Page MenuHomeFreeBSD

rrs (Randall Stewart)
User

Projects

User Details

User Since
Jan 22 2015, 5:22 AM (410 w, 1 d)

Recent Activity

Tue, Nov 22

rrs accepted D37112: tcp: Don't try to disconnect a socket multiple times..
Tue, Nov 22, 2:10 PM

Thu, Nov 3

rrs accepted D37016: tcp: sync with latest linux tcp_info struct.
Thu, Nov 3, 4:00 PM

Oct 24 2022

rrs committed rG31bc602ff811: Rack and BBR broken with the new timewait state purge. (authored by rrs).
Rack and BBR broken with the new timewait state purge.
Oct 24 2022, 7:50 PM
rrs closed D37080: Rack and BBR broken with the new timewait state purge..
Oct 24 2022, 7:50 PM

Oct 21 2022

rrs requested review of D37080: Rack and BBR broken with the new timewait state purge..
Oct 21 2022, 1:19 PM

Oct 4 2022

rrs committed rGcd84e78f09ec: tcp idle reduce does not work for a server. (authored by rrs).
tcp idle reduce does not work for a server.
Oct 4 2022, 11:11 AM
rrs closed D36721: tcp idle reduce does not work for a server..
Oct 4 2022, 11:10 AM

Oct 3 2022

rrs accepted D36858: time(3): Align fast clock times to avoid firing multiple timers..
Oct 3 2022, 3:08 PM
rrs accepted D36721: tcp idle reduce does not work for a server..

I have a vage recollection, of something along these lines having been checked in "recently" (within last 2 years?)

Ah, here - https://reviews.freebsd.org/D25016

was that incomplete, or not implemented properly?

It shouldn't matter that much if we reduce cwnd right at the receipt of a new segment (in tcp_input), or just prior of sending out the next segment (in tcp_output)?

Oct 3 2022, 11:21 AM
rrs accepted D36863: tcp: allow window scale and timestamps to be toggled individually.
Oct 3 2022, 11:16 AM
rrs accepted D36862: Honor the drop_synfin sysctl variable in TIMEWAIT state.
Oct 3 2022, 11:15 AM

Sep 27 2022

rrs committed rG08af8aac2a61: Tcp progress timeout (authored by rrs).
Tcp progress timeout
Sep 27 2022, 5:40 PM
rrs closed D36716: Tcp progress timeout.
Sep 27 2022, 5:40 PM
rrs updated the diff for D36716: Tcp progress timeout.

Add the man page entry that Michael requested :)

Sep 27 2022, 4:34 PM
rrs added a comment to D36716: Tcp progress timeout.

Sure I can do that :)

Sep 27 2022, 4:07 PM
rrs updated the diff for D36716: Tcp progress timeout.

This updates and gets some of the bits I missed previously.

Sep 27 2022, 1:15 PM

Sep 26 2022

rrs added a comment to D36716: Tcp progress timeout.

This will need a few more changes yet.. so its not yet complete.

Sep 26 2022, 7:37 PM
rrs committed rGd1b07f36a2aa: TCP complete end status work. (authored by rrs).
TCP complete end status work.
Sep 26 2022, 7:22 PM
rrs closed D36712: TCP complete end status work..
Sep 26 2022, 7:22 PM
rrs closed D36711: TCP rack does not work properly with cubic..
Sep 26 2022, 7:14 PM
rrs committed rGe5049a173352: TCP rack does not work properly with cubic. (authored by rrs).
TCP rack does not work properly with cubic.
Sep 26 2022, 7:14 PM
rrs requested review of D36721: tcp idle reduce does not work for a server..
Sep 26 2022, 5:39 PM
rrs requested review of D36716: Tcp progress timeout.
Sep 26 2022, 3:17 PM
rrs requested review of D36712: TCP complete end status work..
Sep 26 2022, 12:22 PM
rrs requested review of D36711: TCP rack does not work properly with cubic..
Sep 26 2022, 11:59 AM

Sep 21 2022

rrs accepted D36641: tcp: fix simultaneous open and refine e80062a2d43.
Sep 21 2022, 4:19 PM

Sep 20 2022

rrs committed rGe398922eaf66: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync… (authored by rrs).
Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync…
Sep 20 2022, 8:27 PM
rrs closed D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..
Sep 20 2022, 8:27 PM
rrs committed rG7cc3ea9c6f8b: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it. (authored by rrs).
mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it.
Sep 20 2022, 5:14 PM
rrs closed D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..
Sep 20 2022, 5:14 PM

Sep 19 2022

rrs updated the diff for D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..

Take out the commented out old code.

Sep 19 2022, 10:47 AM
rrs updated the diff for D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..

Address Hans ULL comment for clock freq

Sep 19 2022, 10:45 AM
rrs added inline comments to D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..
Sep 19 2022, 10:06 AM

Sep 9 2022

rrs committed rG81560c55823b: TCP: Rack ends up sending all that is outstanding every timeout. (authored by rrs).
TCP: Rack ends up sending all that is outstanding every timeout.
Sep 9 2022, 1:01 PM
rrs closed D36494: TCP: Rack ends up sending all that is outstanding every timeout..
Sep 9 2022, 1:01 PM
rrs updated the diff for D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..

Addresses comments by jhb & np

Sep 9 2022, 11:50 AM
rrs added a comment to D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..

Thinking about your "bintime" comment, I think this is probably a good idea, but
I suspect it needs to be applied even broader than in here. You have very similar
code in the Mlx driver since this was my original model. I have a review out on
that side as well making the more highly precision timer fix there.

Sep 9 2022, 11:39 AM
rrs added inline comments to D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..
Sep 9 2022, 11:36 AM

Sep 8 2022

rrs accepted D36488: tcp: avoid call to soisconnected() on transition to ESTABLISHED.
Sep 8 2022, 3:17 PM
rrs requested review of D36494: TCP: Rack ends up sending all that is outstanding every timeout..
Sep 8 2022, 12:30 PM

Aug 30 2022

rrs added inline comments to D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..
Aug 30 2022, 5:49 AM

Aug 25 2022

rrs updated the diff for D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..

Fix a comment spelling error as well as reduce out a / and %.

Aug 25 2022, 4:45 PM
rrs updated the diff for D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..

Fix the spelling pointed out by kib. And take Hans suggestion to cut out a / and *.

Aug 25 2022, 4:44 PM
rrs added inline comments to D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..
Aug 25 2022, 3:17 PM
rrs added inline comments to D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..
Aug 25 2022, 3:09 PM

Aug 24 2022

rrs added inline comments to D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..
Aug 24 2022, 8:00 PM
rrs added inline comments to D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..
Aug 24 2022, 7:55 PM
rrs requested review of D36327: mlx5 M_TSTMP accuracy looses quite a bit of precision so lets fix it..
Aug 24 2022, 9:39 AM
rrs updated the diff for D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..

one more extra line

Aug 24 2022, 8:48 AM
rrs updated the diff for D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..

Address Drew's nits and also get rid of the debugging dump out printing fun.

Aug 24 2022, 8:46 AM
rrs added inline comments to D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..
Aug 24 2022, 8:40 AM
rrs updated the diff for D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..

Take out my debug stuff that I forgot to remove .. opps.. this will now
compile properly and no longer has the calibration query or the float up
of the actual hw_timestamp (had to expand the mbuf to have that and
thats not a good idea in general except for debugging).

Aug 24 2022, 8:32 AM

Aug 23 2022

rrs added a reviewer for D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time.: gallatin.
Aug 23 2022, 5:05 PM
rrs requested review of D36315: Enable M_TSTMP in Chelsio cxgbe driver by creating a mechanism that can sync the time..
Aug 23 2022, 4:55 PM
rrs closed D35166: tcp: Rack rwnd collapse..
Aug 23 2022, 1:18 PM
rrs committed rG62ce18fc9a8e: tcp: Rack rwnd collapse. (authored by rrs).
tcp: Rack rwnd collapse.
Aug 23 2022, 1:18 PM
rrs committed rG4e0ce82b53e2: TCP Lro has a loss of timestamp precision and reorders packets. (authored by rrs).
TCP Lro has a loss of timestamp precision and reorders packets.
Aug 23 2022, 1:16 PM
rrs closed D36043: lro has a loss of timestamp precision..
Aug 23 2022, 1:15 PM
rrs updated the summary of D36043: lro has a loss of timestamp precision..
Aug 23 2022, 1:03 PM

Aug 22 2022

rrs updated the diff for D36043: lro has a loss of timestamp precision..

Address the rest of Hans comment less one

Aug 22 2022, 3:39 PM
rrs added inline comments to D36043: lro has a loss of timestamp precision..
Aug 22 2022, 3:39 PM
rrs updated the diff for D36043: lro has a loss of timestamp precision..

Fix a nit from review of Peter Lei

Aug 22 2022, 3:32 PM
rrs updated the diff for D36043: lro has a loss of timestamp precision..

After a bit of work and digging I have found on top of the timestamp
issue where LRO is reordering packets. Some in response to the
compressed ack and mbuf queuing. Others in just the normal
old flush paths. In any event this is not desirable for TCP, we need
to see the packets arrive in order. The most horrible one (for TCP)
is where the peer sends a RST followed quickly by a new SYN. This
will get reordered into SYN first then RST.

Aug 22 2022, 12:21 PM

Aug 11 2022

rrs accepted D36136: LRO: fix BPF filters for lagg ports in the hpts path.
Aug 11 2022, 2:30 PM

Aug 8 2022

rrs accepted D36066: tcp: address a wire level race with 2 ACKs at the end of TCP handshake.
Aug 8 2022, 5:47 PM
rrs added a comment to D36065: tcp_input: remove call to soisconnected() on transition to ESTABLISHED.

So if you are not doing soisconnected() here where does it happen at now?

Aug 8 2022, 5:45 PM
rrs accepted D36064: tcp: utilize new solisten_clone() and solisten_enqueue().

These look ok but I also think it worth having Michael have a look in
case I am missing something ... :)

Aug 8 2022, 5:43 PM
rrs added a reviewer for D36064: tcp: utilize new solisten_clone() and solisten_enqueue(): tuexen.
Aug 8 2022, 5:42 PM
rrs accepted D36062: inpcb: gather v4/v6 handling code into in_pcballoc() from protocols.
Aug 8 2022, 5:40 PM

Aug 4 2022

rrs updated the diff for D36043: lro has a loss of timestamp precision..

Just found an interesting bug when SACKs are mixing in with compressed acks. We can
push the acks in the wrong order by remembering the last compressed ack we were filling.
We should NULL the cmp field when it turns out we have to add a packet that is
not-compressable.

Aug 4 2022, 7:06 PM
rrs updated the diff for D36043: lro has a loss of timestamp precision..

Drew suggested we use the mbufqueue flag as an indication that
we have stacks that are interested in the M_TSTMP_LRO so we only
do it if thats true.

Aug 4 2022, 6:28 PM
rrs added a reviewer for D36043: lro has a loss of timestamp precision.: gallatin.
Aug 4 2022, 5:24 PM
rrs requested review of D36043: lro has a loss of timestamp precision..
Aug 4 2022, 3:33 PM

Aug 1 2022

rrs accepted D35862: tcp rack: cleanup code and fix minor issue.
Aug 1 2022, 6:25 PM

Jul 18 2022

rrs accepted D35731: tcp rack: fix switch to RACK when a FIN has already been sent.
Jul 18 2022, 7:01 PM

Jun 30 2022

rrs accepted D35663: tcp: use a TCP flag to check if connection has been close(2)d.
Jun 30 2022, 3:42 PM

Jun 27 2022

rrs accepted D35166: tcp: Rack rwnd collapse..

Ok I have finished testing this and the pkt drills scripts show it works (one change on line 59 though had
to add the push bit.. which is correct).

Jun 27 2022, 1:38 PM

Jun 2 2022

rrs requested changes to D27275: Create new CC module to track cubic RFC 8312bis changes.

You need to update files also so that you can compile this
as an option as well i.e.

Jun 2 2022, 1:50 PM

May 10 2022

rrs requested review of D35166: tcp: Rack rwnd collapse..
May 10 2022, 3:08 PM
rrs closed D35066: tcp: Rack idle reduce not working..
May 10 2022, 1:48 PM
rrs committed rG04831efd9f6b: tcp: Rack idle reduce not working. (authored by rrs).
tcp: Rack idle reduce not working.
May 10 2022, 1:48 PM

Apr 26 2022

rrs requested review of D35066: tcp: Rack idle reduce not working..
Apr 26 2022, 12:21 PM

Apr 18 2022

rrs requested changes to D31666: Remove compiler hacks.

Please fix what you broke in the LRO sort code above.

Apr 18 2022, 4:35 PM

Apr 14 2022

rrs committed rG6edfc10ca5fb: tcp: adding a functionality to define "trace points" so that BB logging can be… (authored by rrs).
tcp: adding a functionality to define "trace points" so that BB logging can be…
Apr 14 2022, 8:09 PM
rrs closed D34898: tcp: adding a functionality to define "trace points" so that BB logging can be enabled at specific events..
Apr 14 2022, 8:09 PM
rrs committed rG6e6439b238a4: tcp - hpts timing is off when we are above 1200 connections. (authored by rrs).
tcp - hpts timing is off when we are above 1200 connections.
Apr 14 2022, 8:06 PM
rrs closed D34916: tcp - hpts timing is off when we are above 1200 connections..
Apr 14 2022, 8:06 PM
rrs requested review of D34916: tcp - hpts timing is off when we are above 1200 connections..
Apr 14 2022, 5:38 PM

Apr 13 2022

rrs requested review of D34898: tcp: adding a functionality to define "trace points" so that BB logging can be enabled at specific events..
Apr 13 2022, 12:52 PM

Apr 4 2022

rrs accepted D34743: Don't recurse on non-recursive lock.
Apr 4 2022, 6:21 PM

Apr 1 2022

rrs committed rGe88412d89b4e: Opps sorry, typo in the cc_cubic fix when morphing it from nreno. (authored by rrs).
Opps sorry, typo in the cc_cubic fix when morphing it from nreno.
Apr 1 2022, 12:38 PM
rrs committed rG653cf466f085: hystart++ may not properly exit CSS back to slowstart. (authored by rrs).
hystart++ may not properly exit CSS back to slowstart.
Apr 1 2022, 12:36 PM
rrs closed D34573: rack may end up with a struck connection if the rwnd is collapsed on sent data..
Apr 1 2022, 12:32 PM
rrs committed rGee1a08b8da64: rack may end up with a stuck connectin fi the rwnd is colapsed on sent data. (authored by rrs).
rack may end up with a stuck connectin fi the rwnd is colapsed on sent data.
Apr 1 2022, 12:32 PM
rrs accepted D34573: rack may end up with a struck connection if the rwnd is collapsed on sent data..
Apr 1 2022, 11:23 AM

Mar 21 2022

rrs updated the diff for D34573: rack may end up with a struck connection if the rwnd is collapsed on sent data..

update the comment Michael pointed out.

Mar 21 2022, 12:41 PM
rrs added inline comments to D34573: rack may end up with a struck connection if the rwnd is collapsed on sent data..
Mar 21 2022, 12:36 PM

Mar 16 2022

rrs updated the diff for D34573: rack may end up with a struck connection if the rwnd is collapsed on sent data..

Thanks for finding the debugging printf I missed in my cleanup. Lets also
take out the "no cc" printf too while we are at it. Only INVARIANT type
printfs as well as kld load failures should be in rack.

Mar 16 2022, 7:59 AM
rrs added inline comments to D34573: rack may end up with a struck connection if the rwnd is collapsed on sent data..
Mar 16 2022, 7:55 AM

Mar 15 2022

rrs requested review of D34573: rack may end up with a struck connection if the rwnd is collapsed on sent data..
Mar 15 2022, 9:14 PM