Merge tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-nextgrafted

Pull networking updates from Jakub Kicinski:
 "Core:

   - Add dedicated kmem_cache for typical/small skb->head, avoid having
     to access struct page at kfree time, and improve memory use.

   - Introduce sysctl to set default RPS configuration for new netdevs.

   - Define Netlink protocol specification format which can be used to
     describe messages used by each family and auto-generate parsers.
     Add tools for generating kernel data structures and uAPI headers.

   - Expose all net/core sysctls inside netns.

   - Remove 4s sleep in netpoll if carrier is instantly detected on
     boot.

   - Add configurable limit of MDB entries per port, and port-vlan.

   - Continue populating drop reasons throughout the stack.

   - Retire a handful of legacy Qdiscs and classifiers.

  Protocols:

   - Support IPv4 big TCP (TSO frames larger than 64kB).

   - Add IP_LOCAL_PORT_RANGE socket option, to control local port range
     on socket by socket basis.

   - Track and report in procfs number of MPTCP sockets used.

   - Support mixing IPv4 and IPv6 flows in the in-kernel MPTCP path
     manager.

   - IPv6: don't check net.ipv6.route.max_size and rely on garbage
     collection to free memory (similarly to IPv4).

   - Support Penultimate Segment Pop (PSP) flavor in SRv6 (RFC8986).

   - ICMP: add per-rate limit counters.

   - Add support for user scanning requests in ieee802154.

   - Remove static WEP support.

   - Support minimal Wi-Fi 7 Extremely High Throughput (EHT) rate
     reporting.

   - WiFi 7 EHT channel puncturing support (client & AP).

  BPF:

   - Add a rbtree data structure following the "next-gen data structure"
     precedent set by recently added linked list, that is, by using
     kfunc + kptr instead of adding a new BPF map type.

   - Expose XDP hints via kfuncs with initial support for RX hash and
     timestamp metadata.

   - Add BPF_F_NO_TUNNEL_KEY extension to bpf_skb_set_tunnel_key to
     better support decap on GRE tunnel devices not operating in collect
     metadata.

   - Improve x86 JIT's codegen for PROBE_MEM runtime error checks.

   - Remove the need for trace_printk_lock for bpf_trace_printk and
     bpf_trace_vprintk helpers.

   - Extend libbpf's bpf_tracing.h support for tracing arguments of
     kprobes/uprobes and syscall as a special case.

   - Significantly reduce the search time for module symbols by
     livepatch and BPF.

   - Enable cpumasks to be used as kptrs, which is useful for tracing
     programs tracking which tasks end up running on which CPUs in
     different time intervals.

   - Add support for BPF trampoline on s390x and riscv64.

   - Add capability to export the XDP features supported by the NIC.

   - Add __bpf_kfunc tag for marking kernel functions as kfuncs.

   - Add cgroup.memory=nobpf kernel parameter option to disable BPF
     memory accounting for container environments.

  Netfilter:

   - Remove the CLUSTERIP target. It has been marked as obsolete for
     years, and we still have WARN splats wrt races of the out-of-band
     /proc interface installed by this target.

   - Add 'destroy' commands to nf_tables. They are identical to the
     existing 'delete' commands, but do not return an error if the
     referenced object (set, chain, rule...) did not exist.

  Driver API:

   - Improve cpumask_local_spread() locality to help NICs set the right
     IRQ affinity on AMD platforms.

   - Separate C22 and C45 MDIO bus transactions more clearly.

   - Introduce new DCB table to control DSCP rewrite on egress.

   - Support configuration of Physical Layer Collision Avoidance (PLCA)
     Reconciliation Sublayer (RS) (802.3cg-2019). Modern version of
     shared medium Ethernet.

   - Support for MAC Merge layer (IEEE 802.3-2018 clause 99). Allowing
     preemption of low priority frames by high priority frames.

   - Add support for controlling MACSec offload using netlink SET.

   - Rework devlink instance refcounts to allow registration and
     de-registration under the instance lock. Split the code into
     multiple files, drop some of the unnecessarily granular locks and
     factor out common parts of netlink operation handling.

   - Add TX frame aggregation parameters (for USB drivers).

   - Add a new attr TCA_EXT_WARN_MSG to report TC (offload) warning
     messages with notifications for debug.

   - Allow offloading of UDP NEW connections via act_ct.

   - Add support for per action HW stats in TC.

   - Support hardware miss to TC action (continue processing in SW from
     a specific point in the action chain).

   - Warn if old Wireless Extension user space interface is used with
     modern cfg80211/mac80211 drivers. Do not support Wireless
     Extensions for Wi-Fi 7 devices at all. Everyone should switch to
     using nl80211 interface instead.

   - Improve the CAN bit timing configuration. Use extack to return
     error messages directly to user space, update the SJW handling,
     including the definition of a new default value that will benefit
     CAN-FD controllers, by increasing their oscillator tolerance.

  New hardware / drivers:

   - Ethernet:
      - nVidia BlueField-3 support (control traffic driver)
      - Ethernet support for imx93 SoCs
      - Motorcomm yt8531 gigabit Ethernet PHY
      - onsemi NCN26000 10BASE-T1S PHY (with support for PLCA)
      - Microchip LAN8841 PHY (incl. cable diagnostics and PTP)
      - Amlogic gxl MDIO mux

   - WiFi:
      - RealTek RTL8188EU (rtl8xxxu)
      - Qualcomm Wi-Fi 7 devices (ath12k)

   - CAN:
      - Renesas R-Car V4H

  Drivers:

   - Bluetooth:
      - Set Per Platform Antenna Gain (PPAG) for Intel controllers.

   - Ethernet NICs:
      - Intel (1G, igc):
         - support TSN / Qbv / packet scheduling features of i226 model
      - Intel (100G, ice):
         - use GNSS subsystem instead of TTY
         - multi-buffer XDP support
         - extend support for GPIO pins to E823 devices
      - nVidia/Mellanox:
         - update the shared buffer configuration on PFC commands
         - implement PTP adjphase function for HW offset control
         - TC support for Geneve and GRE with VF tunnel offload
         - more efficient crypto key management method
         - multi-port eswitch support
      - Netronome/Corigine:
         - add DCB IEEE support
         - support IPsec offloading for NFP3800
      - Freescale/NXP (enetc):
         - support XDP_REDIRECT for XDP non-linear buffers
         - improve reconfig, avoid link flap and waiting for idle
         - support MAC Merge layer
      - Other NICs:
         - sfc/ef100: add basic devlink support for ef100
         - ionic: rx_push mode operation (writing descriptors via MMIO)
         - bnxt: use the auxiliary bus abstraction for RDMA
         - r8169: disable ASPM and reset bus in case of tx timeout
         - cpsw: support QSGMII mode for J721e CPSW9G
         - cpts: support pulse-per-second output
         - ngbe: add an mdio bus driver
         - usbnet: optimize usbnet_bh() by avoiding unnecessary queuing
         - r8152: handle devices with FW with NCM support
         - amd-xgbe: support 10Mbps, 2.5GbE speeds and rx-adaptation
         - virtio-net: support multi buffer XDP
         - virtio/vsock: replace virtio_vsock_pkt with sk_buff
         - tsnep: XDP support

   - Ethernet high-speed switches:
      - nVidia/Mellanox (mlxsw):
         - add support for latency TLV (in FW control messages)
      - Microchip (sparx5):
         - separate explicit and implicit traffic forwarding rules, make
           the implicit rules always active
         - add support for egress DSCP rewrite
         - IS0 VCAP support (Ingress Classification)
         - IS2 VCAP filters (protos, L3 addrs, L4 ports, flags, ToS
           etc.)
         - ES2 VCAP support (Egress Access Control)
         - support for Per-Stream Filtering and Policing (802.1Q,
           8.6.5.1)

   - Ethernet embedded switches:
      - Marvell (mv88e6xxx):
         - add MAB (port auth) offload support
         - enable PTP receive for mv88e6390
      - NXP (ocelot):
         - support MAC Merge layer
         - support for the the vsc7512 internal copper phys
      - Microchip:
         - lan9303: convert to PHYLINK
         - lan966x: support TC flower filter statistics
         - lan937x: PTP support for KSZ9563/KSZ8563 and LAN937x
         - lan937x: support Credit Based Shaper configuration
         - ksz9477: support Energy Efficient Ethernet
      - other:
         - qca8k: convert to regmap read/write API, use bulk operations
         - rswitch: Improve TX timestamp accuracy

   - Intel WiFi (iwlwifi):
      - EHT (Wi-Fi 7) rate reporting
      - STEP equalizer support: transfer some STEP (connection to radio
        on platforms with integrated wifi) related parameters from the
        BIOS to the firmware.

   - Qualcomm 802.11ax WiFi (ath11k):
      - IPQ5018 support
      - Fine Timing Measurement (FTM) responder role support
      - channel 177 support

   - MediaTek WiFi (mt76):
      - per-PHY LED support
      - mt7996: EHT (Wi-Fi 7) support
      - Wireless Ethernet Dispatch (WED) reset support
      - switch to using page pool allocator

   - RealTek WiFi (rtw89):
      - support new version of Bluetooth co-existance

   - Mobile:
      - rmnet: support TX aggregation"

* tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1872 commits)
  page_pool: add a comment explaining the fragment counter usage
  net: ethtool: fix __ethtool_dev_mm_supported() implementation
  ethtool: pse-pd: Fix double word in comments
  xsk: add linux/vmalloc.h to xsk.c
  sefltests: netdevsim: wait for devlink instance after netns removal
  selftest: fib_tests: Always cleanup before exit
  net/mlx5e: Align IPsec ASO result memory to be as required by hardware
  net/mlx5e: TC, Set CT miss to the specific ct action instance
  net/mlx5e: Rename CHAIN_TO_REG to MAPPED_OBJ_TO_REG
  net/mlx5: Refactor tc miss handling to a single function
  net/mlx5: Kconfig: Make tc offload depend on tc skb extension
  net/sched: flower: Support hardware miss to tc action
  net/sched: flower: Move filter handle initialization earlier
  net/sched: cls_api: Support hardware miss to tc action
  net/sched: Rename user cookie and act cookie
  sfc: fix builds without CONFIG_RTC_LIB
  sfc: clean up some inconsistent indentings
  net/mlx4_en: Introduce flexible array to silence overflow warning
  net: lan966x: Fix possible deadlock inside PTP
  net/ulp: Remove redundant ->clone() test in inet_clone_ulp().
  ...

1 files changed, 739 insertions, 0 deletions

diff --git a/arch/x86/crypto/serpent-sse2-x86_64-asm_64.S b/arch/x86/crypto/serpent-sse2-x86_64-asm_64.S
new file mode 100644
index 000000000..e0998a011
--- /dev/null
+++ b/arch/x86/crypto/serpent-sse2-x86_64-asm_64.S
@@ -0,0 +1,739 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Serpent Cipher 8-way parallel algorithm (x86_64/SSE2)
+ *
+ * Copyright (C) 2011 Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
+ *
+ * Based on crypto/serpent.c by
+ *  Copyright (C) 2002 Dag Arne Osvik <osvik@ii.uib.no>
+ *                2003 Herbert Valerio Riedel <hvr@gnu.org>
+ */
+
+#include <linux/linkage.h>
+
+.file "serpent-sse2-x86_64-asm_64.S"
+.text
+
+#define CTX %rdi
+
+/**********************************************************************
+  8-way SSE2 serpent
+ **********************************************************************/
+#define RA1 %xmm0
+#define RB1 %xmm1
+#define RC1 %xmm2
+#define RD1 %xmm3
+#define RE1 %xmm4
+
+#define RA2 %xmm5
+#define RB2 %xmm6
+#define RC2 %xmm7
+#define RD2 %xmm8
+#define RE2 %xmm9
+
+#define RNOT %xmm10
+
+#define RK0 %xmm11
+#define RK1 %xmm12
+#define RK2 %xmm13
+#define RK3 %xmm14
+
+#define S0_1(x0, x1, x2, x3, x4) \
+	movdqa x3,		x4; \
+	por x0,			x3; \
+	pxor x4,		x0; \
+	pxor x2,		x4; \
+	pxor RNOT,		x4; \
+	pxor x1,		x3; \
+	pand x0,		x1; \
+	pxor x4,		x1; \
+	pxor x0,		x2;
+#define S0_2(x0, x1, x2, x3, x4) \
+	pxor x3,		x0; \
+	por x0,			x4; \
+	pxor x2,		x0; \
+	pand x1,		x2; \
+	pxor x2,		x3; \
+	pxor RNOT,		x1; \
+	pxor x4,		x2; \
+	pxor x2,		x1;
+
+#define S1_1(x0, x1, x2, x3, x4) \
+	movdqa x1,		x4; \
+	pxor x0,		x1; \
+	pxor x3,		x0; \
+	pxor RNOT,		x3; \
+	pand x1,		x4; \
+	por x1,			x0; \
+	pxor x2,		x3; \
+	pxor x3,		x0; \
+	pxor x3,		x1;
+#define S1_2(x0, x1, x2, x3, x4) \
+	pxor x4,		x3; \
+	por x4,			x1; \
+	pxor x2,		x4; \
+	pand x0,		x2; \
+	pxor x1,		x2; \
+	por x0,			x1; \
+	pxor RNOT,		x0; \
+	pxor x2,		x0; \
+	pxor x1,		x4;
+
+#define S2_1(x0, x1, x2, x3, x4) \
+	pxor RNOT,		x3; \
+	pxor x0,		x1; \
+	movdqa x0,		x4; \
+	pand x2,		x0; \
+	pxor x3,		x0; \
+	por x4,			x3; \
+	pxor x1,		x2; \
+	pxor x1,		x3; \
+	pand x0,		x1;
+#define S2_2(x0, x1, x2, x3, x4) \
+	pxor x2,		x0; \
+	pand x3,		x2; \
+	por x1,			x3; \
+	pxor RNOT,		x0; \
+	pxor x0,		x3; \
+	pxor x0,		x4; \
+	pxor x2,		x0; \
+	por x2,			x1;
+
+#define S3_1(x0, x1, x2, x3, x4) \
+	movdqa x1,		x4; \
+	pxor x3,		x1; \
+	por x0,			x3; \
+	pand x0,		x4; \
+	pxor x2,		x0; \
+	pxor x1,		x2; \
+	pand x3,		x1; \
+	pxor x3,		x2; \
+	por x4,			x0; \
+	pxor x3,		x4;
+#define S3_2(x0, x1, x2, x3, x4) \
+	pxor x0,		x1; \
+	pand x3,		x0; \
+	pand x4,		x3; \
+	pxor x2,		x3; \
+	por x1,			x4; \
+	pand x1,		x2; \
+	pxor x3,		x4; \
+	pxor x3,		x0; \
+	pxor x2,		x3;
+
+#define S4_1(x0, x1, x2, x3, x4) \
+	movdqa x3,		x4; \
+	pand x0,		x3; \
+	pxor x4,		x0; \
+	pxor x2,		x3; \
+	por x4,			x2; \
+	pxor x1,		x0; \
+	pxor x3,		x4; \
+	por x0,			x2; \
+	pxor x1,		x2;
+#define S4_2(x0, x1, x2, x3, x4) \
+	pand x0,		x1; \
+	pxor x4,		x1; \
+	pand x2,		x4; \
+	pxor x3,		x2; \
+	pxor x0,		x4; \
+	por x1,			x3; \
+	pxor RNOT,		x1; \
+	pxor x0,		x3;
+
+#define S5_1(x0, x1, x2, x3, x4) \
+	movdqa x1,		x4; \
+	por x0,			x1; \
+	pxor x1,		x2; \
+	pxor RNOT,		x3; \
+	pxor x0,		x4; \
+	pxor x2,		x0; \
+	pand x4,		x1; \
+	por x3,			x4; \
+	pxor x0,		x4;
+#define S5_2(x0, x1, x2, x3, x4) \
+	pand x3,		x0; \
+	pxor x3,		x1; \
+	pxor x2,		x3; \
+	pxor x1,		x0; \
+	pand x4,		x2; \
+	pxor x2,		x1; \
+	pand x0,		x2; \
+	pxor x2,		x3;
+
+#define S6_1(x0, x1, x2, x3, x4) \
+	movdqa x1,		x4; \
+	pxor x0,		x3; \
+	pxor x2,		x1; \
+	pxor x0,		x2; \
+	pand x3,		x0; \
+	por x3,			x1; \
+	pxor RNOT,		x4; \
+	pxor x1,		x0; \
+	pxor x2,		x1;
+#define S6_2(x0, x1, x2, x3, x4) \
+	pxor x4,		x3; \
+	pxor x0,		x4; \
+	pand x0,		x2; \
+	pxor x1,		x4; \
+	pxor x3,		x2; \
+	pand x1,		x3; \
+	pxor x0,		x3; \
+	pxor x2,		x1;
+
+#define S7_1(x0, x1, x2, x3, x4) \
+	pxor RNOT,		x1; \
+	movdqa x1,		x4; \
+	pxor RNOT,		x0; \
+	pand x2,		x1; \
+	pxor x3,		x1; \
+	por x4,			x3; \
+	pxor x2,		x4; \
+	pxor x3,		x2; \
+	pxor x0,		x3; \
+	por x1,			x0;
+#define S7_2(x0, x1, x2, x3, x4) \
+	pand x0,		x2; \
+	pxor x4,		x0; \
+	pxor x3,		x4; \
+	pand x0,		x3; \
+	pxor x1,		x4; \
+	pxor x4,		x2; \
+	pxor x1,		x3; \
+	por x0,			x4; \
+	pxor x1,		x4;
+
+#define SI0_1(x0, x1, x2, x3, x4) \
+	movdqa x3,		x4; \
+	pxor x0,		x1; \
+	por x1,			x3; \
+	pxor x1,		x4; \
+	pxor RNOT,		x0; \
+	pxor x3,		x2; \
+	pxor x0,		x3; \
+	pand x1,		x0; \
+	pxor x2,		x0;
+#define SI0_2(x0, x1, x2, x3, x4) \
+	pand x3,		x2; \
+	pxor x4,		x3; \
+	pxor x3,		x2; \
+	pxor x3,		x1; \
+	pand x0,		x3; \
+	pxor x0,		x1; \
+	pxor x2,		x0; \
+	pxor x3,		x4;
+
+#define SI1_1(x0, x1, x2, x3, x4) \
+	pxor x3,		x1; \
+	movdqa x0,		x4; \
+	pxor x2,		x0; \
+	pxor RNOT,		x2; \
+	por x1,			x4; \
+	pxor x3,		x4; \
+	pand x1,		x3; \
+	pxor x2,		x1; \
+	pand x4,		x2;
+#define SI1_2(x0, x1, x2, x3, x4) \
+	pxor x1,		x4; \
+	por x3,			x1; \
+	pxor x0,		x3; \
+	pxor x0,		x2; \
+	por x4,			x0; \
+	pxor x4,		x2; \
+	pxor x0,		x1; \
+	pxor x1,		x4;
+
+#define SI2_1(x0, x1, x2, x3, x4) \
+	pxor x1,		x2; \
+	movdqa x3,		x4; \
+	pxor RNOT,		x3; \
+	por x2,			x3; \
+	pxor x4,		x2; \
+	pxor x0,		x4; \
+	pxor x1,		x3; \
+	por x2,			x1; \
+	pxor x0,		x2;
+#define SI2_2(x0, x1, x2, x3, x4) \
+	pxor x4,		x1; \
+	por x3,			x4; \
+	pxor x3,		x2; \
+	pxor x2,		x4; \
+	pand x1,		x2; \
+	pxor x3,		x2; \
+	pxor x4,		x3; \
+	pxor x0,		x4;
+
+#define SI3_1(x0, x1, x2, x3, x4) \
+	pxor x1,		x2; \
+	movdqa x1,		x4; \
+	pand x2,		x1; \
+	pxor x0,		x1; \
+	por x4,			x0; \
+	pxor x3,		x4; \
+	pxor x3,		x0; \
+	por x1,			x3; \
+	pxor x2,		x1;
+#define SI3_2(x0, x1, x2, x3, x4) \
+	pxor x3,		x1; \
+	pxor x2,		x0; \
+	pxor x3,		x2; \
+	pand x1,		x3; \
+	pxor x0,		x1; \
+	pand x2,		x0; \
+	pxor x3,		x4; \
+	pxor x0,		x3; \
+	pxor x1,		x0;
+
+#define SI4_1(x0, x1, x2, x3, x4) \
+	pxor x3,		x2; \
+	movdqa x0,		x4; \
+	pand x1,		x0; \
+	pxor x2,		x0; \
+	por x3,			x2; \
+	pxor RNOT,		x4; \
+	pxor x0,		x1; \
+	pxor x2,		x0; \
+	pand x4,		x2;
+#define SI4_2(x0, x1, x2, x3, x4) \
+	pxor x0,		x2; \
+	por x4,			x0; \
+	pxor x3,		x0; \
+	pand x2,		x3; \
+	pxor x3,		x4; \
+	pxor x1,		x3; \
+	pand x0,		x1; \
+	pxor x1,		x4; \
+	pxor x3,		x0;
+
+#define SI5_1(x0, x1, x2, x3, x4) \
+	movdqa x1,		x4; \
+	por x2,			x1; \
+	pxor x4,		x2; \
+	pxor x3,		x1; \
+	pand x4,		x3; \
+	pxor x3,		x2; \
+	por x0,			x3; \
+	pxor RNOT,		x0; \
+	pxor x2,		x3; \
+	por x0,			x2;
+#define SI5_2(x0, x1, x2, x3, x4) \
+	pxor x1,		x4; \
+	pxor x4,		x2; \
+	pand x0,		x4; \
+	pxor x1,		x0; \
+	pxor x3,		x1; \
+	pand x2,		x0; \
+	pxor x3,		x2; \
+	pxor x2,		x0; \
+	pxor x4,		x2; \
+	pxor x3,		x4;
+
+#define SI6_1(x0, x1, x2, x3, x4) \
+	pxor x2,		x0; \
+	movdqa x0,		x4; \
+	pand x3,		x0; \
+	pxor x3,		x2; \
+	pxor x2,		x0; \
+	pxor x1,		x3; \
+	por x4,			x2; \
+	pxor x3,		x2; \
+	pand x0,		x3;
+#define SI6_2(x0, x1, x2, x3, x4) \
+	pxor RNOT,		x0; \
+	pxor x1,		x3; \
+	pand x2,		x1; \
+	pxor x0,		x4; \
+	pxor x4,		x3; \
+	pxor x2,		x4; \
+	pxor x1,		x0; \
+	pxor x0,		x2;
+
+#define SI7_1(x0, x1, x2, x3, x4) \
+	movdqa x3,		x4; \
+	pand x0,		x3; \
+	pxor x2,		x0; \
+	por x4,			x2; \
+	pxor x1,		x4; \
+	pxor RNOT,		x0; \
+	por x3,			x1; \
+	pxor x0,		x4; \
+	pand x2,		x0; \
+	pxor x1,		x0;
+#define SI7_2(x0, x1, x2, x3, x4) \
+	pand x2,		x1; \
+	pxor x2,		x3; \
+	pxor x3,		x4; \
+	pand x3,		x2; \
+	por x0,			x3; \
+	pxor x4,		x1; \
+	pxor x4,		x3; \
+	pand x0,		x4; \
+	pxor x2,		x4;
+
+#define get_key(i, j, t) \
+	movd (4*(i)+(j))*4(CTX), t; \
+	pshufd $0, t, t;
+
+#define K2(x0, x1, x2, x3, x4, i) \
+	get_key(i, 0, RK0); \
+	get_key(i, 1, RK1); \
+	get_key(i, 2, RK2); \
+	get_key(i, 3, RK3); \
+	pxor RK0,		x0 ## 1; \
+	pxor RK1,		x1 ## 1; \
+	pxor RK2,		x2 ## 1; \
+	pxor RK3,		x3 ## 1; \
+		pxor RK0,		x0 ## 2; \
+		pxor RK1,		x1 ## 2; \
+		pxor RK2,		x2 ## 2; \
+		pxor RK3,		x3 ## 2;
+
+#define LK2(x0, x1, x2, x3, x4, i) \
+	movdqa x0 ## 1,		x4 ## 1; \
+	pslld $13,		x0 ## 1; \
+	psrld $(32 - 13),	x4 ## 1; \
+	por x4 ## 1,		x0 ## 1; \
+	pxor x0 ## 1,		x1 ## 1; \
+	movdqa x2 ## 1,		x4 ## 1; \
+	pslld $3,		x2 ## 1; \
+	psrld $(32 - 3),	x4 ## 1; \
+	por x4 ## 1,		x2 ## 1; \
+	pxor x2 ## 1,		x1 ## 1; \
+		movdqa x0 ## 2,		x4 ## 2; \
+		pslld $13,		x0 ## 2; \
+		psrld $(32 - 13),	x4 ## 2; \
+		por x4 ## 2,		x0 ## 2; \
+		pxor x0 ## 2,		x1 ## 2; \
+		movdqa x2 ## 2,		x4 ## 2; \
+		pslld $3,		x2 ## 2; \
+		psrld $(32 - 3),	x4 ## 2; \
+		por x4 ## 2,		x2 ## 2; \
+		pxor x2 ## 2,		x1 ## 2; \
+	movdqa x1 ## 1,		x4 ## 1; \
+	pslld $1,		x1 ## 1; \
+	psrld $(32 - 1),	x4 ## 1; \
+	por x4 ## 1,		x1 ## 1; \
+	movdqa x0 ## 1,		x4 ## 1; \
+	pslld $3,		x4 ## 1; \
+	pxor x2 ## 1,		x3 ## 1; \
+	pxor x4 ## 1,		x3 ## 1; \
+	movdqa x3 ## 1,		x4 ## 1; \
+	get_key(i, 1, RK1); \
+		movdqa x1 ## 2,		x4 ## 2; \
+		pslld $1,		x1 ## 2; \
+		psrld $(32 - 1),	x4 ## 2; \
+		por x4 ## 2,		x1 ## 2; \
+		movdqa x0 ## 2,		x4 ## 2; \
+		pslld $3,		x4 ## 2; \
+		pxor x2 ## 2,		x3 ## 2; \
+		pxor x4 ## 2,		x3 ## 2; \
+		movdqa x3 ## 2,		x4 ## 2; \
+		get_key(i, 3, RK3); \
+	pslld $7,		x3 ## 1; \
+	psrld $(32 - 7),	x4 ## 1; \
+	por x4 ## 1,		x3 ## 1; \
+	movdqa x1 ## 1,		x4 ## 1; \
+	pslld $7,		x4 ## 1; \
+	pxor x1 ## 1,		x0 ## 1; \
+	pxor x3 ## 1,		x0 ## 1; \
+	pxor x3 ## 1,		x2 ## 1; \
+	pxor x4 ## 1,		x2 ## 1; \
+	get_key(i, 0, RK0); \
+		pslld $7,		x3 ## 2; \
+		psrld $(32 - 7),	x4 ## 2; \
+		por x4 ## 2,		x3 ## 2; \
+		movdqa x1 ## 2,		x4 ## 2; \
+		pslld $7,		x4 ## 2; \
+		pxor x1 ## 2,		x0 ## 2; \
+		pxor x3 ## 2,		x0 ## 2; \
+		pxor x3 ## 2,		x2 ## 2; \
+		pxor x4 ## 2,		x2 ## 2; \
+		get_key(i, 2, RK2); \
+	pxor RK1,		x1 ## 1; \
+	pxor RK3,		x3 ## 1; \
+	movdqa x0 ## 1,		x4 ## 1; \
+	pslld $5,		x0 ## 1; \
+	psrld $(32 - 5),	x4 ## 1; \
+	por x4 ## 1,		x0 ## 1; \
+	movdqa x2 ## 1,		x4 ## 1; \
+	pslld $22,		x2 ## 1; \
+	psrld $(32 - 22),	x4 ## 1; \
+	por x4 ## 1,		x2 ## 1; \
+	pxor RK0,		x0 ## 1; \
+	pxor RK2,		x2 ## 1; \
+		pxor RK1,		x1 ## 2; \
+		pxor RK3,		x3 ## 2; \
+		movdqa x0 ## 2,		x4 ## 2; \
+		pslld $5,		x0 ## 2; \
+		psrld $(32 - 5),	x4 ## 2; \
+		por x4 ## 2,		x0 ## 2; \
+		movdqa x2 ## 2,		x4 ## 2; \
+		pslld $22,		x2 ## 2; \
+		psrld $(32 - 22),	x4 ## 2; \
+		por x4 ## 2,		x2 ## 2; \
+		pxor RK0,		x0 ## 2; \
+		pxor RK2,		x2 ## 2;
+
+#define KL2(x0, x1, x2, x3, x4, i) \
+	pxor RK0,		x0 ## 1; \
+	pxor RK2,		x2 ## 1; \
+	movdqa x0 ## 1,		x4 ## 1; \
+	psrld $5,		x0 ## 1; \
+	pslld $(32 - 5),	x4 ## 1; \
+	por x4 ## 1,		x0 ## 1; \
+	pxor RK3,		x3 ## 1; \
+	pxor RK1,		x1 ## 1; \
+	movdqa x2 ## 1,		x4 ## 1; \
+	psrld $22,		x2 ## 1; \
+	pslld $(32 - 22),	x4 ## 1; \
+	por x4 ## 1,		x2 ## 1; \
+	pxor x3 ## 1,		x2 ## 1; \
+		pxor RK0,		x0 ## 2; \
+		pxor RK2,		x2 ## 2; \
+		movdqa x0 ## 2,		x4 ## 2; \
+		psrld $5,		x0 ## 2; \
+		pslld $(32 - 5),	x4 ## 2; \
+		por x4 ## 2,		x0 ## 2; \
+		pxor RK3,		x3 ## 2; \
+		pxor RK1,		x1 ## 2; \
+		movdqa x2 ## 2,		x4 ## 2; \
+		psrld $22,		x2 ## 2; \
+		pslld $(32 - 22),	x4 ## 2; \
+		por x4 ## 2,		x2 ## 2; \
+		pxor x3 ## 2,		x2 ## 2; \
+	pxor x3 ## 1,		x0 ## 1; \
+	movdqa x1 ## 1,		x4 ## 1; \
+	pslld $7,		x4 ## 1; \
+	pxor x1 ## 1,		x0 ## 1; \
+	pxor x4 ## 1,		x2 ## 1; \
+	movdqa x1 ## 1,		x4 ## 1; \
+	psrld $1,		x1 ## 1; \
+	pslld $(32 - 1),	x4 ## 1; \
+	por x4 ## 1,		x1 ## 1; \
+		pxor x3 ## 2,		x0 ## 2; \
+		movdqa x1 ## 2,		x4 ## 2; \
+		pslld $7,		x4 ## 2; \
+		pxor x1 ## 2,		x0 ## 2; \
+		pxor x4 ## 2,		x2 ## 2; \
+		movdqa x1 ## 2,		x4 ## 2; \
+		psrld $1,		x1 ## 2; \
+		pslld $(32 - 1),	x4 ## 2; \
+		por x4 ## 2,		x1 ## 2; \
+	movdqa x3 ## 1,		x4 ## 1; \
+	psrld $7,		x3 ## 1; \
+	pslld $(32 - 7),	x4 ## 1; \
+	por x4 ## 1,		x3 ## 1; \
+	pxor x0 ## 1,		x1 ## 1; \
+	movdqa x0 ## 1,		x4 ## 1; \
+	pslld $3,		x4 ## 1; \
+	pxor x4 ## 1,		x3 ## 1; \
+	movdqa x0 ## 1,		x4 ## 1; \
+		movdqa x3 ## 2,		x4 ## 2; \
+		psrld $7,		x3 ## 2; \
+		pslld $(32 - 7),	x4 ## 2; \
+		por x4 ## 2,		x3 ## 2; \
+		pxor x0 ## 2,		x1 ## 2; \
+		movdqa x0 ## 2,		x4 ## 2; \
+		pslld $3,		x4 ## 2; \
+		pxor x4 ## 2,		x3 ## 2; \
+		movdqa x0 ## 2,		x4 ## 2; \
+	psrld $13,		x0 ## 1; \
+	pslld $(32 - 13),	x4 ## 1; \
+	por x4 ## 1,		x0 ## 1; \
+	pxor x2 ## 1,		x1 ## 1; \
+	pxor x2 ## 1,		x3 ## 1; \
+	movdqa x2 ## 1,		x4 ## 1; \
+	psrld $3,		x2 ## 1; \
+	pslld $(32 - 3),	x4 ## 1; \
+	por x4 ## 1,		x2 ## 1; \
+		psrld $13,		x0 ## 2; \
+		pslld $(32 - 13),	x4 ## 2; \
+		por x4 ## 2,		x0 ## 2; \
+		pxor x2 ## 2,		x1 ## 2; \
+		pxor x2 ## 2,		x3 ## 2; \
+		movdqa x2 ## 2,		x4 ## 2; \
+		psrld $3,		x2 ## 2; \
+		pslld $(32 - 3),	x4 ## 2; \
+		por x4 ## 2,		x2 ## 2;
+
+#define S(SBOX, x0, x1, x2, x3, x4) \
+	SBOX ## _1(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
+	SBOX ## _2(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
+	SBOX ## _1(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2); \
+	SBOX ## _2(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2);
+
+#define SP(SBOX, x0, x1, x2, x3, x4, i) \
+	get_key(i, 0, RK0); \
+	SBOX ## _1(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
+	get_key(i, 2, RK2); \
+	SBOX ## _1(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2); \
+	get_key(i, 3, RK3); \
+	SBOX ## _2(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
+	get_key(i, 1, RK1); \
+	SBOX ## _2(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2); \
+
+#define transpose_4x4(x0, x1, x2, x3, t0, t1, t2) \
+	movdqa x0,		t2; \
+	punpckldq x1,		x0; \
+	punpckhdq x1,		t2; \
+	movdqa x2,		t1; \
+	punpckhdq x3,		x2; \
+	punpckldq x3,		t1; \
+	movdqa x0,		x1; \
+	punpcklqdq t1,		x0; \
+	punpckhqdq t1,		x1; \
+	movdqa t2,		x3; \
+	punpcklqdq x2,		t2; \
+	punpckhqdq x2,		x3; \
+	movdqa t2,		x2;
+
+#define read_blocks(in, x0, x1, x2, x3, t0, t1, t2) \
+	movdqu (0*4*4)(in),	x0; \
+	movdqu (1*4*4)(in),	x1; \
+	movdqu (2*4*4)(in),	x2; \
+	movdqu (3*4*4)(in),	x3; \
+	\
+	transpose_4x4(x0, x1, x2, x3, t0, t1, t2)
+
+#define write_blocks(out, x0, x1, x2, x3, t0, t1, t2) \
+	transpose_4x4(x0, x1, x2, x3, t0, t1, t2) \
+	\
+	movdqu x0,		(0*4*4)(out); \
+	movdqu x1,		(1*4*4)(out); \
+	movdqu x2,		(2*4*4)(out); \
+	movdqu x3,		(3*4*4)(out);
+
+#define xor_blocks(out, x0, x1, x2, x3, t0, t1, t2) \
+	transpose_4x4(x0, x1, x2, x3, t0, t1, t2) \
+	\
+	movdqu (0*4*4)(out),	t0; \
+	pxor t0,		x0; \
+	movdqu x0,		(0*4*4)(out); \
+	movdqu (1*4*4)(out),	t0; \
+	pxor t0,		x1; \
+	movdqu x1,		(1*4*4)(out); \
+	movdqu (2*4*4)(out),	t0; \
+	pxor t0,		x2; \
+	movdqu x2,		(2*4*4)(out); \
+	movdqu (3*4*4)(out),	t0; \
+	pxor t0,		x3; \
+	movdqu x3,		(3*4*4)(out);
+
+SYM_FUNC_START(__serpent_enc_blk_8way)
+	/* input:
+	 *	%rdi: ctx, CTX
+	 *	%rsi: dst
+	 *	%rdx: src
+	 *	%rcx: bool, if true: xor output
+	 */
+
+	pcmpeqd RNOT, RNOT;
+
+	leaq (4*4*4)(%rdx), %rax;
+	read_blocks(%rdx, RA1, RB1, RC1, RD1, RK0, RK1, RK2);
+	read_blocks(%rax, RA2, RB2, RC2, RD2, RK0, RK1, RK2);
+
+						 K2(RA, RB, RC, RD, RE, 0);
+	S(S0, RA, RB, RC, RD, RE);		LK2(RC, RB, RD, RA, RE, 1);
+	S(S1, RC, RB, RD, RA, RE);		LK2(RE, RD, RA, RC, RB, 2);
+	S(S2, RE, RD, RA, RC, RB);		LK2(RB, RD, RE, RC, RA, 3);
+	S(S3, RB, RD, RE, RC, RA);		LK2(RC, RA, RD, RB, RE, 4);
+	S(S4, RC, RA, RD, RB, RE);		LK2(RA, RD, RB, RE, RC, 5);
+	S(S5, RA, RD, RB, RE, RC);		LK2(RC, RA, RD, RE, RB, 6);
+	S(S6, RC, RA, RD, RE, RB);		LK2(RD, RB, RA, RE, RC, 7);
+	S(S7, RD, RB, RA, RE, RC);		LK2(RC, RA, RE, RD, RB, 8);
+	S(S0, RC, RA, RE, RD, RB);		LK2(RE, RA, RD, RC, RB, 9);
+	S(S1, RE, RA, RD, RC, RB);		LK2(RB, RD, RC, RE, RA, 10);
+	S(S2, RB, RD, RC, RE, RA);		LK2(RA, RD, RB, RE, RC, 11);
+	S(S3, RA, RD, RB, RE, RC);		LK2(RE, RC, RD, RA, RB, 12);
+	S(S4, RE, RC, RD, RA, RB);		LK2(RC, RD, RA, RB, RE, 13);
+	S(S5, RC, RD, RA, RB, RE);		LK2(RE, RC, RD, RB, RA, 14);
+	S(S6, RE, RC, RD, RB, RA);		LK2(RD, RA, RC, RB, RE, 15);
+	S(S7, RD, RA, RC, RB, RE);		LK2(RE, RC, RB, RD, RA, 16);
+	S(S0, RE, RC, RB, RD, RA);		LK2(RB, RC, RD, RE, RA, 17);
+	S(S1, RB, RC, RD, RE, RA);		LK2(RA, RD, RE, RB, RC, 18);
+	S(S2, RA, RD, RE, RB, RC);		LK2(RC, RD, RA, RB, RE, 19);
+	S(S3, RC, RD, RA, RB, RE);		LK2(RB, RE, RD, RC, RA, 20);
+	S(S4, RB, RE, RD, RC, RA);		LK2(RE, RD, RC, RA, RB, 21);
+	S(S5, RE, RD, RC, RA, RB);		LK2(RB, RE, RD, RA, RC, 22);
+	S(S6, RB, RE, RD, RA, RC);		LK2(RD, RC, RE, RA, RB, 23);
+	S(S7, RD, RC, RE, RA, RB);		LK2(RB, RE, RA, RD, RC, 24);
+	S(S0, RB, RE, RA, RD, RC);		LK2(RA, RE, RD, RB, RC, 25);
+	S(S1, RA, RE, RD, RB, RC);		LK2(RC, RD, RB, RA, RE, 26);
+	S(S2, RC, RD, RB, RA, RE);		LK2(RE, RD, RC, RA, RB, 27);
+	S(S3, RE, RD, RC, RA, RB);		LK2(RA, RB, RD, RE, RC, 28);
+	S(S4, RA, RB, RD, RE, RC);		LK2(RB, RD, RE, RC, RA, 29);
+	S(S5, RB, RD, RE, RC, RA);		LK2(RA, RB, RD, RC, RE, 30);
+	S(S6, RA, RB, RD, RC, RE);		LK2(RD, RE, RB, RC, RA, 31);
+	S(S7, RD, RE, RB, RC, RA);		 K2(RA, RB, RC, RD, RE, 32);
+
+	leaq (4*4*4)(%rsi), %rax;
+
+	testb %cl, %cl;
+	jnz .L__enc_xor8;
+
+	write_blocks(%rsi, RA1, RB1, RC1, RD1, RK0, RK1, RK2);
+	write_blocks(%rax, RA2, RB2, RC2, RD2, RK0, RK1, RK2);
+
+	RET;
+
+.L__enc_xor8:
+	xor_blocks(%rsi, RA1, RB1, RC1, RD1, RK0, RK1, RK2);
+	xor_blocks(%rax, RA2, RB2, RC2, RD2, RK0, RK1, RK2);
+
+	RET;
+SYM_FUNC_END(__serpent_enc_blk_8way)
+
+SYM_FUNC_START(serpent_dec_blk_8way)
+	/* input:
+	 *	%rdi: ctx, CTX
+	 *	%rsi: dst
+	 *	%rdx: src
+	 */
+
+	pcmpeqd RNOT, RNOT;
+
+	leaq (4*4*4)(%rdx), %rax;
+	read_blocks(%rdx, RA1, RB1, RC1, RD1, RK0, RK1, RK2);
+	read_blocks(%rax, RA2, RB2, RC2, RD2, RK0, RK1, RK2);
+
+						 K2(RA, RB, RC, RD, RE, 32);
+	SP(SI7, RA, RB, RC, RD, RE, 31);	KL2(RB, RD, RA, RE, RC, 31);
+	SP(SI6, RB, RD, RA, RE, RC, 30);	KL2(RA, RC, RE, RB, RD, 30);
+	SP(SI5, RA, RC, RE, RB, RD, 29);	KL2(RC, RD, RA, RE, RB, 29);
+	SP(SI4, RC, RD, RA, RE, RB, 28);	KL2(RC, RA, RB, RE, RD, 28);
+	SP(SI3, RC, RA, RB, RE, RD, 27);	KL2(RB, RC, RD, RE, RA, 27);
+	SP(SI2, RB, RC, RD, RE, RA, 26);	KL2(RC, RA, RE, RD, RB, 26);
+	SP(SI1, RC, RA, RE, RD, RB, 25);	KL2(RB, RA, RE, RD, RC, 25);
+	SP(SI0, RB, RA, RE, RD, RC, 24);	KL2(RE, RC, RA, RB, RD, 24);
+	SP(SI7, RE, RC, RA, RB, RD, 23);	KL2(RC, RB, RE, RD, RA, 23);
+	SP(SI6, RC, RB, RE, RD, RA, 22);	KL2(RE, RA, RD, RC, RB, 22);
+	SP(SI5, RE, RA, RD, RC, RB, 21);	KL2(RA, RB, RE, RD, RC, 21);
+	SP(SI4, RA, RB, RE, RD, RC, 20);	KL2(RA, RE, RC, RD, RB, 20);
+	SP(SI3, RA, RE, RC, RD, RB, 19);	KL2(RC, RA, RB, RD, RE, 19);
+	SP(SI2, RC, RA, RB, RD, RE, 18);	KL2(RA, RE, RD, RB, RC, 18);
+	SP(SI1, RA, RE, RD, RB, RC, 17);	KL2(RC, RE, RD, RB, RA, 17);
+	SP(SI0, RC, RE, RD, RB, RA, 16);	KL2(RD, RA, RE, RC, RB, 16);
+	SP(SI7, RD, RA, RE, RC, RB, 15);	KL2(RA, RC, RD, RB, RE, 15);
+	SP(SI6, RA, RC, RD, RB, RE, 14);	KL2(RD, RE, RB, RA, RC, 14);
+	SP(SI5, RD, RE, RB, RA, RC, 13);	KL2(RE, RC, RD, RB, RA, 13);
+	SP(SI4, RE, RC, RD, RB, RA, 12);	KL2(RE, RD, RA, RB, RC, 12);
+	SP(SI3, RE, RD, RA, RB, RC, 11);	KL2(RA, RE, RC, RB, RD, 11);
+	SP(SI2, RA, RE, RC, RB, RD, 10);	KL2(RE, RD, RB, RC, RA, 10);
+	SP(SI1, RE, RD, RB, RC, RA, 9);		KL2(RA, RD, RB, RC, RE, 9);
+	SP(SI0, RA, RD, RB, RC, RE, 8);		KL2(RB, RE, RD, RA, RC, 8);
+	SP(SI7, RB, RE, RD, RA, RC, 7);		KL2(RE, RA, RB, RC, RD, 7);
+	SP(SI6, RE, RA, RB, RC, RD, 6);		KL2(RB, RD, RC, RE, RA, 6);
+	SP(SI5, RB, RD, RC, RE, RA, 5);		KL2(RD, RA, RB, RC, RE, 5);
+	SP(SI4, RD, RA, RB, RC, RE, 4);		KL2(RD, RB, RE, RC, RA, 4);
+	SP(SI3, RD, RB, RE, RC, RA, 3);		KL2(RE, RD, RA, RC, RB, 3);
+	SP(SI2, RE, RD, RA, RC, RB, 2);		KL2(RD, RB, RC, RA, RE, 2);
+	SP(SI1, RD, RB, RC, RA, RE, 1);		KL2(RE, RB, RC, RA, RD, 1);
+	S(SI0, RE, RB, RC, RA, RD);		 K2(RC, RD, RB, RE, RA, 0);
+
+	leaq (4*4*4)(%rsi), %rax;
+	write_blocks(%rsi, RC1, RD1, RB1, RE1, RK0, RK1, RK2);
+	write_blocks(%rax, RC2, RD2, RB2, RE2, RK0, RK1, RK2);
+
+	RET;
+SYM_FUNC_END(serpent_dec_blk_8way)