Skip to contentSkip to conversation

skb->data bytes

DISCUSSION Sandhya Rallapalli 2 participants 2 messages Feb 12, 2002

lore
Share & tools
lore
#0 lore raw

Hi,
As a part of my research work, I've been working on introduction of
bit errors into packets. The main purpose is to choose a random byte from (skb->head) to (skb->tail) (header & data) and change a bit in the selected byte. And then, I test the throughput variation according to the variation in ber using ttcp.

Here, I face a problem:
If a random byte within (skb->head) and (skb->data + 65) is chosen, the throughput is acceptable. But, if a byte outside (skb->head) to (skb->data + 65) is chosen, the transmission stops. The link becomes congested and netstat shows a non-changing number of bytes in Send-Q section. i.e., (skb->data + 66) to (skb->tail) are not allowed to be erred. Is there any significance to the 65th byte and beyond in skb->data?

Thanks,
-Sandhya.

#1 lore raw

You've corrupted the data therefore you have to correct the checksum in the TCP header. You can't just randomly corrupt data and expect the checksums to still pass.