ip_frag: fix ipv6 when MTU sizes not aligned to 8 bytes

PATCH Applied Chas Williams 4 participants 5 messages Nov 27, 2018 – Dec 19, 2018

lore

Share & tools

lore

Chas Williams

author

2018-11-27 04:56 UTC #0 lore raw

PATCH 1/1 ip_frag: fix ipv6 when MTU sizes not aligned to 8 bytes

From: Chas Williams <[email protected]>

The same issue was fixed on for the ipv4 version of this routine in commit 8d4d3a4f7337 ("ip_frag: handle MTU sizes not aligned to 8 bytes"). Briefly, the size of an ipv6 header is always 40 bytes. With an MTU of 1500, this will never produce a multiple of 8 bytes for the frag_size and this routine can never succeed. Since RTE_ASSERTS are disabled by default, this failure is tpyically ignored.

To fix this, round down to the nearest 8 bytes and use this when producing the fragments.

Fixes 0aa31d7a5929 ("ip_frag: add IPv6 fragmentation support")

Signed-off-by Chas Williams <[email protected]>

---
 lib/librte_ip_frag/rte_ip_frag.h            |  1 +
 lib/librte_ip_frag/rte_ipv6_fragmentation.c | 18 +++++++++++-------
 2 files changed, 12 insertions(+), 7 deletions(-)

lib/librte_ip_frag/rte_ip_frag.h +1 -0

index a4ccaf9d1..04fd9df52 100644
@@ -115,6 +115,7 @@ struct rte_ip_frag_tbl {
 #define	RTE_IPV6_EHDR_MF_MASK			1
 #define	RTE_IPV6_EHDR_FO_SHIFT			3
 #define	RTE_IPV6_EHDR_FO_MASK			(~((1 << RTE_IPV6_EHDR_FO_SHIFT) - 1))
+#define	RTE_IPV6_EHDR_FO_ALIGN			(1 << RTE_IPV6_EHDR_FO_SHIFT)
 
 #define RTE_IPV6_FRAG_USED_MASK			\
 	(RTE_IPV6_EHDR_MF_MASK | RTE_IPV6_EHDR_FO_MASK)

lib/librte_ip_frag/rte_ipv6_fragmentation.c +11 -7

index 62a7e4e83..b9437eb11 100644
@@ -77,11 +77,14 @@ rte_ipv6_fragment_packet(struct rte_mbuf *pkt_in,
 	uint32_t out_pkt_pos, in_seg_data_pos;
 	uint32_t more_in_segs;
 	uint16_t fragment_offset, frag_size;
+	uint64_t frag_bytes_remaining;
 
-	frag_size = (uint16_t)(mtu_size - sizeof(struct ipv6_hdr));
-
-	/* Fragment size should be a multiple of 8. */
-	RTE_ASSERT((frag_size & ~RTE_IPV6_EHDR_FO_MASK) == 0);
+	/*
+	 * Ensure the IP payload length of all fragments (except the
+	 * the last fragment) are a multiple of 8 bytes per RFC2460.
+	 */
+	frag_size = RTE_ALIGN_FLOOR(mtu_size - sizeof(struct ipv6_hdr),
+				    RTE_IPV6_EHDR_FO_ALIGN);
 
 	/* Check that pkts_out is big enough to hold all fragments */
 	if (unlikely (frag_size * nb_pkts_out <

@@ -111,6 +114,7 @@ rte_ipv6_fragment_packet(struct rte_mbuf *pkt_in,
 		/* Reserve space for the IP header that will be built later */
 		out_pkt->data_len = sizeof(struct ipv6_hdr) + sizeof(struct ipv6_extension_fragment);
 		out_pkt->pkt_len  = sizeof(struct ipv6_hdr) + sizeof(struct ipv6_extension_fragment);
+		frag_bytes_remaining = frag_size;
 
 		out_seg_prev = out_pkt;
 		more_out_segs = 1;

@@ -130,7 +134,7 @@ rte_ipv6_fragment_packet(struct rte_mbuf *pkt_in,
 
 			/* Prepare indirect buffer */
 			rte_pktmbuf_attach(out_seg, in_seg);
-			len = mtu_size - out_pkt->pkt_len;
+			len = frag_bytes_remaining;
 			if (len > (in_seg->data_len - in_seg_data_pos)) {
 				len = in_seg->data_len - in_seg_data_pos;
 			}

@@ -140,11 +144,11 @@ rte_ipv6_fragment_packet(struct rte_mbuf *pkt_in,
 			    out_pkt->pkt_len);
 			out_pkt->nb_segs += 1;
 			in_seg_data_pos += len;
+			frag_bytes_remaining -= len;
 
 			/* Current output packet (i.e. fragment) done ? */
-			if (unlikely(out_pkt->pkt_len >= mtu_size)) {
+			if (unlikely(frag_bytes_remaining == 0))
 				more_out_segs = 0;
-			}
 
 			/* Current input segment done ? */
 			if (unlikely(in_seg_data_pos == in_seg->data_len)) {

Signature · 1 line

2.14.4

Luca Boccassi

Acked-by on v1 1/1

2018-11-28 10:15 UTC #1 lore raw

Quoted 14 lines · Chas Williams

From: Chas Williams <[email protected]>
The same issue was fixed on for the ipv4 version of this routine in commit 8d4d3a4f7337 ("ip_frag: handle MTU sizes not aligned to 8 bytes").
Briefly, the size of an ipv6 header is always 40 bytes. With an MTU of
1500, this will never produce a multiple of 8 bytes for the frag_size and this routine can never succeed. Since RTE_ASSERTS are disabled by default, this failure is tpyically ignored.
To fix this, round down to the nearest 8 bytes and use this when producing the fragments.

Quoted 7 lines

Fixes 0aa31d7a5929 ("ip_frag: add IPv6 fragmentation support")
Signed-off-by Chas Williams <[email protected]>
---
lib/librte_ip_frag/rte_ip_frag.h | 1 + lib/librte_ip_frag/rte_ipv6_fragmentation.c | 18 +++++++++++-------
 2 files changed, 12 insertions(+), 7 deletions(-)

Acked-by Luca Boccassi <[email protected]>

Signature · 2 lines

Kind regards,
Luca Boccassi

Thomas Monjalon

2018-12-19 03:47 UTC #2 lore raw

28/11/2018 11:15, Luca Boccassi:

Quoted 17 lines · Chas Williams

Quoted 16 lines
From: Chas Williams <[email protected]>
The same issue was fixed on for the ipv4 version of this routine in commit 8d4d3a4f7337 ("ip_frag: handle MTU sizes not aligned to 8 bytes").
Briefly, the size of an ipv6 header is always 40 bytes. With an MTU of
1500, this will never produce a multiple of 8 bytes for the frag_size and this routine can never succeed. Since RTE_ASSERTS are disabled by default, this failure is tpyically ignored.
To fix this, round down to the nearest 8 bytes and use this when producing the fragments.
Fixes 0aa31d7a5929 ("ip_frag: add IPv6 fragmentation support")

Quoted 7 lines

Quoted 5 lines
Signed-off-by Chas Williams <[email protected]>
---
 lib/librte_ip_frag/rte_ip_frag.h            |  1 +
 lib/librte_ip_frag/rte_ipv6_fragmentation.c | 18 +++++++++++-------
 2 files changed, 12 insertions(+), 7 deletions(-)
Acked-by Luca Boccassi <[email protected]>

Konstantin, no comment?

Ananyev, Konstantin

Acked-by on v1 1/1

2018-12-19 10:12 UTC #3 lore raw

In reply to Chas Williams

Quoted 18 lines

-----Original Message-----
From: Chas Williams [mailto:[email protected]]
Sent: Tuesday, November 27, 2018 4:56 AM
To: [email protected]
Cc: Ananyev, Konstantin <[email protected]>; Chas Williams <[email protected]>
Subject: [PATCH] ip_frag: fix ipv6 when MTU sizes not aligned to 8 bytes
From: Chas Williams <[email protected]>
The same issue was fixed on for the ipv4 version of this routine in commit 8d4d3a4f7337 ("ip_frag: handle MTU sizes not aligned to 8 bytes"). Briefly, the size of an ipv6 header is always 40 bytes. With an MTU of 1500, this will never produce a multiple of 8 bytes for the frag_size and this routine can never succeed. Since RTE_ASSERTS are disabled by default, this failure is tpyically ignored.

Quoted 7 lines

To fix this, round down to the nearest 8 bytes and use this when producing the fragments.
Fixes 0aa31d7a5929 ("ip_frag: add IPv6 fragmentation support")
Signed-off-by Chas Williams <[email protected]>
---

Acked-by Konstantin Ananyev <[email protected]>

Quoted 1 line

2.14.4

Thomas Monjalon

Applied

2018-12-19 21:37 UTC #4 lore raw

Quoted 7 lines

Quoted 7 lines
The same issue was fixed on for the ipv4 version of this routine in commit 8d4d3a4f7337 ("ip_frag: handle MTU sizes not aligned to 8 bytes"). Briefly, the size of an ipv6 header is always 40 bytes. With an MTU of 1500, this will never produce a multiple of 8 bytes for the frag_size and this routine can never succeed. Since RTE_ASSERTS are disabled by default, this failure is tpyically ignored.

Quoted 8 lines

Quoted 6 lines
To fix this, round down to the nearest 8 bytes and use this when producing the fragments.
Fixes 0aa31d7a5929 ("ip_frag: add IPv6 fragmentation support")
Signed-off-by Chas Williams <[email protected]>
Acked-by Konstantin Ananyev <[email protected]>

Applied, thanks