2011-04-20 17:27:32 +08:00
|
|
|
/* bpf_jit.S : BPF JIT helper functions
|
|
|
|
*
|
|
|
|
* Copyright (C) 2011 Eric Dumazet (eric.dumazet@gmail.com)
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public License
|
|
|
|
* as published by the Free Software Foundation; version 2
|
|
|
|
* of the License.
|
|
|
|
*/
|
|
|
|
#include <linux/linkage.h>
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Calling convention :
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
* rbx : skb pointer (callee saved)
|
2011-04-20 17:27:32 +08:00
|
|
|
* esi : offset of byte(s) to fetch in skb (can be scratched)
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
* r10 : copy of skb->data
|
2011-04-20 17:27:32 +08:00
|
|
|
* r9d : hlen = skb->len - skb->data_len
|
|
|
|
*/
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
#define SKBDATA %r10
|
2012-03-30 13:24:05 +08:00
|
|
|
#define SKF_MAX_NEG_OFF $(-0x200000) /* SKF_LL_OFF from filter.h */
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
#define MAX_BPF_STACK (512 /* from filter.h */ + \
|
|
|
|
32 /* space for rbx,r13,r14,r15 */ + \
|
|
|
|
8 /* space for skb_copy_bits */)
|
2011-04-20 17:27:32 +08:00
|
|
|
|
|
|
|
sk_load_word:
|
|
|
|
.globl sk_load_word
|
|
|
|
|
2012-03-30 13:24:05 +08:00
|
|
|
test %esi,%esi
|
|
|
|
js bpf_slow_path_word_neg
|
|
|
|
|
|
|
|
sk_load_word_positive_offset:
|
|
|
|
.globl sk_load_word_positive_offset
|
|
|
|
|
2011-04-20 17:27:32 +08:00
|
|
|
mov %r9d,%eax # hlen
|
|
|
|
sub %esi,%eax # hlen - offset
|
|
|
|
cmp $3,%eax
|
|
|
|
jle bpf_slow_path_word
|
|
|
|
mov (SKBDATA,%rsi),%eax
|
|
|
|
bswap %eax /* ntohl() */
|
|
|
|
ret
|
|
|
|
|
|
|
|
sk_load_half:
|
|
|
|
.globl sk_load_half
|
|
|
|
|
2012-03-30 13:24:05 +08:00
|
|
|
test %esi,%esi
|
|
|
|
js bpf_slow_path_half_neg
|
|
|
|
|
|
|
|
sk_load_half_positive_offset:
|
|
|
|
.globl sk_load_half_positive_offset
|
|
|
|
|
2011-04-20 17:27:32 +08:00
|
|
|
mov %r9d,%eax
|
|
|
|
sub %esi,%eax # hlen - offset
|
|
|
|
cmp $1,%eax
|
|
|
|
jle bpf_slow_path_half
|
|
|
|
movzwl (SKBDATA,%rsi),%eax
|
|
|
|
rol $8,%ax # ntohs()
|
|
|
|
ret
|
|
|
|
|
|
|
|
sk_load_byte:
|
|
|
|
.globl sk_load_byte
|
|
|
|
|
2012-03-30 13:24:05 +08:00
|
|
|
test %esi,%esi
|
|
|
|
js bpf_slow_path_byte_neg
|
|
|
|
|
|
|
|
sk_load_byte_positive_offset:
|
|
|
|
.globl sk_load_byte_positive_offset
|
|
|
|
|
2011-04-20 17:27:32 +08:00
|
|
|
cmp %esi,%r9d /* if (offset >= hlen) goto bpf_slow_path_byte */
|
|
|
|
jle bpf_slow_path_byte
|
|
|
|
movzbl (SKBDATA,%rsi),%eax
|
|
|
|
ret
|
|
|
|
|
|
|
|
/* rsi contains offset and can be scratched */
|
|
|
|
#define bpf_slow_path_common(LEN) \
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
mov %rbx, %rdi; /* arg1 == skb */ \
|
2011-04-20 17:27:32 +08:00
|
|
|
push %r9; \
|
|
|
|
push SKBDATA; \
|
|
|
|
/* rsi already has offset */ \
|
|
|
|
mov $LEN,%ecx; /* len */ \
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
lea - MAX_BPF_STACK + 32(%rbp),%rdx; \
|
2011-04-20 17:27:32 +08:00
|
|
|
call skb_copy_bits; \
|
|
|
|
test %eax,%eax; \
|
|
|
|
pop SKBDATA; \
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
pop %r9;
|
2011-04-20 17:27:32 +08:00
|
|
|
|
|
|
|
|
|
|
|
bpf_slow_path_word:
|
|
|
|
bpf_slow_path_common(4)
|
|
|
|
js bpf_error
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
mov - MAX_BPF_STACK + 32(%rbp),%eax
|
2011-04-20 17:27:32 +08:00
|
|
|
bswap %eax
|
|
|
|
ret
|
|
|
|
|
|
|
|
bpf_slow_path_half:
|
|
|
|
bpf_slow_path_common(2)
|
|
|
|
js bpf_error
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
mov - MAX_BPF_STACK + 32(%rbp),%ax
|
2011-04-20 17:27:32 +08:00
|
|
|
rol $8,%ax
|
|
|
|
movzwl %ax,%eax
|
|
|
|
ret
|
|
|
|
|
|
|
|
bpf_slow_path_byte:
|
|
|
|
bpf_slow_path_common(1)
|
|
|
|
js bpf_error
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
movzbl - MAX_BPF_STACK + 32(%rbp),%eax
|
2011-04-20 17:27:32 +08:00
|
|
|
ret
|
2012-03-30 13:24:05 +08:00
|
|
|
|
|
|
|
#define sk_negative_common(SIZE) \
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
mov %rbx, %rdi; /* arg1 == skb */ \
|
2012-03-30 13:24:05 +08:00
|
|
|
push %r9; \
|
|
|
|
push SKBDATA; \
|
|
|
|
/* rsi already has offset */ \
|
2014-03-11 06:56:51 +08:00
|
|
|
mov $SIZE,%edx; /* size */ \
|
2012-03-30 13:24:05 +08:00
|
|
|
call bpf_internal_load_pointer_neg_helper; \
|
|
|
|
test %rax,%rax; \
|
|
|
|
pop SKBDATA; \
|
|
|
|
pop %r9; \
|
|
|
|
jz bpf_error
|
|
|
|
|
|
|
|
bpf_slow_path_word_neg:
|
|
|
|
cmp SKF_MAX_NEG_OFF, %esi /* test range */
|
|
|
|
jl bpf_error /* offset lower -> error */
|
|
|
|
sk_load_word_negative_offset:
|
|
|
|
.globl sk_load_word_negative_offset
|
|
|
|
sk_negative_common(4)
|
|
|
|
mov (%rax), %eax
|
|
|
|
bswap %eax
|
|
|
|
ret
|
|
|
|
|
|
|
|
bpf_slow_path_half_neg:
|
|
|
|
cmp SKF_MAX_NEG_OFF, %esi
|
|
|
|
jl bpf_error
|
|
|
|
sk_load_half_negative_offset:
|
|
|
|
.globl sk_load_half_negative_offset
|
|
|
|
sk_negative_common(2)
|
|
|
|
mov (%rax),%ax
|
|
|
|
rol $8,%ax
|
|
|
|
movzwl %ax,%eax
|
|
|
|
ret
|
|
|
|
|
|
|
|
bpf_slow_path_byte_neg:
|
|
|
|
cmp SKF_MAX_NEG_OFF, %esi
|
|
|
|
jl bpf_error
|
|
|
|
sk_load_byte_negative_offset:
|
|
|
|
.globl sk_load_byte_negative_offset
|
|
|
|
sk_negative_common(1)
|
|
|
|
movzbl (%rax), %eax
|
|
|
|
ret
|
|
|
|
|
|
|
|
bpf_error:
|
|
|
|
# force a return 0 from jit handler
|
net: filter: x86: internal BPF JIT
Maps all internal BPF instructions into x86_64 instructions.
This patch replaces original BPF x64 JIT with internal BPF x64 JIT.
sysctl net.core.bpf_jit_enable is reused as on/off switch.
Performance:
1. old BPF JIT and internal BPF JIT generate equivalent x86_64 code.
No performance difference is observed for filters that were JIT-able before
Example assembler code for BPF filter "tcpdump port 22"
original BPF -> old JIT: original BPF -> internal BPF -> new JIT:
0: push %rbp 0: push %rbp
1: mov %rsp,%rbp 1: mov %rsp,%rbp
4: sub $0x60,%rsp 4: sub $0x228,%rsp
8: mov %rbx,-0x8(%rbp) b: mov %rbx,-0x228(%rbp) // prologue
12: mov %r13,-0x220(%rbp)
19: mov %r14,-0x218(%rbp)
20: mov %r15,-0x210(%rbp)
27: xor %eax,%eax // clear A
c: xor %ebx,%ebx 29: xor %r13,%r13 // clear X
e: mov 0x68(%rdi),%r9d 2c: mov 0x68(%rdi),%r9d
12: sub 0x6c(%rdi),%r9d 30: sub 0x6c(%rdi),%r9d
16: mov 0xd8(%rdi),%r8 34: mov 0xd8(%rdi),%r10
3b: mov %rdi,%rbx
1d: mov $0xc,%esi 3e: mov $0xc,%esi
22: callq 0xffffffffe1021e15 43: callq 0xffffffffe102bd75
27: cmp $0x86dd,%eax 48: cmp $0x86dd,%rax
2c: jne 0x0000000000000069 4f: jne 0x000000000000009a
2e: mov $0x14,%esi 51: mov $0x14,%esi
33: callq 0xffffffffe1021e31 56: callq 0xffffffffe102bd91
38: cmp $0x84,%eax 5b: cmp $0x84,%rax
3d: je 0x0000000000000049 62: je 0x0000000000000074
3f: cmp $0x6,%eax 64: cmp $0x6,%rax
42: je 0x0000000000000049 68: je 0x0000000000000074
44: cmp $0x11,%eax 6a: cmp $0x11,%rax
47: jne 0x00000000000000c6 6e: jne 0x0000000000000117
49: mov $0x36,%esi 74: mov $0x36,%esi
4e: callq 0xffffffffe1021e15 79: callq 0xffffffffe102bd75
53: cmp $0x16,%eax 7e: cmp $0x16,%rax
56: je 0x00000000000000bf 82: je 0x0000000000000110
58: mov $0x38,%esi 88: mov $0x38,%esi
5d: callq 0xffffffffe1021e15 8d: callq 0xffffffffe102bd75
62: cmp $0x16,%eax 92: cmp $0x16,%rax
65: je 0x00000000000000bf 96: je 0x0000000000000110
67: jmp 0x00000000000000c6 98: jmp 0x0000000000000117
69: cmp $0x800,%eax 9a: cmp $0x800,%rax
6e: jne 0x00000000000000c6 a1: jne 0x0000000000000117
70: mov $0x17,%esi a3: mov $0x17,%esi
75: callq 0xffffffffe1021e31 a8: callq 0xffffffffe102bd91
7a: cmp $0x84,%eax ad: cmp $0x84,%rax
7f: je 0x000000000000008b b4: je 0x00000000000000c2
81: cmp $0x6,%eax b6: cmp $0x6,%rax
84: je 0x000000000000008b ba: je 0x00000000000000c2
86: cmp $0x11,%eax bc: cmp $0x11,%rax
89: jne 0x00000000000000c6 c0: jne 0x0000000000000117
8b: mov $0x14,%esi c2: mov $0x14,%esi
90: callq 0xffffffffe1021e15 c7: callq 0xffffffffe102bd75
95: test $0x1fff,%ax cc: test $0x1fff,%rax
99: jne 0x00000000000000c6 d3: jne 0x0000000000000117
d5: mov %rax,%r14
9b: mov $0xe,%esi d8: mov $0xe,%esi
a0: callq 0xffffffffe1021e44 dd: callq 0xffffffffe102bd91 // MSH
e2: and $0xf,%eax
e5: shl $0x2,%eax
e8: mov %rax,%r13
eb: mov %r14,%rax
ee: mov %r13,%rsi
a5: lea 0xe(%rbx),%esi f1: add $0xe,%esi
a8: callq 0xffffffffe1021e0d f4: callq 0xffffffffe102bd6d
ad: cmp $0x16,%eax f9: cmp $0x16,%rax
b0: je 0x00000000000000bf fd: je 0x0000000000000110
ff: mov %r13,%rsi
b2: lea 0x10(%rbx),%esi 102: add $0x10,%esi
b5: callq 0xffffffffe1021e0d 105: callq 0xffffffffe102bd6d
ba: cmp $0x16,%eax 10a: cmp $0x16,%rax
bd: jne 0x00000000000000c6 10e: jne 0x0000000000000117
bf: mov $0xffff,%eax 110: mov $0xffff,%eax
c4: jmp 0x00000000000000c8 115: jmp 0x000000000000011c
c6: xor %eax,%eax 117: mov $0x0,%eax
c8: mov -0x8(%rbp),%rbx 11c: mov -0x228(%rbp),%rbx // epilogue
cc: leaveq 123: mov -0x220(%rbp),%r13
cd: retq 12a: mov -0x218(%rbp),%r14
131: mov -0x210(%rbp),%r15
138: leaveq
139: retq
On fully cached SKBs both JITed functions take 12 nsec to execute.
BPF interpreter executes the program in 30 nsec.
The difference in generated assembler is due to the following:
Old BPF imlements LDX_MSH instruction via sk_load_byte_msh() helper function
inside bpf_jit.S.
New JIT removes the helper and does it explicitly, so ldx_msh cost
is the same for both JITs, but generated code looks longer.
New JIT has 4 registers to save, so prologue/epilogue are larger,
but the cost is within noise on x64.
Old JIT checks whether first insn clears A and if not emits 'xor %eax,%eax'.
New JIT clears %rax unconditionally.
2. old BPF JIT doesn't support ANC_NLATTR, ANC_PAY_OFFSET, ANC_RANDOM
extensions. New JIT supports all BPF extensions.
Performance of such filters improves 2-4 times depending on a filter.
The longer the filter the higher performance gain.
Synthetic benchmarks with many ancillary loads see 20x speedup
which seems to be the maximum gain from JIT
Notes:
. net.core.bpf_jit_enable=2 + tools/net/bpf_jit_disasm is still functional
and can be used to see generated assembler
. there are two jit_compile() functions and code flow for classic filters is:
sk_attach_filter() - load classic BPF
bpf_jit_compile() - try to JIT from classic BPF
sk_convert_filter() - convert classic to internal
bpf_int_jit_compile() - JIT from internal BPF
seccomp and tracing filters will just call bpf_int_jit_compile()
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-05-14 10:50:46 +08:00
|
|
|
xor %eax,%eax
|
|
|
|
mov - MAX_BPF_STACK(%rbp),%rbx
|
|
|
|
mov - MAX_BPF_STACK + 8(%rbp),%r13
|
|
|
|
mov - MAX_BPF_STACK + 16(%rbp),%r14
|
|
|
|
mov - MAX_BPF_STACK + 24(%rbp),%r15
|
2012-03-30 13:24:05 +08:00
|
|
|
leaveq
|
|
|
|
ret
|