perf trace: Refactor augmented_raw_syscalls using bpf_for

Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
- [High] Removing the (int) cast in the buffer augmentation condition makes it unconditionally false, completely breaking buffer payload collection.
- [Medium] Huge buffer sizes (where the MSB is set) fail to be augmented, whereas previously they were safely clamped to TRACE_AUG_MAX_BUF.
- [Low] Multi-line comment does not follow the required BPF subsystem style.
--

commit 0a122bc3ab16635b5165b20566e6e1890cba855b
Author: Viktor Malik <[email protected]>

perf trace: Refactor augmented_raw_syscalls using bpf_for

This commit replaces the standard loop in augment_sys_enter with a bpf_for loop to prevent BPF verifier -E2BIG errors. It also drops some workarounds previously added to help the verifier.

Quoted 4 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +0 -0

index cbdd5ce19a2f6..60babc06f3812 100644

[ ... ]

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c static inline int augment_arg(struct syscall_enter_args *args, int i, · line 476

Quoted 6 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +1 -1

@@ -464,11 +473,13 @@ static inline int augment_arg(struct syscall_enter_args *args, int i,
 	} else if (size > 0 && size <= value_size) { /* struct */
 		if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, size, arg))
 			augmented = true;
-	} else if ((int)size < 0 && size >= -6) { /* buffer */
+	} else if (size < 0 && size >= -6) { /* buffer */

[Severity: High]
Does removing the (int) cast make this condition unconditionally false?

Since size is an s64 assigned from an unsigned int beauty_map[i], a negative marker (e.g. 0xFFFFFFFF) is zero-extended to a positive s64 value. Without the (int) cast to truncate and sign-extend it back, size < 0 is always false, breaking buffer payload collection.

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c static inline int augment_arg(struct syscall_enter_args *args, int i, · line 482

Quoted 7 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +3 -1

 		index = -(size + 1);
 		barrier_var(index); // Prevent clang (noticed with v18) from removing the &= 7 trick.
 		index &= 7;	    // Satisfy the bounds checking with the verifier in some kernels.
-		aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
+		aug_size = args->args[index];
+		if (aug_size > TRACE_AUG_MAX_BUF)
+			aug_size = TRACE_AUG_MAX_BUF;

[Severity: Medium]
Will huge buffer sizes (where the most significant bit is set) fail to be augmented here?

args->args[index] is an unsigned long. If the most significant bit is set, assigning it to the s64 aug_size makes it negative. The subsequent if (aug_size > TRACE_AUG_MAX_BUF) is a signed comparison and will evaluate to false. The length remains negative, causing the later if (aug_size > 0) check to fail, skipping buffer processing entirely instead of clamping it.

Quoted 3 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +0 -0

 		if (aug_size > 0) {
 			if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, aug_size, arg))

[ ... ]

Quoted 18 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +6 -8

@@ -521,14 +530,29 @@ static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 	/* copy the sys_enter header, which has the syscall_nr */
 	__builtin_memcpy(&payload->args, args, sizeof(struct syscall_enter_args));
 
-	for (int i = 0; i < 6; i++) {
-		written = augment_arg(args, i, beauty_map, payload_offset);
-		if (written < 0)
-			return 1;
-		if (written > 0) {
-			output += written;
-			payload_offset += written;
-			do_output = true;
+	if (bpf_ksym_exists(bpf_iter_num_new)) {
+		bpf_for(i, 0, 6) {
+			written = augment_arg(args, i, beauty_map, payload, output);
+			if (written < 0)
+				return 1;
+			if (written > 0) {

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c static int augment_sys_enter(void *ctx, struct syscall_enter_args *args) · line 542

Quoted 4 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +4 -0

+				output += written;
+				/* guide the verifier to forget range of `output`, which
+				 * helps to prove convergence of the loop
+				 */

[Severity: Low]
This isn't a bug, but this multi-line comment does not follow the preferred BPF subsystem style where the opening /* is on its own line.

Quoted 4 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +4 -0

+				output += ZERO;
+				do_output = true;
+			}
+		}

The BPF program in augmented_raw_syscalls uses a for loop to iterate all syscall arguments. The loop body is quite complex and often poses problems for the BPF verifier. As a preparation step for addressing this issue, factor out the loop body into a separate function.

Signed-off-by Viktor Malik <[email protected]>

Cc [email protected]

---
 .../bpf_skel/augmented_raw_syscalls.bpf.c     | 127 ++++++++++--------
 1 file changed, 72 insertions(+), 55 deletions(-)

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +1 -0

index 2a6e61864ee0..cbdd5ce19a2f 100644
@@ -429,15 +429,79 @@ static bool pid_filter__has(struct pids_filtered *pids, pid_t pid)
 	return bpf_map_lookup_elem(pids, &pid) != NULL;
 }
 
+/*

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +61 -55

+{
+	int index, value_size = sizeof(struct augmented_arg) - offsetof(struct augmented_arg, value);
+	s64 aug_size, size;
+	bool augmented;
+	void *arg;
+
+	arg = (void *)args->args[i];
+	augmented = false;
+	size = beauty_map[i];
+	aug_size = size; /* size of the augmented data read from user space */
+
+	if (size == 0 || arg == NULL)
+		return 0;
+
+	if (size == 1) { /* string */
+		aug_size = bpf_probe_read_user_str(((struct augmented_arg *)payload_offset)->value, value_size, arg);
+		/* minimum of 0 to pass the verifier */
+		if (aug_size < 0)
+			aug_size = 0;
+
+		augmented = true;
+	} else if (size > 0 && size <= value_size) { /* struct */
+		if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, size, arg))
+			augmented = true;
+	} else if ((int)size < 0 && size >= -6) { /* buffer */
+		index = -(size + 1);
+		barrier_var(index); // Prevent clang (noticed with v18) from removing the &= 7 trick.
+		index &= 7;	    // Satisfy the bounds checking with the verifier in some kernels.
+		aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
+
+		if (aug_size > 0) {
+			if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, aug_size, arg))
+				augmented = true;
+		}
+	}
+
+	/* Augmented data size is limited to sizeof(augmented_arg->unnamed union with value field) */
+	if (aug_size > value_size)
+		aug_size = value_size;
+
+	/* write data to payload */
+	if (augmented) {
+		int written = offsetof(struct augmented_arg, value) + aug_size;
+
+		if (written < 0 || written > sizeof(struct augmented_arg))
+			return -1;
+
+		((struct augmented_arg *)payload_offset)->size = aug_size;
+		return written;
+	}
+
+	return 0;
+}
+
 static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 {
-	bool augmented, do_output = false;
-	int zero = 0, index, value_size = sizeof(struct augmented_arg) - offsetof(struct augmented_arg, value);
+	bool do_output = false;
+	int zero = 0, written;
 	u64 output = 0; /* has to be u64, otherwise it won't pass the verifier */
-	s64 aug_size, size;
 	unsigned int nr, *beauty_map;
 	struct beauty_payload_enter *payload;
-	void *arg, *payload_offset;
+	void *payload_offset;
 
 	/* fall back to do predefined tail call */
 	if (args == NULL)

@@ -457,58 +521,11 @@ static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 	/* copy the sys_enter header, which has the syscall_nr */
 	__builtin_memcpy(&payload->args, args, sizeof(struct syscall_enter_args));
 
-	/*
-	 * Determine what type of argument and how many bytes to read from user space, using the
-	 * value in the beauty_map. This is the relation of parameter type and its corresponding
-	 * value in the beauty map, and how many bytes we read eventually:
-	 *
-	 * string: 1			      -> size of string
-	 * struct: size of struct	      -> size of struct
-	 * buffer: -1 * (index of paired len) -> value of paired len (maximum: TRACE_AUG_MAX_BUF)
-	 */
 	for (int i = 0; i < 6; i++) {
-		arg = (void *)args->args[i];
-		augmented = false;
-		size = beauty_map[i];
-		aug_size = size; /* size of the augmented data read from user space */
-
-		if (size == 0 || arg == NULL)
-			continue;
-
-		if (size == 1) { /* string */
-			aug_size = bpf_probe_read_user_str(((struct augmented_arg *)payload_offset)->value, value_size, arg);
-			/* minimum of 0 to pass the verifier */
-			if (aug_size < 0)
-				aug_size = 0;
-
-			augmented = true;
-		} else if (size > 0 && size <= value_size) { /* struct */
-			if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, size, arg))
-				augmented = true;
-		} else if ((int)size < 0 && size >= -6) { /* buffer */
-			index = -(size + 1);
-			barrier_var(index); // Prevent clang (noticed with v18) from removing the &= 7 trick.
-			index &= 7;	    // Satisfy the bounds checking with the verifier in some kernels.
-			aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
-
-			if (aug_size > 0) {
-				if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, aug_size, arg))
-					augmented = true;
-			}
-		}
-
-		/* Augmented data size is limited to sizeof(augmented_arg->unnamed union with value field) */
-		if (aug_size > value_size)
-			aug_size = value_size;
-
-		/* write data to payload */
-		if (augmented) {
-			int written = offsetof(struct augmented_arg, value) + aug_size;
-
-			if (written < 0 || written > sizeof(struct augmented_arg))
-				return 1;
-
-			((struct augmented_arg *)payload_offset)->size = aug_size;
+		written = augment_arg(args, i, beauty_map, payload_offset);
+		if (written < 0)
+			return 1;
+		if (written > 0) {
 			output += written;
 			payload_offset += written;
 			do_output = true;

Signature · 1 line

2.54.0

The loop for processing syscall args in augment_raw_syscalls has a history of breaking with Clang updates, see e.g. commit 013eb043f37b ("perf trace: Fix BPF loading failure (-E2BIG)") from Clang 15 to 16.

Now, a similar thing happened between Clang 21 and 22. While the issue is mitigated on the main line by a recent verifier update, it remains broken on the 6.12 and 6.18 stable branches:

[linux-6.18.y]# sudo perf trace true
libbpf: prog 'sys_enter': BPF program load failed: -E2BIG
libbpf: prog 'sys_enter': -- BEGIN PROG LOAD LOG --
[...]
BPF program is too large. Processed 1000001 insn
processed 1000001 insns (limit 1000000) max_states_per_insn 40 total_states 37941 peak_states 232 mark_read 0 -- END PROG LOAD LOG --
libbpf: prog 'sys_enter': failed to load: -E2BIG
libbpf: failed to load object 'augmented_raw_syscalls_bpf'
libbpf: failed to load BPF skeleton 'augmented_raw_syscalls_bpf': -E2BIG
Error: failed to get syscall or beauty map fd
[...]

The reason is that the loop is quite complex and the BPF verifier often struggles to prove that it terminates.

Fix the issue by replacing the standard for loop by the bpf_for macro, which uses numeric BPF iterator. This should prevent future breakages of this kind since the verifier has much easier job proving that the loop terminates.

Small adjustments were necessary for the loop to make it work. The main problem is that the verifier has sometimes problems with bpf_for loops that use a carry-over state, such as the payload_offset and output vars here, since the verifier tries to track their values too precisely and cannot prove loop convergence. To resolve the issue, we (1) explicitly recompute payload_offset in every iteration and (2) use a trick with adding a global zero to output to help verifier forget its precise state and use a range instead.

In exchange, this also allows to drop a few artificial checks to help the verifier, including the changes introduced by 013eb043f37b.

Finally, to keep backwards compatibility with older kernel versions which do not have bpf_for (i.e. numeric iterators), fall back to standard for loop in such a case.

Signed-off-by Viktor Malik <[email protected]>

Suggested-by Andrii Nakryiko <[email protected]>

Fixes a68fd6a6cdd3 ("perf trace: Collect augmented data using BPF")

Fixes 013eb043f37b ("perf trace: Fix BPF loading failure (-E2BIG)")

Cc [email protected]

---
 .../bpf_skel/augmented_raw_syscalls.bpf.c     | 54 +++++++++++++------
 1 file changed, 39 insertions(+), 15 deletions(-)

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +27 -15

index cbdd5ce19a2f..60babc06f381 100644
@@ -429,6 +429,8 @@ static bool pid_filter__has(struct pids_filtered *pids, pid_t pid)
 	return bpf_map_lookup_elem(pids, &pid) != NULL;
 }
 
+u64 ZERO = 0;
+
 /*
  * Determine what type of argument and how many bytes to read from user space, using the
  * value in the beauty_map. This is the relation of parameter type and its corresponding

@@ -439,12 +441,13 @@ static bool pid_filter__has(struct pids_filtered *pids, pid_t pid)
  * buffer: -1 * (index of paired len) -> value of paired len (maximum: TRACE_AUG_MAX_BUF)
  */
 static inline int augment_arg(struct syscall_enter_args *args, int i,
-			      unsigned int *beauty_map, void *payload_offset)
+			      unsigned int *beauty_map,
+			      struct beauty_payload_enter *payload, u64 offset)
 {
 	int index, value_size = sizeof(struct augmented_arg) - offsetof(struct augmented_arg, value);
 	s64 aug_size, size;
 	bool augmented;
-	void *arg;
+	void *arg, *payload_offset;
 
 	arg = (void *)args->args[i];
 	augmented = false;

@@ -454,6 +457,12 @@ static inline int augment_arg(struct syscall_enter_args *args, int i,
 	if (size == 0 || arg == NULL)
 		return 0;
 
+	/* bounds check for the verifier */
+	if (offset > sizeof(payload->aug_args) - sizeof(payload->aug_args[0]))
+		return -1;
+	barrier_var(offset);
+	payload_offset = (void *)&payload->aug_args + offset;
+
 	if (size == 1) { /* string */
 		aug_size = bpf_probe_read_user_str(((struct augmented_arg *)payload_offset)->value, value_size, arg);
 		/* minimum of 0 to pass the verifier */

@@ -464,11 +473,13 @@ static inline int augment_arg(struct syscall_enter_args *args, int i,
 	} else if (size > 0 && size <= value_size) { /* struct */
 		if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, size, arg))
 			augmented = true;
-	} else if ((int)size < 0 && size >= -6) { /* buffer */
+	} else if (size < 0 && size >= -6) { /* buffer */
 		index = -(size + 1);
 		barrier_var(index); // Prevent clang (noticed with v18) from removing the &= 7 trick.
 		index &= 7;	    // Satisfy the bounds checking with the verifier in some kernels.
-		aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
+		aug_size = args->args[index];
+		if (aug_size > TRACE_AUG_MAX_BUF)
+			aug_size = TRACE_AUG_MAX_BUF;
 
 		if (aug_size > 0) {
 			if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, aug_size, arg))

@@ -497,11 +508,10 @@ static inline int augment_arg(struct syscall_enter_args *args, int i,
 static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 {
 	bool do_output = false;
-	int zero = 0, written;
+	int i, zero = 0, written;
 	u64 output = 0; /* has to be u64, otherwise it won't pass the verifier */
 	unsigned int nr, *beauty_map;
 	struct beauty_payload_enter *payload;
-	void *payload_offset;
 
 	/* fall back to do predefined tail call */
 	if (args == NULL)

@@ -513,7 +523,6 @@ static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 
 	/* set up payload for output */
 	payload        = bpf_map_lookup_elem(&beauty_payload_enter_map, &zero);
-	payload_offset = (void *)&payload->aug_args;
 
 	if (beauty_map == NULL || payload == NULL)
 		return 1;

@@ -521,14 +530,29 @@ static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 	/* copy the sys_enter header, which has the syscall_nr */
 	__builtin_memcpy(&payload->args, args, sizeof(struct syscall_enter_args));
 
-	for (int i = 0; i < 6; i++) {
-		written = augment_arg(args, i, beauty_map, payload_offset);
-		if (written < 0)
-			return 1;
-		if (written > 0) {
-			output += written;
-			payload_offset += written;
-			do_output = true;
+	if (bpf_ksym_exists(bpf_iter_num_new)) {
+		bpf_for(i, 0, 6) {
+			written = augment_arg(args, i, beauty_map, payload, output);
+			if (written < 0)
+				return 1;
+			if (written > 0) {
+				output += written;
+				/* guide the verifier to forget range of `output`, which
+				 * helps to prove convergence of the loop
+				 */
+				output += ZERO;

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +2 -0

+				do_output = true;
+			}
 		}
 	}
 

Signature · 1 line

2.54.0

Hello,

Quoted 19 lines · Viktor Malik

The BPF program in augmented_raw_syscalls uses a for loop to iterate all syscall arguments. The loop body is quite complex and often poses problems for the BPF verifier. As a preparation step for addressing this issue, factor out the loop body into a separate function.

Signed-off-by Viktor Malik <[email protected]>

Cc [email protected]

---
 .../bpf_skel/augmented_raw_syscalls.bpf.c     | 127 ++++++++++--------
 1 file changed, 72 insertions(+), 55 deletions(-)

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +0 -0

index 2a6e61864ee0..cbdd5ce19a2f 100644
@@ -429,15 +429,79 @@ static bool pid_filter__has(struct pids_filtered *pids, pid_t pid)
 	return bpf_map_lookup_elem(pids, &pid) != NULL;
 }
 

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c static bool pid_filter__has(struct pids_filtered *pids, pid_t pid) · line 442

Quoted 11 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +11 -0

+/*
+ * Determine what type of argument and how many bytes to read from user space, using the
+ * value in the beauty_map. This is the relation of parameter type and its corresponding
+ * value in the beauty map, and how many bytes we read eventually:
+ *
+ * string: 1			      -> size of string
+ * struct: size of struct	      -> size of struct
+ * buffer: -1 * (index of paired len) -> value of paired len (maximum: TRACE_AUG_MAX_BUF)
+ */
+static inline int augment_arg(struct syscall_enter_args *args, int i,
+			      unsigned int *beauty_map, void *payload_offset)

Can we make it 'struct augmented_arg *payload_offset' instead?

Thanks,
Namhyung

Quoted 135 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +61 -55

+{
+	int index, value_size = sizeof(struct augmented_arg) - offsetof(struct augmented_arg, value);
+	s64 aug_size, size;
+	bool augmented;
+	void *arg;
+
+	arg = (void *)args->args[i];
+	augmented = false;
+	size = beauty_map[i];
+	aug_size = size; /* size of the augmented data read from user space */
+
+	if (size == 0 || arg == NULL)
+		return 0;
+
+	if (size == 1) { /* string */
+		aug_size = bpf_probe_read_user_str(((struct augmented_arg *)payload_offset)->value, value_size, arg);
+		/* minimum of 0 to pass the verifier */
+		if (aug_size < 0)
+			aug_size = 0;
+
+		augmented = true;
+	} else if (size > 0 && size <= value_size) { /* struct */
+		if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, size, arg))
+			augmented = true;
+	} else if ((int)size < 0 && size >= -6) { /* buffer */
+		index = -(size + 1);
+		barrier_var(index); // Prevent clang (noticed with v18) from removing the &= 7 trick.
+		index &= 7;	    // Satisfy the bounds checking with the verifier in some kernels.
+		aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
+
+		if (aug_size > 0) {
+			if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, aug_size, arg))
+				augmented = true;
+		}
+	}
+
+	/* Augmented data size is limited to sizeof(augmented_arg->unnamed union with value field) */
+	if (aug_size > value_size)
+		aug_size = value_size;
+
+	/* write data to payload */
+	if (augmented) {
+		int written = offsetof(struct augmented_arg, value) + aug_size;
+
+		if (written < 0 || written > sizeof(struct augmented_arg))
+			return -1;
+
+		((struct augmented_arg *)payload_offset)->size = aug_size;
+		return written;
+	}
+
+	return 0;
+}
+
 static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 {
-	bool augmented, do_output = false;
-	int zero = 0, index, value_size = sizeof(struct augmented_arg) - offsetof(struct augmented_arg, value);
+	bool do_output = false;
+	int zero = 0, written;
 	u64 output = 0; /* has to be u64, otherwise it won't pass the verifier */
-	s64 aug_size, size;
 	unsigned int nr, *beauty_map;
 	struct beauty_payload_enter *payload;
-	void *arg, *payload_offset;
+	void *payload_offset;
 
 	/* fall back to do predefined tail call */
 	if (args == NULL)

@@ -457,58 +521,11 @@ static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 	/* copy the sys_enter header, which has the syscall_nr */
 	__builtin_memcpy(&payload->args, args, sizeof(struct syscall_enter_args));
 
-	/*
-	 * Determine what type of argument and how many bytes to read from user space, using the
-	 * value in the beauty_map. This is the relation of parameter type and its corresponding
-	 * value in the beauty map, and how many bytes we read eventually:
-	 *
-	 * string: 1			      -> size of string
-	 * struct: size of struct	      -> size of struct
-	 * buffer: -1 * (index of paired len) -> value of paired len (maximum: TRACE_AUG_MAX_BUF)
-	 */
 	for (int i = 0; i < 6; i++) {
-		arg = (void *)args->args[i];
-		augmented = false;
-		size = beauty_map[i];
-		aug_size = size; /* size of the augmented data read from user space */
-
-		if (size == 0 || arg == NULL)
-			continue;
-
-		if (size == 1) { /* string */
-			aug_size = bpf_probe_read_user_str(((struct augmented_arg *)payload_offset)->value, value_size, arg);
-			/* minimum of 0 to pass the verifier */
-			if (aug_size < 0)
-				aug_size = 0;
-
-			augmented = true;
-		} else if (size > 0 && size <= value_size) { /* struct */
-			if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, size, arg))
-				augmented = true;
-		} else if ((int)size < 0 && size >= -6) { /* buffer */
-			index = -(size + 1);
-			barrier_var(index); // Prevent clang (noticed with v18) from removing the &= 7 trick.
-			index &= 7;	    // Satisfy the bounds checking with the verifier in some kernels.
-			aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
-
-			if (aug_size > 0) {
-				if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, aug_size, arg))
-					augmented = true;
-			}
-		}
-
-		/* Augmented data size is limited to sizeof(augmented_arg->unnamed union with value field) */
-		if (aug_size > value_size)
-			aug_size = value_size;
-
-		/* write data to payload */
-		if (augmented) {
-			int written = offsetof(struct augmented_arg, value) + aug_size;
-
-			if (written < 0 || written > sizeof(struct augmented_arg))
-				return 1;
-
-			((struct augmented_arg *)payload_offset)->size = aug_size;
+		written = augment_arg(args, i, beauty_map, payload_offset);
+		if (written < 0)
+			return 1;
+		if (written > 0) {
 			output += written;
 			payload_offset += written;
 			do_output = true;

Signature · 1 line

2.54.0

Quoted 160 lines · Viktor Malik

The loop for processing syscall args in augment_raw_syscalls has a history of breaking with Clang updates, see e.g. commit 013eb043f37b ("perf trace: Fix BPF loading failure (-E2BIG)") from Clang 15 to 16.

Now, a similar thing happened between Clang 21 and 22. While the issue is mitigated on the main line by a recent verifier update, it remains broken on the 6.12 and 6.18 stable branches:

[linux-6.18.y]# sudo perf trace true
libbpf: prog 'sys_enter': BPF program load failed: -E2BIG
libbpf: prog 'sys_enter': -- BEGIN PROG LOAD LOG --
[...]
BPF program is too large. Processed 1000001 insn
processed 1000001 insns (limit 1000000) max_states_per_insn 40 total_states 37941 peak_states 232 mark_read 0 -- END PROG LOAD LOG --
libbpf: prog 'sys_enter': failed to load: -E2BIG
libbpf: failed to load object 'augmented_raw_syscalls_bpf'
libbpf: failed to load BPF skeleton 'augmented_raw_syscalls_bpf': -E2BIG
Error: failed to get syscall or beauty map fd
[...]

The reason is that the loop is quite complex and the BPF verifier often struggles to prove that it terminates.

Fix the issue by replacing the standard for loop by the bpf_for macro, which uses numeric BPF iterator. This should prevent future breakages of this kind since the verifier has much easier job proving that the loop terminates.

Small adjustments were necessary for the loop to make it work. The main problem is that the verifier has sometimes problems with bpf_for loops that use a carry-over state, such as the payload_offset and output vars here, since the verifier tries to track their values too precisely and cannot prove loop convergence. To resolve the issue, we (1) explicitly recompute payload_offset in every iteration and (2) use a trick with adding a global zero to output to help verifier forget its precise state and use a range instead.

In exchange, this also allows to drop a few artificial checks to help the verifier, including the changes introduced by 013eb043f37b.

Finally, to keep backwards compatibility with older kernel versions which do not have bpf_for (i.e. numeric iterators), fall back to standard for loop in such a case.

Signed-off-by Viktor Malik <[email protected]>

Suggested-by Andrii Nakryiko <[email protected]>

Fixes a68fd6a6cdd3 ("perf trace: Collect augmented data using BPF")

Fixes 013eb043f37b ("perf trace: Fix BPF loading failure (-E2BIG)")

Cc [email protected]

---
 .../bpf_skel/augmented_raw_syscalls.bpf.c     | 54 +++++++++++++------
 1 file changed, 39 insertions(+), 15 deletions(-)

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +31 -15

index cbdd5ce19a2f..60babc06f381 100644
@@ -429,6 +429,8 @@ static bool pid_filter__has(struct pids_filtered *pids, pid_t pid)
 	return bpf_map_lookup_elem(pids, &pid) != NULL;
 }
 
+u64 ZERO = 0;
+
 /*
  * Determine what type of argument and how many bytes to read from user space, using the
  * value in the beauty_map. This is the relation of parameter type and its corresponding

@@ -439,12 +441,13 @@ static bool pid_filter__has(struct pids_filtered *pids, pid_t pid)
  * buffer: -1 * (index of paired len) -> value of paired len (maximum: TRACE_AUG_MAX_BUF)
  */
 static inline int augment_arg(struct syscall_enter_args *args, int i,
-			      unsigned int *beauty_map, void *payload_offset)
+			      unsigned int *beauty_map,
+			      struct beauty_payload_enter *payload, u64 offset)
 {
 	int index, value_size = sizeof(struct augmented_arg) - offsetof(struct augmented_arg, value);
 	s64 aug_size, size;
 	bool augmented;
-	void *arg;
+	void *arg, *payload_offset;
 
 	arg = (void *)args->args[i];
 	augmented = false;

@@ -454,6 +457,12 @@ static inline int augment_arg(struct syscall_enter_args *args, int i,
 	if (size == 0 || arg == NULL)
 		return 0;
 
+	/* bounds check for the verifier */
+	if (offset > sizeof(payload->aug_args) - sizeof(payload->aug_args[0]))
+		return -1;
+	barrier_var(offset);
+	payload_offset = (void *)&payload->aug_args + offset;
+
 	if (size == 1) { /* string */
 		aug_size = bpf_probe_read_user_str(((struct augmented_arg *)payload_offset)->value, value_size, arg);
 		/* minimum of 0 to pass the verifier */

@@ -464,11 +473,13 @@ static inline int augment_arg(struct syscall_enter_args *args, int i,
 	} else if (size > 0 && size <= value_size) { /* struct */
 		if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, size, arg))
 			augmented = true;
-	} else if ((int)size < 0 && size >= -6) { /* buffer */
+	} else if (size < 0 && size >= -6) { /* buffer */
 		index = -(size + 1);
 		barrier_var(index); // Prevent clang (noticed with v18) from removing the &= 7 trick.
 		index &= 7;	    // Satisfy the bounds checking with the verifier in some kernels.
-		aug_size = args->args[index] > TRACE_AUG_MAX_BUF ? TRACE_AUG_MAX_BUF : args->args[index];
+		aug_size = args->args[index];
+		if (aug_size > TRACE_AUG_MAX_BUF)
+			aug_size = TRACE_AUG_MAX_BUF;
 
 		if (aug_size > 0) {
 			if (!bpf_probe_read_user(((struct augmented_arg *)payload_offset)->value, aug_size, arg))

@@ -497,11 +508,10 @@ static inline int augment_arg(struct syscall_enter_args *args, int i,
 static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 {
 	bool do_output = false;
-	int zero = 0, written;
+	int i, zero = 0, written;
 	u64 output = 0; /* has to be u64, otherwise it won't pass the verifier */
 	unsigned int nr, *beauty_map;
 	struct beauty_payload_enter *payload;
-	void *payload_offset;
 
 	/* fall back to do predefined tail call */
 	if (args == NULL)

@@ -513,7 +523,6 @@ static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 
 	/* set up payload for output */
 	payload        = bpf_map_lookup_elem(&beauty_payload_enter_map, &zero);
-	payload_offset = (void *)&payload->aug_args;
 
 	if (beauty_map == NULL || payload == NULL)
 		return 1;

@@ -521,14 +530,29 @@ static int augment_sys_enter(void *ctx, struct syscall_enter_args *args)
 	/* copy the sys_enter header, which has the syscall_nr */
 	__builtin_memcpy(&payload->args, args, sizeof(struct syscall_enter_args));
 
-	for (int i = 0; i < 6; i++) {
-		written = augment_arg(args, i, beauty_map, payload_offset);
-		if (written < 0)
-			return 1;
-		if (written > 0) {
-			output += written;
-			payload_offset += written;
-			do_output = true;
+	if (bpf_ksym_exists(bpf_iter_num_new)) {
+		bpf_for(i, 0, 6) {
+			written = augment_arg(args, i, beauty_map, payload, output);
+			if (written < 0)
+				return 1;
+			if (written > 0) {
+				output += written;
+				/* guide the verifier to forget range of `output`, which
+				 * helps to prove convergence of the loop
+				 */
+				output += ZERO;
+				do_output = true;
+			}
+		}
+	} else {

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c static int augment_sys_enter(void *ctx, struct syscall_enter_args *args) · line 553

Quoted 6 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +6 -0

+		for (i = 0; i < 6; i++) {
+			written = augment_arg(args, i, beauty_map, payload, output);
+			if (written < 0)
+				return 1;
+			if (written > 0) {
+				output += written;

Woundn't it also need '+= ZERO' here?

Thanks,
Namhyung

Quoted 8 lines

tools/perf/util/bpf_skel/augmented_raw_syscalls.bpf.c +2 -0

+				do_output = true;
+			}
 		}
 	}
 

Signature · 1 line

2.54.0