@
visa [Sat, 24 Apr 2021 07:49:11 +0000 (07:49 +0000)]
Add cduart(4), a driver for the Cadence UART found on Xilinx Zynq SoCs.
Tested on Xilinx Zynq-7000.
Input and OK kettenis@
jsg [Sat, 24 Apr 2021 06:08:41 +0000 (06:08 +0000)]
reduce diff to other archs
ok deraadt@ mlarkin@
jsg [Sat, 24 Apr 2021 05:16:27 +0000 (05:16 +0000)]
add gfrtc(4)
jsg [Sat, 24 Apr 2021 05:14:45 +0000 (05:14 +0000)]
Add gfrtc a driver for the real-time clock interface of Google's
Goldfish Android virtual hardware platform. Used for the RTC on
qemu-system-riscv64 -M virt.
feedback and ok kettenis@
inoguchi [Sat, 24 Apr 2021 00:10:43 +0000 (00:10 +0000)]
Remove "-4" option treatment and use it always on s_server test in appstest.sh
bluhm [Fri, 23 Apr 2021 21:55:36 +0000 (21:55 +0000)]
Setting variable arpinit_done is not MP save if we want to execute
arp_rtrequest() in parallel. Move initialization to arpinit()
function.
OK kettenis@ mvs@
bluhm [Fri, 23 Apr 2021 21:47:32 +0000 (21:47 +0000)]
The variable la_hold_total contains the number of packets currently
in the arp queue. So the sysctl net.inet.ip.arpqueued must be read
only. In if_ether.c include the header with the declaration of
la_hold_total to ensure that the definition matches.
OK mvs@
tb [Fri, 23 Apr 2021 21:01:45 +0000 (21:01 +0000)]
Enable test-renegotiation-changed-clienthello.py but skip
"drop extended_master_secret in renegotiation" since we don't
support this extension.
tb [Fri, 23 Apr 2021 18:30:18 +0000 (18:30 +0000)]
whitespace
drahn [Fri, 23 Apr 2021 15:53:07 +0000 (15:53 +0000)]
Add MID_RISCV64 and EM_RISCV to enable riscv64
ok deraadt@
drahn [Fri, 23 Apr 2021 15:31:12 +0000 (15:31 +0000)]
File was accidentally duplicated on inital checkin, remove dup.
jsg [Fri, 23 Apr 2021 13:03:40 +0000 (13:03 +0000)]
enable viornd(4) on riscv64
jsg [Fri, 23 Apr 2021 12:49:53 +0000 (12:49 +0000)]
enable syscon(4) on riscv64
makes reboot and poweroff work with qemu
ok kettenis@
kettenis [Fri, 23 Apr 2021 12:38:00 +0000 (12:38 +0000)]
Adjust support for "syscon-reboot" and "syscon-poweroff" for the updated
bindings in mainline Linux. Necessary to support the QEMY RISCV target.
ok jsg@
jsg [Fri, 23 Apr 2021 12:09:45 +0000 (12:09 +0000)]
build with -Werror
ok mlarkin@
jsg [Fri, 23 Apr 2021 12:07:14 +0000 (12:07 +0000)]
cast TCB_GET() to void * to avoid -Wpointer-integer-compare warning
when building kern_sig.c
ok mlarkin@
jsg [Fri, 23 Apr 2021 12:01:02 +0000 (12:01 +0000)]
reuse arm64 openprom(4) on riscv64
ok kettenis@
bluhm [Fri, 23 Apr 2021 07:21:02 +0000 (07:21 +0000)]
Remove the sysctl kern.allowdt code from kernel if dt(4) is not
configured. This will result in a "value is not available" error
from sysctl when trying to enable dt on a kernel without support.
The variable allowdt should be in the device, not in sysctl source.
We don't need #ifdef for extern and prototypes.
OK mpi@
jmatthew [Fri, 23 Apr 2021 07:00:58 +0000 (07:00 +0000)]
When we have a single tx/rx queue, the cp ring is also used for async
events, so it can't be allocated when the interface is brought up and
freed when it's taken down.
ok dlg@
jsg [Fri, 23 Apr 2021 05:40:06 +0000 (05:40 +0000)]
spelling
jmatthew [Fri, 23 Apr 2021 04:58:34 +0000 (04:58 +0000)]
Create separate structures for rx and tx queues to allow creating multiples
of each of them, and prepare for having multiple completion rings.
ok dlg@
drahn [Fri, 23 Apr 2021 04:36:05 +0000 (04:36 +0000)]
Include sys/systm.h in riscv64 db files,
matches other archs, fixes -Werror failues,
ok jsg@
jsg [Fri, 23 Apr 2021 04:35:54 +0000 (04:35 +0000)]
add missing break
ok drahn@
dlg [Fri, 23 Apr 2021 03:43:19 +0000 (03:43 +0000)]
call klist_invalidate from bpfsdetach to tell kq listeners what happened.
without this, something using a kevent to monitor a bpf fd on an
idle interface never has the event fire, which means it never
realises the interface goes away. with this, the read event goes
off and the next read fails with EIO, like pretty much every other
driver when the underlying device is removed.
ok claudio@ visa@ jmatthew@
dlg [Fri, 23 Apr 2021 03:29:24 +0000 (03:29 +0000)]
only skip pf once for packets that are injected by a divert-packet socket.
when a divert socket gets a packet from userland to send back through
the kernel, it marks it as diverted so pf knows not to divert it
back to userland again. this marking stuck to the packet though,
so if it went through pf again (eg, on the way out of the network
stack) pf would skip it again. this is undesirable if you want pf
to do things to the packet on this outgoing hope, such as nat.
this has pf clear the mark once it's been used, which allows the
next run of a packet through pf to have stuff work on it.
found by some people at parta networks.
ok sashan@ lteo@ bluhm@
bluhm@ also suggested keeping my diff in the same style as the rest of pf.c
jmatthew [Fri, 23 Apr 2021 03:23:44 +0000 (03:23 +0000)]
We're only going to use the first CoS queue, so only store the ID of that
one. Some of this would have got in the way of setting up multiple tx/rx
rings.
ok dlg@
drahn [Fri, 23 Apr 2021 02:42:16 +0000 (02:42 +0000)]
Initial import of OpenBSD/riscv64
This work is based on the effort:
https://www.openbsd.org/papers/Porting_OpenBSD_to_RISCV_FinalReport.pdf
"Porting OpenBSD to RISC-V ISA"
by
Brian Bamsch <bbamsch@google.com>
Wenyan He <wenyan.he@sjsu.edu>
Mars Li <mengshi.li.mars@gmail.com>
Shivam Waghela <shivamwaghela@gmail.com>
With additional work by Dale Rahn <drahn@openbsd.org>
patrick [Thu, 22 Apr 2021 22:14:30 +0000 (22:14 +0000)]
Use the long version of the bwfm(4) firmware path, which includes the
board's compatible string, when printing an error about not being able
to load the firmware. Since most NVRAM files are board- or package-
specific, having the compatible makes it easier for us to find the
correct files, so that we can add them to the bwfm-firmware port.
ok kurt@
lum [Thu, 22 Apr 2021 19:50:55 +0000 (19:50 +0000)]
Add a non-interactive version of query-replace-regexp function called
replace-regexp. Unfortunately query-replace-regexp can't be used in a
startup file.
dv [Thu, 22 Apr 2021 18:40:21 +0000 (18:40 +0000)]
vmd(8): guard against bad virtio drivers
Add protections against guests with bad virtio-{blk,net,scsi}
drivers, specifically avoiding invalid descriptor chains and
invalid vionet packet sizes. This helps prevent possible lockup
of the host vm process due to a spinning device event loop thread.
Also fix an unneeded cast in the vioblk handling in case of invalid
buffer lengths.
OK mlarkin@
tb [Thu, 22 Apr 2021 18:27:53 +0000 (18:27 +0000)]
Only hash known CH extensions
RFC 4.1.2 specifies the ways in which the extensions in the first and
the second ClientHello may differ. It basically says that extensions
not known to a server must not change. This in turn makes it impossible
to introduce new extensions that do change. It makes little sense to
enforce that extensions we don't know and care about aren't modified,
so make the hashing more lenient and restrict it to the extensions we
do care about. Arguably, enforcing no change in an unknown extension
is incompatible with the requirement that it be ignored.
ok bcook jsing
visa [Thu, 22 Apr 2021 15:30:12 +0000 (15:30 +0000)]
kqueue: Make timer re-addition reset existing timer
When an existing EVFILT_TIMER filter is re-added, cancel the existing
timer and any pending event, and restart the timer using the new timeout
period. This makes the new timeout period take effect immediately and
matches the behaviour of FreeBSD. Previously, the new setting was
applied only after the existing timer expired.
The timer rescheduling is done by using an f_modify callback. The
reading of timer events is moved from f_event to f_process. f_event of
timer_filtops becomes redundant. Unlike most other event sources, timers
activate knotes directly without using a klist and knote(9).
OK mpi@
patrick [Thu, 22 Apr 2021 14:06:59 +0000 (14:06 +0000)]
Mark umb(4) as network device instead of a generic one. This also makes
it show up with the proper category in hotplug scripts.
From Tilo Stritzky
ok groth@ sthen@
mpi [Thu, 22 Apr 2021 11:54:32 +0000 (11:54 +0000)]
Keep under #ifdef TMPFS functions to grow/shrink uaobj.
ok patrick@
mpi [Thu, 22 Apr 2021 11:53:13 +0000 (11:53 +0000)]
Indent and simplify the grammar.
mpi [Thu, 22 Apr 2021 11:36:11 +0000 (11:36 +0000)]
Eliminate S/R conflicts and simplify filter grammar.
dv [Thu, 22 Apr 2021 10:45:21 +0000 (10:45 +0000)]
Remove trailing whitespace in comment blocks.
naddy [Thu, 22 Apr 2021 10:23:07 +0000 (10:23 +0000)]
reenable POOL_DEBUG
mpi [Thu, 22 Apr 2021 10:06:52 +0000 (10:06 +0000)]
Simplify token declaration.
mpi [Thu, 22 Apr 2021 09:44:38 +0000 (09:44 +0000)]
Simplify now that TID and PID are now only being parsed as builtin.
mpi [Thu, 22 Apr 2021 09:36:39 +0000 (09:36 +0000)]
Remove support for in-kernel filters.
This might be added back in a future if copying events to userland becomes
a performance issue. However note that it is not always possible to filter
in-kernel if, for example. a variable has to be evaluated when a rule fires.
tb [Wed, 21 Apr 2021 20:15:08 +0000 (20:15 +0000)]
Add a test that roundtrips a bunch of points on all builtin curves
via point2oct and oct2point and that checks the corner case in hybrid
encoding that was fixed in ec2_oct.c r1.13.
jsing [Wed, 21 Apr 2021 19:27:56 +0000 (19:27 +0000)]
Clean up TLSv1.2 certificate request handshake data.
Currently cert_req is used by clients and cert_request is used by servers.
Replace this by a single cert_request used by either client or server.
Remove the certificate types as they are currently unused. This also fixes
a bug whereby if the number of certificate types exceeds SSL3_CT_NUMBER
the number of bytes read in is insufficient, which will break decoding.
ok inoguchi@ tb@
dv [Wed, 21 Apr 2021 18:27:36 +0000 (18:27 +0000)]
Fix packet size checks and remove bad casts.
Because dhcpsz was an uninitialized ssize_t, it was possible that a
garbage "packet" would be queued on the receiving end of the virtio
network device.
Change the type to size_t and add proper checks based on it being
greater than zero. Remove the cast of ssize_t to uint64_t that also
caused garbage sizes when dhcpsz was unintialized and set at runtime
to something < 0.
kettenis [Wed, 21 Apr 2021 18:11:17 +0000 (18:11 +0000)]
Turn panic for unknown event type into a printf and return.
ok patrick@
lum [Wed, 21 Apr 2021 14:45:28 +0000 (14:45 +0000)]
Use the plenteous BUFSIZE found in mg in lieu of 128.
jmc [Wed, 21 Apr 2021 13:23:56 +0000 (13:23 +0000)]
remove -p from usage;
mpi [Wed, 21 Apr 2021 10:53:17 +0000 (10:53 +0000)]
Extend filters to support any conditionnal test including global variables.
Stop using in-kernel filtering for the moment except for not tracing the
tracer.
Keep track of the number of filtered events.
mpi [Wed, 21 Apr 2021 10:34:36 +0000 (10:34 +0000)]
Always initialized min value for an histogram.
mpi [Wed, 21 Apr 2021 10:30:46 +0000 (10:30 +0000)]
Test local variables, boolean and print() of variables other than maps.
mpi [Wed, 21 Apr 2021 10:29:49 +0000 (10:29 +0000)]
typo
mpi [Wed, 21 Apr 2021 10:28:54 +0000 (10:28 +0000)]
Support for local (scratch) variables: "$var_name".
Every rule gets its own list of (local) variables.
mpi [Wed, 21 Apr 2021 10:26:18 +0000 (10:26 +0000)]
Extend print() to support any kind of variable.
mpi [Wed, 21 Apr 2021 10:22:36 +0000 (10:22 +0000)]
Support first shell argument as $1 in order to use it in filters.
Remove '-p' option now that scripts can filter by pid/tid.
bluhm [Wed, 21 Apr 2021 10:18:35 +0000 (10:18 +0000)]
Check that path MTU discovery across routing domains inserts a
dynamic IPv4 route. Add IPv6 test that runs tcpbench over pair
interfaces in different routing domains. For that setup, PMTU
currently does not work. IPv6 TCP packets get fragmented.
mpi [Wed, 21 Apr 2021 10:02:05 +0000 (10:02 +0000)]
Convert remaining uvm_km_zalloc(9) to km_alloc(9).
Tested by bluhm@, jj@, kettenis@ and Scott Bennett.
ok kettenis@
bluhm [Wed, 21 Apr 2021 09:38:11 +0000 (09:38 +0000)]
Improve ntpd offset handling. Call the index of the offset loops
"shift" consistently. Merge the two offset loops in client_update()
into one. Use a simple assignment for the best value instead of
memcpy(). Use the same mechanism to loop over the offset array
everywhere to avoid an invalid best value.
tested by weerd@; OK claudio@
claudio [Wed, 21 Apr 2021 09:36:06 +0000 (09:36 +0000)]
HTTP errors are logged by the http module. No need to repeat them.
Instead add a logx when a fallback from delta sync to a snapshot happens.
claudio [Wed, 21 Apr 2021 09:03:42 +0000 (09:03 +0000)]
When setting RRDP_STATE_PARSE_ERROR a warning is issued so there is no need
to warn a 2nd time here, it adds no additional information.
eric [Wed, 21 Apr 2021 07:54:10 +0000 (07:54 +0000)]
unplug unused certificate verification code, now that this is done by libtls.
ok tb@ millert@
tb [Wed, 21 Apr 2021 00:31:59 +0000 (00:31 +0000)]
Fix const in previous. Pointed out by asou
gkoehler [Wed, 21 Apr 2021 00:22:16 +0000 (00:22 +0000)]
Fix __builtin_bitreverse32 on 32-bit PowerPC
This is a backport from LLVM 11. Before this fix, code using
__builtin_bitreverse32 was crashing SIGILL because clang-10 emitted a
64-bit rldicl/clrldi instruction. The SIGILL only happened on 32-bit
cpus, not on the G5. The code for LLVM 11 uses __builtin_bitreverse,
so clang-10 needs this fix to build clang-11.
https://github.com/llvm/llvm-project/commit/
a5d161c119d5a
https://reviews.llvm.org/D77946
ok kettenis@
dv [Tue, 20 Apr 2021 21:11:56 +0000 (21:11 +0000)]
Move TAILQ initialization to files where they are used.
These priv-sep daemons all follow a similar design and use TAILQs
for tracking control process connections. In most cases, the TAILQs
are initialized separate from where they are used. Since the scope
of use is generally confined to a specific control process file,
this commit also removes any extern definitions and exposing the
TAILQ structures to other compilation units.
ok bluhm@, tb@
kettenis [Tue, 20 Apr 2021 19:33:03 +0000 (19:33 +0000)]
Match on "amlogic,meson-sm1-mmc" compatible.
tb [Tue, 20 Apr 2021 17:38:02 +0000 (17:38 +0000)]
Fix indent of EC_METHODs as requested by jsing.
While there zap trailing whitespace from a KNF approximation gone wrong.
tb [Tue, 20 Apr 2021 17:35:21 +0000 (17:35 +0000)]
Adjust ectest.c for set_compressed_coordinates
tb [Tue, 20 Apr 2021 17:34:33 +0000 (17:34 +0000)]
Compare pointer against NULL and fix a KNF issue.
ok jsing
tb [Tue, 20 Apr 2021 17:32:57 +0000 (17:32 +0000)]
Prepare to provide EC_POINT_set_compressed_coordinates
ok jsing
tb [Tue, 20 Apr 2021 17:30:32 +0000 (17:30 +0000)]
Adjust ectest.c for get_Jprojective coordinate change
tb [Tue, 20 Apr 2021 17:29:21 +0000 (17:29 +0000)]
Compare function pointers against NULL, not 0.
ok jsing
tb [Tue, 20 Apr 2021 17:28:18 +0000 (17:28 +0000)]
Provide EC_POINT_{g,s}et_Jprojective_coordinates for internal use
ok jsing
tb [Tue, 20 Apr 2021 17:23:37 +0000 (17:23 +0000)]
Simplify code after adding EC_POINT_{s,g}et_affine_coordinates()
ok jsing
tb [Tue, 20 Apr 2021 17:21:27 +0000 (17:21 +0000)]
Adjust ecdhtest.c for affine_coordinates change
tb [Tue, 20 Apr 2021 17:19:39 +0000 (17:19 +0000)]
Adjust ectest.c for affine_coordinates change
tb [Tue, 20 Apr 2021 17:17:47 +0000 (17:17 +0000)]
Compare function pointers against NULL, not 0.
ok jsing
tb [Tue, 20 Apr 2021 17:16:37 +0000 (17:16 +0000)]
Prepare to provide EC_POINT_{g,s}et_affine_coordinates
Similar to part of OpenSSL commit
8e3cced75fb5fee5da59ebef9605d403a999391b
ok jsing
tb [Tue, 20 Apr 2021 17:12:43 +0000 (17:12 +0000)]
Simplify after EC_POINT_get_curve() addition
ok jsing
tb [Tue, 20 Apr 2021 17:09:45 +0000 (17:09 +0000)]
Adjust ectest.c for EC_GROUP_{g,s}et_curve change
tb [Tue, 20 Apr 2021 17:08:08 +0000 (17:08 +0000)]
Add prototypes for EC_GROUP_get_curve_{GF2m,GFp}().
These will be removed once EC_GROUP_get_curve() is public.
tb [Tue, 20 Apr 2021 17:06:17 +0000 (17:06 +0000)]
Compare function pointers against NULL, not 0.
ok jsing
tb [Tue, 20 Apr 2021 17:04:13 +0000 (17:04 +0000)]
Prepare to provide EC_GROUP_{get,set}_curve(3)
There are numerous functions in ec/ that exist with _GF2m and _GFp
variants for no good reason. The code of both variants is the same.
The EC_METHODs contain a pointer to the appropriate version. This
commit hides the _GF2m and _GFp variants from internal use and
provides versions that work for both curve types. These will be made
public in an upcoming library bump.
Similar to part of OpenSSL commit
8e3cced75fb5fee5da59ebef9605d403a999391b
ok jsing