From: "Ilpo_Järvinen" <[EMAIL PROTECTED]> Date: Sat, 16 Jun 2007 02:04:25 +0300 (EEST)
> There are still some things I must think carefully in sacktag processing > since it does not validate start_seq and end_seq at all which can be > abused currently at least in tcp-2.6. ...I would rather put end to the > whole russian roulette in tcp-2.6 sacktag rather than fix/think individual > cases and leave future modifications of it similarily hazardous. It's not > very clear to me how to handle all possible cases of invalid SACK blocks > though, perhaps TCP should just ignore such sack blocks without doing > any processing based on them, i.e., ignore them whenever start_seq-end_seq > does not fully fit to snd_una-snd_nxt (expect DSACK of course, which > should be processed if it's between undo_marker-snd_nxt). Do you have any > thoughts about this? I agree. This is a problem that basically every single TCP stack out there right now is vulnerable to, lots of cpu processing for invalid or maliciously created SACK blocks. This is why I even considered the RB-Tree stuff at all. Therefore the earlier we toss out bad SACK blocks the better, and thus I agree with a scheme that does validation at the earliest stage possible as you seem to be suggesting. - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html