- Kernel - mailweb.openeuler.org

[PATCH OLK-5.10 1/5] LoongArch: Optimize memory ops (memset/memcpy/memmove)
by liweihao 19 Jan '24

19 Jan '24

From: WANG Rui <wangrui(a)loongson.cn> LoongArch inclusion bugzilla: https://gitee.com/openeuler/kernel/issues/I76XQZ To optimize memset()/memcpy()/memmove() and so on, we use a jump table to dispatch cases for short data lengths; and for long data lengths, we split the destination into head part (first 8 bytes), tail part (last 8 bytes) and middle part. The head part and tail part may be at unaligned addresses, while the middle part is always aligned (the middle part is allowed to overlap the head/tail part). In this way, the first and last 8 bytes may be unaligned accesses, but we can make sure the data in the middle is processed at an aligned destination address. We have tested micro-bench[1] on a Loongson-3C5000 16-core machine (2.2GHz): 1. memset | length | src offset | dst offset | speed before | speed after | % | |--------|------------|------------|--------------|-------------|---------| | 8 | 0 | 0 | 696.191 | 1518.785 | 118.16% | | 8 | 0 | 1 | 696.325 | 1518.937 | 118.14% | | 50 | 0 | 0 | 969.976 | 8053.902 | 730.32% | | 50 | 0 | 1 | 970.034 | 8058.475 | 730.74% | | 300 | 0 | 0 | 5876.612 | 16544.703 | 181.53% | | 300 | 0 | 1 | 5030.849 | 16549.011 | 228.95% | | 1200 | 0 | 0 | 11797.077 | 16752.137 | 42.00% | | 1200 | 0 | 1 | 5687.141 | 16645.233 | 192.68% | | 4000 | 0 | 0 | 15723.27 | 16761.557 | 6.60% | | 4000 | 0 | 1 | 5906.114 | 16732.316 | 183.30% | | 8000 | 0 | 0 | 16751.403 | 16770.002 | 0.11% | | 8000 | 0 | 1 | 5995.449 | 16754.07 | 179.45% | 2. memcpy | length | src offset | dst offset | speed before | speed after | % | |--------|------------|------------|--------------|-------------|---------| | 8 | 0 | 0 | 696.2 | 1670.605 | 139.96% | | 8 | 0 | 1 | 696.325 | 1671.138 | 139.99% | | 50 | 0 | 0 | 969.974 | 8724.999 | 799.51% | | 50 | 0 | 1 | 970.032 | 8730.138 | 799.98% | | 300 | 0 | 0 | 5564.662 | 16272.652 | 192.43% | | 300 | 0 | 1 | 4670.436 | 14972.842 | 220.59% | | 1200 | 0 | 0 | 10740.23 | 16751.728 | 55.97% | | 1200 | 0 | 1 | 5027.741 | 14874.564 | 195.85% | | 4000 | 0 | 0 | 15122.367 | 16737.642 | 10.68% | | 4000 | 0 | 1 | 5536.918 | 14890.397 | 168.93% | | 8000 | 0 | 0 | 16505.453 | 16553.543 | 0.29% | | 8000 | 0 | 1 | 5821.619 | 14841.804 | 154.94% | 3. memmove | length | src offset | dst offset | speed before | speed after | % | |--------|------------|------------|--------------|-------------|---------| | 8 | 0 | 0 | 982.693 | 1670.568 | 70.00% | | 8 | 0 | 1 | 983.023 | 1671.174 | 70.00% | | 50 | 0 | 0 | 1230.87 | 8727.625 | 609.06% | | 50 | 0 | 1 | 1232.515 | 8730.138 | 608.32% | | 300 | 0 | 0 | 6490.375 | 16296.993 | 151.09% | | 300 | 0 | 1 | 4282.687 | 14972.842 | 249.61% | | 1200 | 0 | 0 | 11742.755 | 16752.546 | 42.66% | | 1200 | 0 | 1 | 5039.338 | 14872.951 | 195.14% | | 4000 | 0 | 0 | 15467.786 | 16737.09 | 8.21% | | 4000 | 0 | 1 | 5009.905 | 14890.542 | 197.22% | | 8000 | 0 | 0 | 16489.664 | 16553.273 | 0.39% | | 8000 | 0 | 1 | 5823.786 | 14858.646 | 155.14% | * speed: MB/s * length: byte [1] https://github.com/heiher/mem-bench Signed-off-by: WANG Rui <wangrui(a)loongson.cn> Signed-off-by: Huacai Chen <chenhuacai(a)loongson.cn> --- arch/loongarch/lib/clear_user.S | 136 +++++++++++++++-- arch/loongarch/lib/copy_user.S | 251 +++++++++++++++++++++++++------- arch/loongarch/lib/memcpy.S | 146 +++++++++++++++---- arch/loongarch/lib/memmove.S | 120 ++++++++------- arch/loongarch/lib/memset.S | 116 ++++++++++++--- 5 files changed, 602 insertions(+), 167 deletions(-) diff --git a/arch/loongarch/lib/clear_user.S b/arch/loongarch/lib/clear_user.S index 2dc48e61a2c8..fd1d62b244f2 100644 --- a/arch/loongarch/lib/clear_user.S +++ b/arch/loongarch/lib/clear_user.S @@ -13,7 +13,14 @@ .irp to, 0, 1, 2, 3, 4, 5, 6, 7 .L_fixup_handle_\to\(): - addi.d a0, a1, (\to) * (-8) + sub.d a0, a2, a0 + addi.d a0, a0, (\to) * (-8) + jr ra +.endr + +.irp to, 0, 2, 4 +.L_fixup_handle_s\to\(): + addi.d a0, a1, -\to jr ra .endr @@ -44,7 +51,7 @@ SYM_FUNC_START(__clear_user_generic) 2: move a0, a1 jr ra - _asm_extable 1b, .L_fixup_handle_0 + _asm_extable 1b, .L_fixup_handle_s0 SYM_FUNC_END(__clear_user_generic) /* @@ -54,12 +61,21 @@ SYM_FUNC_END(__clear_user_generic) * a1: size */ SYM_FUNC_START(__clear_user_fast) - beqz a1, 10f + sltui t0, a1, 9 + bnez t0, .Lsmall - ori a2, zero, 64 - blt a1, a2, 9f + add.d a2, a0, a1 +0: st.d zero, a0, 0 + + /* align up address */ + addi.d a0, a0, 8 + bstrins.d a0, zero, 2, 0 + + addi.d a3, a2, -64 + bgeu a0, a3, .Llt64 /* set 64 bytes at a time */ +.Lloop64: 1: st.d zero, a0, 0 2: st.d zero, a0, 8 3: st.d zero, a0, 16 @@ -68,24 +84,95 @@ SYM_FUNC_START(__clear_user_fast) 6: st.d zero, a0, 40 7: st.d zero, a0, 48 8: st.d zero, a0, 56 - addi.d a0, a0, 64 - addi.d a1, a1, -64 - bge a1, a2, 1b - - beqz a1, 10f + bltu a0, a3, .Lloop64 /* set the remaining bytes */ -9: st.b zero, a0, 0 - addi.d a0, a0, 1 - addi.d a1, a1, -1 - bgt a1, zero, 9b +.Llt64: + addi.d a3, a2, -32 + bgeu a0, a3, .Llt32 +9: st.d zero, a0, 0 +10: st.d zero, a0, 8 +11: st.d zero, a0, 16 +12: st.d zero, a0, 24 + addi.d a0, a0, 32 + +.Llt32: + addi.d a3, a2, -16 + bgeu a0, a3, .Llt16 +13: st.d zero, a0, 0 +14: st.d zero, a0, 8 + addi.d a0, a0, 16 + +.Llt16: + addi.d a3, a2, -8 + bgeu a0, a3, .Llt8 +15: st.d zero, a0, 0 + +.Llt8: +16: st.d zero, a2, -8 /* return */ -10: move a0, a1 + move a0, zero + jr ra + + .align 4 +.Lsmall: + pcaddi t0, 4 + slli.d a2, a1, 4 + add.d t0, t0, a2 + jr t0 + + .align 4 + move a0, zero + jr ra + + .align 4 +17: st.b zero, a0, 0 + move a0, zero + jr ra + + .align 4 +18: st.h zero, a0, 0 + move a0, zero + jr ra + + .align 4 +19: st.h zero, a0, 0 +20: st.b zero, a0, 2 + move a0, zero + jr ra + + .align 4 +21: st.w zero, a0, 0 + move a0, zero + jr ra + + .align 4 +22: st.w zero, a0, 0 +23: st.b zero, a0, 4 + move a0, zero + jr ra + + .align 4 +24: st.w zero, a0, 0 +25: st.h zero, a0, 4 + move a0, zero + jr ra + + .align 4 +26: st.w zero, a0, 0 +27: st.w zero, a0, 3 + move a0, zero + jr ra + + .align 4 +28: st.d zero, a0, 0 + move a0, zero jr ra /* fixup and ex_table */ + _asm_extable 0b, .L_fixup_handle_0 _asm_extable 1b, .L_fixup_handle_0 _asm_extable 2b, .L_fixup_handle_1 _asm_extable 3b, .L_fixup_handle_2 @@ -95,4 +182,23 @@ SYM_FUNC_START(__clear_user_fast) _asm_extable 7b, .L_fixup_handle_6 _asm_extable 8b, .L_fixup_handle_7 _asm_extable 9b, .L_fixup_handle_0 + _asm_extable 10b, .L_fixup_handle_1 + _asm_extable 11b, .L_fixup_handle_2 + _asm_extable 12b, .L_fixup_handle_3 + _asm_extable 13b, .L_fixup_handle_0 + _asm_extable 14b, .L_fixup_handle_1 + _asm_extable 15b, .L_fixup_handle_0 + _asm_extable 16b, .L_fixup_handle_1 + _asm_extable 17b, .L_fixup_handle_s0 + _asm_extable 18b, .L_fixup_handle_s0 + _asm_extable 19b, .L_fixup_handle_s0 + _asm_extable 20b, .L_fixup_handle_s2 + _asm_extable 21b, .L_fixup_handle_s0 + _asm_extable 22b, .L_fixup_handle_s0 + _asm_extable 23b, .L_fixup_handle_s4 + _asm_extable 24b, .L_fixup_handle_s0 + _asm_extable 25b, .L_fixup_handle_s4 + _asm_extable 26b, .L_fixup_handle_s0 + _asm_extable 27b, .L_fixup_handle_s4 + _asm_extable 28b, .L_fixup_handle_s0 SYM_FUNC_END(__clear_user_fast) diff --git a/arch/loongarch/lib/copy_user.S b/arch/loongarch/lib/copy_user.S index 55ac6020a1ad..b21f6d5d38f5 100644 --- a/arch/loongarch/lib/copy_user.S +++ b/arch/loongarch/lib/copy_user.S @@ -13,7 +13,14 @@ .irp to, 0, 1, 2, 3, 4, 5, 6, 7 .L_fixup_handle_\to\(): - addi.d a0, a2, (\to) * (-8) + sub.d a0, a2, a0 + addi.d a0, a0, (\to) * (-8) + jr ra +.endr + +.irp to, 0, 2, 4 +.L_fixup_handle_s\to\(): + addi.d a0, a2, -\to jr ra .endr @@ -47,8 +54,8 @@ SYM_FUNC_START(__copy_user_generic) 3: move a0, a2 jr ra - _asm_extable 1b, .L_fixup_handle_0 - _asm_extable 2b, .L_fixup_handle_0 + _asm_extable 1b, .L_fixup_handle_s0 + _asm_extable 2b, .L_fixup_handle_s0 SYM_FUNC_END(__copy_user_generic) /* @@ -59,65 +66,209 @@ SYM_FUNC_END(__copy_user_generic) * a2: n */ SYM_FUNC_START(__copy_user_fast) - beqz a2, 19f + sltui t0, a2, 9 + bnez t0, .Lsmall - ori a3, zero, 64 - blt a2, a3, 17f + add.d a3, a1, a2 + add.d a2, a0, a2 +0: ld.d t0, a1, 0 +1: st.d t0, a0, 0 - /* copy 64 bytes at a time */ -1: ld.d t0, a1, 0 -2: ld.d t1, a1, 8 -3: ld.d t2, a1, 16 -4: ld.d t3, a1, 24 -5: ld.d t4, a1, 32 -6: ld.d t5, a1, 40 -7: ld.d t6, a1, 48 -8: ld.d t7, a1, 56 -9: st.d t0, a0, 0 -10: st.d t1, a0, 8 -11: st.d t2, a0, 16 -12: st.d t3, a0, 24 -13: st.d t4, a0, 32 -14: st.d t5, a0, 40 -15: st.d t6, a0, 48 -16: st.d t7, a0, 56 + /* align up destination address */ + andi t1, a0, 7 + sub.d t0, zero, t1 + addi.d t0, t0, 8 + add.d a1, a1, t0 + add.d a0, a0, t0 - addi.d a0, a0, 64 - addi.d a1, a1, 64 - addi.d a2, a2, -64 - bge a2, a3, 1b + addi.d a4, a3, -64 + bgeu a1, a4, .Llt64 - beqz a2, 19f + /* copy 64 bytes at a time */ +.Lloop64: +2: ld.d t0, a1, 0 +3: ld.d t1, a1, 8 +4: ld.d t2, a1, 16 +5: ld.d t3, a1, 24 +6: ld.d t4, a1, 32 +7: ld.d t5, a1, 40 +8: ld.d t6, a1, 48 +9: ld.d t7, a1, 56 + addi.d a1, a1, 64 +10: st.d t0, a0, 0 +11: st.d t1, a0, 8 +12: st.d t2, a0, 16 +13: st.d t3, a0, 24 +14: st.d t4, a0, 32 +15: st.d t5, a0, 40 +16: st.d t6, a0, 48 +17: st.d t7, a0, 56 + addi.d a0, a0, 64 + bltu a1, a4, .Lloop64 /* copy the remaining bytes */ -17: ld.b t0, a1, 0 -18: st.b t0, a0, 0 - addi.d a0, a0, 1 - addi.d a1, a1, 1 - addi.d a2, a2, -1 - bgt a2, zero, 17b +.Llt64: + addi.d a4, a3, -32 + bgeu a1, a4, .Llt32 +18: ld.d t0, a1, 0 +19: ld.d t1, a1, 8 +20: ld.d t2, a1, 16 +21: ld.d t3, a1, 24 + addi.d a1, a1, 32 +22: st.d t0, a0, 0 +23: st.d t1, a0, 8 +24: st.d t2, a0, 16 +25: st.d t3, a0, 24 + addi.d a0, a0, 32 + +.Llt32: + addi.d a4, a3, -16 + bgeu a1, a4, .Llt16 +26: ld.d t0, a1, 0 +27: ld.d t1, a1, 8 + addi.d a1, a1, 16 +28: st.d t0, a0, 0 +29: st.d t1, a0, 8 + addi.d a0, a0, 16 + +.Llt16: + addi.d a4, a3, -8 + bgeu a1, a4, .Llt8 +30: ld.d t0, a1, 0 +31: st.d t0, a0, 0 + +.Llt8: +32: ld.d t0, a3, -8 +33: st.d t0, a2, -8 /* return */ -19: move a0, a2 + move a0, zero + jr ra + + .align 5 +.Lsmall: + pcaddi t0, 8 + slli.d a3, a2, 5 + add.d t0, t0, a3 + jr t0 + + .align 5 + move a0, zero + jr ra + + .align 5 +34: ld.b t0, a1, 0 +35: st.b t0, a0, 0 + move a0, zero + jr ra + + .align 5 +36: ld.h t0, a1, 0 +37: st.h t0, a0, 0 + move a0, zero + jr ra + + .align 5 +38: ld.h t0, a1, 0 +39: ld.b t1, a1, 2 +40: st.h t0, a0, 0 +41: st.b t1, a0, 2 + move a0, zero + jr ra + + .align 5 +42: ld.w t0, a1, 0 +43: st.w t0, a0, 0 + move a0, zero + jr ra + + .align 5 +44: ld.w t0, a1, 0 +45: ld.b t1, a1, 4 +46: st.w t0, a0, 0 +47: st.b t1, a0, 4 + move a0, zero + jr ra + + .align 5 +48: ld.w t0, a1, 0 +49: ld.h t1, a1, 4 +50: st.w t0, a0, 0 +51: st.h t1, a0, 4 + move a0, zero + jr ra + + .align 5 +52: ld.w t0, a1, 0 +53: ld.w t1, a1, 3 +54: st.w t0, a0, 0 +55: st.w t1, a0, 3 + move a0, zero + jr ra + + .align 5 +56: ld.d t0, a1, 0 +57: st.d t0, a0, 0 + move a0, zero jr ra /* fixup and ex_table */ + _asm_extable 0b, .L_fixup_handle_0 _asm_extable 1b, .L_fixup_handle_0 - _asm_extable 2b, .L_fixup_handle_1 - _asm_extable 3b, .L_fixup_handle_2 - _asm_extable 4b, .L_fixup_handle_3 - _asm_extable 5b, .L_fixup_handle_4 - _asm_extable 6b, .L_fixup_handle_5 - _asm_extable 7b, .L_fixup_handle_6 - _asm_extable 8b, .L_fixup_handle_7 + _asm_extable 2b, .L_fixup_handle_0 + _asm_extable 3b, .L_fixup_handle_0 + _asm_extable 4b, .L_fixup_handle_0 + _asm_extable 5b, .L_fixup_handle_0 + _asm_extable 6b, .L_fixup_handle_0 + _asm_extable 7b, .L_fixup_handle_0 + _asm_extable 8b, .L_fixup_handle_0 _asm_extable 9b, .L_fixup_handle_0 - _asm_extable 10b, .L_fixup_handle_1 - _asm_extable 11b, .L_fixup_handle_2 - _asm_extable 12b, .L_fixup_handle_3 - _asm_extable 13b, .L_fixup_handle_4 - _asm_extable 14b, .L_fixup_handle_5 - _asm_extable 15b, .L_fixup_handle_6 - _asm_extable 16b, .L_fixup_handle_7 - _asm_extable 17b, .L_fixup_handle_0 + _asm_extable 10b, .L_fixup_handle_0 + _asm_extable 11b, .L_fixup_handle_1 + _asm_extable 12b, .L_fixup_handle_2 + _asm_extable 13b, .L_fixup_handle_3 + _asm_extable 14b, .L_fixup_handle_4 + _asm_extable 15b, .L_fixup_handle_5 + _asm_extable 16b, .L_fixup_handle_6 + _asm_extable 17b, .L_fixup_handle_7 _asm_extable 18b, .L_fixup_handle_0 + _asm_extable 19b, .L_fixup_handle_0 + _asm_extable 20b, .L_fixup_handle_0 + _asm_extable 21b, .L_fixup_handle_0 + _asm_extable 22b, .L_fixup_handle_0 + _asm_extable 23b, .L_fixup_handle_1 + _asm_extable 24b, .L_fixup_handle_2 + _asm_extable 25b, .L_fixup_handle_3 + _asm_extable 26b, .L_fixup_handle_0 + _asm_extable 27b, .L_fixup_handle_0 + _asm_extable 28b, .L_fixup_handle_0 + _asm_extable 29b, .L_fixup_handle_1 + _asm_extable 30b, .L_fixup_handle_0 + _asm_extable 31b, .L_fixup_handle_0 + _asm_extable 32b, .L_fixup_handle_0 + _asm_extable 33b, .L_fixup_handle_1 + _asm_extable 34b, .L_fixup_handle_s0 + _asm_extable 35b, .L_fixup_handle_s0 + _asm_extable 36b, .L_fixup_handle_s0 + _asm_extable 37b, .L_fixup_handle_s0 + _asm_extable 38b, .L_fixup_handle_s0 + _asm_extable 39b, .L_fixup_handle_s0 + _asm_extable 40b, .L_fixup_handle_s0 + _asm_extable 41b, .L_fixup_handle_s2 + _asm_extable 42b, .L_fixup_handle_s0 + _asm_extable 43b, .L_fixup_handle_s0 + _asm_extable 44b, .L_fixup_handle_s0 + _asm_extable 45b, .L_fixup_handle_s0 + _asm_extable 46b, .L_fixup_handle_s0 + _asm_extable 47b, .L_fixup_handle_s4 + _asm_extable 48b, .L_fixup_handle_s0 + _asm_extable 49b, .L_fixup_handle_s0 + _asm_extable 50b, .L_fixup_handle_s0 + _asm_extable 51b, .L_fixup_handle_s4 + _asm_extable 52b, .L_fixup_handle_s0 + _asm_extable 53b, .L_fixup_handle_s0 + _asm_extable 54b, .L_fixup_handle_s0 + _asm_extable 55b, .L_fixup_handle_s4 + _asm_extable 56b, .L_fixup_handle_s0 + _asm_extable 57b, .L_fixup_handle_s0 SYM_FUNC_END(__copy_user_fast) diff --git a/arch/loongarch/lib/memcpy.S b/arch/loongarch/lib/memcpy.S index 7c07d595ee89..612bf620760a 100644 --- a/arch/loongarch/lib/memcpy.S +++ b/arch/loongarch/lib/memcpy.S @@ -42,6 +42,65 @@ SYM_FUNC_START(__memcpy_generic) jr ra SYM_FUNC_END(__memcpy_generic) + .align 5 +SYM_FUNC_START_NOALIGN(__memcpy_small) + pcaddi t0, 8 + slli.d a2, a2, 5 + add.d t0, t0, a2 + jr t0 + + .align 5 +0: jr ra + + .align 5 +1: ld.b t0, a1, 0 + st.b t0, a0, 0 + jr ra + + .align 5 +2: ld.h t0, a1, 0 + st.h t0, a0, 0 + jr ra + + .align 5 +3: ld.h t0, a1, 0 + ld.b t1, a1, 2 + st.h t0, a0, 0 + st.b t1, a0, 2 + jr ra + + .align 5 +4: ld.w t0, a1, 0 + st.w t0, a0, 0 + jr ra + + .align 5 +5: ld.w t0, a1, 0 + ld.b t1, a1, 4 + st.w t0, a0, 0 + st.b t1, a0, 4 + jr ra + + .align 5 +6: ld.w t0, a1, 0 + ld.h t1, a1, 4 + st.w t0, a0, 0 + st.h t1, a0, 4 + jr ra + + .align 5 +7: ld.w t0, a1, 0 + ld.w t1, a1, 3 + st.w t0, a0, 0 + st.w t1, a0, 3 + jr ra + + .align 5 +8: ld.d t0, a1, 0 + st.d t0, a0, 0 + jr ra +SYM_FUNC_END(__memcpy_small) + /* * void *__memcpy_fast(void *dst, const void *src, size_t n) * @@ -50,14 +109,27 @@ SYM_FUNC_END(__memcpy_generic) * a2: n */ SYM_FUNC_START(__memcpy_fast) - move a3, a0 - beqz a2, 3f + sltui t0, a2, 9 + bnez t0, __memcpy_small + + add.d a3, a1, a2 + add.d a2, a0, a2 + ld.d a6, a1, 0 + ld.d a7, a3, -8 + + /* align up destination address */ + andi t1, a0, 7 + sub.d t0, zero, t1 + addi.d t0, t0, 8 + add.d a1, a1, t0 + add.d a5, a0, t0 - ori a4, zero, 64 - blt a2, a4, 2f + addi.d a4, a3, -64 + bgeu a1, a4, .Llt64 /* copy 64 bytes at a time */ -1: ld.d t0, a1, 0 +.Lloop64: + ld.d t0, a1, 0 ld.d t1, a1, 8 ld.d t2, a1, 16 ld.d t3, a1, 24 @@ -65,31 +137,53 @@ SYM_FUNC_START(__memcpy_fast) ld.d t5, a1, 40 ld.d t6, a1, 48 ld.d t7, a1, 56 - st.d t0, a0, 0 - st.d t1, a0, 8 - st.d t2, a0, 16 - st.d t3, a0, 24 - st.d t4, a0, 32 - st.d t5, a0, 40 - st.d t6, a0, 48 - st.d t7, a0, 56 - - addi.d a0, a0, 64 addi.d a1, a1, 64 - addi.d a2, a2, -64 - bge a2, a4, 1b - - beqz a2, 3f + st.d t0, a5, 0 + st.d t1, a5, 8 + st.d t2, a5, 16 + st.d t3, a5, 24 + st.d t4, a5, 32 + st.d t5, a5, 40 + st.d t6, a5, 48 + st.d t7, a5, 56 + addi.d a5, a5, 64 + bltu a1, a4, .Lloop64 /* copy the remaining bytes */ -2: ld.b t0, a1, 0 - st.b t0, a0, 0 - addi.d a0, a0, 1 - addi.d a1, a1, 1 - addi.d a2, a2, -1 - bgt a2, zero, 2b +.Llt64: + addi.d a4, a3, -32 + bgeu a1, a4, .Llt32 + ld.d t0, a1, 0 + ld.d t1, a1, 8 + ld.d t2, a1, 16 + ld.d t3, a1, 24 + addi.d a1, a1, 32 + st.d t0, a5, 0 + st.d t1, a5, 8 + st.d t2, a5, 16 + st.d t3, a5, 24 + addi.d a5, a5, 32 + +.Llt32: + addi.d a4, a3, -16 + bgeu a1, a4, .Llt16 + ld.d t0, a1, 0 + ld.d t1, a1, 8 + addi.d a1, a1, 16 + st.d t0, a5, 0 + st.d t1, a5, 8 + addi.d a5, a5, 16 + +.Llt16: + addi.d a4, a3, -8 + bgeu a1, a4, .Llt8 + ld.d t0, a1, 0 + st.d t0, a5, 0 + +.Llt8: + st.d a6, a0, 0 + st.d a7, a2, -8 /* return */ -3: move a0, a3 jr ra SYM_FUNC_END(__memcpy_fast) diff --git a/arch/loongarch/lib/memmove.S b/arch/loongarch/lib/memmove.S index 6ffdb46da78f..d729e6cf4627 100644 --- a/arch/loongarch/lib/memmove.S +++ b/arch/loongarch/lib/memmove.S @@ -11,23 +11,9 @@ #include <asm/regdef.h> SYM_FUNC_START(memmove) - blt a0, a1, 1f /* dst < src, memcpy */ - blt a1, a0, 3f /* src < dst, rmemcpy */ + blt a0, a1, memcpy /* dst < src, memcpy */ + blt a1, a0, rmemcpy /* src < dst, rmemcpy */ jr ra /* dst == src, return */ - - /* if (src - dst) < 64, copy 1 byte at a time */ -1: ori a3, zero, 64 - sub.d t0, a1, a0 - blt t0, a3, 2f - b memcpy -2: b __memcpy_generic - - /* if (dst - src) < 64, copy 1 byte at a time */ -3: ori a3, zero, 64 - sub.d t0, a0, a1 - blt t0, a3, 4f - b rmemcpy -4: b __rmemcpy_generic SYM_FUNC_END(memmove) EXPORT_SYMBOL(memmove) @@ -73,49 +59,79 @@ SYM_FUNC_END(__rmemcpy_generic) * a2: n */ SYM_FUNC_START(__rmemcpy_fast) - move a3, a0 - beqz a2, 3f + sltui t0, a2, 9 + bnez t0, __memcpy_small - add.d a0, a0, a2 - add.d a1, a1, a2 + add.d a3, a1, a2 + add.d a2, a0, a2 + ld.d a6, a1, 0 + ld.d a7, a3, -8 + + /* align up destination address */ + andi t1, a2, 7 + sub.d a3, a3, t1 + sub.d a5, a2, t1 - ori a4, zero, 64 - blt a2, a4, 2f + addi.d a4, a1, 64 + bgeu a4, a3, .Llt64 /* copy 64 bytes at a time */ -1: ld.d t0, a1, -8 - ld.d t1, a1, -16 - ld.d t2, a1, -24 - ld.d t3, a1, -32 - ld.d t4, a1, -40 - ld.d t5, a1, -48 - ld.d t6, a1, -56 - ld.d t7, a1, -64 - st.d t0, a0, -8 - st.d t1, a0, -16 - st.d t2, a0, -24 - st.d t3, a0, -32 - st.d t4, a0, -40 - st.d t5, a0, -48 - st.d t6, a0, -56 - st.d t7, a0, -64 - - addi.d a0, a0, -64 - addi.d a1, a1, -64 - addi.d a2, a2, -64 - bge a2, a4, 1b - - beqz a2, 3f +.Lloop64: + ld.d t0, a3, -8 + ld.d t1, a3, -16 + ld.d t2, a3, -24 + ld.d t3, a3, -32 + ld.d t4, a3, -40 + ld.d t5, a3, -48 + ld.d t6, a3, -56 + ld.d t7, a3, -64 + addi.d a3, a3, -64 + st.d t0, a5, -8 + st.d t1, a5, -16 + st.d t2, a5, -24 + st.d t3, a5, -32 + st.d t4, a5, -40 + st.d t5, a5, -48 + st.d t6, a5, -56 + st.d t7, a5, -64 + addi.d a5, a5, -64 + bltu a4, a3, .Lloop64 /* copy the remaining bytes */ -2: ld.b t0, a1, -1 - st.b t0, a0, -1 - addi.d a0, a0, -1 - addi.d a1, a1, -1 - addi.d a2, a2, -1 - bgt a2, zero, 2b +.Llt64: + addi.d a4, a1, 32 + bgeu a4, a3, .Llt32 + ld.d t0, a3, -8 + ld.d t1, a3, -16 + ld.d t2, a3, -24 + ld.d t3, a3, -32 + addi.d a3, a3, -32 + st.d t0, a5, -8 + st.d t1, a5, -16 + st.d t2, a5, -24 + st.d t3, a5, -32 + addi.d a5, a5, -32 + +.Llt32: + addi.d a4, a1, 16 + bgeu a4, a3, .Llt16 + ld.d t0, a3, -8 + ld.d t1, a3, -16 + addi.d a3, a3, -16 + st.d t0, a5, -8 + st.d t1, a5, -16 + addi.d a5, a5, -16 + +.Llt16: + addi.d a4, a1, 8 + bgeu a4, a3, .Llt8 + ld.d t0, a3, -8 + st.d t0, a5, -8 + +.Llt8: + st.d a6, a0, 0 + st.d a7, a2, -8 /* return */ -3: move a0, a3 jr ra SYM_FUNC_END(__rmemcpy_fast) diff --git a/arch/loongarch/lib/memset.S b/arch/loongarch/lib/memset.S index e7cb4ea3747d..15dcd077991b 100644 --- a/arch/loongarch/lib/memset.S +++ b/arch/loongarch/lib/memset.S @@ -54,38 +54,106 @@ SYM_FUNC_END(__memset_generic) * a2: n */ SYM_FUNC_START(__memset_fast) - move a3, a0 - beqz a2, 3f - - ori a4, zero, 64 - blt a2, a4, 2f - /* fill a1 to 64 bits */ fill_to_64 a1 - /* set 64 bytes at a time */ -1: st.d a1, a0, 0 - st.d a1, a0, 8 - st.d a1, a0, 16 - st.d a1, a0, 24 - st.d a1, a0, 32 - st.d a1, a0, 40 - st.d a1, a0, 48 - st.d a1, a0, 56 + sltui t0, a2, 9 + bnez t0, .Lsmall - addi.d a0, a0, 64 - addi.d a2, a2, -64 - bge a2, a4, 1b + add.d a2, a0, a2 + st.d a1, a0, 0 - beqz a2, 3f + /* align up address */ + addi.d a3, a0, 8 + bstrins.d a3, zero, 2, 0 + + addi.d a4, a2, -64 + bgeu a3, a4, .Llt64 + + /* set 64 bytes at a time */ +.Lloop64: + st.d a1, a3, 0 + st.d a1, a3, 8 + st.d a1, a3, 16 + st.d a1, a3, 24 + st.d a1, a3, 32 + st.d a1, a3, 40 + st.d a1, a3, 48 + st.d a1, a3, 56 + addi.d a3, a3, 64 + bltu a3, a4, .Lloop64 /* set the remaining bytes */ -2: st.b a1, a0, 0 - addi.d a0, a0, 1 - addi.d a2, a2, -1 - bgt a2, zero, 2b +.Llt64: + addi.d a4, a2, -32 + bgeu a3, a4, .Llt32 + st.d a1, a3, 0 + st.d a1, a3, 8 + st.d a1, a3, 16 + st.d a1, a3, 24 + addi.d a3, a3, 32 + +.Llt32: + addi.d a4, a2, -16 + bgeu a3, a4, .Llt16 + st.d a1, a3, 0 + st.d a1, a3, 8 + addi.d a3, a3, 16 + +.Llt16: + addi.d a4, a2, -8 + bgeu a3, a4, .Llt8 + st.d a1, a3, 0 + +.Llt8: + st.d a1, a2, -8 /* return */ -3: move a0, a3 + jr ra + + .align 4 +.Lsmall: + pcaddi t0, 4 + slli.d a2, a2, 4 + add.d t0, t0, a2 + jr t0 + + .align 4 +0: jr ra + + .align 4 +1: st.b a1, a0, 0 + jr ra + + .align 4 +2: st.h a1, a0, 0 + jr ra + + .align 4 +3: st.h a1, a0, 0 + st.b a1, a0, 2 + jr ra + + .align 4 +4: st.w a1, a0, 0 + jr ra + + .align 4 +5: st.w a1, a0, 0 + st.b a1, a0, 4 + jr ra + + .align 4 +6: st.w a1, a0, 0 + st.h a1, a0, 4 + jr ra + + .align 4 +7: st.w a1, a0, 0 + st.w a1, a0, 3 + jr ra + + .align 4 +8: st.d a1, a0, 0 jr ra SYM_FUNC_END(__memset_fast) -- 2.39.2

2 5

[PATCH 1/5] LoongArch: Optimize memory ops (memset/memcpy/memmove)
by liweihao 19 Jan '24

19 Jan '24

From: WANG Rui <wangrui(a)loongson.cn> LoongArch inclusion bugzilla: https://gitee.com/openeuler/kernel/issues/I76XQZ To optimize memset()/memcpy()/memmove() and so on, we use a jump table to dispatch cases for short data lengths; and for long data lengths, we split the destination into head part (first 8 bytes), tail part (last 8 bytes) and middle part. The head part and tail part may be at unaligned addresses, while the middle part is always aligned (the middle part is allowed to overlap the head/tail part). In this way, the first and last 8 bytes may be unaligned accesses, but we can make sure the data in the middle is processed at an aligned destination address. We have tested micro-bench[1] on a Loongson-3C5000 16-core machine (2.2GHz): 1. memset | length | src offset | dst offset | speed before | speed after | % | |--------|------------|------------|--------------|-------------|---------| | 8 | 0 | 0 | 696.191 | 1518.785 | 118.16% | | 8 | 0 | 1 | 696.325 | 1518.937 | 118.14% | | 50 | 0 | 0 | 969.976 | 8053.902 | 730.32% | | 50 | 0 | 1 | 970.034 | 8058.475 | 730.74% | | 300 | 0 | 0 | 5876.612 | 16544.703 | 181.53% | | 300 | 0 | 1 | 5030.849 | 16549.011 | 228.95% | | 1200 | 0 | 0 | 11797.077 | 16752.137 | 42.00% | | 1200 | 0 | 1 | 5687.141 | 16645.233 | 192.68% | | 4000 | 0 | 0 | 15723.27 | 16761.557 | 6.60% | | 4000 | 0 | 1 | 5906.114 | 16732.316 | 183.30% | | 8000 | 0 | 0 | 16751.403 | 16770.002 | 0.11% | | 8000 | 0 | 1 | 5995.449 | 16754.07 | 179.45% | 2. memcpy | length | src offset | dst offset | speed before | speed after | % | |--------|------------|------------|--------------|-------------|---------| | 8 | 0 | 0 | 696.2 | 1670.605 | 139.96% | | 8 | 0 | 1 | 696.325 | 1671.138 | 139.99% | | 50 | 0 | 0 | 969.974 | 8724.999 | 799.51% | | 50 | 0 | 1 | 970.032 | 8730.138 | 799.98% | | 300 | 0 | 0 | 5564.662 | 16272.652 | 192.43% | | 300 | 0 | 1 | 4670.436 | 14972.842 | 220.59% | | 1200 | 0 | 0 | 10740.23 | 16751.728 | 55.97% | | 1200 | 0 | 1 | 5027.741 | 14874.564 | 195.85% | | 4000 | 0 | 0 | 15122.367 | 16737.642 | 10.68% | | 4000 | 0 | 1 | 5536.918 | 14890.397 | 168.93% | | 8000 | 0 | 0 | 16505.453 | 16553.543 | 0.29% | | 8000 | 0 | 1 | 5821.619 | 14841.804 | 154.94% | 3. memmove | length | src offset | dst offset | speed before | speed after | % | |--------|------------|------------|--------------|-------------|---------| | 8 | 0 | 0 | 982.693 | 1670.568 | 70.00% | | 8 | 0 | 1 | 983.023 | 1671.174 | 70.00% | | 50 | 0 | 0 | 1230.87 | 8727.625 | 609.06% | | 50 | 0 | 1 | 1232.515 | 8730.138 | 608.32% | | 300 | 0 | 0 | 6490.375 | 16296.993 | 151.09% | | 300 | 0 | 1 | 4282.687 | 14972.842 | 249.61% | | 1200 | 0 | 0 | 11742.755 | 16752.546 | 42.66% | | 1200 | 0 | 1 | 5039.338 | 14872.951 | 195.14% | | 4000 | 0 | 0 | 15467.786 | 16737.09 | 8.21% | | 4000 | 0 | 1 | 5009.905 | 14890.542 | 197.22% | | 8000 | 0 | 0 | 16489.664 | 16553.273 | 0.39% | | 8000 | 0 | 1 | 5823.786 | 14858.646 | 155.14% | * speed: MB/s * length: byte [1] https://github.com/heiher/mem-bench Signed-off-by: WANG Rui <wangrui(a)loongson.cn> Signed-off-by: Huacai Chen <chenhuacai(a)loongson.cn> --- arch/loongarch/lib/clear_user.S | 136 +++++++++++++++-- arch/loongarch/lib/copy_user.S | 251 +++++++++++++++++++++++++------- arch/loongarch/lib/memcpy.S | 146 +++++++++++++++---- arch/loongarch/lib/memmove.S | 120 ++++++++------- arch/loongarch/lib/memset.S | 116 ++++++++++++--- 5 files changed, 602 insertions(+), 167 deletions(-) diff --git a/arch/loongarch/lib/clear_user.S b/arch/loongarch/lib/clear_user.S index 2dc48e61a2c8..fd1d62b244f2 100644 --- a/arch/loongarch/lib/clear_user.S +++ b/arch/loongarch/lib/clear_user.S @@ -13,7 +13,14 @@ .irp to, 0, 1, 2, 3, 4, 5, 6, 7 .L_fixup_handle_\to\(): - addi.d a0, a1, (\to) * (-8) + sub.d a0, a2, a0 + addi.d a0, a0, (\to) * (-8) + jr ra +.endr + +.irp to, 0, 2, 4 +.L_fixup_handle_s\to\(): + addi.d a0, a1, -\to jr ra .endr @@ -44,7 +51,7 @@ SYM_FUNC_START(__clear_user_generic) 2: move a0, a1 jr ra - _asm_extable 1b, .L_fixup_handle_0 + _asm_extable 1b, .L_fixup_handle_s0 SYM_FUNC_END(__clear_user_generic) /* @@ -54,12 +61,21 @@ SYM_FUNC_END(__clear_user_generic) * a1: size */ SYM_FUNC_START(__clear_user_fast) - beqz a1, 10f + sltui t0, a1, 9 + bnez t0, .Lsmall - ori a2, zero, 64 - blt a1, a2, 9f + add.d a2, a0, a1 +0: st.d zero, a0, 0 + + /* align up address */ + addi.d a0, a0, 8 + bstrins.d a0, zero, 2, 0 + + addi.d a3, a2, -64 + bgeu a0, a3, .Llt64 /* set 64 bytes at a time */ +.Lloop64: 1: st.d zero, a0, 0 2: st.d zero, a0, 8 3: st.d zero, a0, 16 @@ -68,24 +84,95 @@ SYM_FUNC_START(__clear_user_fast) 6: st.d zero, a0, 40 7: st.d zero, a0, 48 8: st.d zero, a0, 56 - addi.d a0, a0, 64 - addi.d a1, a1, -64 - bge a1, a2, 1b - - beqz a1, 10f + bltu a0, a3, .Lloop64 /* set the remaining bytes */ -9: st.b zero, a0, 0 - addi.d a0, a0, 1 - addi.d a1, a1, -1 - bgt a1, zero, 9b +.Llt64: + addi.d a3, a2, -32 + bgeu a0, a3, .Llt32 +9: st.d zero, a0, 0 +10: st.d zero, a0, 8 +11: st.d zero, a0, 16 +12: st.d zero, a0, 24 + addi.d a0, a0, 32 + +.Llt32: + addi.d a3, a2, -16 + bgeu a0, a3, .Llt16 +13: st.d zero, a0, 0 +14: st.d zero, a0, 8 + addi.d a0, a0, 16 + +.Llt16: + addi.d a3, a2, -8 + bgeu a0, a3, .Llt8 +15: st.d zero, a0, 0 + +.Llt8: +16: st.d zero, a2, -8 /* return */ -10: move a0, a1 + move a0, zero + jr ra + + .align 4 +.Lsmall: + pcaddi t0, 4 + slli.d a2, a1, 4 + add.d t0, t0, a2 + jr t0 + + .align 4 + move a0, zero + jr ra + + .align 4 +17: st.b zero, a0, 0 + move a0, zero + jr ra + + .align 4 +18: st.h zero, a0, 0 + move a0, zero + jr ra + + .align 4 +19: st.h zero, a0, 0 +20: st.b zero, a0, 2 + move a0, zero + jr ra + + .align 4 +21: st.w zero, a0, 0 + move a0, zero + jr ra + + .align 4 +22: st.w zero, a0, 0 +23: st.b zero, a0, 4 + move a0, zero + jr ra + + .align 4 +24: st.w zero, a0, 0 +25: st.h zero, a0, 4 + move a0, zero + jr ra + + .align 4 +26: st.w zero, a0, 0 +27: st.w zero, a0, 3 + move a0, zero + jr ra + + .align 4 +28: st.d zero, a0, 0 + move a0, zero jr ra /* fixup and ex_table */ + _asm_extable 0b, .L_fixup_handle_0 _asm_extable 1b, .L_fixup_handle_0 _asm_extable 2b, .L_fixup_handle_1 _asm_extable 3b, .L_fixup_handle_2 @@ -95,4 +182,23 @@ SYM_FUNC_START(__clear_user_fast) _asm_extable 7b, .L_fixup_handle_6 _asm_extable 8b, .L_fixup_handle_7 _asm_extable 9b, .L_fixup_handle_0 + _asm_extable 10b, .L_fixup_handle_1 + _asm_extable 11b, .L_fixup_handle_2 + _asm_extable 12b, .L_fixup_handle_3 + _asm_extable 13b, .L_fixup_handle_0 + _asm_extable 14b, .L_fixup_handle_1 + _asm_extable 15b, .L_fixup_handle_0 + _asm_extable 16b, .L_fixup_handle_1 + _asm_extable 17b, .L_fixup_handle_s0 + _asm_extable 18b, .L_fixup_handle_s0 + _asm_extable 19b, .L_fixup_handle_s0 + _asm_extable 20b, .L_fixup_handle_s2 + _asm_extable 21b, .L_fixup_handle_s0 + _asm_extable 22b, .L_fixup_handle_s0 + _asm_extable 23b, .L_fixup_handle_s4 + _asm_extable 24b, .L_fixup_handle_s0 + _asm_extable 25b, .L_fixup_handle_s4 + _asm_extable 26b, .L_fixup_handle_s0 + _asm_extable 27b, .L_fixup_handle_s4 + _asm_extable 28b, .L_fixup_handle_s0 SYM_FUNC_END(__clear_user_fast) diff --git a/arch/loongarch/lib/copy_user.S b/arch/loongarch/lib/copy_user.S index 55ac6020a1ad..b21f6d5d38f5 100644 --- a/arch/loongarch/lib/copy_user.S +++ b/arch/loongarch/lib/copy_user.S @@ -13,7 +13,14 @@ .irp to, 0, 1, 2, 3, 4, 5, 6, 7 .L_fixup_handle_\to\(): - addi.d a0, a2, (\to) * (-8) + sub.d a0, a2, a0 + addi.d a0, a0, (\to) * (-8) + jr ra +.endr + +.irp to, 0, 2, 4 +.L_fixup_handle_s\to\(): + addi.d a0, a2, -\to jr ra .endr @@ -47,8 +54,8 @@ SYM_FUNC_START(__copy_user_generic) 3: move a0, a2 jr ra - _asm_extable 1b, .L_fixup_handle_0 - _asm_extable 2b, .L_fixup_handle_0 + _asm_extable 1b, .L_fixup_handle_s0 + _asm_extable 2b, .L_fixup_handle_s0 SYM_FUNC_END(__copy_user_generic) /* @@ -59,65 +66,209 @@ SYM_FUNC_END(__copy_user_generic) * a2: n */ SYM_FUNC_START(__copy_user_fast) - beqz a2, 19f + sltui t0, a2, 9 + bnez t0, .Lsmall - ori a3, zero, 64 - blt a2, a3, 17f + add.d a3, a1, a2 + add.d a2, a0, a2 +0: ld.d t0, a1, 0 +1: st.d t0, a0, 0 - /* copy 64 bytes at a time */ -1: ld.d t0, a1, 0 -2: ld.d t1, a1, 8 -3: ld.d t2, a1, 16 -4: ld.d t3, a1, 24 -5: ld.d t4, a1, 32 -6: ld.d t5, a1, 40 -7: ld.d t6, a1, 48 -8: ld.d t7, a1, 56 -9: st.d t0, a0, 0 -10: st.d t1, a0, 8 -11: st.d t2, a0, 16 -12: st.d t3, a0, 24 -13: st.d t4, a0, 32 -14: st.d t5, a0, 40 -15: st.d t6, a0, 48 -16: st.d t7, a0, 56 + /* align up destination address */ + andi t1, a0, 7 + sub.d t0, zero, t1 + addi.d t0, t0, 8 + add.d a1, a1, t0 + add.d a0, a0, t0 - addi.d a0, a0, 64 - addi.d a1, a1, 64 - addi.d a2, a2, -64 - bge a2, a3, 1b + addi.d a4, a3, -64 + bgeu a1, a4, .Llt64 - beqz a2, 19f + /* copy 64 bytes at a time */ +.Lloop64: +2: ld.d t0, a1, 0 +3: ld.d t1, a1, 8 +4: ld.d t2, a1, 16 +5: ld.d t3, a1, 24 +6: ld.d t4, a1, 32 +7: ld.d t5, a1, 40 +8: ld.d t6, a1, 48 +9: ld.d t7, a1, 56 + addi.d a1, a1, 64 +10: st.d t0, a0, 0 +11: st.d t1, a0, 8 +12: st.d t2, a0, 16 +13: st.d t3, a0, 24 +14: st.d t4, a0, 32 +15: st.d t5, a0, 40 +16: st.d t6, a0, 48 +17: st.d t7, a0, 56 + addi.d a0, a0, 64 + bltu a1, a4, .Lloop64 /* copy the remaining bytes */ -17: ld.b t0, a1, 0 -18: st.b t0, a0, 0 - addi.d a0, a0, 1 - addi.d a1, a1, 1 - addi.d a2, a2, -1 - bgt a2, zero, 17b +.Llt64: + addi.d a4, a3, -32 + bgeu a1, a4, .Llt32 +18: ld.d t0, a1, 0 +19: ld.d t1, a1, 8 +20: ld.d t2, a1, 16 +21: ld.d t3, a1, 24 + addi.d a1, a1, 32 +22: st.d t0, a0, 0 +23: st.d t1, a0, 8 +24: st.d t2, a0, 16 +25: st.d t3, a0, 24 + addi.d a0, a0, 32 + +.Llt32: + addi.d a4, a3, -16 + bgeu a1, a4, .Llt16 +26: ld.d t0, a1, 0 +27: ld.d t1, a1, 8 + addi.d a1, a1, 16 +28: st.d t0, a0, 0 +29: st.d t1, a0, 8 + addi.d a0, a0, 16 + +.Llt16: + addi.d a4, a3, -8 + bgeu a1, a4, .Llt8 +30: ld.d t0, a1, 0 +31: st.d t0, a0, 0 + +.Llt8: +32: ld.d t0, a3, -8 +33: st.d t0, a2, -8 /* return */ -19: move a0, a2 + move a0, zero + jr ra + + .align 5 +.Lsmall: + pcaddi t0, 8 + slli.d a3, a2, 5 + add.d t0, t0, a3 + jr t0 + + .align 5 + move a0, zero + jr ra + + .align 5 +34: ld.b t0, a1, 0 +35: st.b t0, a0, 0 + move a0, zero + jr ra + + .align 5 +36: ld.h t0, a1, 0 +37: st.h t0, a0, 0 + move a0, zero + jr ra + + .align 5 +38: ld.h t0, a1, 0 +39: ld.b t1, a1, 2 +40: st.h t0, a0, 0 +41: st.b t1, a0, 2 + move a0, zero + jr ra + + .align 5 +42: ld.w t0, a1, 0 +43: st.w t0, a0, 0 + move a0, zero + jr ra + + .align 5 +44: ld.w t0, a1, 0 +45: ld.b t1, a1, 4 +46: st.w t0, a0, 0 +47: st.b t1, a0, 4 + move a0, zero + jr ra + + .align 5 +48: ld.w t0, a1, 0 +49: ld.h t1, a1, 4 +50: st.w t0, a0, 0 +51: st.h t1, a0, 4 + move a0, zero + jr ra + + .align 5 +52: ld.w t0, a1, 0 +53: ld.w t1, a1, 3 +54: st.w t0, a0, 0 +55: st.w t1, a0, 3 + move a0, zero + jr ra + + .align 5 +56: ld.d t0, a1, 0 +57: st.d t0, a0, 0 + move a0, zero jr ra /* fixup and ex_table */ + _asm_extable 0b, .L_fixup_handle_0 _asm_extable 1b, .L_fixup_handle_0 - _asm_extable 2b, .L_fixup_handle_1 - _asm_extable 3b, .L_fixup_handle_2 - _asm_extable 4b, .L_fixup_handle_3 - _asm_extable 5b, .L_fixup_handle_4 - _asm_extable 6b, .L_fixup_handle_5 - _asm_extable 7b, .L_fixup_handle_6 - _asm_extable 8b, .L_fixup_handle_7 + _asm_extable 2b, .L_fixup_handle_0 + _asm_extable 3b, .L_fixup_handle_0 + _asm_extable 4b, .L_fixup_handle_0 + _asm_extable 5b, .L_fixup_handle_0 + _asm_extable 6b, .L_fixup_handle_0 + _asm_extable 7b, .L_fixup_handle_0 + _asm_extable 8b, .L_fixup_handle_0 _asm_extable 9b, .L_fixup_handle_0 - _asm_extable 10b, .L_fixup_handle_1 - _asm_extable 11b, .L_fixup_handle_2 - _asm_extable 12b, .L_fixup_handle_3 - _asm_extable 13b, .L_fixup_handle_4 - _asm_extable 14b, .L_fixup_handle_5 - _asm_extable 15b, .L_fixup_handle_6 - _asm_extable 16b, .L_fixup_handle_7 - _asm_extable 17b, .L_fixup_handle_0 + _asm_extable 10b, .L_fixup_handle_0 + _asm_extable 11b, .L_fixup_handle_1 + _asm_extable 12b, .L_fixup_handle_2 + _asm_extable 13b, .L_fixup_handle_3 + _asm_extable 14b, .L_fixup_handle_4 + _asm_extable 15b, .L_fixup_handle_5 + _asm_extable 16b, .L_fixup_handle_6 + _asm_extable 17b, .L_fixup_handle_7 _asm_extable 18b, .L_fixup_handle_0 + _asm_extable 19b, .L_fixup_handle_0 + _asm_extable 20b, .L_fixup_handle_0 + _asm_extable 21b, .L_fixup_handle_0 + _asm_extable 22b, .L_fixup_handle_0 + _asm_extable 23b, .L_fixup_handle_1 + _asm_extable 24b, .L_fixup_handle_2 + _asm_extable 25b, .L_fixup_handle_3 + _asm_extable 26b, .L_fixup_handle_0 + _asm_extable 27b, .L_fixup_handle_0 + _asm_extable 28b, .L_fixup_handle_0 + _asm_extable 29b, .L_fixup_handle_1 + _asm_extable 30b, .L_fixup_handle_0 + _asm_extable 31b, .L_fixup_handle_0 + _asm_extable 32b, .L_fixup_handle_0 + _asm_extable 33b, .L_fixup_handle_1 + _asm_extable 34b, .L_fixup_handle_s0 + _asm_extable 35b, .L_fixup_handle_s0 + _asm_extable 36b, .L_fixup_handle_s0 + _asm_extable 37b, .L_fixup_handle_s0 + _asm_extable 38b, .L_fixup_handle_s0 + _asm_extable 39b, .L_fixup_handle_s0 + _asm_extable 40b, .L_fixup_handle_s0 + _asm_extable 41b, .L_fixup_handle_s2 + _asm_extable 42b, .L_fixup_handle_s0 + _asm_extable 43b, .L_fixup_handle_s0 + _asm_extable 44b, .L_fixup_handle_s0 + _asm_extable 45b, .L_fixup_handle_s0 + _asm_extable 46b, .L_fixup_handle_s0 + _asm_extable 47b, .L_fixup_handle_s4 + _asm_extable 48b, .L_fixup_handle_s0 + _asm_extable 49b, .L_fixup_handle_s0 + _asm_extable 50b, .L_fixup_handle_s0 + _asm_extable 51b, .L_fixup_handle_s4 + _asm_extable 52b, .L_fixup_handle_s0 + _asm_extable 53b, .L_fixup_handle_s0 + _asm_extable 54b, .L_fixup_handle_s0 + _asm_extable 55b, .L_fixup_handle_s4 + _asm_extable 56b, .L_fixup_handle_s0 + _asm_extable 57b, .L_fixup_handle_s0 SYM_FUNC_END(__copy_user_fast) diff --git a/arch/loongarch/lib/memcpy.S b/arch/loongarch/lib/memcpy.S index 7c07d595ee89..612bf620760a 100644 --- a/arch/loongarch/lib/memcpy.S +++ b/arch/loongarch/lib/memcpy.S @@ -42,6 +42,65 @@ SYM_FUNC_START(__memcpy_generic) jr ra SYM_FUNC_END(__memcpy_generic) + .align 5 +SYM_FUNC_START_NOALIGN(__memcpy_small) + pcaddi t0, 8 + slli.d a2, a2, 5 + add.d t0, t0, a2 + jr t0 + + .align 5 +0: jr ra + + .align 5 +1: ld.b t0, a1, 0 + st.b t0, a0, 0 + jr ra + + .align 5 +2: ld.h t0, a1, 0 + st.h t0, a0, 0 + jr ra + + .align 5 +3: ld.h t0, a1, 0 + ld.b t1, a1, 2 + st.h t0, a0, 0 + st.b t1, a0, 2 + jr ra + + .align 5 +4: ld.w t0, a1, 0 + st.w t0, a0, 0 + jr ra + + .align 5 +5: ld.w t0, a1, 0 + ld.b t1, a1, 4 + st.w t0, a0, 0 + st.b t1, a0, 4 + jr ra + + .align 5 +6: ld.w t0, a1, 0 + ld.h t1, a1, 4 + st.w t0, a0, 0 + st.h t1, a0, 4 + jr ra + + .align 5 +7: ld.w t0, a1, 0 + ld.w t1, a1, 3 + st.w t0, a0, 0 + st.w t1, a0, 3 + jr ra + + .align 5 +8: ld.d t0, a1, 0 + st.d t0, a0, 0 + jr ra +SYM_FUNC_END(__memcpy_small) + /* * void *__memcpy_fast(void *dst, const void *src, size_t n) * @@ -50,14 +109,27 @@ SYM_FUNC_END(__memcpy_generic) * a2: n */ SYM_FUNC_START(__memcpy_fast) - move a3, a0 - beqz a2, 3f + sltui t0, a2, 9 + bnez t0, __memcpy_small + + add.d a3, a1, a2 + add.d a2, a0, a2 + ld.d a6, a1, 0 + ld.d a7, a3, -8 + + /* align up destination address */ + andi t1, a0, 7 + sub.d t0, zero, t1 + addi.d t0, t0, 8 + add.d a1, a1, t0 + add.d a5, a0, t0 - ori a4, zero, 64 - blt a2, a4, 2f + addi.d a4, a3, -64 + bgeu a1, a4, .Llt64 /* copy 64 bytes at a time */ -1: ld.d t0, a1, 0 +.Lloop64: + ld.d t0, a1, 0 ld.d t1, a1, 8 ld.d t2, a1, 16 ld.d t3, a1, 24 @@ -65,31 +137,53 @@ SYM_FUNC_START(__memcpy_fast) ld.d t5, a1, 40 ld.d t6, a1, 48 ld.d t7, a1, 56 - st.d t0, a0, 0 - st.d t1, a0, 8 - st.d t2, a0, 16 - st.d t3, a0, 24 - st.d t4, a0, 32 - st.d t5, a0, 40 - st.d t6, a0, 48 - st.d t7, a0, 56 - - addi.d a0, a0, 64 addi.d a1, a1, 64 - addi.d a2, a2, -64 - bge a2, a4, 1b - - beqz a2, 3f + st.d t0, a5, 0 + st.d t1, a5, 8 + st.d t2, a5, 16 + st.d t3, a5, 24 + st.d t4, a5, 32 + st.d t5, a5, 40 + st.d t6, a5, 48 + st.d t7, a5, 56 + addi.d a5, a5, 64 + bltu a1, a4, .Lloop64 /* copy the remaining bytes */ -2: ld.b t0, a1, 0 - st.b t0, a0, 0 - addi.d a0, a0, 1 - addi.d a1, a1, 1 - addi.d a2, a2, -1 - bgt a2, zero, 2b +.Llt64: + addi.d a4, a3, -32 + bgeu a1, a4, .Llt32 + ld.d t0, a1, 0 + ld.d t1, a1, 8 + ld.d t2, a1, 16 + ld.d t3, a1, 24 + addi.d a1, a1, 32 + st.d t0, a5, 0 + st.d t1, a5, 8 + st.d t2, a5, 16 + st.d t3, a5, 24 + addi.d a5, a5, 32 + +.Llt32: + addi.d a4, a3, -16 + bgeu a1, a4, .Llt16 + ld.d t0, a1, 0 + ld.d t1, a1, 8 + addi.d a1, a1, 16 + st.d t0, a5, 0 + st.d t1, a5, 8 + addi.d a5, a5, 16 + +.Llt16: + addi.d a4, a3, -8 + bgeu a1, a4, .Llt8 + ld.d t0, a1, 0 + st.d t0, a5, 0 + +.Llt8: + st.d a6, a0, 0 + st.d a7, a2, -8 /* return */ -3: move a0, a3 jr ra SYM_FUNC_END(__memcpy_fast) diff --git a/arch/loongarch/lib/memmove.S b/arch/loongarch/lib/memmove.S index 6ffdb46da78f..d729e6cf4627 100644 --- a/arch/loongarch/lib/memmove.S +++ b/arch/loongarch/lib/memmove.S @@ -11,23 +11,9 @@ #include <asm/regdef.h> SYM_FUNC_START(memmove) - blt a0, a1, 1f /* dst < src, memcpy */ - blt a1, a0, 3f /* src < dst, rmemcpy */ + blt a0, a1, memcpy /* dst < src, memcpy */ + blt a1, a0, rmemcpy /* src < dst, rmemcpy */ jr ra /* dst == src, return */ - - /* if (src - dst) < 64, copy 1 byte at a time */ -1: ori a3, zero, 64 - sub.d t0, a1, a0 - blt t0, a3, 2f - b memcpy -2: b __memcpy_generic - - /* if (dst - src) < 64, copy 1 byte at a time */ -3: ori a3, zero, 64 - sub.d t0, a0, a1 - blt t0, a3, 4f - b rmemcpy -4: b __rmemcpy_generic SYM_FUNC_END(memmove) EXPORT_SYMBOL(memmove) @@ -73,49 +59,79 @@ SYM_FUNC_END(__rmemcpy_generic) * a2: n */ SYM_FUNC_START(__rmemcpy_fast) - move a3, a0 - beqz a2, 3f + sltui t0, a2, 9 + bnez t0, __memcpy_small - add.d a0, a0, a2 - add.d a1, a1, a2 + add.d a3, a1, a2 + add.d a2, a0, a2 + ld.d a6, a1, 0 + ld.d a7, a3, -8 + + /* align up destination address */ + andi t1, a2, 7 + sub.d a3, a3, t1 + sub.d a5, a2, t1 - ori a4, zero, 64 - blt a2, a4, 2f + addi.d a4, a1, 64 + bgeu a4, a3, .Llt64 /* copy 64 bytes at a time */ -1: ld.d t0, a1, -8 - ld.d t1, a1, -16 - ld.d t2, a1, -24 - ld.d t3, a1, -32 - ld.d t4, a1, -40 - ld.d t5, a1, -48 - ld.d t6, a1, -56 - ld.d t7, a1, -64 - st.d t0, a0, -8 - st.d t1, a0, -16 - st.d t2, a0, -24 - st.d t3, a0, -32 - st.d t4, a0, -40 - st.d t5, a0, -48 - st.d t6, a0, -56 - st.d t7, a0, -64 - - addi.d a0, a0, -64 - addi.d a1, a1, -64 - addi.d a2, a2, -64 - bge a2, a4, 1b - - beqz a2, 3f +.Lloop64: + ld.d t0, a3, -8 + ld.d t1, a3, -16 + ld.d t2, a3, -24 + ld.d t3, a3, -32 + ld.d t4, a3, -40 + ld.d t5, a3, -48 + ld.d t6, a3, -56 + ld.d t7, a3, -64 + addi.d a3, a3, -64 + st.d t0, a5, -8 + st.d t1, a5, -16 + st.d t2, a5, -24 + st.d t3, a5, -32 + st.d t4, a5, -40 + st.d t5, a5, -48 + st.d t6, a5, -56 + st.d t7, a5, -64 + addi.d a5, a5, -64 + bltu a4, a3, .Lloop64 /* copy the remaining bytes */ -2: ld.b t0, a1, -1 - st.b t0, a0, -1 - addi.d a0, a0, -1 - addi.d a1, a1, -1 - addi.d a2, a2, -1 - bgt a2, zero, 2b +.Llt64: + addi.d a4, a1, 32 + bgeu a4, a3, .Llt32 + ld.d t0, a3, -8 + ld.d t1, a3, -16 + ld.d t2, a3, -24 + ld.d t3, a3, -32 + addi.d a3, a3, -32 + st.d t0, a5, -8 + st.d t1, a5, -16 + st.d t2, a5, -24 + st.d t3, a5, -32 + addi.d a5, a5, -32 + +.Llt32: + addi.d a4, a1, 16 + bgeu a4, a3, .Llt16 + ld.d t0, a3, -8 + ld.d t1, a3, -16 + addi.d a3, a3, -16 + st.d t0, a5, -8 + st.d t1, a5, -16 + addi.d a5, a5, -16 + +.Llt16: + addi.d a4, a1, 8 + bgeu a4, a3, .Llt8 + ld.d t0, a3, -8 + st.d t0, a5, -8 + +.Llt8: + st.d a6, a0, 0 + st.d a7, a2, -8 /* return */ -3: move a0, a3 jr ra SYM_FUNC_END(__rmemcpy_fast) diff --git a/arch/loongarch/lib/memset.S b/arch/loongarch/lib/memset.S index e7cb4ea3747d..15dcd077991b 100644 --- a/arch/loongarch/lib/memset.S +++ b/arch/loongarch/lib/memset.S @@ -54,38 +54,106 @@ SYM_FUNC_END(__memset_generic) * a2: n */ SYM_FUNC_START(__memset_fast) - move a3, a0 - beqz a2, 3f - - ori a4, zero, 64 - blt a2, a4, 2f - /* fill a1 to 64 bits */ fill_to_64 a1 - /* set 64 bytes at a time */ -1: st.d a1, a0, 0 - st.d a1, a0, 8 - st.d a1, a0, 16 - st.d a1, a0, 24 - st.d a1, a0, 32 - st.d a1, a0, 40 - st.d a1, a0, 48 - st.d a1, a0, 56 + sltui t0, a2, 9 + bnez t0, .Lsmall - addi.d a0, a0, 64 - addi.d a2, a2, -64 - bge a2, a4, 1b + add.d a2, a0, a2 + st.d a1, a0, 0 - beqz a2, 3f + /* align up address */ + addi.d a3, a0, 8 + bstrins.d a3, zero, 2, 0 + + addi.d a4, a2, -64 + bgeu a3, a4, .Llt64 + + /* set 64 bytes at a time */ +.Lloop64: + st.d a1, a3, 0 + st.d a1, a3, 8 + st.d a1, a3, 16 + st.d a1, a3, 24 + st.d a1, a3, 32 + st.d a1, a3, 40 + st.d a1, a3, 48 + st.d a1, a3, 56 + addi.d a3, a3, 64 + bltu a3, a4, .Lloop64 /* set the remaining bytes */ -2: st.b a1, a0, 0 - addi.d a0, a0, 1 - addi.d a2, a2, -1 - bgt a2, zero, 2b +.Llt64: + addi.d a4, a2, -32 + bgeu a3, a4, .Llt32 + st.d a1, a3, 0 + st.d a1, a3, 8 + st.d a1, a3, 16 + st.d a1, a3, 24 + addi.d a3, a3, 32 + +.Llt32: + addi.d a4, a2, -16 + bgeu a3, a4, .Llt16 + st.d a1, a3, 0 + st.d a1, a3, 8 + addi.d a3, a3, 16 + +.Llt16: + addi.d a4, a2, -8 + bgeu a3, a4, .Llt8 + st.d a1, a3, 0 + +.Llt8: + st.d a1, a2, -8 /* return */ -3: move a0, a3 + jr ra + + .align 4 +.Lsmall: + pcaddi t0, 4 + slli.d a2, a2, 4 + add.d t0, t0, a2 + jr t0 + + .align 4 +0: jr ra + + .align 4 +1: st.b a1, a0, 0 + jr ra + + .align 4 +2: st.h a1, a0, 0 + jr ra + + .align 4 +3: st.h a1, a0, 0 + st.b a1, a0, 2 + jr ra + + .align 4 +4: st.w a1, a0, 0 + jr ra + + .align 4 +5: st.w a1, a0, 0 + st.b a1, a0, 4 + jr ra + + .align 4 +6: st.w a1, a0, 0 + st.h a1, a0, 4 + jr ra + + .align 4 +7: st.w a1, a0, 0 + st.w a1, a0, 3 + jr ra + + .align 4 +8: st.d a1, a0, 0 jr ra SYM_FUNC_END(__memset_fast) -- 2.39.2

1 4

[PATCH OLK-5.10 v2 0/2] Fix io pressure when memory insufficient
by Liu Shixin 19 Jan '24

19 Jan '24

When memory is only a little and there are some pagecache can be reclaimed to avoid triggering OOM, but these pagecache is just right what the program is currently reading. The pagecache will be reclaimed and read from disk again, which will keeping a long time before OOM. Due to read-ahead, there are large of redundant read of disk which aggravate this problem. These two patch lessen the problem by early break read-ahead. Liu Shixin (2): mm/readahead: stop readahead loop if memcg charge fails mm/readahead: limit sync readahead while too many active refault include/linux/fs.h | 1 + include/linux/pagemap.h | 1 + mm/filemap.c | 16 ++++++++++++++++ mm/readahead.c | 21 ++++++++++++++++----- 4 files changed, 34 insertions(+), 5 deletions(-) -- 2.25.1

2 3

[PATCH openEuler-1.0-LTS 0/2] rtnetlink: Reject negative ifindexes in RTM_NEWLINK
by Wang Yufen 18 Jan '24

18 Jan '24

Greg Kroah-Hartman (1): Revert "rtnetlink: Reject negative ifindexes in RTM_NEWLINK" Ido Schimmel (1): rtnetlink: Reject negative ifindexes in RTM_NEWLINK net/core/rtnetlink.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) -- 1.8.3.1

2 3

[PATCH openEuler-1.0-LTS 1/2] Revert "rtnetlink: Reject negative ifindexes in RTM_NEWLINK"
by Wang Yufen 18 Jan '24

18 Jan '24

From: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> stable inclusion from stable-v4.19.296 commit 88f8a01c42535285bcf2666d1f7eb9ada0de88cc category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8XA2F CVE: NA -------------------------------- This reverts commit 42c8406432e730cb7442d97ecfdbf47084a5af4d which is commit 30188bd7838c16a98a520db1fe9df01ffc6ed368 upstream. It was improperly backported to 4.19.y, and applied to the wrong function, which obviously causes problems. A fixed version will be applied as a separate commit later. Reported-by: Boris Ostrovsky <boris.ostrovsky(a)oracle.com> Link: https://lore.kernel.org/r/ZSQeA8fhUT++iZvz@ostr-mac Cc: Ido Schimmel <idosch(a)nvidia.com> Cc: Jiri Pirko <jiri(a)nvidia.com> Cc: Jakub Kicinski <kuba(a)kernel.org> Cc: Paolo Abeni <pabeni(a)redhat.com> Cc: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Wang Yufen <wangyufen(a)huawei.com> --- net/core/rtnetlink.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 794db63..79f6251 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -2702,10 +2702,7 @@ static int rtnl_setlink(struct sk_buff *skb, struct nlmsghdr *nlh, ifm = nlmsg_data(nlh); if (ifm->ifi_index > 0) dev = __dev_get_by_index(net, ifm->ifi_index); - else if (ifm->ifi_index < 0) { - NL_SET_ERR_MSG(extack, "ifindex can't be negative"); - return -EINVAL; - } else if (tb[IFLA_IFNAME]) + else if (tb[IFLA_IFNAME]) dev = __dev_get_by_name(net, ifname); else goto errout; -- 1.8.3.1

2 3

[PATCH OLK-6.6 00/12] aarch64 KABI reserved for module
by Yuntao Liu 18 Jan '24

18 Jan '24

reserve space for arm64 related structures. Yuntao Liu (12): kabi: reserve space for struct drm_fb_helper.h:drm_fb_helper drm_fb_helper.h:drm_fb_helper_funcs kabi: reserve space for struct drm_file.h:drm_minor drm_file.h:drm_file kabi: reserve space for struct drm_fourcc.h:drm_format_info kabi: reserve space for struct drm_gem.h:vm_operations_struct drm_gem.h:drm_gem_object kabi: reserve space for struct efi.h:efi kabi: reserve space for struct extable.h:exception_table_entry kabi: reserve space for struct fb.h:fb_deferred_io fb.h:fb_info kabi: reserve space for struct irq_work.h:irq_work kabi: reserve space for struct irqdesc.h:irq_desc kabi: reserve space for enum irqdomain_defs.h:irq_domain_bus_token kabi: reserve space for struct irqdomain.h:irq_domain kabi: reserve space for struct processor.h:thread_struct arch/arm64/include/asm/extable.h | 2 ++ arch/arm64/include/asm/processor.h | 9 +++++++++ include/drm/drm_fb_helper.h | 7 +++++++ include/drm/drm_file.h | 7 +++++++ include/drm/drm_fourcc.h | 3 +++ include/drm/drm_gem.h | 13 +++++++++++++ include/linux/efi.h | 3 +++ include/linux/fb.h | 7 +++++++ include/linux/irq_work.h | 4 ++++ include/linux/irqdesc.h | 3 +++ include/linux/irqdomain.h | 9 +++++++++ include/linux/irqdomain_defs.h | 12 ++++++++++++ 12 files changed, 79 insertions(+) -- 2.34.1

2 13

[PATCH OLK-5.10 0/3] x86/quirks: Add parameter to clear MSIs early on
by Zheng Zengkai 18 Jan '24

18 Jan '24

A crash kernel boot failure issue is reported in openeuler 5.10 series kernel(in x86 virtual machine with Hi1822 virtual nic). Finally the issue was narrowed down to irq vector conflict. Backport following patchset from kernel maillist to fix this issue. https://lore.kernel.org/linux-pci/20181018183721.27467-3-gpiccoli@canonical… Guilherme G. Piccoli (3): x86/quirks: Scan all busses for early PCI quirks x86/PCI: Export find_cap() to be used in early PCI code x86/quirks: Add parameter to clear MSIs early on boot .../admin-guide/kernel-parameters.txt | 6 +++ arch/x86/include/asm/pci-direct.h | 2 + arch/x86/kernel/aperture_64.c | 30 +------------ arch/x86/kernel/early-quirks.c | 45 ++++++++++++++++--- arch/x86/pci/common.c | 4 ++ arch/x86/pci/early.c | 25 +++++++++++ 6 files changed, 77 insertions(+), 35 deletions(-) -- 2.20.1

2 4

[PATCH OLK-6.6 0/9] Backport etmem feature to OLK 6.6
by Yuchen Tang 18 Jan '24

18 Jan '24

v4 -> v5 - fix redundant kvm config check v3 -> v4 - backport only core etmem feature - modify kernel swap enabled toggle implementation v2 -> v3 - fix code style and commit messages v1 -> v2 - make independent patches to export symbols we need - make independent patches for etmem scan and swap config - fix indentation and code style problems Yuchen Tang (9): etmem: add ETMEM feature CONFIG to mm/Kconfig etmem: add ETMEM scan feature CONFIG to mm/Kconfig mm: Export symbol __pud_trans_huge_lock mm: Export symbol walk_page_range etmem: add etmem scan feature mm: Export symbol reclaim_pages etmem: add etmem swap feature etmem: add original kernel swap enabled options etmem: enable etmem configurations arch/arm64/configs/openeuler_defconfig | 3 + arch/x86/configs/openeuler_defconfig | 3 + fs/proc/Makefile | 3 + fs/proc/base.c | 8 + fs/proc/etmem_proc.c | 216 ++++ fs/proc/etmem_scan.c | 1383 ++++++++++++++++++++++++ fs/proc/etmem_scan.h | 149 +++ fs/proc/etmem_swap.c | 109 ++ fs/proc/internal.h | 4 + include/linux/etmem.h | 47 + include/linux/mm.h | 4 + include/linux/mm_types.h | 4 + include/linux/swap.h | 1 + include/uapi/asm-generic/mman-common.h | 4 + mm/Kconfig | 33 + mm/Makefile | 1 + mm/etmem.c | 95 ++ mm/huge_memory.c | 1 + mm/internal.h | 1 - mm/madvise.c | 16 + mm/pagewalk.c | 1 + mm/swap_state.c | 4 + mm/vmscan.c | 8 + virt/kvm/kvm_main.c | 8 + 24 files changed, 2105 insertions(+), 1 deletion(-) create mode 100644 fs/proc/etmem_proc.c create mode 100644 fs/proc/etmem_scan.c create mode 100644 fs/proc/etmem_scan.h create mode 100644 fs/proc/etmem_swap.c create mode 100644 include/linux/etmem.h create mode 100644 mm/etmem.c -- 2.33.0

2 10

[PATCH OLK-5.10] tipc: fix a potential deadlock on &tx->lock
by Zhengchao Shao 18 Jan '24

18 Jan '24

From: Chengfeng Ye <dg573847474(a)gmail.com> stable inclusion from stable-v5.10.198 commit 6a24d0661fa389c241d935da38e0f6a5ee8eb1ae category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I8WZL1 CVE: CVE-2024-0641 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 08e50cf071847323414df0835109b6f3560d44f5 ] It seems that tipc_crypto_key_revoke() could be be invoked by wokequeue tipc_crypto_work_rx() under process context and timer/rx callback under softirq context, thus the lock acquisition on &tx->lock seems better use spin_lock_bh() to prevent possible deadlock. This flaw was found by an experimental static analysis tool I am developing for irq-related deadlock. tipc_crypto_work_rx() <workqueue> --> tipc_crypto_key_distr() --> tipc_bcast_xmit() --> tipc_bcbase_xmit() --> tipc_bearer_bc_xmit() --> tipc_crypto_xmit() --> tipc_ehdr_build() --> tipc_crypto_key_revoke() --> spin_lock(&tx->lock) <timer interrupt> --> tipc_disc_timeout() --> tipc_bearer_xmit_skb() --> tipc_crypto_xmit() --> tipc_ehdr_build() --> tipc_crypto_key_revoke() --> spin_lock(&tx->lock) <deadlock here> Signed-off-by: Chengfeng Ye <dg573847474(a)gmail.com> Reviewed-by: Jacob Keller <jacob.e.keller(a)intel.com> Acked-by: Jon Maloy <jmaloy(a)redhat.com> Fixes: fc1b6d6de220 ("tipc: introduce TIPC encryption & authentication") Link: https://lore.kernel.org/r/20230927181414.59928-1-dg573847474@gmail.com Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Zhengchao Shao <shaozhengchao(a)huawei.com> --- net/tipc/crypto.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/tipc/crypto.c b/net/tipc/crypto.c index de63d6d41645..7d26d0cb207a 100644 --- a/net/tipc/crypto.c +++ b/net/tipc/crypto.c @@ -1445,14 +1445,14 @@ static int tipc_crypto_key_revoke(struct net *net, u8 tx_key) struct tipc_crypto *tx = tipc_net(net)->crypto_tx; struct tipc_key key; - spin_lock(&tx->lock); + spin_lock_bh(&tx->lock); key = tx->key; WARN_ON(!key.active || tx_key != key.active); /* Free the active key */ tipc_crypto_key_set_state(tx, key.passive, 0, key.pending); tipc_crypto_key_detach(tx->aead[key.active], &tx->lock); - spin_unlock(&tx->lock); + spin_unlock_bh(&tx->lock); pr_warn("%s: key is revoked\n", tx->name); return -EKEYREVOKED; -- 2.34.1

2 1

[PATCH OLK-6.6 v2] mm/oom_kill: fix NULL pointer dereference in memcg_print_bad_task()
by Jinjiang Tu 18 Jan '24

18 Jan '24

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8WVL2 CVE: NA -------------------------------- When memcg oom occurs, commit be8d95530886 ("memcg: support priority for oom") calls memcg_print_bad_task() to print oc->chosen when the chosen task comes from low-priority memcg. oc->chosen may be assigned with -1 in oom_evaluate_task(). Howerver, memcg_print_bad_task() does't check if oc->chosen is -1 and leads to NULL pointer dereference. To fix it, check if oc->chosen is -1 in memcg_print_bad_task(). Fixes: be8d95530886 ("memcg: support priority for oom") Signed-off-by: Jinjiang Tu <tujinjiang(a)huawei.com> --- mm/memcontrol.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 7ab0eec942ba..8327cef9d53a 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -4219,7 +4219,7 @@ void memcg_print_bad_task(struct oom_control *oc) if (memcg_oom_prio_disabled()) return; - if (oc->chosen) { + if (oc->chosen && oc->chosen != (void *)-1UL) { struct mem_cgroup *memcg; rcu_read_lock(); -- 2.25.1

2 1