aboutsummaryrefslogtreecommitdiff
path: root/arch/mips/cavium-octeon/octeon-memcpy.S
diff options
context:
space:
mode:
authorLibravatar Linus Torvalds <torvalds@linux-foundation.org>2023-02-21 18:24:12 -0800
committerLibravatar Linus Torvalds <torvalds@linux-foundation.org>2023-02-21 18:24:12 -0800
commit5b7c4cabbb65f5c469464da6c5f614cbd7f730f2 (patch)
treecc5c2d0a898769fd59549594fedb3ee6f84e59a0 /arch/mips/cavium-octeon/octeon-memcpy.S
downloadlinux-5b7c4cabbb65f5c469464da6c5f614cbd7f730f2.tar.gz
linux-5b7c4cabbb65f5c469464da6c5f614cbd7f730f2.zip
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(). ...
Diffstat (limited to '')
-rw-r--r--arch/mips/cavium-octeon/octeon-memcpy.S484
1 files changed, 484 insertions, 0 deletions
diff --git a/arch/mips/cavium-octeon/octeon-memcpy.S b/arch/mips/cavium-octeon/octeon-memcpy.S
new file mode 100644
index 000000000..25860fba6
--- /dev/null
+++ b/arch/mips/cavium-octeon/octeon-memcpy.S
@@ -0,0 +1,484 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License. See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Unified implementation of memcpy, memmove and the __copy_user backend.
+ *
+ * Copyright (C) 1998, 99, 2000, 01, 2002 Ralf Baechle (ralf@gnu.org)
+ * Copyright (C) 1999, 2000, 01, 2002 Silicon Graphics, Inc.
+ * Copyright (C) 2002 Broadcom, Inc.
+ * memcpy/copy_user author: Mark Vandevoorde
+ *
+ * Mnemonic names for arguments to memcpy/__copy_user
+ */
+
+#include <asm/asm.h>
+#include <asm/asm-offsets.h>
+#include <asm/export.h>
+#include <asm/regdef.h>
+
+#define dst a0
+#define src a1
+#define len a2
+
+/*
+ * Spec
+ *
+ * memcpy copies len bytes from src to dst and sets v0 to dst.
+ * It assumes that
+ * - src and dst don't overlap
+ * - src is readable
+ * - dst is writable
+ * memcpy uses the standard calling convention
+ *
+ * __copy_user copies up to len bytes from src to dst and sets a2 (len) to
+ * the number of uncopied bytes due to an exception caused by a read or write.
+ * __copy_user assumes that src and dst don't overlap, and that the call is
+ * implementing one of the following:
+ * copy_to_user
+ * - src is readable (no exceptions when reading src)
+ * copy_from_user
+ * - dst is writable (no exceptions when writing dst)
+ * __copy_user uses a non-standard calling convention; see
+ * arch/mips/include/asm/uaccess.h
+ *
+ * When an exception happens on a load, the handler must
+ # ensure that all of the destination buffer is overwritten to prevent
+ * leaking information to user mode programs.
+ */
+
+/*
+ * Implementation
+ */
+
+/*
+ * The exception handler for loads requires that:
+ * 1- AT contain the address of the byte just past the end of the source
+ * of the copy,
+ * 2- src_entry <= src < AT, and
+ * 3- (dst - src) == (dst_entry - src_entry),
+ * The _entry suffix denotes values when __copy_user was called.
+ *
+ * (1) is set up up by uaccess.h and maintained by not writing AT in copy_user
+ * (2) is met by incrementing src by the number of bytes copied
+ * (3) is met by not doing loads between a pair of increments of dst and src
+ *
+ * The exception handlers for stores adjust len (if necessary) and return.
+ * These handlers do not need to overwrite any data.
+ *
+ * For __rmemcpy and memmove an exception is always a kernel bug, therefore
+ * they're not protected.
+ */
+
+#define EXC(inst_reg,addr,handler) \
+9: inst_reg, addr; \
+ .section __ex_table,"a"; \
+ PTR_WD 9b, handler; \
+ .previous
+
+/*
+ * Only on the 64-bit kernel we can made use of 64-bit registers.
+ */
+
+#define LOAD ld
+#define LOADL ldl
+#define LOADR ldr
+#define STOREL sdl
+#define STORER sdr
+#define STORE sd
+#define ADD daddu
+#define SUB dsubu
+#define SRL dsrl
+#define SRA dsra
+#define SLL dsll
+#define SLLV dsllv
+#define SRLV dsrlv
+#define NBYTES 8
+#define LOG_NBYTES 3
+
+/*
+ * As we are sharing code base with the mips32 tree (which use the o32 ABI
+ * register definitions). We need to redefine the register definitions from
+ * the n64 ABI register naming to the o32 ABI register naming.
+ */
+#undef t0
+#undef t1
+#undef t2
+#undef t3
+#define t0 $8
+#define t1 $9
+#define t2 $10
+#define t3 $11
+#define t4 $12
+#define t5 $13
+#define t6 $14
+#define t7 $15
+
+#ifdef CONFIG_CPU_LITTLE_ENDIAN
+#define LDFIRST LOADR
+#define LDREST LOADL
+#define STFIRST STORER
+#define STREST STOREL
+#define SHIFT_DISCARD SLLV
+#else
+#define LDFIRST LOADL
+#define LDREST LOADR
+#define STFIRST STOREL
+#define STREST STORER
+#define SHIFT_DISCARD SRLV
+#endif
+
+#define FIRST(unit) ((unit)*NBYTES)
+#define REST(unit) (FIRST(unit)+NBYTES-1)
+#define UNIT(unit) FIRST(unit)
+
+#define ADDRMASK (NBYTES-1)
+
+ .text
+ .set noreorder
+ .set noat
+
+/*
+ * A combined memcpy/__copy_user
+ * __copy_user sets len to 0 for success; else to an upper bound of
+ * the number of uncopied bytes.
+ * memcpy sets v0 to dst.
+ */
+ .align 5
+LEAF(memcpy) /* a0=dst a1=src a2=len */
+EXPORT_SYMBOL(memcpy)
+ move v0, dst /* return value */
+__memcpy:
+FEXPORT(__raw_copy_from_user)
+EXPORT_SYMBOL(__raw_copy_from_user)
+FEXPORT(__raw_copy_to_user)
+EXPORT_SYMBOL(__raw_copy_to_user)
+ /*
+ * Note: dst & src may be unaligned, len may be 0
+ * Temps
+ */
+ #
+ # Octeon doesn't care if the destination is unaligned. The hardware
+ # can fix it faster than we can special case the assembly.
+ #
+ pref 0, 0(src)
+ sltu t0, len, NBYTES # Check if < 1 word
+ bnez t0, copy_bytes_checklen
+ and t0, src, ADDRMASK # Check if src unaligned
+ bnez t0, src_unaligned
+ sltu t0, len, 4*NBYTES # Check if < 4 words
+ bnez t0, less_than_4units
+ sltu t0, len, 8*NBYTES # Check if < 8 words
+ bnez t0, less_than_8units
+ sltu t0, len, 16*NBYTES # Check if < 16 words
+ bnez t0, cleanup_both_aligned
+ sltu t0, len, 128+1 # Check if len < 129
+ bnez t0, 1f # Skip prefetch if len is too short
+ sltu t0, len, 256+1 # Check if len < 257
+ bnez t0, 1f # Skip prefetch if len is too short
+ pref 0, 128(src) # We must not prefetch invalid addresses
+ #
+ # This is where we loop if there is more than 128 bytes left
+2: pref 0, 256(src) # We must not prefetch invalid addresses
+ #
+ # This is where we loop if we can't prefetch anymore
+1:
+EXC( LOAD t0, UNIT(0)(src), l_exc)
+EXC( LOAD t1, UNIT(1)(src), l_exc_copy)
+EXC( LOAD t2, UNIT(2)(src), l_exc_copy)
+EXC( LOAD t3, UNIT(3)(src), l_exc_copy)
+ SUB len, len, 16*NBYTES
+EXC( STORE t0, UNIT(0)(dst), s_exc_p16u)
+EXC( STORE t1, UNIT(1)(dst), s_exc_p15u)
+EXC( STORE t2, UNIT(2)(dst), s_exc_p14u)
+EXC( STORE t3, UNIT(3)(dst), s_exc_p13u)
+EXC( LOAD t0, UNIT(4)(src), l_exc_copy)
+EXC( LOAD t1, UNIT(5)(src), l_exc_copy)
+EXC( LOAD t2, UNIT(6)(src), l_exc_copy)
+EXC( LOAD t3, UNIT(7)(src), l_exc_copy)
+EXC( STORE t0, UNIT(4)(dst), s_exc_p12u)
+EXC( STORE t1, UNIT(5)(dst), s_exc_p11u)
+EXC( STORE t2, UNIT(6)(dst), s_exc_p10u)
+ ADD src, src, 16*NBYTES
+EXC( STORE t3, UNIT(7)(dst), s_exc_p9u)
+ ADD dst, dst, 16*NBYTES
+EXC( LOAD t0, UNIT(-8)(src), l_exc_copy_rewind16)
+EXC( LOAD t1, UNIT(-7)(src), l_exc_copy_rewind16)
+EXC( LOAD t2, UNIT(-6)(src), l_exc_copy_rewind16)
+EXC( LOAD t3, UNIT(-5)(src), l_exc_copy_rewind16)
+EXC( STORE t0, UNIT(-8)(dst), s_exc_p8u)
+EXC( STORE t1, UNIT(-7)(dst), s_exc_p7u)
+EXC( STORE t2, UNIT(-6)(dst), s_exc_p6u)
+EXC( STORE t3, UNIT(-5)(dst), s_exc_p5u)
+EXC( LOAD t0, UNIT(-4)(src), l_exc_copy_rewind16)
+EXC( LOAD t1, UNIT(-3)(src), l_exc_copy_rewind16)
+EXC( LOAD t2, UNIT(-2)(src), l_exc_copy_rewind16)
+EXC( LOAD t3, UNIT(-1)(src), l_exc_copy_rewind16)
+EXC( STORE t0, UNIT(-4)(dst), s_exc_p4u)
+EXC( STORE t1, UNIT(-3)(dst), s_exc_p3u)
+EXC( STORE t2, UNIT(-2)(dst), s_exc_p2u)
+EXC( STORE t3, UNIT(-1)(dst), s_exc_p1u)
+ sltu t0, len, 256+1 # See if we can prefetch more
+ beqz t0, 2b
+ sltu t0, len, 128 # See if we can loop more time
+ beqz t0, 1b
+ nop
+ #
+ # Jump here if there are less than 16*NBYTES left.
+ #
+cleanup_both_aligned:
+ beqz len, done
+ sltu t0, len, 8*NBYTES
+ bnez t0, less_than_8units
+ nop
+EXC( LOAD t0, UNIT(0)(src), l_exc)
+EXC( LOAD t1, UNIT(1)(src), l_exc_copy)
+EXC( LOAD t2, UNIT(2)(src), l_exc_copy)
+EXC( LOAD t3, UNIT(3)(src), l_exc_copy)
+ SUB len, len, 8*NBYTES
+EXC( STORE t0, UNIT(0)(dst), s_exc_p8u)
+EXC( STORE t1, UNIT(1)(dst), s_exc_p7u)
+EXC( STORE t2, UNIT(2)(dst), s_exc_p6u)
+EXC( STORE t3, UNIT(3)(dst), s_exc_p5u)
+EXC( LOAD t0, UNIT(4)(src), l_exc_copy)
+EXC( LOAD t1, UNIT(5)(src), l_exc_copy)
+EXC( LOAD t2, UNIT(6)(src), l_exc_copy)
+EXC( LOAD t3, UNIT(7)(src), l_exc_copy)
+EXC( STORE t0, UNIT(4)(dst), s_exc_p4u)
+EXC( STORE t1, UNIT(5)(dst), s_exc_p3u)
+EXC( STORE t2, UNIT(6)(dst), s_exc_p2u)
+EXC( STORE t3, UNIT(7)(dst), s_exc_p1u)
+ ADD src, src, 8*NBYTES
+ beqz len, done
+ ADD dst, dst, 8*NBYTES
+ #
+ # Jump here if there are less than 8*NBYTES left.
+ #
+less_than_8units:
+ sltu t0, len, 4*NBYTES
+ bnez t0, less_than_4units
+ nop
+EXC( LOAD t0, UNIT(0)(src), l_exc)
+EXC( LOAD t1, UNIT(1)(src), l_exc_copy)
+EXC( LOAD t2, UNIT(2)(src), l_exc_copy)
+EXC( LOAD t3, UNIT(3)(src), l_exc_copy)
+ SUB len, len, 4*NBYTES
+EXC( STORE t0, UNIT(0)(dst), s_exc_p4u)
+EXC( STORE t1, UNIT(1)(dst), s_exc_p3u)
+EXC( STORE t2, UNIT(2)(dst), s_exc_p2u)
+EXC( STORE t3, UNIT(3)(dst), s_exc_p1u)
+ ADD src, src, 4*NBYTES
+ beqz len, done
+ ADD dst, dst, 4*NBYTES
+ #
+ # Jump here if there are less than 4*NBYTES left. This means
+ # we may need to copy up to 3 NBYTES words.
+ #
+less_than_4units:
+ sltu t0, len, 1*NBYTES
+ bnez t0, copy_bytes_checklen
+ nop
+ #
+ # 1) Copy NBYTES, then check length again
+ #
+EXC( LOAD t0, 0(src), l_exc)
+ SUB len, len, NBYTES
+ sltu t1, len, 8
+EXC( STORE t0, 0(dst), s_exc_p1u)
+ ADD src, src, NBYTES
+ bnez t1, copy_bytes_checklen
+ ADD dst, dst, NBYTES
+ #
+ # 2) Copy NBYTES, then check length again
+ #
+EXC( LOAD t0, 0(src), l_exc)
+ SUB len, len, NBYTES
+ sltu t1, len, 8
+EXC( STORE t0, 0(dst), s_exc_p1u)
+ ADD src, src, NBYTES
+ bnez t1, copy_bytes_checklen
+ ADD dst, dst, NBYTES
+ #
+ # 3) Copy NBYTES, then check length again
+ #
+EXC( LOAD t0, 0(src), l_exc)
+ SUB len, len, NBYTES
+ ADD src, src, NBYTES
+ ADD dst, dst, NBYTES
+ b copy_bytes_checklen
+EXC( STORE t0, -8(dst), s_exc_p1u)
+
+src_unaligned:
+#define rem t8
+ SRL t0, len, LOG_NBYTES+2 # +2 for 4 units/iter
+ beqz t0, cleanup_src_unaligned
+ and rem, len, (4*NBYTES-1) # rem = len % 4*NBYTES
+1:
+/*
+ * Avoid consecutive LD*'s to the same register since some mips
+ * implementations can't issue them in the same cycle.
+ * It's OK to load FIRST(N+1) before REST(N) because the two addresses
+ * are to the same unit (unless src is aligned, but it's not).
+ */
+EXC( LDFIRST t0, FIRST(0)(src), l_exc)
+EXC( LDFIRST t1, FIRST(1)(src), l_exc_copy)
+ SUB len, len, 4*NBYTES
+EXC( LDREST t0, REST(0)(src), l_exc_copy)
+EXC( LDREST t1, REST(1)(src), l_exc_copy)
+EXC( LDFIRST t2, FIRST(2)(src), l_exc_copy)
+EXC( LDFIRST t3, FIRST(3)(src), l_exc_copy)
+EXC( LDREST t2, REST(2)(src), l_exc_copy)
+EXC( LDREST t3, REST(3)(src), l_exc_copy)
+ ADD src, src, 4*NBYTES
+EXC( STORE t0, UNIT(0)(dst), s_exc_p4u)
+EXC( STORE t1, UNIT(1)(dst), s_exc_p3u)
+EXC( STORE t2, UNIT(2)(dst), s_exc_p2u)
+EXC( STORE t3, UNIT(3)(dst), s_exc_p1u)
+ bne len, rem, 1b
+ ADD dst, dst, 4*NBYTES
+
+cleanup_src_unaligned:
+ beqz len, done
+ and rem, len, NBYTES-1 # rem = len % NBYTES
+ beq rem, len, copy_bytes
+ nop
+1:
+EXC( LDFIRST t0, FIRST(0)(src), l_exc)
+EXC( LDREST t0, REST(0)(src), l_exc_copy)
+ SUB len, len, NBYTES
+EXC( STORE t0, 0(dst), s_exc_p1u)
+ ADD src, src, NBYTES
+ bne len, rem, 1b
+ ADD dst, dst, NBYTES
+
+copy_bytes_checklen:
+ beqz len, done
+ nop
+copy_bytes:
+ /* 0 < len < NBYTES */
+#define COPY_BYTE(N) \
+EXC( lb t0, N(src), l_exc); \
+ SUB len, len, 1; \
+ beqz len, done; \
+EXC( sb t0, N(dst), s_exc_p1)
+
+ COPY_BYTE(0)
+ COPY_BYTE(1)
+ COPY_BYTE(2)
+ COPY_BYTE(3)
+ COPY_BYTE(4)
+ COPY_BYTE(5)
+EXC( lb t0, NBYTES-2(src), l_exc)
+ SUB len, len, 1
+ jr ra
+EXC( sb t0, NBYTES-2(dst), s_exc_p1)
+done:
+ jr ra
+ nop
+ END(memcpy)
+
+l_exc_copy_rewind16:
+ /* Rewind src and dst by 16*NBYTES for l_exc_copy */
+ SUB src, src, 16*NBYTES
+ SUB dst, dst, 16*NBYTES
+l_exc_copy:
+ /*
+ * Copy bytes from src until faulting load address (or until a
+ * lb faults)
+ *
+ * When reached by a faulting LDFIRST/LDREST, THREAD_BUADDR($28)
+ * may be more than a byte beyond the last address.
+ * Hence, the lb below may get an exception.
+ *
+ * Assumes src < THREAD_BUADDR($28)
+ */
+ LOAD t0, TI_TASK($28)
+ LOAD t0, THREAD_BUADDR(t0)
+1:
+EXC( lb t1, 0(src), l_exc)
+ ADD src, src, 1
+ sb t1, 0(dst) # can't fault -- we're copy_from_user
+ bne src, t0, 1b
+ ADD dst, dst, 1
+l_exc:
+ LOAD t0, TI_TASK($28)
+ LOAD t0, THREAD_BUADDR(t0) # t0 is just past last good address
+ SUB len, AT, t0 # len number of uncopied bytes
+ jr ra
+ nop
+
+
+#define SEXC(n) \
+s_exc_p ## n ## u: \
+ jr ra; \
+ ADD len, len, n*NBYTES
+
+SEXC(16)
+SEXC(15)
+SEXC(14)
+SEXC(13)
+SEXC(12)
+SEXC(11)
+SEXC(10)
+SEXC(9)
+SEXC(8)
+SEXC(7)
+SEXC(6)
+SEXC(5)
+SEXC(4)
+SEXC(3)
+SEXC(2)
+SEXC(1)
+
+s_exc_p1:
+ jr ra
+ ADD len, len, 1
+s_exc:
+ jr ra
+ nop
+
+ .align 5
+LEAF(memmove)
+EXPORT_SYMBOL(memmove)
+ ADD t0, a0, a2
+ ADD t1, a1, a2
+ sltu t0, a1, t0 # dst + len <= src -> memcpy
+ sltu t1, a0, t1 # dst >= src + len -> memcpy
+ and t0, t1
+ beqz t0, __memcpy
+ move v0, a0 /* return value */
+ beqz a2, r_out
+ END(memmove)
+
+ /* fall through to __rmemcpy */
+LEAF(__rmemcpy) /* a0=dst a1=src a2=len */
+ sltu t0, a1, a0
+ beqz t0, r_end_bytes_up # src >= dst
+ nop
+ ADD a0, a2 # dst = dst + len
+ ADD a1, a2 # src = src + len
+
+r_end_bytes:
+ lb t0, -1(a1)
+ SUB a2, a2, 0x1
+ sb t0, -1(a0)
+ SUB a1, a1, 0x1
+ bnez a2, r_end_bytes
+ SUB a0, a0, 0x1
+
+r_out:
+ jr ra
+ move a2, zero
+
+r_end_bytes_up:
+ lb t0, (a1)
+ SUB a2, a2, 0x1
+ sb t0, (a0)
+ ADD a1, a1, 0x1
+ bnez a2, r_end_bytes_up
+ ADD a0, a0, 0x1
+
+ jr ra
+ move a2, zero
+ END(__rmemcpy)