Both BBR and Rack have the ability to log socket options, which is currently disabled. Rack
has an experimental SaD (Sack Attack Detection) algorithm that should be made available. Also
there is a t_maxpeak_rate that needs to be removed (its un-used).
Details
- Reviewers
tuexen rrs cc - Group Reviewers
transport - Commits
- rG945f9a7cc9dc: tcp: misc cleanup of options for rack as well as socket option logging.
Build and play with rack and bbr after cleaning up the issues. Also lets
make sure to compile in TCP_SAD_DETECTION to make sure that all
works properly for those wanting to experiment with it.
Diff Detail
- Lint
Lint Skipped - Unit
Tests Skipped
Event Timeline
sys/conf/options | ||
---|---|---|
233 | Should the word "Sad" be "SAD" and the words "Sack attack Detection" be "Sack Attack Detection"? | |
sys/netinet/tcp.h | ||
431 | Need to reduce padding. | |
sys/netinet/tcp_log_buf.h | ||
258 | Please adjust the alignment of the comment. | |
sys/netinet/tcp_subr.c | ||
4585 | Please adjust alignment here and below. | |
sys/netinet/tcp_var.h | ||
1087 | Please add a new line and comment note above these two new variables as "TLP related stats" | |
1098 | Need to reduce padding. | |
1403 | The alignments of some functions from above look like already a mess. So I can't ask for any better alignment. |
sys/netinet/tcp.h | ||
---|---|---|
431 | Last time we discussed this on the transport call, the statement was to NOT reduce the passing in head, but to do that when MFCing. |
sys/netinet/tcp.h | ||
---|---|---|
431 | OK. I got it. Thanks for letting me know this. |
sys/conf/options | ||
---|---|---|
233 | Actually I have almost always labeled it SaD just because.. so I should fix that :) | |
sys/netinet/tcp.h | ||
431 | Yeah I can't make the transport call.. sorry.. unless I try by car (I will do that) like Gleb does :)) Basically you don't change padding unless its an MFC thing and these changes are not planned to MFC. | |
sys/netinet/tcp_subr.c | ||
4585 | Sure | |
sys/netinet/tcp_var.h | ||
1087 | Sure | |
1098 | No these changes are *not* to be MFC'd and so we do not reduce padding in such a case. |
Update comments and spacing per cc's request. I do not update padding since the set of rack
changes is *not* intended to ever be MFC'd.