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, 605 insertions, 0 deletions

diff --git a/arch/alpha/lib/ev6-memset.S b/arch/alpha/lib/ev6-memset.S
new file mode 100644
index 000000000..1cfcfbbea
--- /dev/null
+++ b/arch/alpha/lib/ev6-memset.S
@@ -0,0 +1,605 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * arch/alpha/lib/ev6-memset.S
+ *
+ * This is an efficient (and relatively small) implementation of the C library
+ * "memset()" function for the 21264 implementation of Alpha.
+ *
+ * 21264 version  contributed by Rick Gorton <rick.gorton@alpha-processor.com>
+ *
+ * Much of the information about 21264 scheduling/coding comes from:
+ *	Compiler Writer's Guide for the Alpha 21264
+ *	abbreviated as 'CWG' in other comments here
+ *	ftp.digital.com/pub/Digital/info/semiconductor/literature/dsc-library.html
+ * Scheduling notation:
+ *	E	- either cluster
+ *	U	- upper subcluster; U0 - subcluster U0; U1 - subcluster U1
+ *	L	- lower subcluster; L0 - subcluster L0; L1 - subcluster L1
+ * The algorithm for the leading and trailing quadwords remains the same,
+ * however the loop has been unrolled to enable better memory throughput,
+ * and the code has been replicated for each of the entry points: __memset
+ * and __memset16 to permit better scheduling to eliminate the stalling
+ * encountered during the mask replication.
+ * A future enhancement might be to put in a byte store loop for really
+ * small (say < 32 bytes) memset()s.  Whether or not that change would be
+ * a win in the kernel would depend upon the contextual usage.
+ * WARNING: Maintaining this is going to be more work than the above version,
+ * as fixes will need to be made in multiple places.  The performance gain
+ * is worth it.
+ */
+#include <asm/export.h>
+	.set noat
+	.set noreorder
+.text
+	.globl memset
+	.globl __memset
+	.globl ___memset
+	.globl __memset16
+	.globl __constant_c_memset
+
+	.ent ___memset
+.align 5
+___memset:
+	.frame $30,0,$26,0
+	.prologue 0
+
+	/*
+	 * Serious stalling happens.  The only way to mitigate this is to
+	 * undertake a major re-write to interleave the constant materialization
+	 * with other parts of the fall-through code.  This is important, even
+	 * though it makes maintenance tougher.
+	 * Do this later.
+	 */
+	and $17,255,$1		# E : 00000000000000ch
+	insbl $17,1,$2		# U : 000000000000ch00
+	bis $16,$16,$0		# E : return value
+	ble $18,end_b		# U : zero length requested?
+
+	addq $18,$16,$6		# E : max address to write to
+	bis	$1,$2,$17	# E : 000000000000chch
+	insbl	$1,2,$3		# U : 0000000000ch0000
+	insbl	$1,3,$4		# U : 00000000ch000000
+
+	or	$3,$4,$3	# E : 00000000chch0000
+	inswl	$17,4,$5	# U : 0000chch00000000
+	xor	$16,$6,$1	# E : will complete write be within one quadword?
+	inswl	$17,6,$2	# U : chch000000000000
+
+	or	$17,$3,$17	# E : 00000000chchchch
+	or	$2,$5,$2	# E : chchchch00000000
+	bic	$1,7,$1		# E : fit within a single quadword?
+	and	$16,7,$3	# E : Target addr misalignment
+
+	or	$17,$2,$17	# E : chchchchchchchch
+	beq	$1,within_quad_b # U :
+	nop			# E :
+	beq	$3,aligned_b	# U : target is 0mod8
+
+	/*
+	 * Target address is misaligned, and won't fit within a quadword
+	 */
+	ldq_u $4,0($16)		# L : Fetch first partial
+	bis $16,$16,$5		# E : Save the address
+	insql $17,$16,$2	# U : Insert new bytes
+	subq $3,8,$3		# E : Invert (for addressing uses)
+
+	addq $18,$3,$18		# E : $18 is new count ($3 is negative)
+	mskql $4,$16,$4		# U : clear relevant parts of the quad
+	subq $16,$3,$16		# E : $16 is new aligned destination
+	bis $2,$4,$1		# E : Final bytes
+
+	nop
+	stq_u $1,0($5)		# L : Store result
+	nop
+	nop
+
+.align 4
+aligned_b:
+	/*
+	 * We are now guaranteed to be quad aligned, with at least
+	 * one partial quad to write.
+	 */
+
+	sra $18,3,$3		# U : Number of remaining quads to write
+	and $18,7,$18		# E : Number of trailing bytes to write
+	bis $16,$16,$5		# E : Save dest address
+	beq $3,no_quad_b	# U : tail stuff only
+
+	/*
+	 * it's worth the effort to unroll this and use wh64 if possible
+	 * Lifted a bunch of code from clear_user.S
+	 * At this point, entry values are:
+	 * $16	Current destination address
+	 * $5	A copy of $16
+	 * $6	The max quadword address to write to
+	 * $18	Number trailer bytes
+	 * $3	Number quads to write
+	 */
+
+	and	$16, 0x3f, $2	# E : Forward work (only useful for unrolled loop)
+	subq	$3, 16, $4	# E : Only try to unroll if > 128 bytes
+	subq	$2, 0x40, $1	# E : bias counter (aligning stuff 0mod64)
+	blt	$4, loop_b	# U :
+
+	/*
+	 * We know we've got at least 16 quads, minimum of one trip
+	 * through unrolled loop.  Do a quad at a time to get us 0mod64
+	 * aligned.
+	 */
+
+	nop			# E :
+	nop			# E :
+	nop			# E :
+	beq	$1, $bigalign_b	# U :
+
+$alignmod64_b:
+	stq	$17, 0($5)	# L :
+	subq	$3, 1, $3	# E : For consistency later
+	addq	$1, 8, $1	# E : Increment towards zero for alignment
+	addq	$5, 8, $4	# E : Initial wh64 address (filler instruction)
+
+	nop
+	nop
+	addq	$5, 8, $5	# E : Inc address
+	blt	$1, $alignmod64_b # U :
+
+$bigalign_b:
+	/*
+	 * $3 - number quads left to go
+	 * $5 - target address (aligned 0mod64)
+	 * $17 - mask of stuff to store
+	 * Scratch registers available: $7, $2, $4, $1
+	 * we know that we'll be taking a minimum of one trip through
+ 	 * CWG Section 3.7.6: do not expect a sustained store rate of > 1/cycle
+	 * Assumes the wh64 needs to be for 2 trips through the loop in the future
+	 * The wh64 is issued on for the starting destination address for trip +2
+	 * through the loop, and if there are less than two trips left, the target
+	 * address will be for the current trip.
+	 */
+
+$do_wh64_b:
+	wh64	($4)		# L1 : memory subsystem write hint
+	subq	$3, 24, $2	# E : For determining future wh64 addresses
+	stq	$17, 0($5)	# L :
+	nop			# E :
+
+	addq	$5, 128, $4	# E : speculative target of next wh64
+	stq	$17, 8($5)	# L :
+	stq	$17, 16($5)	# L :
+	addq	$5, 64, $7	# E : Fallback address for wh64 (== next trip addr)
+
+	stq	$17, 24($5)	# L :
+	stq	$17, 32($5)	# L :
+	cmovlt	$2, $7, $4	# E : Latency 2, extra mapping cycle
+	nop
+
+	stq	$17, 40($5)	# L :
+	stq	$17, 48($5)	# L :
+	subq	$3, 16, $2	# E : Repeat the loop at least once more?
+	nop
+
+	stq	$17, 56($5)	# L :
+	addq	$5, 64, $5	# E :
+	subq	$3, 8, $3	# E :
+	bge	$2, $do_wh64_b	# U :
+
+	nop
+	nop
+	nop
+	beq	$3, no_quad_b	# U : Might have finished already
+
+.align 4
+	/*
+	 * Simple loop for trailing quadwords, or for small amounts
+	 * of data (where we can't use an unrolled loop and wh64)
+	 */
+loop_b:
+	stq $17,0($5)		# L :
+	subq $3,1,$3		# E : Decrement number quads left
+	addq $5,8,$5		# E : Inc address
+	bne $3,loop_b		# U : more?
+
+no_quad_b:
+	/*
+	 * Write 0..7 trailing bytes.
+	 */
+	nop			# E :
+	beq $18,end_b		# U : All done?
+	ldq $7,0($5)		# L :
+	mskqh $7,$6,$2		# U : Mask final quad
+
+	insqh $17,$6,$4		# U : New bits
+	bis $2,$4,$1		# E : Put it all together
+	stq $1,0($5)		# L : And back to memory
+	ret $31,($26),1		# L0 :
+
+within_quad_b:
+	ldq_u $1,0($16)		# L :
+	insql $17,$16,$2	# U : New bits
+	mskql $1,$16,$4		# U : Clear old
+	bis $2,$4,$2		# E : New result
+
+	mskql $2,$6,$4		# U :
+	mskqh $1,$6,$2		# U :
+	bis $2,$4,$1		# E :
+	stq_u $1,0($16)		# L :
+
+end_b:
+	nop
+	nop
+	nop
+	ret $31,($26),1		# L0 :
+	.end ___memset
+	EXPORT_SYMBOL(___memset)
+
+	/*
+	 * This is the original body of code, prior to replication and
+	 * rescheduling.  Leave it here, as there may be calls to this
+	 * entry point.
+	 */
+.align 4
+	.ent __constant_c_memset
+__constant_c_memset:
+	.frame $30,0,$26,0
+	.prologue 0
+
+	addq $18,$16,$6		# E : max address to write to
+	bis $16,$16,$0		# E : return value
+	xor $16,$6,$1		# E : will complete write be within one quadword?
+	ble $18,end		# U : zero length requested?
+
+	bic $1,7,$1		# E : fit within a single quadword
+	beq $1,within_one_quad	# U :
+	and $16,7,$3		# E : Target addr misalignment
+	beq $3,aligned		# U : target is 0mod8
+
+	/*
+	 * Target address is misaligned, and won't fit within a quadword
+	 */
+	ldq_u $4,0($16)		# L : Fetch first partial
+	bis $16,$16,$5		# E : Save the address
+	insql $17,$16,$2	# U : Insert new bytes
+	subq $3,8,$3		# E : Invert (for addressing uses)
+
+	addq $18,$3,$18		# E : $18 is new count ($3 is negative)
+	mskql $4,$16,$4		# U : clear relevant parts of the quad
+	subq $16,$3,$16		# E : $16 is new aligned destination
+	bis $2,$4,$1		# E : Final bytes
+
+	nop
+	stq_u $1,0($5)		# L : Store result
+	nop
+	nop
+
+.align 4
+aligned:
+	/*
+	 * We are now guaranteed to be quad aligned, with at least
+	 * one partial quad to write.
+	 */
+
+	sra $18,3,$3		# U : Number of remaining quads to write
+	and $18,7,$18		# E : Number of trailing bytes to write
+	bis $16,$16,$5		# E : Save dest address
+	beq $3,no_quad		# U : tail stuff only
+
+	/*
+	 * it's worth the effort to unroll this and use wh64 if possible
+	 * Lifted a bunch of code from clear_user.S
+	 * At this point, entry values are:
+	 * $16	Current destination address
+	 * $5	A copy of $16
+	 * $6	The max quadword address to write to
+	 * $18	Number trailer bytes
+	 * $3	Number quads to write
+	 */
+
+	and	$16, 0x3f, $2	# E : Forward work (only useful for unrolled loop)
+	subq	$3, 16, $4	# E : Only try to unroll if > 128 bytes
+	subq	$2, 0x40, $1	# E : bias counter (aligning stuff 0mod64)
+	blt	$4, loop	# U :
+
+	/*
+	 * We know we've got at least 16 quads, minimum of one trip
+	 * through unrolled loop.  Do a quad at a time to get us 0mod64
+	 * aligned.
+	 */
+
+	nop			# E :
+	nop			# E :
+	nop			# E :
+	beq	$1, $bigalign	# U :
+
+$alignmod64:
+	stq	$17, 0($5)	# L :
+	subq	$3, 1, $3	# E : For consistency later
+	addq	$1, 8, $1	# E : Increment towards zero for alignment
+	addq	$5, 8, $4	# E : Initial wh64 address (filler instruction)
+
+	nop
+	nop
+	addq	$5, 8, $5	# E : Inc address
+	blt	$1, $alignmod64	# U :
+
+$bigalign:
+	/*
+	 * $3 - number quads left to go
+	 * $5 - target address (aligned 0mod64)
+	 * $17 - mask of stuff to store
+	 * Scratch registers available: $7, $2, $4, $1
+	 * we know that we'll be taking a minimum of one trip through
+ 	 * CWG Section 3.7.6: do not expect a sustained store rate of > 1/cycle
+	 * Assumes the wh64 needs to be for 2 trips through the loop in the future
+	 * The wh64 is issued on for the starting destination address for trip +2
+	 * through the loop, and if there are less than two trips left, the target
+	 * address will be for the current trip.
+	 */
+
+$do_wh64:
+	wh64	($4)		# L1 : memory subsystem write hint
+	subq	$3, 24, $2	# E : For determining future wh64 addresses
+	stq	$17, 0($5)	# L :
+	nop			# E :
+
+	addq	$5, 128, $4	# E : speculative target of next wh64
+	stq	$17, 8($5)	# L :
+	stq	$17, 16($5)	# L :
+	addq	$5, 64, $7	# E : Fallback address for wh64 (== next trip addr)
+
+	stq	$17, 24($5)	# L :
+	stq	$17, 32($5)	# L :
+	cmovlt	$2, $7, $4	# E : Latency 2, extra mapping cycle
+	nop
+
+	stq	$17, 40($5)	# L :
+	stq	$17, 48($5)	# L :
+	subq	$3, 16, $2	# E : Repeat the loop at least once more?
+	nop
+
+	stq	$17, 56($5)	# L :
+	addq	$5, 64, $5	# E :
+	subq	$3, 8, $3	# E :
+	bge	$2, $do_wh64	# U :
+
+	nop
+	nop
+	nop
+	beq	$3, no_quad	# U : Might have finished already
+
+.align 4
+	/*
+	 * Simple loop for trailing quadwords, or for small amounts
+	 * of data (where we can't use an unrolled loop and wh64)
+	 */
+loop:
+	stq $17,0($5)		# L :
+	subq $3,1,$3		# E : Decrement number quads left
+	addq $5,8,$5		# E : Inc address
+	bne $3,loop		# U : more?
+
+no_quad:
+	/*
+	 * Write 0..7 trailing bytes.
+	 */
+	nop			# E :
+	beq $18,end		# U : All done?
+	ldq $7,0($5)		# L :
+	mskqh $7,$6,$2		# U : Mask final quad
+
+	insqh $17,$6,$4		# U : New bits
+	bis $2,$4,$1		# E : Put it all together
+	stq $1,0($5)		# L : And back to memory
+	ret $31,($26),1		# L0 :
+
+within_one_quad:
+	ldq_u $1,0($16)		# L :
+	insql $17,$16,$2	# U : New bits
+	mskql $1,$16,$4		# U : Clear old
+	bis $2,$4,$2		# E : New result
+
+	mskql $2,$6,$4		# U :
+	mskqh $1,$6,$2		# U :
+	bis $2,$4,$1		# E :
+	stq_u $1,0($16)		# L :
+
+end:
+	nop
+	nop
+	nop
+	ret $31,($26),1		# L0 :
+	.end __constant_c_memset
+	EXPORT_SYMBOL(__constant_c_memset)
+
+	/*
+	 * This is a replicant of the __constant_c_memset code, rescheduled
+	 * to mask stalls.  Note that entry point names also had to change
+	 */
+	.align 5
+	.ent __memset16
+
+__memset16:
+	.frame $30,0,$26,0
+	.prologue 0
+
+	inswl $17,0,$5		# U : 000000000000c1c2
+	inswl $17,2,$2		# U : 00000000c1c20000
+	bis $16,$16,$0		# E : return value
+	addq	$18,$16,$6	# E : max address to write to
+
+	ble $18, end_w		# U : zero length requested?
+	inswl	$17,4,$3	# U : 0000c1c200000000
+	inswl	$17,6,$4	# U : c1c2000000000000
+	xor	$16,$6,$1	# E : will complete write be within one quadword?
+
+	or	$2,$5,$2	# E : 00000000c1c2c1c2
+	or	$3,$4,$17	# E : c1c2c1c200000000
+	bic	$1,7,$1		# E : fit within a single quadword
+	and	$16,7,$3	# E : Target addr misalignment
+
+	or	$17,$2,$17	# E : c1c2c1c2c1c2c1c2
+	beq $1,within_quad_w	# U :
+	nop
+	beq $3,aligned_w	# U : target is 0mod8
+
+	/*
+	 * Target address is misaligned, and won't fit within a quadword
+	 */
+	ldq_u $4,0($16)		# L : Fetch first partial
+	bis $16,$16,$5		# E : Save the address
+	insql $17,$16,$2	# U : Insert new bytes
+	subq $3,8,$3		# E : Invert (for addressing uses)
+
+	addq $18,$3,$18		# E : $18 is new count ($3 is negative)
+	mskql $4,$16,$4		# U : clear relevant parts of the quad
+	subq $16,$3,$16		# E : $16 is new aligned destination
+	bis $2,$4,$1		# E : Final bytes
+
+	nop
+	stq_u $1,0($5)		# L : Store result
+	nop
+	nop
+
+.align 4
+aligned_w:
+	/*
+	 * We are now guaranteed to be quad aligned, with at least
+	 * one partial quad to write.
+	 */
+
+	sra $18,3,$3		# U : Number of remaining quads to write
+	and $18,7,$18		# E : Number of trailing bytes to write
+	bis $16,$16,$5		# E : Save dest address
+	beq $3,no_quad_w	# U : tail stuff only
+
+	/*
+	 * it's worth the effort to unroll this and use wh64 if possible
+	 * Lifted a bunch of code from clear_user.S
+	 * At this point, entry values are:
+	 * $16	Current destination address
+	 * $5	A copy of $16
+	 * $6	The max quadword address to write to
+	 * $18	Number trailer bytes
+	 * $3	Number quads to write
+	 */
+
+	and	$16, 0x3f, $2	# E : Forward work (only useful for unrolled loop)
+	subq	$3, 16, $4	# E : Only try to unroll if > 128 bytes
+	subq	$2, 0x40, $1	# E : bias counter (aligning stuff 0mod64)
+	blt	$4, loop_w	# U :
+
+	/*
+	 * We know we've got at least 16 quads, minimum of one trip
+	 * through unrolled loop.  Do a quad at a time to get us 0mod64
+	 * aligned.
+	 */
+
+	nop			# E :
+	nop			# E :
+	nop			# E :
+	beq	$1, $bigalign_w	# U :
+
+$alignmod64_w:
+	stq	$17, 0($5)	# L :
+	subq	$3, 1, $3	# E : For consistency later
+	addq	$1, 8, $1	# E : Increment towards zero for alignment
+	addq	$5, 8, $4	# E : Initial wh64 address (filler instruction)
+
+	nop
+	nop
+	addq	$5, 8, $5	# E : Inc address
+	blt	$1, $alignmod64_w	# U :
+
+$bigalign_w:
+	/*
+	 * $3 - number quads left to go
+	 * $5 - target address (aligned 0mod64)
+	 * $17 - mask of stuff to store
+	 * Scratch registers available: $7, $2, $4, $1
+	 * we know that we'll be taking a minimum of one trip through
+ 	 * CWG Section 3.7.6: do not expect a sustained store rate of > 1/cycle
+	 * Assumes the wh64 needs to be for 2 trips through the loop in the future
+	 * The wh64 is issued on for the starting destination address for trip +2
+	 * through the loop, and if there are less than two trips left, the target
+	 * address will be for the current trip.
+	 */
+
+$do_wh64_w:
+	wh64	($4)		# L1 : memory subsystem write hint
+	subq	$3, 24, $2	# E : For determining future wh64 addresses
+	stq	$17, 0($5)	# L :
+	nop			# E :
+
+	addq	$5, 128, $4	# E : speculative target of next wh64
+	stq	$17, 8($5)	# L :
+	stq	$17, 16($5)	# L :
+	addq	$5, 64, $7	# E : Fallback address for wh64 (== next trip addr)
+
+	stq	$17, 24($5)	# L :
+	stq	$17, 32($5)	# L :
+	cmovlt	$2, $7, $4	# E : Latency 2, extra mapping cycle
+	nop
+
+	stq	$17, 40($5)	# L :
+	stq	$17, 48($5)	# L :
+	subq	$3, 16, $2	# E : Repeat the loop at least once more?
+	nop
+
+	stq	$17, 56($5)	# L :
+	addq	$5, 64, $5	# E :
+	subq	$3, 8, $3	# E :
+	bge	$2, $do_wh64_w	# U :
+
+	nop
+	nop
+	nop
+	beq	$3, no_quad_w	# U : Might have finished already
+
+.align 4
+	/*
+	 * Simple loop for trailing quadwords, or for small amounts
+	 * of data (where we can't use an unrolled loop and wh64)
+	 */
+loop_w:
+	stq $17,0($5)		# L :
+	subq $3,1,$3		# E : Decrement number quads left
+	addq $5,8,$5		# E : Inc address
+	bne $3,loop_w		# U : more?
+
+no_quad_w:
+	/*
+	 * Write 0..7 trailing bytes.
+	 */
+	nop			# E :
+	beq $18,end_w		# U : All done?
+	ldq $7,0($5)		# L :
+	mskqh $7,$6,$2		# U : Mask final quad
+
+	insqh $17,$6,$4		# U : New bits
+	bis $2,$4,$1		# E : Put it all together
+	stq $1,0($5)		# L : And back to memory
+	ret $31,($26),1		# L0 :
+
+within_quad_w:
+	ldq_u $1,0($16)		# L :
+	insql $17,$16,$2	# U : New bits
+	mskql $1,$16,$4		# U : Clear old
+	bis $2,$4,$2		# E : New result
+
+	mskql $2,$6,$4		# U :
+	mskqh $1,$6,$2		# U :
+	bis $2,$4,$1		# E :
+	stq_u $1,0($16)		# L :
+
+end_w:
+	nop
+	nop
+	nop
+	ret $31,($26),1		# L0 :
+
+	.end __memset16
+	EXPORT_SYMBOL(__memset16)
+
+memset = ___memset
+__memset = ___memset
+	EXPORT_SYMBOL(memset)
+	EXPORT_SYMBOL(__memset)