Thread (16 messages) 16 messages, 5 authors, 2012-11-27

[net-next 10/10] igb: Do not parse past IP header on fragments beyond the first

flat view
STALE5075d SUPERSEDED: 1 (0M)

From: Jeff Kirsher <hidden>
Date: 2012-11-21 10:47:47
Subsystem: intel ethernet drivers, networking drivers, the rest · Maintainers: Tony Nguyen, Przemek Kitszel, Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds

Revision v1 of 113 in this series; 1 review trailer.

Revisions (113)
  1. v1
  2. v2 [diff vs current]
  3. v1
  4. v1
  5. v1
  6. v1
  7. v1
  8. v1
  9. v1
  10. v1
  11. v3 [diff vs current]
  12. v1
  13. v1
  14. v2 [diff vs current]
  15. v2 [diff vs current]
  16. v1
  17. v1
  18. v1
  19. v1
  20. v1
  21. v1
  22. v1
  23. v1
  24. v1
  25. v1
  26. v1
  27. v1
  28. v1
  29. v1
  30. v1
  31. v1
  32. v1
  33. v1
  34. v1
  35. v2 [diff vs current]
  36. v1
  37. v1
  38. v1 current
  39. v1
  40. v2 [diff vs current]
  41. v1
  42. v2 [diff vs current]
  43. v1
  44. v1
  45. v1
  46. v1
  47. v1
  48. v1
  49. v1
  50. v1
  51. v1
  52. v1
  53. v2 [diff vs current]
  54. v1
  55. v1
  56. v1
  57. v1
  58. v2 [diff vs current]
  59. v1
  60. v2 [diff vs current]
  61. v1
  62. v2 [diff vs current]
  63. v1
  64. v1
  65. v2 [diff vs current]
  66. v1
  67. v1
  68. v1
  69. v1
  70. v1
  71. v1
  72. v2 [diff vs current]
  73. v1
  74. v1
  75. v2 [diff vs current]
  76. v3 [diff vs current]
  77. v4 [diff vs current]
  78. v2 [diff vs current]
  79. v5 [diff vs current]
  80. v1
  81. v1
  82. v1
  83. v2 [diff vs current]
  84. v1
  85. v1
  86. v2 [diff vs current]
  87. v3 [diff vs current]
  88. v1
  89. v2 [diff vs current]
  90. v1
  91. v1
  92. v2 [diff vs current]
  93. v1
  94. v2 [diff vs current]
  95. v1
  96. v2 [diff vs current]
  97. v3 [diff vs current]
  98. v1
  99. v1
  100. v1
  101. v2 [diff vs current]
  102. v1
  103. v1
  104. v1
  105. v1
  106. v1
  107. v2 [diff vs current]
  108. v1
  109. v1
  110. v1
  111. v2 [diff vs current]
  112. v1
  113. v1 
From: Alexander Duyck <redacted>

This change makes it so that only the first fragment in a series of fragments
will have the L4 header pulled.  Previously we were always pulling the L4
header as well and in the case of UDP this can harm performance since only the
first fragment will have the header, the rest just contain data which should
be left in the paged portion of the packet.

Signed-off-by: Alexander Duyck <redacted>
Tested-by: Aaron Brown <redacted>
Signed-off-by: Jeff Kirsher <redacted>
---
 drivers/net/ethernet/intel/igb/igb_main.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 0fe2521..0ce145e 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -6145,20 +6145,23 @@ static unsigned int igb_get_headlen(unsigned char *data,
 		if (hlen < sizeof(struct iphdr))
 			return hdr.network - data;
 
-		/* record next protocol */
-		nexthdr = hdr.ipv4->protocol;
-		hdr.network += hlen;
+		/* record next protocol if header is present */
+		if (!hdr.ipv4->frag_off)
+			nexthdr = hdr.ipv4->protocol;
 	} else if (protocol == __constant_htons(ETH_P_IPV6)) {
 		if ((hdr.network - data) > (max_len - sizeof(struct ipv6hdr)))
 			return max_len;
 
 		/* record next protocol */
 		nexthdr = hdr.ipv6->nexthdr;
-		hdr.network += sizeof(struct ipv6hdr);
+		hlen = sizeof(struct ipv6hdr);
 	} else {
 		return hdr.network - data;
 	}
 
+	/* relocate pointer to start of L4 header */
+	hdr.network += hlen;
+
 	/* finally sort out TCP */
 	if (nexthdr == IPPROTO_TCP) {
 		if ((hdr.network - data) > (max_len - sizeof(struct tcphdr)))
-- 
1.7.11.7
Other recent patches touching these files
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help