sparc64: Add SPARC-T4 optimized memcpy.
Before After
-------------- --------------
bw_tcp: 1288.53 MB/sec 1637.77 MB/sec
bw_pipe: 1517.18 MB/sec 2107.61 MB/sec
bw_unix: 1838.38 MB/sec 2640.91 MB/sec
make -s -j128
allmodconfig 5min 49sec 5min 31sec
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/arch/sparc/lib/NG4memcpy.S b/arch/sparc/lib/NG4memcpy.S
new file mode 100644
index 0000000..6f4b526
--- /dev/null
+++ b/arch/sparc/lib/NG4memcpy.S
@@ -0,0 +1,360 @@
+/* NG4memcpy.S: Niagara-4 optimized memcpy.
+ *
+ * Copyright (C) 2012 David S. Miller (davem@davemloft.net)
+ */
+
+#ifdef __KERNEL__
+#include <asm/visasm.h>
+#include <asm/asi.h>
+#define GLOBAL_SPARE %g7
+#else
+#define ASI_BLK_INIT_QUAD_LDD_P 0xe2
+#define FPRS_FEF 0x04
+
+/* On T4 it is very expensive to access ASRs like %fprs and
+ * %asi, avoiding a read or a write can save ~50 cycles.
+ */
+#define FPU_ENTER \
+ rd %fprs, %o5; \
+ andcc %o5, FPRS_FEF, %g0; \
+ be,a,pn %icc, 999f; \
+ wr %g0, FPRS_FEF, %fprs; \
+ 999:
+
+#ifdef MEMCPY_DEBUG
+#define VISEntryHalf FPU_ENTER; \
+ clr %g1; clr %g2; clr %g3; clr %g5; subcc %g0, %g0, %g0;
+#define VISExitHalf and %o5, FPRS_FEF, %o5; wr %o5, 0x0, %fprs
+#else
+#define VISEntryHalf FPU_ENTER
+#define VISExitHalf and %o5, FPRS_FEF, %o5; wr %o5, 0x0, %fprs
+#endif
+
+#define GLOBAL_SPARE %g5
+#endif
+
+#ifndef STORE_ASI
+#ifndef SIMULATE_NIAGARA_ON_NON_NIAGARA
+#define STORE_ASI ASI_BLK_INIT_QUAD_LDD_P
+#else
+#define STORE_ASI 0x80 /* ASI_P */
+#endif
+#endif
+
+#ifndef EX_LD
+#define EX_LD(x) x
+#endif
+
+#ifndef EX_ST
+#define EX_ST(x) x
+#endif
+
+#ifndef EX_RETVAL
+#define EX_RETVAL(x) x
+#endif
+
+#ifndef LOAD
+#define LOAD(type,addr,dest) type [addr], dest
+#endif
+
+#ifndef STORE
+#ifndef MEMCPY_DEBUG
+#define STORE(type,src,addr) type src, [addr]
+#else
+#define STORE(type,src,addr) type##a src, [addr] %asi
+#endif
+#endif
+
+#ifndef STORE_INIT
+#define STORE_INIT(src,addr) stxa src, [addr] STORE_ASI
+#endif
+
+#ifndef FUNC_NAME
+#define FUNC_NAME NG4memcpy
+#endif
+#ifndef PREAMBLE
+#define PREAMBLE
+#endif
+
+#ifndef XCC
+#define XCC xcc
+#endif
+
+ .register %g2,#scratch
+ .register %g3,#scratch
+
+ .text
+ .align 64
+
+ .globl FUNC_NAME
+ .type FUNC_NAME,#function
+FUNC_NAME: /* %o0=dst, %o1=src, %o2=len */
+#ifdef MEMCPY_DEBUG
+ wr %g0, 0x80, %asi
+#endif
+ srlx %o2, 31, %g2
+ cmp %g2, 0
+ tne %XCC, 5
+ PREAMBLE
+ mov %o0, %o3
+ brz,pn %o2, .Lexit
+ cmp %o2, 3
+ ble,pn %icc, .Ltiny
+ cmp %o2, 19
+ ble,pn %icc, .Lsmall
+ or %o0, %o1, %g2
+ cmp %o2, 128
+ bl,pn %icc, .Lmedium
+ nop
+
+.Llarge:/* len >= 0x80 */
+ /* First get dest 8 byte aligned. */
+ sub %g0, %o0, %g1
+ and %g1, 0x7, %g1
+ brz,pt %g1, 51f
+ sub %o2, %g1, %o2
+
+1: EX_LD(LOAD(ldub, %o1 + 0x00, %g2))
+ add %o1, 1, %o1
+ subcc %g1, 1, %g1
+ add %o0, 1, %o0
+ bne,pt %icc, 1b
+ EX_ST(STORE(stb, %g2, %o0 - 0x01))
+
+51: LOAD(prefetch, %o1 + 0x040, #n_reads_strong)
+ LOAD(prefetch, %o1 + 0x080, #n_reads_strong)
+ LOAD(prefetch, %o1 + 0x0c0, #n_reads_strong)
+ LOAD(prefetch, %o1 + 0x100, #n_reads_strong)
+ LOAD(prefetch, %o1 + 0x140, #n_reads_strong)
+ LOAD(prefetch, %o1 + 0x180, #n_reads_strong)
+ LOAD(prefetch, %o1 + 0x1c0, #n_reads_strong)
+ LOAD(prefetch, %o1 + 0x200, #n_reads_strong)
+
+ /* Check if we can use the straight fully aligned
+ * loop, or we require the alignaddr/faligndata variant.
+ */
+ andcc %o1, 0x7, %o5
+ bne,pn %icc, .Llarge_src_unaligned
+ sub %g0, %o0, %g1
+
+ /* Legitimize the use of initializing stores by getting dest
+ * to be 64-byte aligned.
+ */
+ and %g1, 0x3f, %g1
+ brz,pt %g1, .Llarge_aligned
+ sub %o2, %g1, %o2
+
+1: EX_LD(LOAD(ldx, %o1 + 0x00, %g2))
+ add %o1, 8, %o1
+ subcc %g1, 8, %g1
+ add %o0, 8, %o0
+ bne,pt %icc, 1b
+ EX_ST(STORE(stx, %g2, %o0 - 0x08))
+
+.Llarge_aligned:
+ /* len >= 0x80 && src 8-byte aligned && dest 8-byte aligned */
+ andn %o2, 0x3f, %o4
+ sub %o2, %o4, %o2
+
+1: EX_LD(LOAD(ldx, %o1 + 0x00, %g1))
+ add %o1, 0x40, %o1
+ EX_LD(LOAD(ldx, %o1 - 0x38, %g2))
+ subcc %o4, 0x40, %o4
+ EX_LD(LOAD(ldx, %o1 - 0x30, %g3))
+ EX_LD(LOAD(ldx, %o1 - 0x28, GLOBAL_SPARE))
+ EX_LD(LOAD(ldx, %o1 - 0x20, %o5))
+ EX_ST(STORE_INIT(%g1, %o0))
+ add %o0, 0x08, %o0
+ EX_ST(STORE_INIT(%g2, %o0))
+ add %o0, 0x08, %o0
+ EX_LD(LOAD(ldx, %o1 - 0x18, %g2))
+ EX_ST(STORE_INIT(%g3, %o0))
+ add %o0, 0x08, %o0
+ EX_LD(LOAD(ldx, %o1 - 0x10, %g3))
+ EX_ST(STORE_INIT(GLOBAL_SPARE, %o0))
+ add %o0, 0x08, %o0
+ EX_LD(LOAD(ldx, %o1 - 0x08, GLOBAL_SPARE))
+ EX_ST(STORE_INIT(%o5, %o0))
+ add %o0, 0x08, %o0
+ EX_ST(STORE_INIT(%g2, %o0))
+ add %o0, 0x08, %o0
+ EX_ST(STORE_INIT(%g3, %o0))
+ add %o0, 0x08, %o0
+ EX_ST(STORE_INIT(GLOBAL_SPARE, %o0))
+ add %o0, 0x08, %o0
+ bne,pt %icc, 1b
+ LOAD(prefetch, %o1 + 0x200, #n_reads_strong)
+
+ membar #StoreLoad | #StoreStore
+
+ brz,pn %o2, .Lexit
+ cmp %o2, 19
+ ble,pn %icc, .Lsmall_unaligned
+ nop
+ ba,a,pt %icc, .Lmedium_noprefetch
+
+.Lexit: retl
+ mov EX_RETVAL(%o3), %o0
+
+.Llarge_src_unaligned:
+ andn %o2, 0x3f, %o4
+ sub %o2, %o4, %o2
+ VISEntryHalf
+ alignaddr %o1, %g0, %g1
+ add %o1, %o4, %o1
+ EX_LD(LOAD(ldd, %g1 + 0x00, %f0))
+1: EX_LD(LOAD(ldd, %g1 + 0x08, %f2))
+ subcc %o4, 0x40, %o4
+ EX_LD(LOAD(ldd, %g1 + 0x10, %f4))
+ EX_LD(LOAD(ldd, %g1 + 0x18, %f6))
+ EX_LD(LOAD(ldd, %g1 + 0x20, %f8))
+ EX_LD(LOAD(ldd, %g1 + 0x28, %f10))
+ EX_LD(LOAD(ldd, %g1 + 0x30, %f12))
+ EX_LD(LOAD(ldd, %g1 + 0x38, %f14))
+ faligndata %f0, %f2, %f16
+ EX_LD(LOAD(ldd, %g1 + 0x40, %f0))
+ faligndata %f2, %f4, %f18
+ add %g1, 0x40, %g1
+ faligndata %f4, %f6, %f20
+ faligndata %f6, %f8, %f22
+ faligndata %f8, %f10, %f24
+ faligndata %f10, %f12, %f26
+ faligndata %f12, %f14, %f28
+ faligndata %f14, %f0, %f30
+ EX_ST(STORE(std, %f16, %o0 + 0x00))
+ EX_ST(STORE(std, %f18, %o0 + 0x08))
+ EX_ST(STORE(std, %f20, %o0 + 0x10))
+ EX_ST(STORE(std, %f22, %o0 + 0x18))
+ EX_ST(STORE(std, %f24, %o0 + 0x20))
+ EX_ST(STORE(std, %f26, %o0 + 0x28))
+ EX_ST(STORE(std, %f28, %o0 + 0x30))
+ EX_ST(STORE(std, %f30, %o0 + 0x38))
+ add %o0, 0x40, %o0
+ bne,pt %icc, 1b
+ LOAD(prefetch, %g1 + 0x200, #n_reads_strong)
+ VISExitHalf
+
+ brz,pn %o2, .Lexit
+ cmp %o2, 19
+ ble,pn %icc, .Lsmall_unaligned
+ nop
+ ba,a,pt %icc, .Lmedium_unaligned
+
+.Lmedium:
+ LOAD(prefetch, %o1 + 0x40, #n_reads_strong)
+ andcc %g2, 0x7, %g0
+ bne,pn %icc, .Lmedium_unaligned
+ nop
+.Lmedium_noprefetch:
+ andncc %o2, 0x20 - 1, %o5
+ be,pn %icc, 2f
+ sub %o2, %o5, %o2
+1: EX_LD(LOAD(ldx, %o1 + 0x00, %g1))
+ EX_LD(LOAD(ldx, %o1 + 0x08, %g2))
+ EX_LD(LOAD(ldx, %o1 + 0x10, GLOBAL_SPARE))
+ EX_LD(LOAD(ldx, %o1 + 0x18, %o4))
+ add %o1, 0x20, %o1
+ subcc %o5, 0x20, %o5
+ EX_ST(STORE(stx, %g1, %o0 + 0x00))
+ EX_ST(STORE(stx, %g2, %o0 + 0x08))
+ EX_ST(STORE(stx, GLOBAL_SPARE, %o0 + 0x10))
+ EX_ST(STORE(stx, %o4, %o0 + 0x18))
+ bne,pt %icc, 1b
+ add %o0, 0x20, %o0
+2: andcc %o2, 0x18, %o5
+ be,pt %icc, 3f
+ sub %o2, %o5, %o2
+1: EX_LD(LOAD(ldx, %o1 + 0x00, %g1))
+ add %o1, 0x08, %o1
+ add %o0, 0x08, %o0
+ subcc %o5, 0x08, %o5
+ bne,pt %icc, 1b
+ EX_ST(STORE(stx, %g1, %o0 - 0x08))
+3: brz,pt %o2, .Lexit
+ cmp %o2, 0x04
+ bl,pn %icc, .Ltiny
+ nop
+ EX_LD(LOAD(lduw, %o1 + 0x00, %g1))
+ add %o1, 0x04, %o1
+ add %o0, 0x04, %o0
+ subcc %o2, 0x04, %o2
+ bne,pn %icc, .Ltiny
+ EX_ST(STORE(stw, %g1, %o0 - 0x04))
+ ba,a,pt %icc, .Lexit
+.Lmedium_unaligned:
+ /* First get dest 8 byte aligned. */
+ sub %g0, %o0, %g1
+ and %g1, 0x7, %g1
+ brz,pt %g1, 2f
+ sub %o2, %g1, %o2
+
+1: EX_LD(LOAD(ldub, %o1 + 0x00, %g2))
+ add %o1, 1, %o1
+ subcc %g1, 1, %g1
+ add %o0, 1, %o0
+ bne,pt %icc, 1b
+ EX_ST(STORE(stb, %g2, %o0 - 0x01))
+2:
+ and %o1, 0x7, %g1
+ brz,pn %g1, .Lmedium_noprefetch
+ sll %g1, 3, %g1
+ mov 64, %g2
+ sub %g2, %g1, %g2
+ andn %o1, 0x7, %o1
+ EX_LD(LOAD(ldx, %o1 + 0x00, %o4))
+ sllx %o4, %g1, %o4
+ andn %o2, 0x08 - 1, %o5
+ sub %o2, %o5, %o2
+1: EX_LD(LOAD(ldx, %o1 + 0x08, %g3))
+ add %o1, 0x08, %o1
+ subcc %o5, 0x08, %o5
+ srlx %g3, %g2, GLOBAL_SPARE
+ or GLOBAL_SPARE, %o4, GLOBAL_SPARE
+ EX_ST(STORE(stx, GLOBAL_SPARE, %o0 + 0x00))
+ add %o0, 0x08, %o0
+ bne,pt %icc, 1b
+ sllx %g3, %g1, %o4
+ srl %g1, 3, %g1
+ add %o1, %g1, %o1
+ brz,pn %o2, .Lexit
+ nop
+ ba,pt %icc, .Lsmall_unaligned
+
+.Ltiny:
+ EX_LD(LOAD(ldub, %o1 + 0x00, %g1))
+ subcc %o2, 1, %o2
+ be,pn %icc, .Lexit
+ EX_ST(STORE(stb, %g1, %o0 + 0x00))
+ EX_LD(LOAD(ldub, %o1 + 0x01, %g1))
+ subcc %o2, 1, %o2
+ be,pn %icc, .Lexit
+ EX_ST(STORE(stb, %g1, %o0 + 0x01))
+ EX_LD(LOAD(ldub, %o1 + 0x02, %g1))
+ ba,pt %icc, .Lexit
+ EX_ST(STORE(stb, %g1, %o0 + 0x02))
+
+.Lsmall:
+ andcc %g2, 0x3, %g0
+ bne,pn %icc, .Lsmall_unaligned
+ andn %o2, 0x4 - 1, %o5
+ sub %o2, %o5, %o2
+1:
+ EX_LD(LOAD(lduw, %o1 + 0x00, %g1))
+ add %o1, 0x04, %o1
+ subcc %o5, 0x04, %o5
+ add %o0, 0x04, %o0
+ bne,pt %icc, 1b
+ EX_ST(STORE(stw, %g1, %o0 - 0x04))
+ brz,pt %o2, .Lexit
+ nop
+ ba,a,pt %icc, .Ltiny
+
+.Lsmall_unaligned:
+1: EX_LD(LOAD(ldub, %o1 + 0x00, %g1))
+ add %o1, 1, %o1
+ add %o0, 1, %o0
+ subcc %o2, 1, %o2
+ bne,pt %icc, 1b
+ EX_ST(STORE(stb, %g1, %o0 - 0x01))
+ ba,a,pt %icc, .Lexit
+ .size FUNC_NAME, .-FUNC_NAME