Skip to contentSkip to conversation

mptcp: Remove all traces of checksum support

PATCH Superseded Peter Krystad 3 participants 5 messages Sep 16, 2019 – Sep 24, 2019

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Peter Krystad

author
#0 lore raw

PATCH 1/1 mptcp: Remove all traces of checksum support

Since checksum support is not planned remove any references that look it may be supported.

Also re-factor DSS option size calculation and make DSS option size check for exactly correct lengths.

Signed-off-by Peter Krystad <peter.krystad(a)linux.intel.com>
---
 net/mptcp/options.c  | 69 +++++++++++++++++---------------------------
 net/mptcp/protocol.c | 13 ++++-----
 net/mptcp/protocol.h |  4 +--
 net/mptcp/subflow.c  |  5 ----
 4 files changed, 32 insertions(+), 59 deletions(-)
net/mptcp/options.c +26 -43
index b82557af3fae..ec56a25e6948 100644
@@ -130,7 +130,6 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 	 */
 	case MPTCPOPT_DSS:
 		pr_debug("DSS");
-		mp_opt->dss = 1;
 		ptr++;
 
 		mp_opt->dss_flags = (*ptr++) & MPTCP_DSS_FLAG_MASK;
@@ -152,10 +151,27 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK64;
 			else
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK32;
+		}
+
+		if (mp_opt->use_map) {
+			if (mp_opt->dsn64)
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP64;
+			else
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP32;
+		}
 
-			if (opsize < expected_opsize)
-				break;
+		/* RFC 6824, Section 3.3:
+		 * If a checksum is present, but its use had
+		 * not been negotiated in the MP_CAPABLE handshake,
+		 * the checksum field MUST be ignored.
+		 */
+		if (opsize != expected_opsize &&
+		    opsize != expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM)
+			break;
 
+		mp_opt->dss = 1;
+
+		if (mp_opt->use_ack) {
 			if (mp_opt->ack64) {
 				mp_opt->data_ack = get_unaligned_be64(ptr);
 				ptr += 8;
@@ -163,19 +179,10 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 				mp_opt->data_ack = get_unaligned_be32(ptr);
 				ptr += 4;
 			}
-
 			pr_debug("data_ack=%llu", mp_opt->data_ack);
 		}
 
 		if (mp_opt->use_map) {
-			if (mp_opt->dsn64)
-				expected_opsize += TCPOLEN_MPTCP_DSS_MAP64;
-			else
-				expected_opsize += TCPOLEN_MPTCP_DSS_MAP32;
-
-			if (opsize < expected_opsize)
-				break;
-
 			if (mp_opt->dsn64) {
 				mp_opt->data_seq = get_unaligned_be64(ptr);
 				ptr += 8;
@@ -190,13 +197,11 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 			mp_opt->data_len = get_unaligned_be16(ptr);
 			ptr += 2;
 
-			/* Checksum not currently supported */
-			mp_opt->checksum = 0;
-
-			pr_debug("data_seq=%llu subflow_seq=%u data_len=%u ck=%u",
+			pr_debug("data_seq=%llu subflow_seq=%u data_len=%u",
 				 mp_opt->data_seq, mp_opt->subflow_seq,
-				 mp_opt->data_len, mp_opt->checksum);
+				 mp_opt->data_len);
 		}
+
 		break;
 
 	/* MPTCPOPT_ADD_ADDR
@@ -390,12 +395,8 @@ static bool mptcp_established_options_dss(struct sock *sk, struct sk_buff *skb,
 
 	if (!skb || (mpext && mpext->use_map)) {
 		unsigned int map_size;
-		bool use_csum;
 
 		map_size = TCPOLEN_MPTCP_DSS_BASE + TCPOLEN_MPTCP_DSS_MAP64;
-		use_csum = subflow_ctx(sk)->use_checksum;
-		if (use_csum)
-			map_size += TCPOLEN_MPTCP_DSS_CHECKSUM;
 
 		if (map_size <= remaining) {
 			remaining -= map_size;
@@ -404,10 +405,8 @@ static bool mptcp_established_options_dss(struct sock *sk, struct sk_buff *skb,
 				opts->ext_copy.data_seq = mpext->data_seq;
 				opts->ext_copy.subflow_seq = mpext->subflow_seq;
 				opts->ext_copy.data_len = mpext->data_len;
-				opts->ext_copy.checksum = mpext->checksum;
 				opts->ext_copy.use_map = 1;
 				opts->ext_copy.dsn64 = mpext->dsn64;
-				opts->ext_copy.use_checksum = use_csum;
 			}
 		} else {
 			opts->ext_copy.use_map = 0;
@@ -573,10 +572,8 @@ void mptcp_incoming_options(struct sock *sk, struct sk_buff *skb,
 		mpext->data_seq = mp_opt->data_seq;
 		mpext->subflow_seq = mp_opt->subflow_seq;
 		mpext->data_len = mp_opt->data_len;
-		mpext->checksum = mp_opt->checksum;
 		mpext->use_map = 1;
 		mpext->dsn64 = mp_opt->dsn64;
-		mpext->use_checksum = mp_opt->use_checksum;
 	}
 
 	if (mp_opt->use_ack) {
@@ -676,12 +673,8 @@ void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
 		}
 
 		if (mpext->use_map) {
-			pr_debug("Updating DSS length and flags for map");
 			len += TCPOLEN_MPTCP_DSS_MAP64;
 
-			if (mpext->use_checksum)
-				len += TCPOLEN_MPTCP_DSS_CHECKSUM;
-
 			/* Use only 64-bit mapping flags for now, add
 			 * support for optional 32-bit mappings later.
 			 */
@@ -690,10 +683,7 @@ void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
 				flags |= MPTCP_DSS_DATA_FIN;
 		}
 
-		*ptr++ = htonl((TCPOPT_MPTCP << 24) |
-			       (len  << 16) |
-			       (MPTCPOPT_DSS << 12) |
-			       (flags));
+		*ptr++ = mptcp_option(MPTCPOPT_DSS, len, 0, flags);
 
 		if (mpext->use_ack) {
 			put_unaligned_be64(mpext->data_ack, ptr);
@@ -701,18 +691,11 @@ void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
 		}
 
 		if (mpext->use_map) {
-			__sum16 checksum;
-
-			pr_debug("Writing map values");
 			put_unaligned_be64(mpext->data_seq, ptr);
 			ptr += 2;
-			*ptr++ = htonl(mpext->subflow_seq);
-
-			if (mpext->use_checksum)
-				checksum = mpext->checksum;
-			else
-				checksum = TCPOPT_NOP << 8 | TCPOPT_NOP;
-			*ptr = htonl(mpext->data_len << 16 | checksum);
+			put_unaligned_be32(mpext->subflow_seq, ptr);
+			ptr += 1;
+			put_unaligned_be16(mpext->data_len, ptr);
 		}
 	}
 }
net/mptcp/protocol.c +5 -8
index 9c019e696fc7..5c6295611d2d 100644
@@ -156,13 +156,12 @@ static int mptcp_sendmsg_frag(struct sock *sk, struct sock *ssk,
 		mpext->data_seq = msk->write_seq;
 		mpext->subflow_seq = subflow_ctx(ssk)->rel_write_seq;
 		mpext->data_len = ret;
-		mpext->checksum = 0xbeef;
 		mpext->use_map = 1;
 		mpext->dsn64 = 1;
 
-		pr_debug("data_seq=%llu subflow_seq=%u data_len=%u checksum=%u, dsn64=%d",
+		pr_debug("data_seq=%llu subflow_seq=%u data_len=%u dsn64=%d",
 			 mpext->data_seq, mpext->subflow_seq, mpext->data_len,
-			 mpext->checksum, mpext->dsn64);
+			 mpext->dsn64);
 	}
 	/* TODO: else fallback; allocation can fail, but we can't easily retire
 	 * skbs from the write_queue, as we need to roll-back TCP status
@@ -337,9 +336,8 @@ static enum mapping_status mptcp_get_mapping(struct sock *ssk)
 		goto del_out;
 	}
 
-	pr_debug("seq=%llu is64=%d ssn=%u data_len=%u ck=%u",
-		 mpext->data_seq, mpext->dsn64, mpext->subflow_seq,
-		 mpext->data_len, mpext->checksum);
+	pr_debug("seq=%llu is64=%d ssn=%u data_len=%u", mpext->data_seq,
+		 mpext->dsn64, mpext->subflow_seq, mpext->data_len);
 
 	if (mpext->data_len == 0) {
 		pr_err("Infinite mapping not handled");
@@ -912,8 +910,7 @@ static struct socket *mptcp_socket_create_get(struct mptcp_sock *msk)
 	msk->subflow = ssock;
 	subflow = subflow_ctx(msk->subflow->sk);
 	subflow->request_mptcp = 1; /* @@ if MPTCP enabled */
-	subflow->request_cksum = 0; /* checksum not supported */
-	subflow->request_version = 0; /* only v0 supported */
+	subflow->request_version = 0; /* currently only v0 supported */
 
 	sock_hold(ssock->sk);
 
net/mptcp/protocol.h +1 -3
index 40f0efaef55e..d263b64cfcfd 100644
@@ -140,7 +140,7 @@ struct subflow_request_sock {
 	struct	tcp_request_sock sk;
 	u8	mp_capable : 1,
 		mp_join : 1,
-		checksum : 1,
+		unused : 1,
 		backup : 1,
 		version : 4;
 	u8	local_id;
@@ -175,14 +175,12 @@ struct subflow_context {
 	u16	map_data_len;
 	u16	request_mptcp : 1,  /* send MP_CAPABLE */
 		request_join : 1,   /* send MP_JOIN */
-		request_cksum : 1,
 		request_bkup : 1,
 		request_version : 4,
 		mp_capable : 1,     /* remote is MPTCP capable */
 		mp_join : 1,        /* remote is JOINing */
 		fourth_ack : 1,     /* send initial DSS */
 		conn_finished : 1,
-		use_checksum : 1,
 		map_valid : 1,
 		backup : 1;
 	u32	remote_nonce;
net/mptcp/subflow.c +0 -5
index a49cc5ed1b27..9b39c0e26b26 100644
@@ -117,9 +117,6 @@ static void subflow_v4_init_req(struct request_sock *req,
 			subflow_req->version = listener->request_version;
 		else
 			subflow_req->version = rx_opt.mptcp.version;
-		if ((rx_opt.mptcp.flags & MPTCP_CAP_CHECKSUM_REQD) ||
-		    listener->request_cksum)
-			subflow_req->checksum = 1;
 		subflow_req->remote_key = rx_opt.mptcp.sndr_key;
 		pr_debug("syn seq=%u", TCP_SKB_CB(skb)->seq);
 		subflow_req->ssn_offset = TCP_SKB_CB(skb)->seq;
@@ -435,7 +432,6 @@ static int subflow_ulp_init(struct sock *sk)
 	tp->is_mptcp = 1;
 	icsk->icsk_af_ops = &subflow_specific;
 	ctx->tcp_sk_data_ready = sk->sk_data_ready;
-	ctx->use_checksum = 0;
 	sk->sk_data_ready = subflow_data_ready;
 out:
 	return err;
@@ -470,7 +466,6 @@ static void subflow_ulp_clone(const struct request_sock *req,
 	new_ctx->conn = NULL;
 	new_ctx->conn_finished = 1;
 	new_ctx->tcp_sk_data_ready = old_ctx->tcp_sk_data_ready;
-	new_ctx->use_checksum = old_ctx->use_checksum;
 
 	if (subflow_req->mp_capable) {
 		new_ctx->mp_capable = 1;
Signature · 1 line
2.17.2
#1 lore raw

Hi,

Quoted 13 lines · Peter Krystad
net/mptcp/options.c +8 -0
@@ -152,10 +151,27 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK64;
 			else
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK32;
+		}
+
+		if (mp_opt->use_map) {
+			if (mp_opt->dsn64)
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP64;
+			else
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP32;
+		}
 

net/mptcp/options.c void mptcp_parse_option(const unsigned char *ptr, int opsize, · line 170

Quoted 10 lines
net/mptcp/options.c +8 -2
-			if (opsize < expected_opsize)
-				break;
+		/* RFC 6824, Section 3.3:
+		 * If a checksum is present, but its use had
+		 * not been negotiated in the MP_CAPABLE handshake,
+		 * the checksum field MUST be ignored.
+		 */
+		if (opsize != expected_opsize &&
+		    opsize != expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM)
+			break;

My understanding is that we should look for RFC 6824 bis compliance - is that correct?

If so, than we should reset the subflow connection if opsize == expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM, and just ignore other bad len. Is that correct? (looks a bit weird to me)

I'm unsure we can reset the stream from mptcp_parse_option() even providing it the relevant sock variable (and that would be a quite invasive change).

Perhaps we can mark the pkts/DSS/MPTCP ext as 'bad' using an additional flag and send the reset later? (even at rcvmsg time)

Quoted 9 lines
net/mptcp/options.c +0 -3
@@ -701,18 +691,11 @@ void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
 		}
 
 		if (mpext->use_map) {
-			__sum16 checksum;
-
-			pr_debug("Writing map values");
 			put_unaligned_be64(mpext->data_seq, ptr);
 			ptr += 2;

net/mptcp/options.c void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts) · line 698

Quoted 10 lines
net/mptcp/options.c +3 -7
-			*ptr++ = htonl(mpext->subflow_seq);
-
-			if (mpext->use_checksum)
-				checksum = mpext->checksum;
-			else
-				checksum = TCPOPT_NOP << 8 | TCPOPT_NOP;
-			*ptr = htonl(mpext->data_len << 16 | checksum);
+			put_unaligned_be32(mpext->subflow_seq, ptr);
+			ptr += 1;
+			put_unaligned_be16(mpext->data_len, ptr);

I think we should still set 4 bytes here:

	put_unaligned_be32(mpext->subflow_seq << 16 | TCPOPT_NOP << 8 |

TCPOPT_NOP, ptr);

or we can hit some 'uninit value used' error triggered by syzbot or the like.

Cheers,

Paolo

author
#2 lore raw
Quoted 16 lines · Paolo Abeni

Hi,

Quoted 13 lines · Peter Krystad
net/mptcp/options.c +8 -0
@@ -152,10 +151,27 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK64;
 			else
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK32;
+		}
+
+		if (mp_opt->use_map) {
+			if (mp_opt->dsn64)
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP64;
+			else
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP32;
+		}
 
Quoted 13 lines

net/mptcp/options.c void mptcp_parse_option(const unsigned char *ptr, int opsize, · line 170

Quoted 10 lines
net/mptcp/options.c +8 -2
-			if (opsize < expected_opsize)
-				break;
+		/* RFC 6824, Section 3.3:
+		 * If a checksum is present, but its use had
+		 * not been negotiated in the MP_CAPABLE handshake,
+		 * the checksum field MUST be ignored.
+		 */
+		if (opsize != expected_opsize &&
+		    opsize != expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM)
+			break;

My understanding is that we should look for RFC 6824 bis compliance - is that correct?

That's the feedback we got.

Quoted 4 lines

If so, than we should reset the subflow connection if opsize == expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM, and just ignore other bad len. Is that correct? (looks a bit weird to me)

And that's my reading of the 6824bis too. But the code base is v0 right now, so I think this is OK going in like this, the purpose of this patch is just to make it clear we don't support checksum. Maybe add a comment here that it will need revision for v1? I think it will make sense to move to v1 in one (or more) coordinated patchsets instead of trickling support in over time.

Quoted 6 lines

I'm unsure we can reset the stream from mptcp_parse_option() even providing it the relevant sock variable (and that would be a quite invasive change).

Perhaps we can mark the pkts/DSS/MPTCP ext as 'bad' using an additional flag and send the reset later? (even at rcvmsg time)

It is going to have to be something like that, but out-of-scope for this patch.

Peter.

Quoted 32 lines
Quoted 9 lines
net/mptcp/options.c +0 -3
@@ -701,18 +691,11 @@ void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
 		}
 
 		if (mpext->use_map) {
-			__sum16 checksum;
-
-			pr_debug("Writing map values");
 			put_unaligned_be64(mpext->data_seq, ptr);
 			ptr += 2;

net/mptcp/options.c void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts) · line 698

Quoted 10 lines
net/mptcp/options.c +3 -7
-			*ptr++ = htonl(mpext->subflow_seq);
-
-			if (mpext->use_checksum)
-				checksum = mpext->checksum;
-			else
-				checksum = TCPOPT_NOP << 8 | TCPOPT_NOP;
-			*ptr = htonl(mpext->data_len << 16 | checksum);
+			put_unaligned_be32(mpext->subflow_seq, ptr);
+			ptr += 1;
+			put_unaligned_be16(mpext->data_len, ptr);

I think we should still set 4 bytes here:

	put_unaligned_be32(mpext->subflow_seq << 16 | TCPOPT_NOP << 8 |

TCPOPT_NOP, ptr);

or we can hit some 'uninit value used' error triggered by syzbot or the like.

Cheers,

Paolo

#3 lore raw
Quoted 37 lines · Peter Krystad
Quoted 16 lines · Paolo Abeni

Hi,

Quoted 13 lines · Peter Krystad
net/mptcp/options.c +8 -0
@@ -152,10 +151,27 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK64;
 			else
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK32;
+		}
+
+		if (mp_opt->use_map) {
+			if (mp_opt->dsn64)
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP64;
+			else
+				expected_opsize += TCPOLEN_MPTCP_DSS_MAP32;
+		}
 
Quoted 13 lines

net/mptcp/options.c void mptcp_parse_option(const unsigned char *ptr, int opsize, · line 170

Quoted 10 lines
net/mptcp/options.c +8 -2
-			if (opsize < expected_opsize)
-				break;
+		/* RFC 6824, Section 3.3:
+		 * If a checksum is present, but its use had
+		 * not been negotiated in the MP_CAPABLE handshake,
+		 * the checksum field MUST be ignored.
+		 */
+		if (opsize != expected_opsize &&
+		    opsize != expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM)
+			break;

My understanding is that we should look for RFC 6824 bis compliance - is that correct?

That's the feedback we got.

Quoted 4 lines

If so, than we should reset the subflow connection if opsize == expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM, and just ignore other bad len. Is that correct? (looks a bit weird to me)

Quoted 5 lines

And that's my reading of the 6824bis too. But the code base is v0 right now, so I think this is OK going in like this, the purpose of this patch is just to make it clear we don't support checksum. Maybe add a comment here that it will need revision for v1? I think it will make sense to move to v1 in one (or more) coordinated patchsets instead of trickling support in over time.

I agree that we should switch from v0 to v1 all at once - preferably when we have something to test against! I think we would benefit from interoperating with the v0.95 (protocol v0) kernel with the multiple-subflow transmission and reassembly changes and DATA_FIN before jumping in to v1 compliance.

Mat

Quoted 44 lines
Quoted 7 lines

I'm unsure we can reset the stream from mptcp_parse_option() even providing it the relevant sock variable (and that would be a quite invasive change).

Perhaps we can mark the pkts/DSS/MPTCP ext as 'bad' using an additional flag and send the reset later? (even at rcvmsg time)

It is going to have to be something like that, but out-of-scope for this patch.

Peter.

Quoted 30 lines
Quoted 9 lines
net/mptcp/options.c +0 -3
@@ -701,18 +691,11 @@ void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
 		}
 
 		if (mpext->use_map) {
-			__sum16 checksum;
-
-			pr_debug("Writing map values");
 			put_unaligned_be64(mpext->data_seq, ptr);
 			ptr += 2;

net/mptcp/options.c void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts) · line 698

Quoted 10 lines
net/mptcp/options.c +3 -7
-			*ptr++ = htonl(mpext->subflow_seq);
-
-			if (mpext->use_checksum)
-				checksum = mpext->checksum;
-			else
-				checksum = TCPOPT_NOP << 8 | TCPOPT_NOP;
-			*ptr = htonl(mpext->data_len << 16 | checksum);
+			put_unaligned_be32(mpext->subflow_seq, ptr);
+			ptr += 1;
+			put_unaligned_be16(mpext->data_len, ptr);

I think we should still set 4 bytes here:

	put_unaligned_be32(mpext->subflow_seq << 16 | TCPOPT_NOP << 8 |

TCPOPT_NOP, ptr);

or we can hit some 'uninit value used' error triggered by syzbot or the like.

Cheers,

Quoted 8 lines
Quoted 2 lines

Paolo

_______________________________________________
mptcp mailing list
mptcp(a)lists.01.org
https://lists.01.org/mailman/listinfo/mptcp

--
Mat Martineau
Intel

author
#4 lore raw
In reply to Paolo Abeni
Quoted 6 lines · Paolo Abeni

Hi,

Quoted 3 lines · Peter Krystad
net/mptcp/options.c +0 -0
@@ -152,10 +151,27 @@ void mptcp_parse_option(const unsigned char *ptr, int opsize,
 				expected_opsize += TCPOLEN_MPTCP_DSS_ACK64;
 

...

Quoted 23 lines
Quoted 10 lines
net/mptcp/options.c +0 -3
@@ -701,18 +691,11 @@ void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
 		}
 
 		if (mpext->use_map) {
-			__sum16 checksum;
-
-			pr_debug("Writing map values");
 			put_unaligned_be64(mpext->data_seq, ptr);
 			ptr += 2;

net/mptcp/options.c void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts) · line 698

Quoted 10 lines
net/mptcp/options.c +3 -7
-			*ptr++ = htonl(mpext->subflow_seq);
-
-			if (mpext->use_checksum)
-				checksum = mpext->checksum;
-			else
-				checksum = TCPOPT_NOP << 8 | TCPOPT_NOP;
-			*ptr = htonl(mpext->data_len << 16 | checksum);
+			put_unaligned_be32(mpext->subflow_seq, ptr);
+			ptr += 1;
+			put_unaligned_be16(mpext->data_len, ptr);

I think we should still set 4 bytes here:

Quoted 5 lines
	put_unaligned_be32(mpext->subflow_seq << 16 | TCPOPT_NOP << 8 |

TCPOPT_NOP, ptr);

or we can hit some 'uninit value used' error triggered by syzbot or the like.

Sorry Paolo I missed this comment before. I'll send a v2.

Peter.

Quoted 4 lines

Cheers,

Paolo