Merge branch 'filter-next'

Alexei Starovoitov says:

====================
BPF testsuite and cleanup

This patchset adds BPF testsuite and improves readability of classic
to internal BPF converter.

The testsuite helped to find 'negative offset bug' in x64 JIT that was
fixed by commit fdfaf64e ("x86: bpf_jit: support negative offsets")

It can be very useful for classic and internal JIT compiler developers.
Also it serves as performance benchmark.

x86_64/i386 pass all tests with and without JIT. arm32 JIT is failing
negative offset tests which are unsupported.

Internal BPF tests are much larger than classic tests to cover different
combinations of registers. Negative tests check correctness of classic
BPF verifier which must reject them.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

include/linux/filter.h

@@ -79,6 +79,57 @@ enum {
 /* BPF program can access up to 512 bytes of stack space. */
 #define MAX_BPF_STACK	512
 
+/* bpf_add|sub|...: a += x, bpf_mov: a = x */
+#define BPF_ALU64_REG(op, a, x) \
+	((struct sock_filter_int) {BPF_ALU64|BPF_OP(op)|BPF_X, a, x, 0, 0})
+#define BPF_ALU32_REG(op, a, x) \
+	((struct sock_filter_int) {BPF_ALU|BPF_OP(op)|BPF_X, a, x, 0, 0})
+
+/* bpf_add|sub|...: a += imm, bpf_mov: a = imm */
+#define BPF_ALU64_IMM(op, a, imm) \
+	((struct sock_filter_int) {BPF_ALU64|BPF_OP(op)|BPF_K, a, 0, 0, imm})
+#define BPF_ALU32_IMM(op, a, imm) \
+	((struct sock_filter_int) {BPF_ALU|BPF_OP(op)|BPF_K, a, 0, 0, imm})
+
+/* R0 = *(uint *) (skb->data + off) */
+#define BPF_LD_ABS(size, off) \
+	((struct sock_filter_int) {BPF_LD|BPF_SIZE(size)|BPF_ABS, 0, 0, 0, off})
+
+/* R0 = *(uint *) (skb->data + x + off) */
+#define BPF_LD_IND(size, x, off) \
+	((struct sock_filter_int) {BPF_LD|BPF_SIZE(size)|BPF_IND, 0, x, 0, off})
+
+/* a = *(uint *) (x + off) */
+#define BPF_LDX_MEM(sz, a, x, off) \
+	((struct sock_filter_int) {BPF_LDX|BPF_SIZE(sz)|BPF_MEM, a, x, off, 0})
+
+/* if (a 'op' x) goto pc+off */
+#define BPF_JMP_REG(op, a, x, off) \
+	((struct sock_filter_int) {BPF_JMP|BPF_OP(op)|BPF_X, a, x, off, 0})
+
+/* if (a 'op' imm) goto pc+off */
+#define BPF_JMP_IMM(op, a, imm, off) \
+	((struct sock_filter_int) {BPF_JMP|BPF_OP(op)|BPF_K, a, 0, off, imm})
+
+#define BPF_EXIT_INSN() \
+	((struct sock_filter_int) {BPF_JMP|BPF_EXIT, 0, 0, 0, 0})
+
+static inline int size_to_bpf(int size)
+{
+	switch (size) {
+	case 1:
+		return BPF_B;
+	case 2:
+		return BPF_H;
+	case 4:
+		return BPF_W;
+	case 8:
+		return BPF_DW;
+	default:
+		return -EINVAL;
+	}
+}
+
 /* Macro to invoke filter function. */
 #define SK_RUN_FILTER(filter, ctx)  (*filter->bpf_func)(ctx, filter->insnsi)
 

lib/Kconfig.debug

@@ -1620,6 +1620,19 @@ config TEST_USER_COPY
 
 	  If unsure, say N.
 
+config TEST_BPF
+	tristate "Test BPF filter functionality"
+	default n
+	depends on m
+	help
+	  This builds the "test_bpf" module that runs various test vectors
+	  against the BPF interpreter or BPF JIT compiler depending on the
+	  current setting. This is in particular useful for BPF JIT compiler
+	  development, but also to run regression tests against changes in
+	  the interpreter code.
+
+	  If unsure, say N.
+
 source "samples/Kconfig"
 
 source "lib/Kconfig.kgdb"

lib/Makefile

@@ -33,6 +33,7 @@ obj-y += kstrtox.o
 obj-$(CONFIG_TEST_KSTRTOX) += test-kstrtox.o
 obj-$(CONFIG_TEST_MODULE) += test_module.o
 obj-$(CONFIG_TEST_USER_COPY) += test_user_copy.o
+obj-$(CONFIG_TEST_BPF) += test_bpf.o
 
 ifeq ($(CONFIG_DEBUG_KOBJECT),y)
 CFLAGS_kobject.o += -DDEBUG

net/core/filter.c

@@ -668,10 +668,9 @@ static bool convert_bpf_extensions(struct sock_filter *fp,
 	case SKF_AD_OFF + SKF_AD_PROTOCOL:
 		BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, protocol) != 2);
 
-		insn->code = BPF_LDX | BPF_MEM | BPF_H;
-		insn->a_reg = BPF_REG_A;
-		insn->x_reg = BPF_REG_CTX;
-		insn->off = offsetof(struct sk_buff, protocol);
+		/* A = *(u16 *) (ctx + offsetof(protocol)) */
+		*insn = BPF_LDX_MEM(BPF_H, BPF_REG_A, BPF_REG_CTX,
+				    offsetof(struct sk_buff, protocol));
 		insn++;
 
 		/* A = ntohs(A) [emitting a nop or swap16] */
@@ -681,37 +680,27 @@ static bool convert_bpf_extensions(struct sock_filter *fp,
 		break;
 
 	case SKF_AD_OFF + SKF_AD_PKTTYPE:
-		insn->code = BPF_LDX | BPF_MEM | BPF_B;
-		insn->a_reg = BPF_REG_A;
-		insn->x_reg = BPF_REG_CTX;
-		insn->off = pkt_type_offset();
+		*insn = BPF_LDX_MEM(BPF_B, BPF_REG_A, BPF_REG_CTX,
+				    pkt_type_offset());
 		if (insn->off < 0)
 			return false;
 		insn++;
 
-		insn->code = BPF_ALU | BPF_AND | BPF_K;
-		insn->a_reg = BPF_REG_A;
-		insn->imm = PKT_TYPE_MAX;
+		*insn = BPF_ALU32_IMM(BPF_AND, BPF_REG_A, PKT_TYPE_MAX);
 		break;
 
 	case SKF_AD_OFF + SKF_AD_IFINDEX:
 	case SKF_AD_OFF + SKF_AD_HATYPE:
-		if (FIELD_SIZEOF(struct sk_buff, dev) == 8)
-			insn->code = BPF_LDX | BPF_MEM | BPF_DW;
-		else
-			insn->code = BPF_LDX | BPF_MEM | BPF_W;
-		insn->a_reg = BPF_REG_TMP;
-		insn->x_reg = BPF_REG_CTX;
-		insn->off = offsetof(struct sk_buff, dev);
+		*insn = BPF_LDX_MEM(size_to_bpf(FIELD_SIZEOF(struct sk_buff, dev)),
+				    BPF_REG_TMP, BPF_REG_CTX,
+				    offsetof(struct sk_buff, dev));
 		insn++;
 
-		insn->code = BPF_JMP | BPF_JNE | BPF_K;
-		insn->a_reg = BPF_REG_TMP;
-		insn->imm = 0;
-		insn->off = 1;
+		/* if (tmp != 0) goto pc+1 */
+		*insn = BPF_JMP_IMM(BPF_JNE, BPF_REG_TMP, 0, 1);
 		insn++;
 
-		insn->code = BPF_JMP | BPF_EXIT;
+		*insn = BPF_EXIT_INSN();
 		insn++;
 
 		BUILD_BUG_ON(FIELD_SIZEOF(struct net_device, ifindex) != 4);
@@ -732,55 +721,45 @@ static bool convert_bpf_extensions(struct sock_filter *fp,
 	case SKF_AD_OFF + SKF_AD_MARK:
 		BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, mark) != 4);
 
-		insn->code = BPF_LDX | BPF_MEM | BPF_W;
-		insn->a_reg = BPF_REG_A;
-		insn->x_reg = BPF_REG_CTX;
-		insn->off = offsetof(struct sk_buff, mark);
+		*insn = BPF_LDX_MEM(BPF_W, BPF_REG_A, BPF_REG_CTX,
+				    offsetof(struct sk_buff, mark));
 		break;
 
 	case SKF_AD_OFF + SKF_AD_RXHASH:
 		BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, hash) != 4);
 
-		insn->code = BPF_LDX | BPF_MEM | BPF_W;
-		insn->a_reg = BPF_REG_A;
-		insn->x_reg = BPF_REG_CTX;
-		insn->off = offsetof(struct sk_buff, hash);
+		*insn = BPF_LDX_MEM(BPF_W, BPF_REG_A, BPF_REG_CTX,
+				    offsetof(struct sk_buff, hash));
 		break;
 
 	case SKF_AD_OFF + SKF_AD_QUEUE:
 		BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, queue_mapping) != 2);
 
-		insn->code = BPF_LDX | BPF_MEM | BPF_H;
-		insn->a_reg = BPF_REG_A;
-		insn->x_reg = BPF_REG_CTX;
-		insn->off = offsetof(struct sk_buff, queue_mapping);
+		*insn = BPF_LDX_MEM(BPF_H, BPF_REG_A, BPF_REG_CTX,
+				    offsetof(struct sk_buff, queue_mapping));
 		break;
 
 	case SKF_AD_OFF + SKF_AD_VLAN_TAG:
 	case SKF_AD_OFF + SKF_AD_VLAN_TAG_PRESENT:
 		BUILD_BUG_ON(FIELD_SIZEOF(struct sk_buff, vlan_tci) != 2);
 
-		insn->code = BPF_LDX | BPF_MEM | BPF_H;
-		insn->a_reg = BPF_REG_A;
-		insn->x_reg = BPF_REG_CTX;
-		insn->off = offsetof(struct sk_buff, vlan_tci);
+		/* A = *(u16 *) (ctx + offsetof(vlan_tci)) */
+		*insn = BPF_LDX_MEM(BPF_H, BPF_REG_A, BPF_REG_CTX,
+				    offsetof(struct sk_buff, vlan_tci));
 		insn++;
 
 		BUILD_BUG_ON(VLAN_TAG_PRESENT != 0x1000);
 
 		if (fp->k == SKF_AD_OFF + SKF_AD_VLAN_TAG) {
-			insn->code = BPF_ALU | BPF_AND | BPF_K;
-			insn->a_reg = BPF_REG_A;
-			insn->imm = ~VLAN_TAG_PRESENT;
+			*insn = BPF_ALU32_IMM(BPF_AND, BPF_REG_A,
+					      ~VLAN_TAG_PRESENT);
 		} else {
-			insn->code = BPF_ALU | BPF_RSH | BPF_K;
-			insn->a_reg = BPF_REG_A;
-			insn->imm = 12;
+			/* A >>= 12 */
+			*insn = BPF_ALU32_IMM(BPF_RSH, BPF_REG_A, 12);
 			insn++;
 
-			insn->code = BPF_ALU | BPF_AND | BPF_K;
-			insn->a_reg = BPF_REG_A;
-			insn->imm = 1;
+			/* A &= 1 */
+			*insn = BPF_ALU32_IMM(BPF_AND, BPF_REG_A, 1);
 		}
 		break;
 
@@ -790,21 +769,15 @@ static bool convert_bpf_extensions(struct sock_filter *fp,
 	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 = BPF_REG_ARG1;
-		insn->x_reg = BPF_REG_CTX;
+		*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_ARG1, BPF_REG_CTX);
 		insn++;
 
 		/* arg2 = A */
-		insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-		insn->a_reg = BPF_REG_ARG2;
-		insn->x_reg = BPF_REG_A;
+		*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_ARG2, BPF_REG_A);
 		insn++;
 
 		/* arg3 = X */
-		insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-		insn->a_reg = BPF_REG_ARG3;
-		insn->x_reg = BPF_REG_X;
+		*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_ARG3, BPF_REG_X);
 		insn++;
 
 		/* Emit call(ctx, arg2=A, arg3=X) */
@@ -829,9 +802,8 @@ static bool convert_bpf_extensions(struct sock_filter *fp,
 		break;
 
 	case SKF_AD_OFF + SKF_AD_ALU_XOR_X:
-		insn->code = BPF_ALU | BPF_XOR | BPF_X;
-		insn->a_reg = BPF_REG_A;
-		insn->x_reg = BPF_REG_X;
+		/* A ^= X */
+		*insn = BPF_ALU32_REG(BPF_XOR, BPF_REG_A, BPF_REG_X);
 		break;
 
 	default:
@@ -897,9 +869,7 @@ int sk_convert_filter(struct sock_filter *prog, int len,
 	fp = prog;
 
 	if (new_insn) {
-		new_insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-		new_insn->a_reg = BPF_REG_CTX;
-		new_insn->x_reg = BPF_REG_ARG1;
+		*new_insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_CTX, BPF_REG_ARG1);
 	}
 	new_insn++;
 
@@ -1027,34 +997,28 @@ int sk_convert_filter(struct sock_filter *prog, int len,
 
 		/* ldxb 4 * ([14] & 0xf) is remaped into 6 insns. */
 		case BPF_LDX | BPF_MSH | BPF_B:
-			insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-			insn->a_reg = BPF_REG_TMP;
-			insn->x_reg = BPF_REG_A;
+			/* tmp = A */
+			*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_TMP, BPF_REG_A);
 			insn++;
 
-			insn->code = BPF_LD | BPF_ABS | BPF_B;
-			insn->a_reg = BPF_REG_A;
-			insn->imm = fp->k;
+			/* A = R0 = *(u8 *) (skb->data + K) */
+			*insn = BPF_LD_ABS(BPF_B, fp->k);
 			insn++;
 
-			insn->code = BPF_ALU | BPF_AND | BPF_K;
-			insn->a_reg = BPF_REG_A;
-			insn->imm = 0xf;
+			/* A &= 0xf */
+			*insn = BPF_ALU32_IMM(BPF_AND, BPF_REG_A, 0xf);
 			insn++;
 
-			insn->code = BPF_ALU | BPF_LSH | BPF_K;
-			insn->a_reg = BPF_REG_A;
-			insn->imm = 2;
+			/* A <<= 2 */
+			*insn = BPF_ALU32_IMM(BPF_LSH, BPF_REG_A, 2);
 			insn++;
 
-			insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-			insn->a_reg = BPF_REG_X;
-			insn->x_reg = BPF_REG_A;
+			/* X = A */
+			*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_X, BPF_REG_A);
 			insn++;
 
-			insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-			insn->a_reg = BPF_REG_A;
-			insn->x_reg = BPF_REG_TMP;
+			/* A = tmp */
+			*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_A, BPF_REG_TMP);
 			break;
 
 		/* RET_K, RET_A are remaped into 2 insns. */
@@ -1068,7 +1032,7 @@ int sk_convert_filter(struct sock_filter *prog, int len,
 			insn->imm = fp->k;
 			insn++;
 
-			insn->code = BPF_JMP | BPF_EXIT;
+			*insn = BPF_EXIT_INSN();
 			break;
 
 		/* Store to stack. */
@@ -1102,16 +1066,12 @@ int sk_convert_filter(struct sock_filter *prog, int len,
 
 		/* X = A */
 		case BPF_MISC | BPF_TAX:
-			insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-			insn->a_reg = BPF_REG_X;
-			insn->x_reg = BPF_REG_A;
+			*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_X, BPF_REG_A);
 			break;
 
 		/* A = X */
 		case BPF_MISC | BPF_TXA:
-			insn->code = BPF_ALU64 | BPF_MOV | BPF_X;
-			insn->a_reg = BPF_REG_A;
-			insn->x_reg = BPF_REG_X;
+			*insn = BPF_ALU64_REG(BPF_MOV, BPF_REG_A, BPF_REG_X);
 			break;
 
 		/* A = skb->len or X = skb->len */
@@ -1126,10 +1086,8 @@ int sk_convert_filter(struct sock_filter *prog, int len,
 
 		/* access seccomp_data fields */
 		case BPF_LDX | BPF_ABS | BPF_W:
-			insn->code = BPF_LDX | BPF_MEM | BPF_W;
-			insn->a_reg = BPF_REG_A;
-			insn->x_reg = BPF_REG_CTX;
-			insn->off = fp->k;
+			/* A = *(u32 *) (ctx + K) */
+			*insn = BPF_LDX_MEM(BPF_W, BPF_REG_A, BPF_REG_CTX, fp->k);
 			break;
 
 		default:

tools/testing/selftests/net/Makefile

@@ -14,6 +14,12 @@ all: $(NET_PROGS)
 run_tests: all
 	@/bin/sh ./run_netsocktests || echo "sockettests: [FAIL]"
 	@/bin/sh ./run_afpackettests || echo "afpackettests: [FAIL]"
-
+	@if /sbin/modprobe test_bpf ; then \
+		/sbin/rmmod test_bpf; \
+		echo "test_bpf: ok"; \
+	else \
+		echo "test_bpf: [FAIL]"; \
+		exit 1; \
+	fi
 clean:
 	$(RM) $(NET_PROGS)