Skip to contentSkip to conversation

selftests/bpf: Use %p for pointer formatting

PATCH In review Sebastian Andrzej Siewior 2 participants 3 messages Sep 18

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 selftests/bpf: Use %p for pointer formatting

Networking used to print the socket pointer with %p as output in the /proc interface. This was later changed to %pK in order not to reveal the actual pointer value. The output format has then been copied to bpf tests which produce the same format including the %pK format modifier.

Networking recently replaced the socket output with a plain 0 because the socket pointer added no value to the output and the %pK should be removed from library handling. The format remained otherwise unchanged in order not to break any tools which parsing this information.

This change removes the %pK modifier from bpf, following the change in networking.

---
v1…v2: https://lore.kernel.org/all/[email protected]/
- sync dump_udp4(), dump_udp6() with net (0, spacing). The columns are (were) not aligned.
- Update comment in snprintf.c

 tools/testing/selftests/bpf/README.rst               |  2 +-
 tools/testing/selftests/bpf/prog_tests/snprintf.c    |  4 ++--
 tools/testing/selftests/bpf/progs/bpf_iter_netlink.c |  4 ++--
 tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c    | 11 +++++------
 tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c    | 11 +++++------
 tools/testing/selftests/bpf/progs/bpf_iter_udp4.c    |  4 ++--
 tools/testing/selftests/bpf/progs/bpf_iter_udp6.c    |  4 ++--
 tools/testing/selftests/bpf/progs/bpf_iter_unix.c    |  5 ++---
 tools/testing/selftests/bpf/progs/test_snprintf.c    |  4 ++--
 9 files changed, 23 insertions(+), 26 deletions(-)
tools/testing/selftests/bpf/README.rst +1 -1
index 07c834433b388..75ff1142c1857 100644
@@ -187,7 +187,7 @@ The symptom for ``bpf_iter/netlink`` looks like
   ...
   15: (bf) r2 = r7
   16: (0f) r2 += r1
-  ; BPF_SEQ_PRINTF(seq, "%pK %-3d ", s, s->sk_protocol);
+  ; BPF_SEQ_PRINTF(seq, "%p %-3d ", s, s->sk_protocol);
   17: (7b) *(u64 *)(r7 +0) = r2
   only read is supported
 
tools/testing/selftests/bpf/prog_tests/snprintf.c +2 -2
index 4e4a82d54f799..ccfb02bd4e7cd 100644
@@ -15,8 +15,8 @@
 #define EXP_SYM_OUT  "schedule schedule+0x0/"
 #define MIN_SYM_RET  sizeof(EXP_SYM_OUT)
 
-/* The third specifier, %p, is a hashed pointer which changes on every reboot */
-#define EXP_ADDR_OUT "0000000000000000 ffff00000add4e55 "
+/* The second specifier, %p, is a hashed pointer which changes on every reboot */
+#define EXP_ADDR_OUT "ffff00000add4e55 "
 #define EXP_ADDR_RET sizeof(EXP_ADDR_OUT "unknownhashedptr")
 
 #define EXP_STR_OUT  "str1         a  b c      d e longstr"
tools/testing/selftests/bpf/progs/bpf_iter_netlink.c +2 -2
index 00b2ceae81fb0..b9107d3ab6039 100644
@@ -25,12 +25,12 @@ int dump_netlink(struct bpf_iter__netlink *ctx)
 		return 0;
 
 	if (ctx->meta->seq_num == 0)
-		BPF_SEQ_PRINTF(seq, "sk               Eth Pid        Groups   "
+		BPF_SEQ_PRINTF(seq, "sk Eth Pid        Groups   "
 				    "Rmem     Wmem     Dump  Locks    Drops    "
 				    "Inode\n");
 
 	s = &nlk->sk;
-	BPF_SEQ_PRINTF(seq, "%pK %-3d ", s, s->sk_protocol);
+	BPF_SEQ_PRINTF(seq, "0  %-3d ", s->sk_protocol);
 
 	if (!nlk->groups)  {
 		group = 0;
tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c +5 -6
index b1e509b231cd9..d149cf0d07be6 100644
@@ -131,8 +131,7 @@ static int dump_tcp_sock(struct seq_file *seq, struct tcp_sock *tp,
 		       icsk->icsk_probes_out,
 		       sock_i_ino(sp),
 		       sp->sk_refcnt.refs.counter);
-	BPF_SEQ_PRINTF(seq, "%pK %lu %lu %u %u %d\n",
-		       tp,
+	BPF_SEQ_PRINTF(seq, "0 %lu %lu %u %u %d\n",
 		       jiffies_to_clock_t(icsk->icsk_rto),
 		       jiffies_to_clock_t(icsk->icsk_ack.ato),
 		       (icsk->icsk_ack.quick << 1) | inet_csk_in_pingpong_mode(icsk),
@@ -161,10 +160,10 @@ static int dump_tw_sock(struct seq_file *seq, struct tcp_timewait_sock *ttw,
 	BPF_SEQ_PRINTF(seq, "%4d: %08X:%04X %08X:%04X ",
 		       seq_num, src, srcp, dest, destp);
 
-	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %pK\n",
+	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d 0\n",
 		       tw->tw_substate, 0, 0,
 		       3, jiffies_delta_to_clock_t(delta), 0, 0, 0, 0,
-		       tw->tw_refcnt.refs.counter, tw);
+		       tw->tw_refcnt.refs.counter);
 
 	return 0;
 }
@@ -185,9 +184,9 @@ static int dump_req_sock(struct seq_file *seq, struct tcp_request_sock *treq,
 		       seq_num, irsk->ir_loc_addr,
 		       irsk->ir_num, irsk->ir_rmt_addr,
 		       bpf_ntohs(irsk->ir_rmt_port));
-	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %pK\n",
+	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d 0\n",
 		       TCP_SYN_RECV, 0, 0, 1, jiffies_to_clock_t(ttd),
-		       req->num_timeout, uid, 0, 0, 0, req);
+		       req->num_timeout, uid, 0, 0, 0);
 
 	return 0;
 }
tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c +5 -6
index dbc7166aee91f..cd65fa81f1219 100644
@@ -135,8 +135,7 @@ static int dump_tcp6_sock(struct seq_file *seq, struct tcp6_sock *tp,
 		       icsk->icsk_probes_out,
 		       sock_i_ino(sp),
 		       sp->sk_refcnt.refs.counter);
-	BPF_SEQ_PRINTF(seq, "%pK %lu %lu %u %u %d\n",
-		       tp,
+	BPF_SEQ_PRINTF(seq, "0 %lu %lu %u %u %d\n",
 		       jiffies_to_clock_t(icsk->icsk_rto),
 		       jiffies_to_clock_t(icsk->icsk_ack.ato),
 		       (icsk->icsk_ack.quick << 1) | inet_csk_in_pingpong_mode(icsk),
@@ -170,10 +169,10 @@ static int dump_tw_sock(struct seq_file *seq, struct tcp_timewait_sock *ttw,
 		       dest->s6_addr32[0], dest->s6_addr32[1],
 		       dest->s6_addr32[2], dest->s6_addr32[3], destp);
 
-	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %pK\n",
+	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d 0\n",
 		       tw->tw_substate, 0, 0,
 		       3, jiffies_delta_to_clock_t(delta), 0, 0, 0, 0,
-		       tw->tw_refcnt.refs.counter, tw);
+		       tw->tw_refcnt.refs.counter);
 
 	return 0;
 }
@@ -201,9 +200,9 @@ static int dump_req_sock(struct seq_file *seq, struct tcp_request_sock *treq,
 		       dest->s6_addr32[0], dest->s6_addr32[1],
 		       dest->s6_addr32[2], dest->s6_addr32[3],
 		       bpf_ntohs(irsk->ir_rmt_port));
-	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %pK\n",
+	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d 0\n",
 		       TCP_SYN_RECV, 0, 0, 1, jiffies_to_clock_t(ttd),
-		       req->num_timeout, uid, 0, 0, 0, req);
+		       req->num_timeout, uid, 0, 0, 0);
 
 	return 0;
 }
tools/testing/selftests/bpf/progs/bpf_iter_udp4.c +2 -2
index 23b2aa2604de2..d4079a619da66 100644
@@ -57,13 +57,13 @@ int dump_udp4(struct bpf_iter__udp *ctx)
 	BPF_SEQ_PRINTF(seq, "%5d: %08X:%04X %08X:%04X ",
 		       ctx->bucket, src, srcp, dest, destp);
 
-	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d %pK %u\n",
+	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d 0 %u\n",
 		       inet->sk.sk_state,
 		       inet->sk.sk_wmem_alloc.refs.counter - 1,
 		       rqueue,
 		       0, 0L, 0, ctx->uid, 0,
 		       sock_i_ino(&inet->sk),
-		       inet->sk.sk_refcnt.refs.counter, udp_sk,
+		       inet->sk.sk_refcnt.refs.counter,
 		       udp_sk->drop_counters.drops0.counter +
 		       udp_sk->drop_counters.drops1.counter);
 
tools/testing/selftests/bpf/progs/bpf_iter_udp6.c +2 -2
index c48b05aa2a4b2..dadc5f73ff098 100644
@@ -65,13 +65,13 @@ int dump_udp6(struct bpf_iter__udp *ctx)
 		       dest->s6_addr32[0], dest->s6_addr32[1],
 		       dest->s6_addr32[2], dest->s6_addr32[3], destp);
 
-	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d %pK %u\n",
+	BPF_SEQ_PRINTF(seq, "%02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d 0 %u\n",
 		       inet->sk.sk_state,
 		       inet->sk.sk_wmem_alloc.refs.counter - 1,
 		       rqueue,
 		       0, 0L, 0, ctx->uid, 0,
 		       sock_i_ino(&inet->sk),
-		       inet->sk.sk_refcnt.refs.counter, udp_sk,
+		       inet->sk.sk_refcnt.refs.counter,
 		       udp_sk->drop_counters.drops0.counter +
 		       udp_sk->drop_counters.drops1.counter);
 	return 0;
tools/testing/selftests/bpf/progs/bpf_iter_unix.c +2 -3
index a2652c8c3616f..df3d971b21e8a 100644
@@ -42,10 +42,9 @@ int dump_unix(struct bpf_iter__unix *ctx)
 	seq = ctx->meta->seq;
 	seq_num = ctx->meta->seq_num;
 	if (seq_num == 0)
-		BPF_SEQ_PRINTF(seq, "Num               RefCount Protocol Flags    Type St    Inode Path\n");
+		BPF_SEQ_PRINTF(seq, "Num       RefCount Protocol Flags    Type St    Inode Path\n");
 
-	BPF_SEQ_PRINTF(seq, "%pK: %08X %08X %08X %04X %02X %8lu",
-		       unix_sk,
+	BPF_SEQ_PRINTF(seq, "%08u: %08X %08X %08X %04X %02X %8lu", 0,
 		       sk->sk_refcnt.refs.counter,
 		       0,
 		       sk->sk_state == TCP_LISTEN ? __SO_ACCEPTCON : 0,
tools/testing/selftests/bpf/progs/test_snprintf.c +2 -2
index 8fda075440235..48a0025cf2ca8 100644
@@ -57,8 +57,8 @@ int handler(const void *ctx)
 	sym_ret  = BPF_SNPRINTF(sym_out,  sizeof(sym_out), "%ps %pS %pB",
 				&schedule, &schedule, &schedule);
 	/* Kernel pointers */
-	addr_ret = BPF_SNPRINTF(addr_out, sizeof(addr_out), "%pK %px %p",
-				0, 0xFFFF00000ADD4E55, 0xFFFF00000ADD4E55);
+	addr_ret = BPF_SNPRINTF(addr_out, sizeof(addr_out), "%px %p",
+				0xFFFF00000ADD4E55, 0xFFFF00000ADD4E55);
 	/* Strings and single-byte character embedding */
 	str_ret  = BPF_SNPRINTF(str_out, sizeof(str_out), "%s % 9c %+2c %-3c %04c %0c %+05s",
 				str1, 'a', 'b', 'c', 'd', 'e', longstr);
Signature · 1 line
2.55.0
#1 lore raw
Does this subject line accurately describe what the patch does?
Quoted 7 lines

selftests/bpf: Use %p for pointer formatting

Networking used to print the socket pointer with %p as output in the /proc interface. This was later changed to %pK in order not to reveal the actual pointer value. The output format has then been copied to bpf tests which produce the same format including the %pK format modifier.

Quoted 7 lines

Networking recently replaced the socket output with a plain 0 because the socket pointer added no value to the output and the %pK should be removed from library handling. The format remained otherwise unchanged in order not to break any tools which parsing this information.

This change removes the %pK modifier from bpf, following the change in networking.

Does this subject line accurately describe what the patch does?

Looking at the actual changes, no %p format specifier is introduced anywhere in the test code:

- bpf_iter_netlink.c gets a literal "0 " string
- bpf_iter_tcp4.c / bpf_iter_tcp6.c / bpf_iter_udp4.c / bpf_iter_udp6.c get a literal "0" string
- bpf_iter_unix.c gets "%08u" with a constant 0 argument
- test_snprintf.c simply drops the %pK specifier (its remaining %p was already present before this patch)

The only place where %p appears in the patch is in README.rst, in a historical verifier log excerpt.

The body text says "This change removes the %pK modifier from bpf", which matches the v1 subject "selftests/bpf: Replace %pK output with 0" rather than this v2 subject.

A few additional points:

The changelog is unchanged from v1 and doesn't mention the header/column-width edits this version adds:
- bpf_iter_netlink.c "sk" column changed from 17 to 3 chars
- bpf_iter_unix.c "Num" column changed from 18 to 10 chars

This seems relevant since the body asserts "The format remained otherwise unchanged in order not to break any tools".

The referenced networking commit isn't named in the message. Looking at the v1 discussion, commit 7c0ec6288b497 ("net: Replace %pK output with 0") was cited below the --- separator, so it's absent from the committed message. That commit doesn't appear to be in this tree yet (checking net/netlink/af_netlink.c, net/unix/af_unix.c, net/ipv4/udp.c and the tcp_ipv[46].c files still use %pK), which would make a Link: or commit reference useful for readers.

Quoted 1 line

order not to break any tools which parsing this information.

This isn't a bug, but there's a grammar error here ("which parsing" should be "which are parsing").

---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/35336391118

#2 lore raw
Quoted 12 lines · On 2026-09-18 11:38:34 [+0000], [email protected] wrote:
Quoted 12 lines

selftests/bpf: Use %p for pointer formatting

Networking used to print the socket pointer with %p as output in the /proc interface. This was later changed to %pK in order not to reveal the actual pointer value. The output format has then been copied to bpf tests which produce the same format including the %pK format modifier.

Networking recently replaced the socket output with a plain 0 because the socket pointer added no value to the output and the %pK should be removed from library handling. The format remained otherwise unchanged in order not to break any tools which parsing this information.

Quoted 4 lines
Quoted 2 lines

This change removes the %pK modifier from bpf, following the change in networking.

Does this subject line accurately describe what the patch does?

Oh. The %pK gets removed, not replaced with %p.

Quoted 10 lines

Looking at the actual changes, no %p format specifier is introduced anywhere in the test code:

- bpf_iter_netlink.c gets a literal "0 " string
- bpf_iter_tcp4.c / bpf_iter_tcp6.c / bpf_iter_udp4.c / bpf_iter_udp6.c get a literal "0" string
- bpf_iter_unix.c gets "%08u" with a constant 0 argument
- test_snprintf.c simply drops the %pK specifier (its remaining %p was already present before this patch)

Quoted 6 lines

The only place where %p appears in the patch is in README.rst, in a historical verifier log excerpt.

The body text says "This change removes the %pK modifier from bpf", which matches the v1 subject "selftests/bpf: Replace %pK output with 0" rather than this v2 subject.

Indeed.

Quoted 9 lines

A few additional points:

The changelog is unchanged from v1 and doesn't mention the header/column-width edits this version adds:
- bpf_iter_netlink.c "sk" column changed from 17 to 3 chars
- bpf_iter_unix.c "Num" column changed from 18 to 10 chars

This seems relevant since the body asserts "The format remained otherwise unchanged in order not to break any tools".

This updates the output so the column width aligns with the output. "Format unchanged" means no items were removed or reordered.

Quoted 8 lines

The referenced networking commit isn't named in the message. Looking at the v1 discussion, commit 7c0ec6288b497 ("net: Replace %pK output with 0") was cited below the --- separator, so it's absent from the committed message. That commit doesn't appear to be in this tree yet (checking net/netlink/af_netlink.c, net/unix/af_unix.c, net/ipv4/udp.c and the tcp_ipv[46].c files still use %pK), which would make a Link: or commit reference useful for readers.

Quoted 4 lines
Quoted 1 line

order not to break any tools which parsing this information.

This isn't a bug, but there's a grammar error here ("which parsing" should be "which are parsing").

Okay.

Sebastian