This module was added with kernel 4.15, but is was removed again with
kernel version 5.3. OpenWrt does not support specifying a kernel version
range so just break it with kernel 4.14 and only support recent kernel
versions.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
With kernel 5.4 kmod-sound-hda-intel also needs snd-intel-nhlt.ko, but
this kernel module is only build on x86, make the OpenWrt kmod depend on
TARGET_x86.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
nf_reset() was renamed to nf_reset_ct() in upstream Linux commit
895b5c9f206e ("netfilter: drop bridge nf reset from nf_reset)"
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
gigaset was moved to staging in kernel 5.4, just deactivate it on
recent kernel versions instead of adapting it.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
In kernel 5.3 and 5.4 some crypto modules were split into two modules,
one implementing the crypto algorithm and the other integrating it
into the Linux crypto framework.
Adapt OpenWrt to support this split.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This adds the new xfrm4_mode_beet, xfrm4_mode_transport,
xfrm4_mode_tunnel and their IPv6 versions on kernel 5.4. These modules
were newly added in kernel 5.2.
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
CONFIG_INCLUDE_CONFIG option is helpful for being able to rebuild the
exact same firmware as you see on a live OpenWRT instance, but it's
crucially missing feeds information, so we can't rebuild the exact same
package versions. This commit fixes this by adding the remaining feeds
(and version) buildinfo files to the image.
Signed-off-by: Xu Wang <xwang1498@gmx.com>
This reverts commit 215598fd03 since it
didn't contain a reference to the CVE it addresses. The next commit
will re-add the commit including a CVE reference in its commit message.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Old MikroTik devices have the RLE-encoded radio calibration data
directly stored in the art (hard_config) partition, without LZO
compression nor any preceding ERD magic bytes. This commit adds
a fallback for these devices.
Tested on the ath79 target with a MikroTik SXT 5nD r2 (SXT Lite5),
only locally --not yet merged upstream--.
Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
Add a radio_config_id property. If the radio config changes return an error
upon receiving the reconf call.
Signed-off-by: John Crispin <john@phrozen.org>
For devices without a dedicated 'diag' LED, we use sometimes one of
other LEDs for indicating at least 'boot', 'failsafe' and 'upgrade'
stages. In some cases, at the same time these LEDs have defined default
triggers in DTS using 'linux,default-trigger' property. Current 'diag'
setup removes the trigger and turns off 'boot' LED after bootup.
One of the examples of such device is TP-Link TL-WR841N v14 (ramips)
which uses 'wlan' LED with defined 'linux,default-trigger' for 'diag':
aliases {
led-boot = &led_wlan;
led-failsafe = &led_wlan;
led-upgrade = &led_wlan;
};
[...]
led_wlan: wlan {
label = "tl-wr841n-v14:green:wlan";
gpios = <&gpio1 9 GPIO_ACTIVE_LOW>;
linux,default-trigger = "phy0tpt";
};
This patch extends 'diag.sh' and 'leds.sh' scripts to make sure default
trigger defined in DTS is restored for 'diag' LED which isn't used for
indicating 'running' stage.
Acked-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
c888e17e06 ("hostapd: manage instances via procd instead of pidfile")
added procd support for managing hostapd and wpa_supplicant daemons
but at the same time limited wiphy names to 'phy*'.
This brings back initial behaviour (introduced in 60fb4c92b6 ("hostapd:
add ubus reload") and makes procd manage daemons for any wiphy device
found in '/sys/class/ieee80211'.
CC: Felix Fietkau <nbd@nbd.name>
CC: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
License "GPL-2.0+" is deprecated License Identifier according to
SPDX License list [1]. The correct one is GPL-2.0-or-later.
While at it, also add the License file.
[1] https://spdx.org/licenses/
Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
The function trace_on_exit() is given to atexit() as a parameter, but
atexit() only takes a function pointer to a function with a void
parameter.
This problem was introduced when the on_exit() function was incompletely
replaced by atexit().
Fixes: ba6c8bd614 ("linux-atm: add portability fixes")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
ath10k-ct supports the combination to select IBSS (ADHOC) mode and
different beacon intervals together. mac80211 does not like this
combination, but Ben says this is ok, so remove this check.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
> Avoid reuse of PKG_NAME in call, define and eval lines for consistency and
> readability. Write the full name instead.
Ref: https://openwrt.org/docs/guide-developer/packages
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
> Avoid reuse of PKG_NAME in call, define and eval lines for consistency and
> readability. Write the full name instead.
Ref: https://openwrt.org/docs/guide-developer/packages
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
As 07e1d88d7b ("kernel: avoid underscore in *6lowpan package names") shows,
underscores might cause build failures. Replace underscore with dash.
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
The current version of ATF does not support power off for SGMII
COMPHY. Update to latest ATF to resolve this issue.
Signed-off-by: Scott Roberts <ttocsr@gmail.com>
The SUBDIRS variable has been removed in kernel 5.4, and was deprecated
since the beginnig of kernel git history in favour of M or KBUILD_EXTMOD.
Signed-off-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
Removed all upstream patches.
Added PKG_BUILD_PARALLEL for faster compilation.
Small Makefile rearrangements for consistency between packages.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
This adds the direct dependencies introduced by TARGET_LDFLAGS
to the package's DEPENDS variable.
This was found by accidentally building rssileds on octeon, which
resulted in:
"Package rssileds is missing dependencies for the following libraries:
libnl-tiny.so"
Though the dependencies are provided when building for the
relevant targets ar71xx, ath79 and ramips, it seems more tidy to
specify them explicitly.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
f4415afce213 mt76: mt76u: loop over all possible rx queues in mt76u_rx_tasklet
5b9f949cb760 mt76: mt76u: fix a possible memory leak in mt76u_init
fd892bc033fb mt76: mt76u: rely only on data buffer for usb control messagges
Signed-off-by: Felix Fietkau <nbd@nbd.name>
8d45443bb5c9 pppd: Ignore received EAP messages when not doing EAP
8d7970b8f3db pppd: Fix bounds check in EAP code
858976b1fc31 radius: Prevent buffer overflow in rc_mksid()
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Update util-linux to 2.35.1 and refresh patches.
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
[commit subject and description tweaks]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
bd0df1b017a8 mt76: avoid extra RCU synchronization on station removal
d5a5e97b67c7 mt76: mt76x2: avoid starting the MAC too early
a67e42990d8a mt76: mt7615: fix msdu_id endianness in mt7615_write_hw_txp
d3af8bd3c722 mt76: mt7615: set proper length in strncmp
9c43417db17c mt76: mt7615: fix max_nss in mt7615_eeprom_parse_hw_cap
764e1d208a06 mt76: mt7615: fix tx power reporting
1881241c7ee5 mt76: fix rounding issues on converting per-chain and combined txpower
fa14e7f33199 mt76: mt7615: rework rx phy index handling
a205ce3e3e2d mt76: mt7615: fix ext_phy flag for stations
457a93203690 mt76: mt7615: fix MT_TX_HW_QUEUE_EXT_PHY to deal with mac80211 changes
c75cf513c674 mt76: do not set HOST_BROADCAST_PS_BUFFERING for mt7615
cc56c400167c mt76: fix LED link time failure
4dbd56b86970 mt76: mt76x0u: add support to TP-Link T2UHP
e226309c4bc4 mt76: mt7615: rely on mt76_queues_read for mt7622
c6a025318075 mt76: mt76u: extend RX scatter gather number
dfc24bc504e3 mt76: mt76u: rename stat_wq in wq
2bbffd2cb37b mt76: mt7615: remove rx_mask in mt7615_eeprom_parse_hw_cap
f408a2b7566c mt76: Introduce mt76_mcu data structure
17ecf0762542 mt76: mt76x02: fix handling MCU timeouts during hw restart
284e9fd72912 mt76: mt7615: fix monitor injection of beacon frames
8f8e9161b355 mt76: fix array overflow on receiving too many fragments for a packet
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Packages kmod-bluetooth_6lowpan and kmod-ieee802154_6lowpan contain an
underscore in the package name. This causes problems in package/install
because when building a list of package files to install offline using
opkg, it uses a wildcard of the form $(dir)/$(pkg)_*.ipk.
If you were to select kmod-bluetooth=y, but kmod-bluetooth_6lowpan=m,
the latter would be picked up by that wildcard, and make package/install
would fail:
Collected errors:
* satisfy_dependencies_for: Cannot satisfy the following dependencies
* for kmod-bluetooth_6lowpan:
* kmod-6lowpan
* opkg_install_cmd: Cannot install package kmod-bluetooth_6lowpan.
Changing the wildcard pattern is not trivial, and there may be other
places in the build system making this assumption about the package name
format.
Using a dash in place of the underscore avoids the issue.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
A file, package/network/services/dnsmasq/files/50-dnsmasq-migrate-resolv-conf-auto.sh,
was added in commit 6a28552120, but it
does not exit in a way that tells the uci-defaults mechanism that it
succeeded, and so it is not cleaned up after running successfully. Add
an exit 0 to the end to correct that.
Signed-off-by: Russell Senior <russell@personaltelco.net>
* send: cleanup skb padding calculation
* socket: remove useless synchronize_net
Sorry for the back-to-back releases. This fixes a regression spotted by Eric
Dumazet.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
This change makes the names of Broadcom targets consistent by using
the common notation based on SoC/CPU ID (which is used internally
anyway), bcmXXXX instead of brcmXXXX.
This is even used for target TITLE in make menuconfig already,
only the short target name used brcm so far.
Despite, since subtargets range from bcm2708 to bcm2711, it seems
appropriate to use bcm27xx instead of bcm2708 (again, as already done
for BOARDNAME).
This also renames the packages brcm2708-userland and brcm2708-gpu-fw.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Álvaro Fernández Rojas <noltari@gmail.com>
This change makes the names of Broadcom targets consistent by using
the common notation based on SoC/CPU ID (which is used internally
anyway), bcmXXXX instead of brcmXXXX.
This is even used for target TITLE in make menuconfig already,
only the short target name used brcm so far.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This change makes the names of Broadcom targets consistent by using
the common notation based on SoC/CPU ID (which is used internally
anyway), bcmXXXX instead of brcmXXXX.
This is even used for target TITLE in make menuconfig already,
only the short target name used brcm so far.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This supports better per-chain noise floor reporting, which in turn allows for
better RSSI reporting in the driver.
Wave-2 fixes a long-standing rate-ctrl problem when connected to xbox (and probably other devices).
Wave-2 has fix for crash likely related to rekeying.
Wave-1 has some debugging code added where a user reported a crash.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [ipq806x+qca9984,ipq4019+qca9986]
Signed-off-by: Michael Yartys <michael.yartys@protonmail.com>
Changes:
ath10k-ct: Support better RSSI measurements.
When used with recent firmware, these changes allow the driver to
query per-chain noise-floor from the radio to better calculate the
per-chain RSSI. The per-chain RSSI is then summed to provide the
'combined RSSI'. This gives better per-chain RSSI as well as combined
RSSI, especially when running with more than 20Mhz bandwidths.
Refresh patches.
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [ipq806x+qca9984,ipq4019+qca9986]
Signed-off-by: Michael Yartys <michael.yartys@protonmail.com>
* wg-quick: android: split uids into multiple commands
Newer android's ndc implementations have limits on uid size, so we have to
break these into several lists.
* man: document dynamic debug trick for Linux
This comes up occasionally, so it may be useful to mention its
possibility in the man page. At least the Arch Linux and Ubuntu kernels
support dynamic debugging, so this advice will at least help somebody. So that
you don't have to go digging into the commit, this adds this helpful tidbit
to the man page for getting debug logs on Linux:
# modprobe wireguard && echo module wireguard +p > /sys/kernel/debug/dynamic_debug/control
* extract-{handshakes,keys}: rework for upstream kernel
These tools will now use the source code from the running kernel instead of
from the old monolithic repo. Essential for the functioning of Wireshark.
* netlink: remove libmnl requirement
We no longer require libmnl. It turns out that inlining the small subset of
libmnl that we actually use results in a smaller binary than the overhead of
linking to the external library. And we intend to gradually morph this code
into something domain specific as a libwg emerges. Performance has also
increased, thanks to the inliner. On all platforms, wg(8) only needs a normal
libc. Compile time on my system is still less than one second. So all in all
we have: smaller binary, zero dependencies, faster performance.
Packagers should no longer have their wireguard-tools package depend on
libmnl.
* embeddable-wg-library: use newer string_list
* netlink: don't pretend that sysconf isn't a function
Small cleanups.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Most of the broadcom packaging codes were moved to broadcom.mk in commit
7f984dab1c ("mac80211: move broadcom packaging code to broadcom.mk"),
but b43/install still remained. Move it now.
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
The $(space) definition in the hostapd Makefile ceased to work with
GNU Make 4.3 and later, leading to syntax errors in the generated
Kconfig files.
Drop the superfluous redefinition and reuse the working $(space)
declaration from rules.mk to fix this issue.
Fixes: GH#2713
Ref: https://github.com/openwrt/openwrt/pull/2713#issuecomment-583722469
Reported-by: Karel Kočí <cynerd@email.cz>
Suggested-by: Jonas Gorski <jonas.gorski@gmail.com>
Tested-by: Shaleen Jain <shaleen@jain.sh>
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
* compat: support building for RHEL-8.2
* compat: remove RHEL-7.6 workaround
Bleeding edge RHEL users should be content now (which includes the actual
RedHat employees I've been talking to about getting this into the RHEL kernel
itself). Also, we remove old hacks for versions we no longer support anyway.
* allowedips: remove previously added list item when OOM fail
* noise: reject peers with low order public keys
With this now being upstream, we benefit from increased fuzzing coverage of
the code, uncovering these two bugs.
* netns: ensure non-addition of peers with failed precomputation
* netns: tie socket waiting to target pid
An added test to our test suite for the above and a small fix for high-load CI
scenarios.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
This reverts commit 96424c143d.
The commit changed libiwinfo's internal ABI which breaks a number of
downstream projects, including LuCI and rpcd-mod-iwinfo.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Bump procd package to reduce log spam related to missing jail binaries
in a non-jail capable system.
bcb8655 instance: add 'requirejail' attribute
An additional jail attribute 'requirejail' can now be used to indicate
mandatory use of a jailed environment and hence prevent process startup
in the event that the jail subsystem is unavailable.
Procd will now only log errors if jail is unavailable and 1) is a mandatory
requirement or 2) a procd debug level of at least 2 is in use.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
eba5a20 iwinfo: add device id for BCM43602
a6914dc iwinfo: add BSS load element to scan result
bb21698 iwinfo: add device id for Atheros AR9287
7483398 iwinfo: add device id for MediaTek MT7615E
Signed-off-by: David Bauer <mail@david-bauer.net>
Add Netgear WNDR3700v2 to the list of supported boards.
Signed-off-by: Michal Cieslakiewicz <michal.cieslakiewicz@wp.pl>
[rebase, adjusted commit title]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Package kernel module for Linear Technology MCP3021/3221 I2C connected
current and voltage monitor chips.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
I'm having another attempt at trying to getting the 'store dscp into
conntrack connmark' functionality into upstream kernel, since the
restore function (act_ctinfo) has been accepted.
The syntax has changed from 'savedscp' to 'set-dscpmark' since that
conforms more closely with existing functionality.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Fixes handling CSA when using AP+STA or AP+Mesh
This change was accidentally dropped in commit 167028b75
("hostapd: Update to version 2.9 (2019-08-08)")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Without this change, wpa-cli features depend on which wpad build variant was
used to build the wpa-cli package
Signed-off-by: Felix Fietkau <nbd@nbd.name>
It is generally more desirable to use older kernel versions for
dependencies, as this will require less changes when newer kernels
are added (they will by default select the newer packages).
Since we currently only have two kernels (4.14 and 4.19) in master,
this patch applies this logic by converting all LINUX_4_19 symbols
to their inverted LINUX_4_14 equivalents.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
"[[" is a bash extension for test. As the ash-implementation is not
fully compatible we drop its usage.
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
[remove shebang, slightly facelift commit title/message]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
"[[" is a bash extension for test. As the ash-implementation is not
fully compatible we drop its usage.
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
"[[" is a bash extension for test. As the ash-implementation is not
fully compatible we drop its usage.
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
"[[" is a bash extension for test. As the ash-implementation is
not fully compatible we drop its usage.
Also change to "=" for simple test, which is sufficient. (see d6ac8ca76c)
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
[split patch, removed shebang]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Call skb_orphan(skb) to call the owner's destructor function and make
the skb unowned.
This is necessary to prevent sk_wmem_alloc of a socket from overflowing,
which leads to ENOBUFS errors on application level.
Signed-off-by: Martin Schiller <ms@dev.tdt.de>
This patch converts the Qxwlan E2600AC image away from
the deprecated .bin file and to the new .qca4019 method.
As a result, we no longer need to carry around the
legacy support for handling .bin files.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
SOC: IPQ4018 / QCA Dakota
CPU: Quad-Core ARMv7 Processor rev 5 (v7l) Cortex-A7
DRAM: 256 MiB
NOR: 32 MiB
ETH: Qualcomm Atheros QCA8072 (1 port)
WLAN1: Qualcomm Atheros QCA4018 2.4GHz 802.11bgn 2:2x2
WLAN2: Qualcomm Atheros QCA4018 5GHz 802.11a/n/ac 2:2x2
INPUT: RESET Button
LEDS: White, Blue, Red, Orange
Flash instruction:
From EnGenius firmware to OpenWrt firmware:
In Firmware Upgrade page, upgrade your openwrt-ipq40xx-generic-engenius_emd1-squashfs-factory.bin directly.
From OpenWrt firmware to EnGenius firmware:
1. Setup a TFTP server on your computer and configure static IP to 192.168.99.8
Put the EnGenius firmware in the TFTP server directory on your computer.
2. Power up EMD1. Press 4 and then press any key to enter u-boot.
3. Download EnGenius firmware
(IPQ40xx) # tftpboot 0x84000000 openwrt-ipq40xx-emd1-nor-fw-s.img
4. Flash the firmware
(IPQ40xx) # imgaddr=0x84000000 && source 0x84000000:script
5. Reboot
(IPQ40xx) # reset
Signed-off-by: Yen-Ting-Shen <frank.shen@senao.com>
[removed BOARD_NAME]
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Specifications
==============
- SOC: IPQ4018
- RAM: DDR3 256MB
- Flash: SPI NOR 16MB
- WiFi:
- 2.4GHz: IPQ4018, 2x2, front end SKY85303-11
- 5GHz: IPQ4018, 2x2, front end SKY85717-21
- Ethernet: 1x 10/100/1000Mbps, POE 802.3af
- PHY: QCA8072
- UART: GND, blocked, 3.3V, RX, TX / 115200 8N1
- LED: 1x red / green
- Button: 1x reset / factory default
- U-Boot bootloader with tftp and "emergency web server" accessible
using serial port.
Installation
============
Flash factory image from D-Link web UI. Constraints in the D-Link web UI
makes the factory image unnecessarily large. Flash again using
sysupgrade from inside OpenWrt to reclaim some flash space.
Return to stock D-Link firmware
===============================
Partition layout is preserved, and it is possible to return to the stock
firmware simply by downloading it from D-Link and writing it to the
firmware partition.
# mtd -r write dap2610-firmware.bin firmware
Quirks
======
To be flashable from the D-Link http server, the firmware must be larger
then 6MB, and the size in the firmware header must match the actual file
size. Also, the boot loader verifies the checksum of the firmware before
each boot, thus the jffs2 must be after the checksum covered part. This
is solved in the factory image by having the rootfs at the very end of
the image (without pad-rootfs).
The sysupgrade image which does not have to be flashable from the D-Link
web UI may be smaller, and the checksum in the firmware header only
covers the kernel part of the image.
Signed-off-by: Fredrik Olofsson <fredrik.olofsson@anyfinetworks.com>
[added WRGG Variables to DEVICE_VARS, squashed spi pinconf/mux,
added emd1's gmac0 config,fix dtc warnings]
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
* Makefile: remove pwd from compile output
* Makefile: add standard 'all' target
* Makefile: evaluate git version lazily
Quality of life improvements for packagers.
* ipc: simplify inflatable buffer and add fuzzer
* fuzz: add generic command argument fuzzer
* fuzz: add set and setconf fuzzers
More fuzzers and a slicker string list implementation. These fuzzers now find
themselves configuring wireguard interfaces from scratch after several million
mutations, which is fun to watch.
* netlink: make sure to clear return value when trying again
Prior, if a dump was interrupted by a concurrent set operation, we'd try
again, but forget to reset an error flag, so we'd keep trying again forever.
Now we do the right thing and succeed when we succeed.
* Makefile: sort inputs to linker so that build is reproducible
Earlier versions of make(1) passed GLOB_NOSORT to glob(3), resulting in the
linker receiving its inputs in a filesystem-dependent order. This screwed up
reproducible builds.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
* Makefile: strip prefixed v from version.h
This fixes a mistake in dmesg output and when parsing the sysfs entry in the
filesystem.
* device: skb_list_walk_safe moved upstream
This is a 5.6 change, which we won't support here, but it does make the code
cleaner, so we make this change to keep things in sync.
* curve25519: x86_64: replace with formally verified implementation
This comes from INRIA's HACL*/Vale. It implements the same algorithm and
implementation strategy as the code it replaces, only this code has been
formally verified, sans the base point multiplication, which uses code
similar to prior, only it uses the formally verified field arithmetic
alongside reproducable ladder generation steps. This doesn't have a
pure-bmi2 version, which means haswell no longer benefits, but the
increased (doubled) code complexity is not worth it for a single
generation of chips that's already old.
Performance-wise, this is around 1% slower on older microarchitectures,
and slightly faster on newer microarchitectures, mainly 10nm ones or
backports of 10nm to 14nm. This implementation is "everest" below:
Xeon E5-2680 v4 (Broadwell)
armfazh: 133340 cycles per call
everest: 133436 cycles per call
Xeon Gold 5120 (Sky Lake Server)
armfazh: 112636 cycles per call
everest: 113906 cycles per call
Core i5-6300U (Sky Lake Client)
armfazh: 116810 cycles per call
everest: 117916 cycles per call
Core i7-7600U (Kaby Lake)
armfazh: 119523 cycles per call
everest: 119040 cycles per call
Core i7-8750H (Coffee Lake)
armfazh: 113914 cycles per call
everest: 113650 cycles per call
Core i9-9880H (Coffee Lake Refresh)
armfazh: 112616 cycles per call
everest: 114082 cycles per call
Core i3-8121U (Cannon Lake)
armfazh: 113202 cycles per call
everest: 111382 cycles per call
Core i7-8265U (Whiskey Lake)
armfazh: 127307 cycles per call
everest: 127697 cycles per call
Core i7-8550U (Kaby Lake Refresh)
armfazh: 127522 cycles per call
everest: 127083 cycles per call
Xeon Platinum 8275CL (Cascade Lake)
armfazh: 114380 cycles per call
everest: 114656 cycles per call
Achieving these kind of results with formally verified code is quite
remarkable, especialy considering that performance is favorable for
newer chips.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
