[RFC PATCH] net: wget: Take in account tcp sequence number wrap around
Michael Trimarchi
michael at amarulasolutions.com
Sun Jan 7 17:34:24 CET 2024
Coming from some discussion on mailing about wget unconsistent. It's
just and idea to play around
Signed-off-by: Michael Trimarchi <michael at amarulasolutions.com>
---
I'm on a train ;). No board to test
---
include/net/tcp.h | 10 ++++++++++
net/wget.c | 8 ++++----
2 files changed, 14 insertions(+), 4 deletions(-)
diff --git a/include/net/tcp.h b/include/net/tcp.h
index c29d4ce24a..584813b637 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -88,6 +88,16 @@ struct ip_tcp_hdr {
#define TCP_MSS 1460 /* Max segment size */
#define TCP_SCALE 0x01 /* Scale */
+/*
+ * The next routines deal with comparing 32 bit unsigned ints
+ * and worry about wraparound (automatic with unsigned arithmetic).
+ */
+static inline bool before(__u32 seq1, __u32 seq2)
+{
+ return (__s32)(seq1-seq2) < 0;
+}
+#define after(seq2, seq1) before(seq1, seq2)
+
/**
* struct tcp_mss - TCP option structure for MSS (Max segment size)
* @kind: Field ID
diff --git a/net/wget.c b/net/wget.c
index 8bb4d72db1..99a688846f 100644
--- a/net/wget.c
+++ b/net/wget.c
@@ -260,8 +260,8 @@ static void wget_connected(uchar *pkt, unsigned int tcp_seq_num,
(phys_addr_t)(pkt_q[i].pkt),
pkt_q[i].len);
store_block(ptr1,
- pkt_q[i].tcp_seq_num -
- initial_data_seq_num,
+ (unsigned int)((s32)pkt_q[i].tcp_seq_num -
+ (s32)initial_data_seq_num),
pkt_q[i].len);
unmap_sysmem(ptr1);
debug_cond(DEBUG_WGET,
@@ -333,8 +333,8 @@ static void wget_handler(uchar *pkt, u16 dport,
"wget: Transferring, seq=%x, ack=%x,len=%x\n",
tcp_seq_num, tcp_ack_num, len);
- if (tcp_seq_num >= initial_data_seq_num &&
- store_block(pkt, tcp_seq_num - initial_data_seq_num,
+ if (!before(initial_data_seq_num, tcp_seq_num) &&
+ store_block(pkt, (unsigned int)((s32)tcp_seq_num - (s32)initial_data_seq_num),
len) != 0) {
wget_fail("wget: store error\n",
tcp_seq_num, tcp_ack_num, action);
--
2.40.1
More information about the U-Boot
mailing list