filter: added BPF random opcode

Added a new ancillary load (bpf call in eBPF parlance) that produces
a 32-bit random number. We are implementing it as an ancillary load
(instead of an ISA opcode) because (a) it is simpler, (b) allows easy
JITing, and (c) seems more in line with generic ISAs that do not have
"get a random number" as a instruction, but as an OS call.

The main use for this ancillary load is to perform random packet sampling.

Signed-off-by: Chema Gonzalez <chema@google.com>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Acked-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 39 insertions, 2 deletions

diff --git a/Documentation/networking/filter.txt b/Documentation/networking/filter.txt
index 81f940f4e884..82e1cb0b3da8 100644
--- a/Documentation/networking/filter.txt
+++ b/Documentation/networking/filter.txt
@@ -281,6 +281,7 @@ Possible BPF extensions are shown in the following table:
   cpu                                   raw_smp_processor_id()
   vlan_tci                              vlan_tx_tag_get(skb)
   vlan_pr                               vlan_tx_tag_present(skb)
+  rand                                  prandom_u32()
 
 These extensions can also be prefixed with '#'.
 Examples for low-level BPF:
@@ -308,6 +309,18 @@ Examples for low-level BPF:
   ret #-1
   drop: ret #0
 
+** icmp random packet sampling, 1 in 4
+  ldh [12]
+  jne #0x800, drop
+  ldb [23]
+  jneq #1, drop
+  # get a random uint32 number
+  ld rand
+  mod #4
+  jneq #1, drop
+  ret #-1
+  drop: ret #0
+
 ** SECCOMP filter example:
 
   ld [4]                  /* offsetof(struct seccomp_data, arch) */
diff --git a/include/linux/filter.h b/include/linux/filter.h
index 024fd03e5d18..759abf78dd61 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -223,6 +223,7 @@ enum {
 	BPF_S_ANC_VLAN_TAG,
 	BPF_S_ANC_VLAN_TAG_PRESENT,
 	BPF_S_ANC_PAY_OFFSET,
+	BPF_S_ANC_RANDOM,
 };
 
 #endif /* __LINUX_FILTER_H__ */
diff --git a/include/uapi/linux/filter.h b/include/uapi/linux/filter.h
index 8eb9ccaa5b48..253b4d42cf2b 100644
--- a/include/uapi/linux/filter.h
+++ b/include/uapi/linux/filter.h
@@ -130,7 +130,8 @@ struct sock_fprog {	/* Required for SO_ATTACH_FILTER. */
 #define SKF_AD_VLAN_TAG	44
 #define SKF_AD_VLAN_TAG_PRESENT 48
 #define SKF_AD_PAY_OFFSET	52
-#define SKF_AD_MAX	56
+#define SKF_AD_RANDOM	56
+#define SKF_AD_MAX	60
 #define SKF_NET_OFF   (-0x100000)
 #define SKF_LL_OFF    (-0x200000)
 
diff --git a/net/core/filter.c b/net/core/filter.c
index cd58614660cf..78a636e60a0b 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -643,6 +643,12 @@ static u64 __get_raw_cpu_id(u64 ctx, u64 A, u64 X, u64 r4, u64 r5)
 	return raw_smp_processor_id();
 }
 
+/* note that this only generates 32-bit random numbers */
+static u64 __get_random_u32(u64 ctx, u64 A, u64 X, u64 r4, u64 r5)
+{
+	return (u64)prandom_u32();
+}
+
 /* Register mappings for user programs. */
 #define A_REG		0
 #define X_REG		7
@@ -779,6 +785,7 @@ static bool convert_bpf_extensions(struct sock_filter *fp,
 	case SKF_AD_OFF + SKF_AD_NLATTR:
 	case SKF_AD_OFF + SKF_AD_NLATTR_NEST:
 	case SKF_AD_OFF + SKF_AD_CPU:
+	case SKF_AD_OFF + SKF_AD_RANDOM:
 		/* arg1 = ctx */
 		insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
 		insn->a_reg = ARG1_REG;
@@ -812,6 +819,9 @@ static bool convert_bpf_extensions(struct sock_filter *fp,
 		case SKF_AD_OFF + SKF_AD_CPU:
 			insn->imm = __get_raw_cpu_id - __bpf_call_base;
 			break;
+		case SKF_AD_OFF + SKF_AD_RANDOM:
+			insn->imm = __get_random_u32 - __bpf_call_base;
+			break;
 		}
 		break;
 
@@ -1362,6 +1372,7 @@ int sk_chk_filter(struct sock_filter *filter, unsigned int flen)
 			ANCILLARY(VLAN_TAG);
 			ANCILLARY(VLAN_TAG_PRESENT);
 			ANCILLARY(PAY_OFFSET);
+			ANCILLARY(RANDOM);
 			}
 
 			/* ancillary operation unknown or unsupported */
@@ -1746,6 +1757,7 @@ void sk_decode_filter(struct sock_filter *filt, struct sock_filter *to)
 		[BPF_S_ANC_VLAN_TAG]	= BPF_LD|BPF_B|BPF_ABS,
 		[BPF_S_ANC_VLAN_TAG_PRESENT] = BPF_LD|BPF_B|BPF_ABS,
 		[BPF_S_ANC_PAY_OFFSET]	= BPF_LD|BPF_B|BPF_ABS,
+		[BPF_S_ANC_RANDOM]	= BPF_LD|BPF_B|BPF_ABS,
 		[BPF_S_LD_W_LEN]	= BPF_LD|BPF_W|BPF_LEN,
 		[BPF_S_LD_W_IND]	= BPF_LD|BPF_W|BPF_IND,
 		[BPF_S_LD_H_IND]	= BPF_LD|BPF_H|BPF_IND,
diff --git a/tools/net/bpf_exp.l b/tools/net/bpf_exp.l
index bf7be77ddd62..833a96611da6 100644
--- a/tools/net/bpf_exp.l
+++ b/tools/net/bpf_exp.l
@@ -92,6 +92,7 @@ extern void yyerror(const char *str);
 "#"?("cpu")	{ return K_CPU; }
 "#"?("vlan_tci") { return K_VLANT; }
 "#"?("vlan_pr")	{ return K_VLANP; }
+"#"?("rand")	{ return K_RAND; }
 
 ":"		{ return ':'; }
 ","		{ return ','; }
diff --git a/tools/net/bpf_exp.y b/tools/net/bpf_exp.y
index d15efc989ef5..e6306c51c26f 100644
--- a/tools/net/bpf_exp.y
+++ b/tools/net/bpf_exp.y
@@ -56,7 +56,7 @@ static void bpf_set_jmp_label(char *label, enum jmp_type type);
 %token OP_LDXI
 
 %token K_PKT_LEN K_PROTO K_TYPE K_NLATTR K_NLATTR_NEST K_MARK K_QUEUE K_HATYPE
-%token K_RXHASH K_CPU K_IFIDX K_VLANT K_VLANP K_POFF
+%token K_RXHASH K_CPU K_IFIDX K_VLANT K_VLANP K_POFF K_RAND
 
 %token ':' ',' '[' ']' '(' ')' 'x' 'a' '+' 'M' '*' '&' '#' '%'
 
@@ -164,6 +164,9 @@ ldb
 	| OP_LDB K_POFF {
 		bpf_set_curr_instr(BPF_LD | BPF_B | BPF_ABS, 0, 0,
 				   SKF_AD_OFF + SKF_AD_PAY_OFFSET); }
+	| OP_LDB K_RAND {
+		bpf_set_curr_instr(BPF_LD | BPF_B | BPF_ABS, 0, 0,
+				   SKF_AD_OFF + SKF_AD_RANDOM); }
 	;
 
 ldh
@@ -212,6 +215,9 @@ ldh
 	| OP_LDH K_POFF {
 		bpf_set_curr_instr(BPF_LD | BPF_H | BPF_ABS, 0, 0,
 				   SKF_AD_OFF + SKF_AD_PAY_OFFSET); }
+	| OP_LDH K_RAND {
+		bpf_set_curr_instr(BPF_LD | BPF_H | BPF_ABS, 0, 0,
+				   SKF_AD_OFF + SKF_AD_RANDOM); }
 	;
 
 ldi
@@ -265,6 +271,9 @@ ld
 	| OP_LD K_POFF {
 		bpf_set_curr_instr(BPF_LD | BPF_W | BPF_ABS, 0, 0,
 				   SKF_AD_OFF + SKF_AD_PAY_OFFSET); }
+	| OP_LD K_RAND {
+		bpf_set_curr_instr(BPF_LD | BPF_W | BPF_ABS, 0, 0,
+				   SKF_AD_OFF + SKF_AD_RANDOM); }
 	| OP_LD 'M' '[' number ']' {
 		bpf_set_curr_instr(BPF_LD | BPF_MEM, 0, 0, $4); }
 	| OP_LD '[' 'x' '+' number ']' {