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>

arch/x86/net/bpf_jit.S

@@ -12,13 +12,16 @@
 
 /*
  * Calling convention :
- * rdi : skb pointer
+ * rbx : skb pointer (callee saved)
  * esi : offset of byte(s) to fetch in skb (can be scratched)
- * r8  : copy of skb->data
+ * r10 : copy of skb->data
  * r9d : hlen = skb->len - skb->data_len
  */
-#define SKBDATA	%r8
+#define SKBDATA	%r10
 #define SKF_MAX_NEG_OFF    $(-0x200000) /* SKF_LL_OFF from filter.h */
+#define MAX_BPF_STACK (512 /* from filter.h */ + \
+	32 /* space for rbx,r13,r14,r15 */ + \
+	8 /* space for skb_copy_bits */)
 
 sk_load_word:
 	.globl	sk_load_word
@@ -68,53 +71,31 @@ sk_load_byte_positive_offset:
 	movzbl	(SKBDATA,%rsi),%eax
 	ret
 
-/**
- * sk_load_byte_msh - BPF_S_LDX_B_MSH helper
- *
- * Implements BPF_S_LDX_B_MSH : ldxb  4*([offset]&0xf)
- * Must preserve A accumulator (%eax)
- * Inputs : %esi is the offset value
- */
-sk_load_byte_msh:
-	.globl	sk_load_byte_msh
-	test	%esi,%esi
-	js	bpf_slow_path_byte_msh_neg
-
-sk_load_byte_msh_positive_offset:
-	.globl	sk_load_byte_msh_positive_offset
-	cmp	%esi,%r9d      /* if (offset >= hlen) goto bpf_slow_path_byte_msh */
-	jle	bpf_slow_path_byte_msh
-	movzbl	(SKBDATA,%rsi),%ebx
-	and	$15,%bl
-	shl	$2,%bl
-	ret
-
 /* rsi contains offset and can be scratched */
 #define bpf_slow_path_common(LEN)		\
-	push	%rdi;    /* save skb */		\
+	mov	%rbx, %rdi; /* arg1 == skb */	\
 	push	%r9;				\
 	push	SKBDATA;			\
 /* rsi already has offset */			\
 	mov	$LEN,%ecx;	/* len */	\
-	lea	-12(%rbp),%rdx;			\
+	lea	- MAX_BPF_STACK + 32(%rbp),%rdx;			\
 	call	skb_copy_bits;			\
 	test    %eax,%eax;			\
 	pop	SKBDATA;			\
-	pop	%r9;				\
-	pop	%rdi
+	pop	%r9;
 
 
 bpf_slow_path_word:
 	bpf_slow_path_common(4)
 	js	bpf_error
-	mov	-12(%rbp),%eax
+	mov	- MAX_BPF_STACK + 32(%rbp),%eax
 	bswap	%eax
 	ret
 
 bpf_slow_path_half:
 	bpf_slow_path_common(2)
 	js	bpf_error
-	mov	-12(%rbp),%ax
+	mov	- MAX_BPF_STACK + 32(%rbp),%ax
 	rol	$8,%ax
 	movzwl	%ax,%eax
 	ret
@@ -122,21 +103,11 @@ bpf_slow_path_half:
 bpf_slow_path_byte:
 	bpf_slow_path_common(1)
 	js	bpf_error
-	movzbl	-12(%rbp),%eax
-	ret
-
-bpf_slow_path_byte_msh:
-	xchg	%eax,%ebx /* dont lose A , X is about to be scratched */
-	bpf_slow_path_common(1)
-	js	bpf_error
-	movzbl	-12(%rbp),%eax
-	and	$15,%al
-	shl	$2,%al
-	xchg	%eax,%ebx
+	movzbl	- MAX_BPF_STACK + 32(%rbp),%eax
 	ret
 
 #define sk_negative_common(SIZE)				\
-	push	%rdi;	/* save skb */				\
+	mov	%rbx, %rdi; /* arg1 == skb */			\
 	push	%r9;						\
 	push	SKBDATA;					\
 /* rsi already has offset */					\
@@ -145,10 +116,8 @@ bpf_slow_path_byte_msh:
 	test	%rax,%rax;					\
 	pop	SKBDATA;					\
 	pop	%r9;						\
-	pop	%rdi;						\
 	jz	bpf_error
 
-
 bpf_slow_path_word_neg:
 	cmp	SKF_MAX_NEG_OFF, %esi	/* test range */
 	jl	bpf_error	/* offset lower -> error  */
@@ -179,22 +148,12 @@ sk_load_byte_negative_offset:
 	movzbl	(%rax), %eax
 	ret
 
-bpf_slow_path_byte_msh_neg:
-	cmp	SKF_MAX_NEG_OFF, %esi
-	jl	bpf_error
-sk_load_byte_msh_negative_offset:
-	.globl	sk_load_byte_msh_negative_offset
-	xchg	%eax,%ebx /* dont lose A , X is about to be scratched */
-	sk_negative_common(1)
-	movzbl	(%rax),%eax
-	and	$15,%al
-	shl	$2,%al
-	xchg	%eax,%ebx
-	ret
-
 bpf_error:
 # force a return 0 from jit handler
 	xor	%eax,%eax
-	mov		-8(%rbp),%rbx
+	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
 	leaveq
 	ret

include/linux/filter.h

@@ -207,6 +207,9 @@ void sk_decode_filter(struct sock_filter *filt, struct sock_filter *to);
 void sk_filter_charge(struct sock *sk, struct sk_filter *fp);
 void sk_filter_uncharge(struct sock *sk, struct sk_filter *fp);
 
+u64 __bpf_call_base(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5);
+void bpf_int_jit_compile(struct sk_filter *fp);
+
 #ifdef CONFIG_BPF_JIT
 #include <stdarg.h>
 #include <linux/linkage.h>

net/core/filter.c

@@ -1524,6 +1524,10 @@ static struct sk_filter *__sk_migrate_filter(struct sk_filter *fp,
 	return ERR_PTR(err);
 }
 
+void __weak bpf_int_jit_compile(struct sk_filter *prog)
+{
+}
+
 static struct sk_filter *__sk_prepare_filter(struct sk_filter *fp,
 					     struct sock *sk)
 {
@@ -1544,9 +1548,12 @@ static struct sk_filter *__sk_prepare_filter(struct sk_filter *fp,
 	/* JIT compiler couldn't process this filter, so do the
 	 * internal BPF translation for the optimized interpreter.
 	 */
-	if (!fp->jited)
+	if (!fp->jited) {
 		fp = __sk_migrate_filter(fp, sk);
 
+		/* Probe if internal BPF can be jit-ed */
+		bpf_int_jit_compile(fp);
+	}
 	return fp;
 }