Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2017-12-18

The following pull-request contains BPF updates for your *net-next* tree.

The main changes are:

1) Allow arbitrary function calls from one BPF function to another BPF function.
   As of today when writing BPF programs, __always_inline had to be used in
   the BPF C programs for all functions, unnecessarily causing LLVM to inflate
   code size. Handle this more naturally with support for BPF to BPF calls
   such that this __always_inline restriction can be overcome. As a result,
   it allows for better optimized code and finally enables to introduce core
   BPF libraries in the future that can be reused out of different projects.
   x86 and arm64 JIT support was added as well, from Alexei.

2) Add infrastructure for tagging functions as error injectable and allow for
   BPF to return arbitrary error values when BPF is attached via kprobes on
   those. This way of injecting errors generically eases testing and debugging
   without having to recompile or restart the kernel. Tags for opting-in for
   this facility are added with BPF_ALLOW_ERROR_INJECTION(), from Josef.

3) For BPF offload via nfp JIT, add support for bpf_xdp_adjust_head() helper
   call for XDP programs. First part of this work adds handling of BPF
   capabilities included in the firmware, and the later patches add support
   to the nfp verifier part and JIT as well as some small optimizations,
   from Jakub.

4) The bpftool now also gets support for basic cgroup BPF operations such
   as attaching, detaching and listing current BPF programs. As a requirement
   for the attach part, bpftool can now also load object files through
   'bpftool prog load'. This reuses libbpf which we have in the kernel tree
   as well. bpftool-cgroup man page is added along with it, from Roman.

5) Back then commit e87c6bc385 ("bpf: permit multiple bpf attachments for
   a single perf event") added support for attaching multiple BPF programs
   to a single perf event. Given they are configured through perf's ioctl()
   interface, the interface has been extended with a PERF_EVENT_IOC_QUERY_BPF
   command in this work in order to return an array of one or multiple BPF
   prog ids that are currently attached, from Yonghong.

6) Various minor fixes and cleanups to the bpftool's Makefile as well
   as a new 'uninstall' and 'doc-uninstall' target for removing bpftool
   itself or prior installed documentation related to it, from Quentin.

7) Add CONFIG_CGROUP_BPF=y to the BPF kernel selftest config file which is
   required for the test_dev_cgroup test case to run, from Naresh.

8) Fix reporting of XDP prog_flags for nfp driver, from Jakub.

9) Fix libbpf's exit code from the Makefile when libelf was not found in
   the system, also from Jakub.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

arch/Kconfig

@@ -196,6 +196,9 @@ config HAVE_OPTPROBES
 config HAVE_KPROBES_ON_FTRACE
 	bool
 
+config HAVE_KPROBE_OVERRIDE
+	bool
+
 config HAVE_NMI
 	bool
 

arch/arm/net/bpf_jit_32.c

@@ -1824,7 +1824,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 	/* If BPF JIT was not enabled then we must fall back to
 	 * the interpreter.
 	 */
-	if (!bpf_jit_enable)
+	if (!prog->jit_requested)
 		return orig_prog;
 
 	/* If constant blinding was enabled and we failed during blinding

arch/arm64/net/bpf_jit_comp.c

@@ -99,6 +99,20 @@ static inline void emit_a64_mov_i64(const int reg, const u64 val,
 	}
 }
 
+static inline void emit_addr_mov_i64(const int reg, const u64 val,
+				     struct jit_ctx *ctx)
+{
+	u64 tmp = val;
+	int shift = 0;
+
+	emit(A64_MOVZ(1, reg, tmp & 0xffff, shift), ctx);
+	for (;shift < 48;) {
+		tmp >>= 16;
+		shift += 16;
+		emit(A64_MOVK(1, reg, tmp & 0xffff, shift), ctx);
+	}
+}
+
 static inline void emit_a64_mov_i(const int is64, const int reg,
 				  const s32 val, struct jit_ctx *ctx)
 {
@@ -603,7 +617,10 @@ emit_cond_jmp:
 		const u8 r0 = bpf2a64[BPF_REG_0];
 		const u64 func = (u64)__bpf_call_base + imm;
 
-		emit_a64_mov_i64(tmp, func, ctx);
+		if (ctx->prog->is_func)
+			emit_addr_mov_i64(tmp, func, ctx);
+		else
+			emit_a64_mov_i64(tmp, func, ctx);
 		emit(A64_BLR(tmp), ctx);
 		emit(A64_MOV(1, r0, A64_R(0)), ctx);
 		break;
@@ -835,16 +852,24 @@ static inline void bpf_flush_icache(void *start, void *end)
 	flush_icache_range((unsigned long)start, (unsigned long)end);
 }
 
+struct arm64_jit_data {
+	struct bpf_binary_header *header;
+	u8 *image;
+	struct jit_ctx ctx;
+};
+
 struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 {
 	struct bpf_prog *tmp, *orig_prog = prog;
 	struct bpf_binary_header *header;
+	struct arm64_jit_data *jit_data;
 	bool tmp_blinded = false;
+	bool extra_pass = false;
 	struct jit_ctx ctx;
 	int image_size;
 	u8 *image_ptr;
 
-	if (!bpf_jit_enable)
+	if (!prog->jit_requested)
 		return orig_prog;
 
 	tmp = bpf_jit_blind_constants(prog);
@@ -858,13 +883,29 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 		prog = tmp;
 	}
 
+	jit_data = prog->aux->jit_data;
+	if (!jit_data) {
+		jit_data = kzalloc(sizeof(*jit_data), GFP_KERNEL);
+		if (!jit_data) {
+			prog = orig_prog;
+			goto out;
+		}
+		prog->aux->jit_data = jit_data;
+	}
+	if (jit_data->ctx.offset) {
+		ctx = jit_data->ctx;
+		image_ptr = jit_data->image;
+		header = jit_data->header;
+		extra_pass = true;
+		goto skip_init_ctx;
+	}
 	memset(&ctx, 0, sizeof(ctx));
 	ctx.prog = prog;
 
 	ctx.offset = kcalloc(prog->len, sizeof(int), GFP_KERNEL);
 	if (ctx.offset == NULL) {
 		prog = orig_prog;
-		goto out;
+		goto out_off;
 	}
 
 	/* 1. Initial fake pass to compute ctx->idx. */
@@ -895,6 +936,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 	/* 2. Now, the actual pass. */
 
 	ctx.image = (__le32 *)image_ptr;
+skip_init_ctx:
 	ctx.idx = 0;
 
 	build_prologue(&ctx);
@@ -920,13 +962,31 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 
 	bpf_flush_icache(header, ctx.image + ctx.idx);
 
-	bpf_jit_binary_lock_ro(header);
+	if (!prog->is_func || extra_pass) {
+		if (extra_pass && ctx.idx != jit_data->ctx.idx) {
+			pr_err_once("multi-func JIT bug %d != %d\n",
+				    ctx.idx, jit_data->ctx.idx);
+			bpf_jit_binary_free(header);
+			prog->bpf_func = NULL;
+			prog->jited = 0;
+			goto out_off;
+		}
+		bpf_jit_binary_lock_ro(header);
+	} else {
+		jit_data->ctx = ctx;
+		jit_data->image = image_ptr;
+		jit_data->header = header;
+	}
 	prog->bpf_func = (void *)ctx.image;
 	prog->jited = 1;
 	prog->jited_len = image_size;
 
+	if (!prog->is_func || extra_pass) {
 out_off:
-	kfree(ctx.offset);
+		kfree(ctx.offset);
+		kfree(jit_data);
+		prog->aux->jit_data = NULL;
+	}
 out:
 	if (tmp_blinded)
 		bpf_jit_prog_release_other(prog, prog == orig_prog ?

arch/mips/net/ebpf_jit.c

@@ -1869,7 +1869,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 	unsigned int image_size;
 	u8 *image_ptr;
 
-	if (!bpf_jit_enable || !cpu_has_mips64r2)
+	if (!prog->jit_requested || !cpu_has_mips64r2)
 		return prog;
 
 	tmp = bpf_jit_blind_constants(prog);

arch/powerpc/net/bpf_jit_comp64.c

@@ -993,7 +993,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *fp)
 	struct bpf_prog *tmp_fp;
 	bool bpf_blinded = false;
 
-	if (!bpf_jit_enable)
+	if (!fp->jit_requested)
 		return org_fp;
 
 	tmp_fp = bpf_jit_blind_constants(org_fp);

arch/s390/net/bpf_jit_comp.c

@@ -1300,7 +1300,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *fp)
 	struct bpf_jit jit;
 	int pass;
 
-	if (!bpf_jit_enable)
+	if (!fp->jit_requested)
 		return orig_fp;
 
 	tmp = bpf_jit_blind_constants(fp);

arch/sparc/net/bpf_jit_comp_64.c

@@ -1517,7 +1517,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 	u8 *image_ptr;
 	int pass;
 
-	if (!bpf_jit_enable)
+	if (!prog->jit_requested)
 		return orig_prog;
 
 	tmp = bpf_jit_blind_constants(prog);

arch/x86/Kconfig

@@ -154,6 +154,7 @@ config X86
 	select HAVE_KERNEL_XZ
 	select HAVE_KPROBES
 	select HAVE_KPROBES_ON_FTRACE
+	select HAVE_KPROBE_OVERRIDE
 	select HAVE_KRETPROBES
 	select HAVE_KVM
 	select HAVE_LIVEPATCH			if X86_64

arch/x86/include/asm/kprobes.h

@@ -67,6 +67,10 @@ extern const int kretprobe_blacklist_size;
 void arch_remove_kprobe(struct kprobe *p);
 asmlinkage void kretprobe_trampoline(void);
 
+#ifdef CONFIG_KPROBES_ON_FTRACE
+extern void arch_ftrace_kprobe_override_function(struct pt_regs *regs);
+#endif
+
 /* Architecture specific copy of original instruction*/
 struct arch_specific_insn {
 	/* copy of the original instruction */

arch/x86/include/asm/ptrace.h

@@ -109,6 +109,11 @@ static inline unsigned long regs_return_value(struct pt_regs *regs)
 	return regs->ax;
 }
 
+static inline void regs_set_return_value(struct pt_regs *regs, unsigned long rc)
+{
+	regs->ax = rc;
+}
+
 /*
  * user_mode(regs) determines whether a register set came from user
  * mode.  On x86_32, this is true if V8086 mode was enabled OR if the

arch/x86/kernel/kprobes/ftrace.c

@@ -97,3 +97,17 @@ int arch_prepare_kprobe_ftrace(struct kprobe *p)
 	p->ainsn.boostable = false;
 	return 0;
 }
+
+asmlinkage void override_func(void);
+asm(
+	".type override_func, @function\n"
+	"override_func:\n"
+	"	ret\n"
+	".size override_func, .-override_func\n"
+);
+
+void arch_ftrace_kprobe_override_function(struct pt_regs *regs)
+{
+	regs->ip = (unsigned long)&override_func;
+}
+NOKPROBE_SYMBOL(arch_ftrace_kprobe_override_function);

arch/x86/net/bpf_jit_comp.c

@@ -1109,19 +1109,29 @@ common_load:
 	return proglen;
 }
 
+struct x64_jit_data {
+	struct bpf_binary_header *header;
+	int *addrs;
+	u8 *image;
+	int proglen;
+	struct jit_context ctx;
+};
+
 struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 {
 	struct bpf_binary_header *header = NULL;
 	struct bpf_prog *tmp, *orig_prog = prog;
+	struct x64_jit_data *jit_data;
 	int proglen, oldproglen = 0;
 	struct jit_context ctx = {};
 	bool tmp_blinded = false;
+	bool extra_pass = false;
 	u8 *image = NULL;
 	int *addrs;
 	int pass;
 	int i;
 
-	if (!bpf_jit_enable)
+	if (!prog->jit_requested)
 		return orig_prog;
 
 	tmp = bpf_jit_blind_constants(prog);
@@ -1135,10 +1145,28 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 		prog = tmp;
 	}
 
+	jit_data = prog->aux->jit_data;
+	if (!jit_data) {
+		jit_data = kzalloc(sizeof(*jit_data), GFP_KERNEL);
+		if (!jit_data) {
+			prog = orig_prog;
+			goto out;
+		}
+		prog->aux->jit_data = jit_data;
+	}
+	addrs = jit_data->addrs;
+	if (addrs) {
+		ctx = jit_data->ctx;
+		oldproglen = jit_data->proglen;
+		image = jit_data->image;
+		header = jit_data->header;
+		extra_pass = true;
+		goto skip_init_addrs;
+	}
 	addrs = kmalloc(prog->len * sizeof(*addrs), GFP_KERNEL);
 	if (!addrs) {
 		prog = orig_prog;
-		goto out;
+		goto out_addrs;
 	}
 
 	/* Before first pass, make a rough estimation of addrs[]
@@ -1149,6 +1177,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 		addrs[i] = proglen;
 	}
 	ctx.cleanup_addr = proglen;
+skip_init_addrs:
 
 	/* JITed image shrinks with every pass and the loop iterates
 	 * until the image stops shrinking. Very large bpf programs
@@ -1189,7 +1218,15 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 
 	if (image) {
 		bpf_flush_icache(header, image + proglen);
-		bpf_jit_binary_lock_ro(header);
+		if (!prog->is_func || extra_pass) {
+			bpf_jit_binary_lock_ro(header);
+		} else {
+			jit_data->addrs = addrs;
+			jit_data->ctx = ctx;
+			jit_data->proglen = proglen;
+			jit_data->image = image;
+			jit_data->header = header;
+		}
 		prog->bpf_func = (void *)image;
 		prog->jited = 1;
 		prog->jited_len = proglen;
@@ -1197,8 +1234,12 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 		prog = orig_prog;
 	}
 
+	if (!prog->is_func || extra_pass) {
 out_addrs:
-	kfree(addrs);
+		kfree(addrs);
+		kfree(jit_data);
+		prog->aux->jit_data = NULL;
+	}
 out:
 	if (tmp_blinded)
 		bpf_jit_prog_release_other(prog, prog == orig_prog ?

drivers/net/ethernet/netronome/nfp/bpf/fw.h

@@ -0,0 +1,54 @@
+/*
+ * Copyright (C) 2017 Netronome Systems, Inc.
+ *
+ * This software is dual licensed under the GNU General License Version 2,
+ * June 1991 as shown in the file COPYING in the top-level directory of this
+ * source tree or the BSD 2-Clause License provided below.  You have the
+ * option to license this software under the complete terms of either license.
+ *
+ * The BSD 2-Clause License:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      1. Redistributions of source code must retain the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer.
+ *
+ *      2. Redistributions in binary form must reproduce the above
+ *         copyright notice, this list of conditions and the following
+ *         disclaimer in the documentation and/or other materials
+ *         provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef NFP_BPF_FW_H
+#define NFP_BPF_FW_H 1
+
+#include <linux/bitops.h>
+#include <linux/types.h>
+
+enum bpf_cap_tlv_type {
+	NFP_BPF_CAP_TYPE_ADJUST_HEAD	= 2,
+};
+
+struct nfp_bpf_cap_tlv_adjust_head {
+	__le32 flags;
+	__le32 off_min;
+	__le32 off_max;
+	__le32 guaranteed_sub;
+	__le32 guaranteed_add;
+};
+
+#define NFP_BPF_ADJUST_HEAD_NO_META	BIT(0)
+
+#endif

drivers/net/ethernet/netronome/nfp/bpf/jit.c

@@ -33,6 +33,7 @@
 
 #define pr_fmt(fmt)	"NFP net bpf: " fmt
 
+#include <linux/bug.h>
 #include <linux/kernel.h>
 #include <linux/bpf.h>
 #include <linux/filter.h>
@@ -87,6 +88,18 @@ static unsigned int nfp_prog_current_offset(struct nfp_prog *nfp_prog)
 	return nfp_prog->start_off + nfp_prog->prog_len;
 }
 
+static bool
+nfp_prog_confirm_current_offset(struct nfp_prog *nfp_prog, unsigned int off)
+{
+	/* If there is a recorded error we may have dropped instructions;
+	 * that doesn't have to be due to translator bug, and the translation
+	 * will fail anyway, so just return OK.
+	 */
+	if (nfp_prog->error)
+		return true;
+	return !WARN_ON_ONCE(nfp_prog_current_offset(nfp_prog) != off);
+}
+
 static unsigned int
 nfp_prog_offset_to_index(struct nfp_prog *nfp_prog, unsigned int offset)
 {
@@ -1196,6 +1209,86 @@ static void wrp_end32(struct nfp_prog *nfp_prog, swreg reg_in, u8 gpr_out)
 		      SHF_SC_R_ROT, 16);
 }
 
+static int adjust_head(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+{
+	swreg tmp = imm_a(nfp_prog), tmp_len = imm_b(nfp_prog);
+	struct nfp_bpf_cap_adjust_head *adjust_head;
+	u32 ret_einval, end;
+
+	adjust_head = &nfp_prog->bpf->adjust_head;
+
+	/* Optimized version - 5 vs 14 cycles */
+	if (nfp_prog->adjust_head_location != UINT_MAX) {
+		if (WARN_ON_ONCE(nfp_prog->adjust_head_location != meta->n))
+			return -EINVAL;
+
+		emit_alu(nfp_prog, pptr_reg(nfp_prog),
+			 reg_a(2 * 2), ALU_OP_ADD, pptr_reg(nfp_prog));
+		emit_alu(nfp_prog, plen_reg(nfp_prog),
+			 plen_reg(nfp_prog), ALU_OP_SUB, reg_a(2 * 2));
+		emit_alu(nfp_prog, pv_len(nfp_prog),
+			 pv_len(nfp_prog), ALU_OP_SUB, reg_a(2 * 2));
+
+		wrp_immed(nfp_prog, reg_both(0), 0);
+		wrp_immed(nfp_prog, reg_both(1), 0);
+
+		/* TODO: when adjust head is guaranteed to succeed we can
+		 * also eliminate the following if (r0 == 0) branch.
+		 */
+
+		return 0;
+	}
+
+	ret_einval = nfp_prog_current_offset(nfp_prog) + 14;
+	end = ret_einval + 2;
+
+	/* We need to use a temp because offset is just a part of the pkt ptr */
+	emit_alu(nfp_prog, tmp,
+		 reg_a(2 * 2), ALU_OP_ADD_2B, pptr_reg(nfp_prog));
+
+	/* Validate result will fit within FW datapath constraints */
+	emit_alu(nfp_prog, reg_none(),
+		 tmp, ALU_OP_SUB, reg_imm(adjust_head->off_min));
+	emit_br(nfp_prog, BR_BLO, ret_einval, 0);
+	emit_alu(nfp_prog, reg_none(),
+		 reg_imm(adjust_head->off_max), ALU_OP_SUB, tmp);
+	emit_br(nfp_prog, BR_BLO, ret_einval, 0);
+
+	/* Validate the length is at least ETH_HLEN */
+	emit_alu(nfp_prog, tmp_len,
+		 plen_reg(nfp_prog), ALU_OP_SUB, reg_a(2 * 2));
+	emit_alu(nfp_prog, reg_none(),
+		 tmp_len, ALU_OP_SUB, reg_imm(ETH_HLEN));
+	emit_br(nfp_prog, BR_BMI, ret_einval, 0);
+
+	/* Load the ret code */
+	wrp_immed(nfp_prog, reg_both(0), 0);
+	wrp_immed(nfp_prog, reg_both(1), 0);
+
+	/* Modify the packet metadata */
+	emit_ld_field(nfp_prog, pptr_reg(nfp_prog), 0x3, tmp, SHF_SC_NONE, 0);
+
+	/* Skip over the -EINVAL ret code (defer 2) */
+	emit_br_def(nfp_prog, end, 2);
+
+	emit_alu(nfp_prog, plen_reg(nfp_prog),
+		 plen_reg(nfp_prog), ALU_OP_SUB, reg_a(2 * 2));
+	emit_alu(nfp_prog, pv_len(nfp_prog),
+		 pv_len(nfp_prog), ALU_OP_SUB, reg_a(2 * 2));
+
+	/* return -EINVAL target */
+	if (!nfp_prog_confirm_current_offset(nfp_prog, ret_einval))
+		return -EINVAL;
+
+	wrp_immed(nfp_prog, reg_both(0), -22);
+	wrp_immed(nfp_prog, reg_both(1), ~0);
+
+	if (!nfp_prog_confirm_current_offset(nfp_prog, end))
+		return -EINVAL;
+
+	return 0;
+}
+
 /* --- Callbacks --- */
 static int mov_reg64(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
 {
@@ -1930,6 +2023,17 @@ static int jne_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
 	return wrp_test_reg(nfp_prog, meta, ALU_OP_XOR, BR_BNE);
 }
 
+static int call(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+{
+	switch (meta->insn.imm) {
+	case BPF_FUNC_xdp_adjust_head:
+		return adjust_head(nfp_prog, meta);
+	default:
+		WARN_ONCE(1, "verifier allowed unsupported function\n");
+		return -EOPNOTSUPP;
+	}
+}
+
 static int goto_out(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
 {
 	wrp_br_special(nfp_prog, BR_UNC, OP_BR_GO_OUT);
@@ -2002,6 +2106,7 @@ static const instr_cb_t instr_cb[256] = {
 	[BPF_JMP | BPF_JLE | BPF_X] =	jle_reg,
 	[BPF_JMP | BPF_JSET | BPF_X] =	jset_reg,
 	[BPF_JMP | BPF_JNE | BPF_X] =	jne_reg,
+	[BPF_JMP | BPF_CALL] =		call,
 	[BPF_JMP | BPF_EXIT] =		goto_out,
 };
 
@@ -2026,6 +2131,8 @@ static int nfp_fixup_branches(struct nfp_prog *nfp_prog)
 	list_for_each_entry(meta, &nfp_prog->insns, l) {
 		if (meta->skip)
 			continue;
+		if (meta->insn.code == (BPF_JMP | BPF_CALL))
+			continue;
 		if (BPF_CLASS(meta->insn.code) != BPF_JMP)
 			continue;
 

drivers/net/ethernet/netronome/nfp/bpf/main.c

@@ -34,10 +34,12 @@
 #include <net/pkt_cls.h>
 
 #include "../nfpcore/nfp_cpp.h"
+#include "../nfpcore/nfp_nffw.h"
 #include "../nfp_app.h"
 #include "../nfp_main.h"
 #include "../nfp_net.h"
 #include "../nfp_port.h"
+#include "fw.h"
 #include "main.h"
 
 static bool nfp_net_ebpf_capable(struct nfp_net *nn)
@@ -155,10 +157,123 @@ static bool nfp_bpf_tc_busy(struct nfp_app *app, struct nfp_net *nn)
 	return nn->dp.ctrl & NFP_NET_CFG_CTRL_BPF;
 }
 
+static int
+nfp_bpf_parse_cap_adjust_head(struct nfp_app_bpf *bpf, void __iomem *value,
+			      u32 length)
+{
+	struct nfp_bpf_cap_tlv_adjust_head __iomem *cap = value;
+	struct nfp_cpp *cpp = bpf->app->pf->cpp;
+
+	if (length < sizeof(*cap)) {
+		nfp_err(cpp, "truncated adjust_head TLV: %d\n", length);
+		return -EINVAL;
+	}
+
+	bpf->adjust_head.flags = readl(&cap->flags);
+	bpf->adjust_head.off_min = readl(&cap->off_min);
+	bpf->adjust_head.off_max = readl(&cap->off_max);
+	bpf->adjust_head.guaranteed_sub = readl(&cap->guaranteed_sub);
+	bpf->adjust_head.guaranteed_add = readl(&cap->guaranteed_add);
+
+	if (bpf->adjust_head.off_min > bpf->adjust_head.off_max) {
+		nfp_err(cpp, "invalid adjust_head TLV: min > max\n");
+		return -EINVAL;
+	}
+	if (!FIELD_FIT(UR_REG_IMM_MAX, bpf->adjust_head.off_min) ||
+	    !FIELD_FIT(UR_REG_IMM_MAX, bpf->adjust_head.off_max)) {
+		nfp_warn(cpp, "disabling adjust_head - driver expects min/max to fit in as immediates\n");
+		memset(&bpf->adjust_head, 0, sizeof(bpf->adjust_head));
+		return 0;
+	}
+
+	return 0;
+}
+
+static int nfp_bpf_parse_capabilities(struct nfp_app *app)
+{
+	struct nfp_cpp *cpp = app->pf->cpp;
+	struct nfp_cpp_area *area;
+	u8 __iomem *mem, *start;
+
+	mem = nfp_rtsym_map(app->pf->rtbl, "_abi_bpf_capabilities", "bpf.cap",
+			    8, &area);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem) == -ENOENT ? 0 : PTR_ERR(mem);
+
+	start = mem;
+	while (mem - start + 8 < nfp_cpp_area_size(area)) {
+		u8 __iomem *value;
+		u32 type, length;
+
+		type = readl(mem);
+		length = readl(mem + 4);
+		value = mem + 8;
+
+		mem += 8 + length;
+		if (mem - start > nfp_cpp_area_size(area))
+			goto err_release_free;
+
+		switch (type) {
+		case NFP_BPF_CAP_TYPE_ADJUST_HEAD:
+			if (nfp_bpf_parse_cap_adjust_head(app->priv, value,
+							  length))
+				goto err_release_free;
+			break;
+		default:
+			nfp_dbg(cpp, "unknown BPF capability: %d\n", type);
+			break;
+		}
+	}
+	if (mem - start != nfp_cpp_area_size(area)) {
+		nfp_err(cpp, "BPF capabilities left after parsing, parsed:%zd total length:%zu\n",
+			mem - start, nfp_cpp_area_size(area));
+		goto err_release_free;
+	}
+
+	nfp_cpp_area_release_free(area);
+
+	return 0;
+
+err_release_free:
+	nfp_err(cpp, "invalid BPF capabilities at offset:%zd\n", mem - start);
+	nfp_cpp_area_release_free(area);
+	return -EINVAL;
+}
+
+static int nfp_bpf_init(struct nfp_app *app)
+{
+	struct nfp_app_bpf *bpf;
+	int err;
+
+	bpf = kzalloc(sizeof(*bpf), GFP_KERNEL);
+	if (!bpf)
+		return -ENOMEM;
+	bpf->app = app;
+	app->priv = bpf;
+
+	err = nfp_bpf_parse_capabilities(app);
+	if (err)
+		goto err_free_bpf;
+
+	return 0;
+
+err_free_bpf:
+	kfree(bpf);
+	return err;
+}
+
+static void nfp_bpf_clean(struct nfp_app *app)
+{
+	kfree(app->priv);
+}
+
 const struct nfp_app_type app_bpf = {
 	.id		= NFP_APP_BPF_NIC,
 	.name		= "ebpf",
 
+	.init		= nfp_bpf_init,
+	.clean		= nfp_bpf_clean,
+
 	.extra_cap	= nfp_bpf_extra_cap,
 
 	.vnic_alloc	= nfp_app_nic_vnic_alloc,

drivers/net/ethernet/netronome/nfp/bpf/main.h

@@ -78,6 +78,29 @@ enum pkt_vec {
 #define NFP_BPF_ABI_FLAGS	reg_imm(0)
 #define   NFP_BPF_ABI_FLAG_MARK	1
 
+/**
+ * struct nfp_app_bpf - bpf app priv structure
+ * @app:		backpointer to the app
+ *
+ * @adjust_head:	adjust head capability
+ * @flags:		extra flags for adjust head
+ * @off_min:		minimal packet offset within buffer required
+ * @off_max:		maximum packet offset within buffer required
+ * @guaranteed_sub:	amount of negative adjustment guaranteed possible
+ * @guaranteed_add:	amount of positive adjustment guaranteed possible
+ */
+struct nfp_app_bpf {
+	struct nfp_app *app;
+
+	struct nfp_bpf_cap_adjust_head {
+		u32 flags;
+		int off_min;
+		int off_max;
+		int guaranteed_sub;
+		int guaranteed_add;
+	} adjust_head;
+};
+
 struct nfp_prog;
 struct nfp_insn_meta;
 typedef int (*instr_cb_t)(struct nfp_prog *, struct nfp_insn_meta *);
@@ -97,6 +120,7 @@ typedef int (*instr_cb_t)(struct nfp_prog *, struct nfp_insn_meta *);
  * @ptr: pointer type for memory operations
  * @ldst_gather_len: memcpy length gathered from load/store sequence
  * @paired_st: the paired store insn at the head of the sequence
+ * @arg2: arg2 for call instructions
  * @ptr_not_const: pointer is not always constant
  * @jmp_dst: destination info for jump instructions
  * @off: index of first generated machine instruction (in nfp_prog.prog)
@@ -116,6 +140,7 @@ struct nfp_insn_meta {
 			bool ptr_not_const;
 		};
 		struct nfp_insn_meta *jmp_dst;
+		struct bpf_reg_state arg2;
 	};
 	unsigned int off;
 	unsigned short n;
@@ -160,6 +185,7 @@ static inline bool is_mbpf_store(const struct nfp_insn_meta *meta)
 
 /**
  * struct nfp_prog - nfp BPF program
+ * @bpf: backpointer to the bpf app priv structure
  * @prog: machine code
  * @prog_len: number of valid instructions in @prog array
  * @__prog_alloc_len: alloc size of @prog array
@@ -173,9 +199,12 @@ static inline bool is_mbpf_store(const struct nfp_insn_meta *meta)
  * @n_translated: number of successfully translated instructions (for errors)
  * @error: error code if something went wrong
  * @stack_depth: max stack depth from the verifier
+ * @adjust_head_location: if program has single adjust head call - the insn no.
  * @insns: list of BPF instruction wrappers (struct nfp_insn_meta)
  */
 struct nfp_prog {
+	struct nfp_app_bpf *bpf;
+
 	u64 *prog;
 	unsigned int prog_len;
 	unsigned int __prog_alloc_len;
@@ -194,6 +223,7 @@ struct nfp_prog {
 	int error;
 
 	unsigned int stack_depth;
+	unsigned int adjust_head_location;
 
 	struct list_head insns;
 };

drivers/net/ethernet/netronome/nfp/bpf/offload.c

@@ -48,6 +48,7 @@
 #include <net/tc_act/tc_mirred.h>
 
 #include "main.h"
+#include "../nfp_app.h"
 #include "../nfp_net_ctrl.h"
 #include "../nfp_net.h"
 
@@ -115,6 +116,7 @@ int nfp_bpf_verifier_prep(struct nfp_app *app, struct nfp_net *nn,
 
 	INIT_LIST_HEAD(&nfp_prog->insns);
 	nfp_prog->type = prog->type;
+	nfp_prog->bpf = app->priv;
 
 	ret = nfp_prog_prepare(nfp_prog, prog->insnsi, prog->len);
 	if (ret)

drivers/net/ethernet/netronome/nfp/bpf/verifier.c

@@ -38,6 +38,7 @@
 #include <linux/kernel.h>
 #include <linux/pkt_cls.h>
 
+#include "fw.h"
 #include "main.h"
 
 struct nfp_insn_meta *
@@ -68,6 +69,73 @@ nfp_bpf_goto_meta(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta,
 	return meta;
 }
 
+static void
+nfp_record_adjust_head(struct nfp_app_bpf *bpf, struct nfp_prog *nfp_prog,
+		       struct nfp_insn_meta *meta,
+		       const struct bpf_reg_state *reg2)
+{
+	unsigned int location =	UINT_MAX;
+	int imm;
+
+	/* Datapath usually can give us guarantees on how much adjust head
+	 * can be done without the need for any checks.  Optimize the simple
+	 * case where there is only one adjust head by a constant.
+	 */
+	if (reg2->type != SCALAR_VALUE || !tnum_is_const(reg2->var_off))
+		goto exit_set_location;
+	imm = reg2->var_off.value;
+	/* Translator will skip all checks, we need to guarantee min pkt len */
+	if (imm > ETH_ZLEN - ETH_HLEN)
+		goto exit_set_location;
+	if (imm > (int)bpf->adjust_head.guaranteed_add ||
+	    imm < -bpf->adjust_head.guaranteed_sub)
+		goto exit_set_location;
+
+	if (nfp_prog->adjust_head_location) {
+		/* Only one call per program allowed */
+		if (nfp_prog->adjust_head_location != meta->n)
+			goto exit_set_location;
+
+		if (meta->arg2.var_off.value != imm)
+			goto exit_set_location;
+	}
+
+	location = meta->n;
+exit_set_location:
+	nfp_prog->adjust_head_location = location;
+}
+
+static int
+nfp_bpf_check_call(struct nfp_prog *nfp_prog, struct bpf_verifier_env *env,
+		   struct nfp_insn_meta *meta)
+{
+	const struct bpf_reg_state *reg2 = cur_regs(env) + BPF_REG_2;
+	struct nfp_app_bpf *bpf = nfp_prog->bpf;
+	u32 func_id = meta->insn.imm;
+
+	switch (func_id) {
+	case BPF_FUNC_xdp_adjust_head:
+		if (!bpf->adjust_head.off_max) {
+			pr_warn("adjust_head not supported by FW\n");
+			return -EOPNOTSUPP;
+		}
+		if (!(bpf->adjust_head.flags & NFP_BPF_ADJUST_HEAD_NO_META)) {
+			pr_warn("adjust_head: FW requires shifting metadata, not supported by the driver\n");
+			return -EOPNOTSUPP;
+		}
+
+		nfp_record_adjust_head(bpf, nfp_prog, meta, reg2);
+		break;
+	default:
+		pr_warn("unsupported function id: %d\n", func_id);
+		return -EOPNOTSUPP;
+	}
+
+	meta->arg2 = *reg2;
+
+	return 0;
+}
+
 static int
 nfp_bpf_check_exit(struct nfp_prog *nfp_prog,
 		   struct bpf_verifier_env *env)
@@ -177,6 +245,8 @@ nfp_verify_insn(struct bpf_verifier_env *env, int insn_idx, int prev_insn_idx)
 		return -EINVAL;
 	}
 
+	if (meta->insn.code == (BPF_JMP | BPF_CALL))
+		return nfp_bpf_check_call(nfp_prog, env, meta);
 	if (meta->insn.code == (BPF_JMP | BPF_EXIT))
 		return nfp_bpf_check_exit(nfp_prog, env);
 

drivers/net/ethernet/netronome/nfp/nfp_asm.h

@@ -77,6 +77,7 @@
 enum br_mask {
 	BR_BEQ = 0x00,
 	BR_BNE = 0x01,
+	BR_BMI = 0x02,
 	BR_BHS = 0x04,
 	BR_BLO = 0x05,
 	BR_BGE = 0x08,
@@ -175,6 +176,7 @@ enum alu_op {
 	ALU_OP_NONE	= 0x00,
 	ALU_OP_ADD	= 0x01,
 	ALU_OP_NOT	= 0x04,
+	ALU_OP_ADD_2B	= 0x05,
 	ALU_OP_AND	= 0x08,
 	ALU_OP_SUB_C	= 0x0d,
 	ALU_OP_ADD_C	= 0x11,

drivers/net/ethernet/netronome/nfp/nfp_net_common.c

@@ -3392,7 +3392,7 @@ static int nfp_net_xdp(struct net_device *netdev, struct netdev_bpf *xdp)
 		if (nn->dp.bpf_offload_xdp)
 			xdp->prog_attached = XDP_ATTACHED_HW;
 		xdp->prog_id = nn->xdp_prog ? nn->xdp_prog->aux->id : 0;
-		xdp->flags = nn->xdp_prog ? nn->xdp_flags : 0;
+		xdp->prog_flags = nn->xdp_prog ? nn->xdp_flags : 0;
 		return 0;
 	case BPF_OFFLOAD_VERIFIER_PREP:
 		return nfp_app_bpf_verifier_prep(nn->app, nn, xdp);