Merge tag 'bpf-next-6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next

Pull bpf updates from Alexei Starovoitov:

 - Add BPF uprobe session support (Jiri Olsa)

 - Optimize uprobe performance (Andrii Nakryiko)

 - Add bpf_fastcall support to helpers and kfuncs (Eduard Zingerman)

 - Avoid calling free_htab_elem() under hash map bucket lock (Hou Tao)

 - Prevent tailcall infinite loop caused by freplace (Leon Hwang)

 - Mark raw_tracepoint arguments as nullable (Kumar Kartikeya Dwivedi)

 - Introduce uptr support in the task local storage map (Martin KaFai
   Lau)

 - Stringify errno log messages in libbpf (Mykyta Yatsenko)

 - Add kmem_cache BPF iterator for perf's lock profiling (Namhyung Kim)

 - Support BPF objects of either endianness in libbpf (Tony Ambardar)

 - Add ksym to struct_ops trampoline to fix stack trace (Xu Kuohai)

 - Introduce private stack for eligible BPF programs (Yonghong Song)

 - Migrate samples/bpf tests to selftests/bpf test_progs (Daniel T. Lee)

 - Migrate test_sock to selftests/bpf test_progs (Jordan Rife)

* tag 'bpf-next-6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (152 commits)
  libbpf: Change hash_combine parameters from long to unsigned long
  selftests/bpf: Fix build error with llvm 19
  libbpf: Fix memory leak in bpf_program__attach_uprobe_multi
  bpf: use common instruction history across all states
  bpf: Add necessary migrate_disable to range_tree.
  bpf: Do not alloc arena on unsupported arches
  selftests/bpf: Set test path for token/obj_priv_implicit_token_envvar
  selftests/bpf: Add a test for arena range tree algorithm
  bpf: Introduce range_tree data structure and use it in bpf arena
  samples/bpf: Remove unused variable in xdp2skb_meta_kern.c
  samples/bpf: Remove unused variables in tc_l2_redirect_kern.c
  bpftool: Cast variable `var` to long long
  bpf, x86: Propagate tailcall info only for subprogs
  bpf: Add kernel symbol for struct_ops trampoline
  bpf: Use function pointers count as struct_ops links count
  bpf: Remove unused member rcu from bpf_struct_ops_map
  selftests/bpf: Add struct_ops prog private stack tests
  bpf: Support private stack for struct_ops progs
  selftests/bpf: Add tracing prog private stack tests
  bpf, x86: Support private stack in jit
  ...

2 files changed, 175 insertions, 21 deletions

diff --git a/arch/arm64/net/bpf_jit_comp.c b/arch/arm64/net/bpf_jit_comp.c
index 5db82bfc9dc1..27ef366363e4 100644
--- a/arch/arm64/net/bpf_jit_comp.c
+++ b/arch/arm64/net/bpf_jit_comp.c
@@ -2094,6 +2094,12 @@ static void restore_args(struct jit_ctx *ctx, int args_off, int nregs)
 	}
 }
 
+static bool is_struct_ops_tramp(const struct bpf_tramp_links *fentry_links)
+{
+	return fentry_links->nr_links == 1 &&
+		fentry_links->links[0]->link.type == BPF_LINK_TYPE_STRUCT_OPS;
+}
+
 /* Based on the x86's implementation of arch_prepare_bpf_trampoline().
  *
  * bpf prog and function entry before bpf trampoline hooked:
@@ -2123,6 +2129,7 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im,
 	struct bpf_tramp_links *fmod_ret = &tlinks[BPF_TRAMP_MODIFY_RETURN];
 	bool save_ret;
 	__le32 **branches = NULL;
+	bool is_struct_ops = is_struct_ops_tramp(fentry);
 
 	/* trampoline stack layout:
 	 *                  [ parent ip         ]
@@ -2191,11 +2198,14 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im,
 	 */
 	emit_bti(A64_BTI_JC, ctx);
 
-	/* frame for parent function */
-	emit(A64_PUSH(A64_FP, A64_R(9), A64_SP), ctx);
-	emit(A64_MOV(1, A64_FP, A64_SP), ctx);
+	/* x9 is not set for struct_ops */
+	if (!is_struct_ops) {
+		/* frame for parent function */
+		emit(A64_PUSH(A64_FP, A64_R(9), A64_SP), ctx);
+		emit(A64_MOV(1, A64_FP, A64_SP), ctx);
+	}
 
-	/* frame for patched function */
+	/* frame for patched function for tracing, or caller for struct_ops */
 	emit(A64_PUSH(A64_FP, A64_LR, A64_SP), ctx);
 	emit(A64_MOV(1, A64_FP, A64_SP), ctx);
 
@@ -2289,19 +2299,24 @@ static int prepare_trampoline(struct jit_ctx *ctx, struct bpf_tramp_image *im,
 	/* reset SP  */
 	emit(A64_MOV(1, A64_SP, A64_FP), ctx);
 
-	/* pop frames  */
-	emit(A64_POP(A64_FP, A64_LR, A64_SP), ctx);
-	emit(A64_POP(A64_FP, A64_R(9), A64_SP), ctx);
-
-	if (flags & BPF_TRAMP_F_SKIP_FRAME) {
-		/* skip patched function, return to parent */
-		emit(A64_MOV(1, A64_LR, A64_R(9)), ctx);
-		emit(A64_RET(A64_R(9)), ctx);
+	if (is_struct_ops) {
+		emit(A64_POP(A64_FP, A64_LR, A64_SP), ctx);
+		emit(A64_RET(A64_LR), ctx);
 	} else {
-		/* return to patched function */
-		emit(A64_MOV(1, A64_R(10), A64_LR), ctx);
-		emit(A64_MOV(1, A64_LR, A64_R(9)), ctx);
-		emit(A64_RET(A64_R(10)), ctx);
+		/* pop frames */
+		emit(A64_POP(A64_FP, A64_LR, A64_SP), ctx);
+		emit(A64_POP(A64_FP, A64_R(9), A64_SP), ctx);
+
+		if (flags & BPF_TRAMP_F_SKIP_FRAME) {
+			/* skip patched function, return to parent */
+			emit(A64_MOV(1, A64_LR, A64_R(9)), ctx);
+			emit(A64_RET(A64_R(9)), ctx);
+		} else {
+			/* return to patched function */
+			emit(A64_MOV(1, A64_R(10), A64_LR), ctx);
+			emit(A64_MOV(1, A64_LR, A64_R(9)), ctx);
+			emit(A64_RET(A64_R(10)), ctx);
+		}
 	}
 
 	kfree(branches);
diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
index 06b080b61aa5..a43fc5af973d 100644
--- a/arch/x86/net/bpf_jit_comp.c
+++ b/arch/x86/net/bpf_jit_comp.c
@@ -325,6 +325,22 @@ struct jit_context {
 /* Number of bytes that will be skipped on tailcall */
 #define X86_TAIL_CALL_OFFSET	(12 + ENDBR_INSN_SIZE)
 
+static void push_r9(u8 **pprog)
+{
+	u8 *prog = *pprog;
+
+	EMIT2(0x41, 0x51);   /* push r9 */
+	*pprog = prog;
+}
+
+static void pop_r9(u8 **pprog)
+{
+	u8 *prog = *pprog;
+
+	EMIT2(0x41, 0x59);   /* pop r9 */
+	*pprog = prog;
+}
+
 static void push_r12(u8 **pprog)
 {
 	u8 *prog = *pprog;
@@ -1404,6 +1420,24 @@ static void emit_shiftx(u8 **pprog, u32 dst_reg, u8 src_reg, bool is64, u8 op)
 	*pprog = prog;
 }
 
+static void emit_priv_frame_ptr(u8 **pprog, void __percpu *priv_frame_ptr)
+{
+	u8 *prog = *pprog;
+
+	/* movabs r9, priv_frame_ptr */
+	emit_mov_imm64(&prog, X86_REG_R9, (__force long) priv_frame_ptr >> 32,
+		       (u32) (__force long) priv_frame_ptr);
+
+#ifdef CONFIG_SMP
+	/* add <r9>, gs:[<off>] */
+	EMIT2(0x65, 0x4c);
+	EMIT3(0x03, 0x0c, 0x25);
+	EMIT((u32)(unsigned long)&this_cpu_off, 4);
+#endif
+
+	*pprog = prog;
+}
+
 #define INSN_SZ_DIFF (((addrs[i] - addrs[i - 1]) - (prog - temp)))
 
 #define __LOAD_TCC_PTR(off)			\
@@ -1412,6 +1446,10 @@ static void emit_shiftx(u8 **pprog, u32 dst_reg, u8 src_reg, bool is64, u8 op)
 #define LOAD_TAIL_CALL_CNT_PTR(stack)				\
 	__LOAD_TCC_PTR(BPF_TAIL_CALL_CNT_PTR_STACK_OFF(stack))
 
+/* Memory size/value to protect private stack overflow/underflow */
+#define PRIV_STACK_GUARD_SZ    8
+#define PRIV_STACK_GUARD_VAL   0xEB9F12345678eb9fULL
+
 static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, u8 *rw_image,
 		  int oldproglen, struct jit_context *ctx, bool jmp_padding)
 {
@@ -1421,18 +1459,28 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, u8 *rw_image
 	int insn_cnt = bpf_prog->len;
 	bool seen_exit = false;
 	u8 temp[BPF_MAX_INSN_SIZE + BPF_INSN_SAFETY];
+	void __percpu *priv_frame_ptr = NULL;
 	u64 arena_vm_start, user_vm_start;
+	void __percpu *priv_stack_ptr;
 	int i, excnt = 0;
 	int ilen, proglen = 0;
 	u8 *prog = temp;
+	u32 stack_depth;
 	int err;
 
+	stack_depth = bpf_prog->aux->stack_depth;
+	priv_stack_ptr = bpf_prog->aux->priv_stack_ptr;
+	if (priv_stack_ptr) {
+		priv_frame_ptr = priv_stack_ptr + PRIV_STACK_GUARD_SZ + round_up(stack_depth, 8);
+		stack_depth = 0;
+	}
+
 	arena_vm_start = bpf_arena_get_kern_vm_start(bpf_prog->aux->arena);
 	user_vm_start = bpf_arena_get_user_vm_start(bpf_prog->aux->arena);
 
 	detect_reg_usage(insn, insn_cnt, callee_regs_used);
 
-	emit_prologue(&prog, bpf_prog->aux->stack_depth,
+	emit_prologue(&prog, stack_depth,
 		      bpf_prog_was_classic(bpf_prog), tail_call_reachable,
 		      bpf_is_subprog(bpf_prog), bpf_prog->aux->exception_cb);
 	/* Exception callback will clobber callee regs for its own use, and
@@ -1454,6 +1502,9 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, u8 *rw_image
 		emit_mov_imm64(&prog, X86_REG_R12,
 			       arena_vm_start >> 32, (u32) arena_vm_start);
 
+	if (priv_frame_ptr)
+		emit_priv_frame_ptr(&prog, priv_frame_ptr);
+
 	ilen = prog - temp;
 	if (rw_image)
 		memcpy(rw_image + proglen, temp, ilen);
@@ -1473,6 +1524,14 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, u8 *rw_image
 		u8 *func;
 		int nops;
 
+		if (priv_frame_ptr) {
+			if (src_reg == BPF_REG_FP)
+				src_reg = X86_REG_R9;
+
+			if (dst_reg == BPF_REG_FP)
+				dst_reg = X86_REG_R9;
+		}
+
 		switch (insn->code) {
 			/* ALU */
 		case BPF_ALU | BPF_ADD | BPF_X:
@@ -2127,15 +2186,21 @@ populate_extable:
 			u8 *ip = image + addrs[i - 1];
 
 			func = (u8 *) __bpf_call_base + imm32;
-			if (tail_call_reachable) {
-				LOAD_TAIL_CALL_CNT_PTR(bpf_prog->aux->stack_depth);
+			if (src_reg == BPF_PSEUDO_CALL && tail_call_reachable) {
+				LOAD_TAIL_CALL_CNT_PTR(stack_depth);
 				ip += 7;
 			}
 			if (!imm32)
 				return -EINVAL;
+			if (priv_frame_ptr) {
+				push_r9(&prog);
+				ip += 2;
+			}
 			ip += x86_call_depth_emit_accounting(&prog, func, ip);
 			if (emit_call(&prog, func, ip))
 				return -EINVAL;
+			if (priv_frame_ptr)
+				pop_r9(&prog);
 			break;
 		}
 
@@ -2145,13 +2210,13 @@ populate_extable:
 							  &bpf_prog->aux->poke_tab[imm32 - 1],
 							  &prog, image + addrs[i - 1],
 							  callee_regs_used,
-							  bpf_prog->aux->stack_depth,
+							  stack_depth,
 							  ctx);
 			else
 				emit_bpf_tail_call_indirect(bpf_prog,
 							    &prog,
 							    callee_regs_used,
-							    bpf_prog->aux->stack_depth,
+							    stack_depth,
 							    image + addrs[i - 1],
 							    ctx);
 			break;
@@ -3303,6 +3368,42 @@ int arch_prepare_bpf_dispatcher(void *image, void *buf, s64 *funcs, int num_func
 	return emit_bpf_dispatcher(&prog, 0, num_funcs - 1, funcs, image, buf);
 }
 
+static const char *bpf_get_prog_name(struct bpf_prog *prog)
+{
+	if (prog->aux->ksym.prog)
+		return prog->aux->ksym.name;
+	return prog->aux->name;
+}
+
+static void priv_stack_init_guard(void __percpu *priv_stack_ptr, int alloc_size)
+{
+	int cpu, underflow_idx = (alloc_size - PRIV_STACK_GUARD_SZ) >> 3;
+	u64 *stack_ptr;
+
+	for_each_possible_cpu(cpu) {
+		stack_ptr = per_cpu_ptr(priv_stack_ptr, cpu);
+		stack_ptr[0] = PRIV_STACK_GUARD_VAL;
+		stack_ptr[underflow_idx] = PRIV_STACK_GUARD_VAL;
+	}
+}
+
+static void priv_stack_check_guard(void __percpu *priv_stack_ptr, int alloc_size,
+				   struct bpf_prog *prog)
+{
+	int cpu, underflow_idx = (alloc_size - PRIV_STACK_GUARD_SZ) >> 3;
+	u64 *stack_ptr;
+
+	for_each_possible_cpu(cpu) {
+		stack_ptr = per_cpu_ptr(priv_stack_ptr, cpu);
+		if (stack_ptr[0] != PRIV_STACK_GUARD_VAL ||
+		    stack_ptr[underflow_idx] != PRIV_STACK_GUARD_VAL) {
+			pr_err("BPF private stack overflow/underflow detected for prog %sx\n",
+			       bpf_get_prog_name(prog));
+			break;
+		}
+	}
+}
+
 struct x64_jit_data {
 	struct bpf_binary_header *rw_header;
 	struct bpf_binary_header *header;
@@ -3320,7 +3421,9 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 	struct bpf_binary_header *rw_header = NULL;
 	struct bpf_binary_header *header = NULL;
 	struct bpf_prog *tmp, *orig_prog = prog;
+	void __percpu *priv_stack_ptr = NULL;
 	struct x64_jit_data *jit_data;
+	int priv_stack_alloc_sz;
 	int proglen, oldproglen = 0;
 	struct jit_context ctx = {};
 	bool tmp_blinded = false;
@@ -3356,6 +3459,23 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
 		}
 		prog->aux->jit_data = jit_data;
 	}
+	priv_stack_ptr = prog->aux->priv_stack_ptr;
+	if (!priv_stack_ptr && prog->aux->jits_use_priv_stack) {
+		/* Allocate actual private stack size with verifier-calculated
+		 * stack size plus two memory guards to protect overflow and
+		 * underflow.
+		 */
+		priv_stack_alloc_sz = round_up(prog->aux->stack_depth, 8) +
+				      2 * PRIV_STACK_GUARD_SZ;
+		priv_stack_ptr = __alloc_percpu_gfp(priv_stack_alloc_sz, 8, GFP_KERNEL);
+		if (!priv_stack_ptr) {
+			prog = orig_prog;
+			goto out_priv_stack;
+		}
+
+		priv_stack_init_guard(priv_stack_ptr, priv_stack_alloc_sz);
+		prog->aux->priv_stack_ptr = priv_stack_ptr;
+	}
 	addrs = jit_data->addrs;
 	if (addrs) {
 		ctx = jit_data->ctx;
@@ -3491,6 +3611,11 @@ out_image:
 			bpf_prog_fill_jited_linfo(prog, addrs + 1);
 out_addrs:
 		kvfree(addrs);
+		if (!image && priv_stack_ptr) {
+			free_percpu(priv_stack_ptr);
+			prog->aux->priv_stack_ptr = NULL;
+		}
+out_priv_stack:
 		kfree(jit_data);
 		prog->aux->jit_data = NULL;
 	}
@@ -3529,6 +3654,8 @@ void bpf_jit_free(struct bpf_prog *prog)
 	if (prog->jited) {
 		struct x64_jit_data *jit_data = prog->aux->jit_data;
 		struct bpf_binary_header *hdr;
+		void __percpu *priv_stack_ptr;
+		int priv_stack_alloc_sz;
 
 		/*
 		 * If we fail the final pass of JIT (from jit_subprogs),
@@ -3544,6 +3671,13 @@ void bpf_jit_free(struct bpf_prog *prog)
 		prog->bpf_func = (void *)prog->bpf_func - cfi_get_offset();
 		hdr = bpf_jit_binary_pack_hdr(prog);
 		bpf_jit_binary_pack_free(hdr, NULL);
+		priv_stack_ptr = prog->aux->priv_stack_ptr;
+		if (priv_stack_ptr) {
+			priv_stack_alloc_sz = round_up(prog->aux->stack_depth, 8) +
+					      2 * PRIV_STACK_GUARD_SZ;
+			priv_stack_check_guard(priv_stack_ptr, priv_stack_alloc_sz, prog);
+			free_percpu(prog->aux->priv_stack_ptr);
+		}
 		WARN_ON_ONCE(!bpf_prog_kallsyms_verify_off(prog));
 	}
 
@@ -3559,6 +3693,11 @@ bool bpf_jit_supports_exceptions(void)
 	return IS_ENABLED(CONFIG_UNWINDER_ORC);
 }
 
+bool bpf_jit_supports_private_stack(void)
+{
+	return true;
+}
+
 void arch_bpf_stack_walk(bool (*consume_fn)(void *cookie, u64 ip, u64 sp, u64 bp), void *cookie)
 {
 #if defined(CONFIG_UNWINDER_ORC)