[TCP]: secure_tcp_sequence_number() should not use a too fast clock
authorEric Dumazet <dada1@cosmosbay.com>
Mon, 1 Oct 2007 20:58:36 +0000 (13:58 -0700)
committerDavid S. Miller <davem@sunset.davemloft.net>
Tue, 2 Oct 2007 04:01:24 +0000 (21:01 -0700)
commit9b42c336d06411e6463949d2dac63949f66ff70b
tree41fbce14cd5a217341649a025622750f71ec71cd
parent32740ddc1095e5e320bf961dda146bf97bc28adb
[TCP]: secure_tcp_sequence_number() should not use a too fast clock

TCP V4 sequence numbers are 32bits, and RFC 793 assumed a 250 KHz clock.
In order to follow network speed increase, we can use a faster clock, but
we should limit this clock so that the delay between two rollovers is
greater than MSL (TCP Maximum Segment Lifetime : 2 minutes)

Choosing a 64 nsec clock should be OK, since the rollovers occur every
274 seconds.

Problem spotted by Denys Fedoryshchenko

[ This bug was introduced by f85958151900f9d30fa5ff941b0ce71eaa45a7de ]

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/char/random.c