Page MenuHomeFreeBSD

rrs (Randall Stewart)
User

Projects

User Details

User Since
Jan 22 2015, 5:22 AM (477 w, 5 d)

Recent Activity

Today

rrs requested review of D44420: Optimize HPTS so that little work is done until we have a hpts thread that is over the connection threshold.
Tue, Mar 19, 10:18 AM

Fri, Mar 15

rrs closed D43986: Update to bring the rack stack with all its fixes in..
Fri, Mar 15, 5:15 PM

Tue, Mar 12

rrs committed rGe18b97bd63a8: Update to bring the rack stack with all its fixes in. (authored by rrs).
Update to bring the rack stack with all its fixes in.
Tue, Mar 12, 11:57 AM

Mon, Mar 11

rrs accepted D43986: Update to bring the rack stack with all its fixes in..
Mon, Mar 11, 3:46 PM
rrs committed rGf6d489f402c3: Update to bring the rack stack with all its fixes in. (authored by rrs).
Update to bring the rack stack with all its fixes in.
Mon, Mar 11, 11:39 AM

Thu, Mar 7

rrs accepted D44250: Deep copy tls_enable struct in copyin_tls_enable.
Thu, Mar 7, 3:45 PM · NetApp
rrs accepted D44259: e1000: Make masks for TCP flag handling during TSO sysctl'able.
Thu, Mar 7, 3:20 PM
rrs accepted D44258: ixgbe: Make masks for TCP flag handling during TSO sysctl'able.
Thu, Mar 7, 3:20 PM
rrs accepted D43727: TCP LRO: add dtrace probe points.
Thu, Mar 7, 3:16 PM
rrs accepted D43769: TCP LRO: disable mbuf queuing when packet filter hooks are in place.
Thu, Mar 7, 3:10 PM

Fri, Mar 1

rrs committed rG638b5ae1c785: HTPS has actually three states not two so the macro needs to account for that. (authored by rrs).
HTPS has actually three states not two so the macro needs to account for that.
Fri, Mar 1, 8:24 PM
rrs closed D44157: HTPS has actually three states not two so the macro needs to account for that..
Fri, Mar 1, 8:23 PM
rrs updated the diff for D44157: HTPS has actually three states not two so the macro needs to account for that..

Add in the additional MPASS as Gleb suggested.

Fri, Mar 1, 12:59 PM
rrs added a comment to D44157: HTPS has actually three states not two so the macro needs to account for that..

Good catch Gleb, I missed the hpts.c changes.. I will update it :=)

Fri, Mar 1, 12:53 PM

Thu, Feb 29

rrs added inline comments to D44157: HTPS has actually three states not two so the macro needs to account for that..
Thu, Feb 29, 7:58 PM
rrs updated the summary of D44157: HTPS has actually three states not two so the macro needs to account for that..
Thu, Feb 29, 7:56 PM
rrs added inline comments to D44157: HTPS has actually three states not two so the macro needs to account for that..
Thu, Feb 29, 5:47 PM
rrs updated the diff for D43986: Update to bring the rack stack with all its fixes in..

So in testing I found a slight bug in that we were on the connect() side setting up the PCM max seg to
early. This means we end up with 10 x 512 not 10 x mss. Lets fix it so just like the listening side things get
deferred until we have gotten to the established state.

Thu, Feb 29, 4:31 PM
rrs added a comment to D43986: Update to bring the rack stack with all its fixes in..

The hpts change is now in https://reviews.freebsd.org/D44157

Thu, Feb 29, 4:07 PM
rrs requested review of D44157: HTPS has actually three states not two so the macro needs to account for that..
Thu, Feb 29, 4:07 PM
rrs updated the diff for D43986: Update to bring the rack stack with all its fixes in..

This separates out the hpts fix. Will add that as a new review after I drop this in. I have
tested the combined hpts fix and this with the current head and everything works as
expected :)

Thu, Feb 29, 4:05 PM
rrs added a comment to D43986: Update to bring the rack stack with all its fixes in..

I still need to re-test this after the last sync I did.. will do that and then split out the hpts change and update the diff before landing..

Thu, Feb 29, 3:10 PM

Wed, Feb 28

rrs accepted D43986: Update to bring the rack stack with all its fixes in..
Wed, Feb 28, 6:56 PM

Mon, Feb 26

rrs updated the diff for D43986: Update to bring the rack stack with all its fixes in..

Ok update to after the sync. Also I forgot to add in the inherit calls in the syncache and the tcp_usr_attach

Mon, Feb 26, 2:25 PM

Tue, Feb 20

rrs added a comment to D43986: Update to bring the rack stack with all its fixes in..
In D43986#1003519, @rrs wrote:

Did you try git add tcp_pcm.c? Please note that this week is stabilization week, meaning only bug fixes should go in...

No I did not, but once I do that will it show up in the diff?

I will give it a try and see.

Tue, Feb 20, 5:50 PM
rrs added a comment to D43986: Update to bring the rack stack with all its fixes in..

Did you try git add tcp_pcm.c? Please note that this week is stabilization week, meaning only bug fixes should go in...

Tue, Feb 20, 5:46 PM
rrs updated the diff for D43986: Update to bring the rack stack with all its fixes in..

Fix a bug I found in testing (and I think Gleb just pointed out)

Tue, Feb 20, 5:44 PM
rrs requested review of D43986: Update to bring the rack stack with all its fixes in..
Tue, Feb 20, 3:48 PM

Jan 24 2024

rrs accepted D43546: callout: retire callout_async_drain().

Glad to see this :)

Jan 24 2024, 5:08 PM

Dec 7 2023

rrs accepted D42946: tcp_hpts: let tcp_hpts_init() set a random CPU only once.
Dec 7 2023, 11:23 AM
rrs accepted D42917: tcp: stop stack timers in tcp_switch_back_to_default().
Dec 7 2023, 11:22 AM

Dec 4 2023

rrs accepted D42859: hpts: remove from opt_inet.h.
Dec 4 2023, 5:11 PM
rrs accepted D42858: hpts/lro: make tcp_lro_flush_tcphpts() and tcp_run_hpts() pointers.
Dec 4 2023, 5:11 PM
rrs accepted D42857: tcp/hpts: make stacks responsible for clearing themselves out HPTS.
Dec 4 2023, 5:10 PM
rrs accepted D42855: hpts: don't ifdef tcp_in_hpts().
Dec 4 2023, 5:08 PM
rrs accepted D42856: hpts: make stacks responsible for tcp_hpts_init().
Dec 4 2023, 5:08 PM
rrs accepted D42854: lro: separate HPTS specific code into tcp_lro_hpts.c.

One comment, Hans had a hand in restructuring the code a while ago too :)

Dec 4 2023, 5:07 PM
rrs accepted D42860: kern/subr_trap.c: repair the HPTS performance hack in userret().
Dec 4 2023, 5:04 PM

Nov 30 2023

rrs accepted D42851: tcp: move sysctl from tcp.do_lrd tp tcp.sack.lrd, remove sockopt for LRD.
Nov 30 2023, 3:47 PM

Nov 27 2023

rrs committed rG6a79e48076bc: Fix two latent bugs in hpts. One where a static is put on (authored by rrs).
Fix two latent bugs in hpts. One where a static is put on
Nov 27 2023, 7:40 PM

Nov 16 2023

rrs accepted D28822: Add PRR 6937bis heuristic and remove prr_conservative sysctl.
Nov 16 2023, 3:31 PM
rrs accepted D41672: tcp: add cubic_compute_pipe.
Nov 16 2023, 3:16 PM

Oct 5 2023

rrs accepted D42031: Improve sending UDP messages using IPv6 mapped addresses.
Oct 5 2023, 3:24 PM
rrs accepted D39299: tcp: include RFC6675 IsLost() in pipe calculation.
Oct 5 2023, 3:10 PM

Oct 4 2023

rrs committed rG8818f0f1124e: TCP: Fix a rack bug that skyzall found which results in a crash. (authored by rrs).
TCP: Fix a rack bug that skyzall found which results in a crash.
Oct 4 2023, 7:20 PM

Sep 7 2023

rrs accepted D41715: tcp: Align cubic cc with rfc9438.
Sep 7 2023, 3:35 PM

Jul 13 2023

rrs accepted D41010: Fix building of TCP CC modules depending on TCP_HHOOK.
Jul 13 2023, 3:19 PM

Jun 28 2023

rrs committed rG96eb0d90cd75: tcp: With the right options in the kernel cc_cubic stays in slowstart always. (authored by rrs).
tcp: With the right options in the kernel cc_cubic stays in slowstart always.
Jun 28 2023, 3:21 PM
rrs closed D40761: tcp: With the right options in the kernel cc_cubic stays in slowstart always..
Jun 28 2023, 3:20 PM

Jun 27 2023

rrs updated the diff for D40761: tcp: With the right options in the kernel cc_cubic stays in slowstart always..

After looking at this and thinking on it changing the structure is the *wrong* approach since it leaves a landmine in place
that can get someone if they don't include opt_inet.h. The better choice is to make all things that ifdef in the tcppcb to
be in opt_global.h so that the entire kernel always knows what tcpcb looks like.

Jun 27 2023, 11:13 AM

Jun 26 2023

rrs requested review of D40761: tcp: With the right options in the kernel cc_cubic stays in slowstart always..
Jun 26 2023, 9:56 PM

Jun 9 2023

rrs closed D40391: tcp: Rack fixes and misc updates.
Jun 9 2023, 2:27 PM
rrs committed rGe022f2b0131a: tcp: Rack fixes and misc updates (authored by rrs).
tcp: Rack fixes and misc updates
Jun 9 2023, 2:27 PM

Jun 2 2023

rrs requested review of D40391: tcp: Rack fixes and misc updates.
Jun 2 2023, 1:51 PM

May 24 2023

rrs closed D40229: tcp: request tracking is not http specific..
May 24 2023, 2:35 PM
rrs committed rG57a3a161a92f: tcp: request tracking is not http specific. (authored by rrs).
tcp: request tracking is not http specific.
May 24 2023, 1:30 PM

May 23 2023

rrs requested review of D40229: tcp: request tracking is not http specific..
May 23 2023, 1:58 PM

May 19 2023

rrs committed rG72ae93825244: Add a comment to the new tcp_get_srtt method to clarify that ticks (authored by rrs).
Add a comment to the new tcp_get_srtt method to clarify that ticks
May 19 2023, 7:55 PM
rrs committed rGec6d620b197e: There are congestion control algorithms will that pull in srtt, and this can… (authored by rrs).
There are congestion control algorithms will that pull in srtt, and this can…
May 19 2023, 3:18 PM
rrs closed D40146: There are congestion control algorithms will that pull in srtt, and this can cause issues with rack..
May 19 2023, 3:18 PM

May 18 2023

rrs requested review of D40146: There are congestion control algorithms will that pull in srtt, and this can cause issues with rack..
May 18 2023, 5:31 PM
rrs accepted D40061: Under RSS, assign a TCP flow's inp_flowid anyway..
May 18 2023, 3:10 PM

Apr 21 2023

rrs committed rG01216268f89e: tcp: hpts needs to still call output even after input. (authored by rrs).
tcp: hpts needs to still call output even after input.
Apr 21 2023, 11:13 AM
rrs closed D39738: tcp: hpts needs to still call output even after input..
Apr 21 2023, 11:12 AM

Apr 20 2023

rrs requested review of D39738: tcp: hpts needs to still call output even after input..
Apr 20 2023, 7:45 PM
rrs accepted D39736: netstat: fix printing of TCP pcbs with -A.
Apr 20 2023, 7:38 PM
rrs accepted D39688: netstat: cleanup.
Apr 20 2023, 3:07 PM
rrs accepted D39698: tcp: move HPTS/LRO flags out of inpcb to tcpcb.

Why if 0x10000 is not used would it matter?

Apr 20 2023, 1:06 PM
rrs accepted D39697: tcp_hpts: move HPTS related fields from inpcb to tcpcb.
Apr 20 2023, 1:03 PM
rrs accepted D39696: tcp: purge the input queue from tcp_discardcb().
Apr 20 2023, 1:02 PM

Apr 19 2023

rrs closed D39699: tcp: rack the request level logging is a bit too noisy when doing point logging..
Apr 19 2023, 6:02 PM
rrs committed rG4e8a20a7642a: tcp: rack the request level logging is a bit too noisy when doing point logging. (authored by rrs).
tcp: rack the request level logging is a bit too noisy when doing point logging.
Apr 19 2023, 6:02 PM
rrs requested review of D39699: tcp: rack the request level logging is a bit too noisy when doing point logging..
Apr 19 2023, 5:53 PM
rrs committed rG7a842346c3cd: tcp: Rack can crash with the new non-TSO fix.. (authored by rrs).
tcp: Rack can crash with the new non-TSO fix..
Apr 19 2023, 5:17 PM
rrs closed D39690: tcp: Rack can crash with the new non-TSO fix...
Apr 19 2023, 5:17 PM
rrs committed rG303246dcdf44: We have a TCP_LOG_CONNEND log that should come out at the very last log of… (authored by rrs).
We have a TCP_LOG_CONNEND log that should come out at the very last log of…
Apr 19 2023, 4:54 PM
rrs closed D39693: tcp: Ending BB logs not always coming out.
Apr 19 2023, 4:54 PM
rrs updated the diff for D39693: tcp: Ending BB logs not always coming out.

Ensure it compiles when we don't have TCP_ACCOUNTING on

Apr 19 2023, 3:47 PM
rrs updated the diff for D39693: tcp: Ending BB logs not always coming out.

fix stray space

Apr 19 2023, 3:30 PM
rrs requested review of D39693: tcp: Ending BB logs not always coming out.
Apr 19 2023, 3:27 PM
rrs requested review of D39690: tcp: Rack can crash with the new non-TSO fix...
Apr 19 2023, 1:03 PM
rrs accepted D39436: trpt: fix logging.
Apr 19 2023, 11:15 AM
rrs accepted D39591: tcp: improve logging for PRUs .
Apr 19 2023, 11:13 AM
rrs committed rG960985a209f2: tcp: bbr.c is non-capable of doing ECN and sets an INP flag to fend off ECN… (authored by rrs).
tcp: bbr.c is non-capable of doing ECN and sets an INP flag to fend off ECN…
Apr 19 2023, 11:10 AM
rrs closed D39672: tcp: bbr.c is non-capable of doing ECN and sets an INP flag to fend off ECN however our syncache is not aware of that flag..
Apr 19 2023, 11:10 AM

Apr 18 2023

rrs requested review of D39672: tcp: bbr.c is non-capable of doing ECN and sets an INP flag to fend off ECN however our syncache is not aware of that flag..
Apr 18 2023, 8:18 PM
rrs closed D39653: tcp: Inconsistent use of hpts_calling flag.
Apr 18 2023, 4:02 PM
rrs committed rG2ad584c5551d: tcp: Inconsistent use of hpts_calling flag (authored by rrs).
tcp: Inconsistent use of hpts_calling flag
Apr 18 2023, 3:58 PM
rrs updated the diff for D39653: tcp: Inconsistent use of hpts_calling flag.

Almost forgot, when going through the compressed ack path we will need to possibly nuke the flag too if
output did not occur...

Apr 18 2023, 2:13 PM
rrs requested review of D39653: tcp: Inconsistent use of hpts_calling flag.
Apr 18 2023, 1:08 PM
rrs committed rG37229fed38c6: tcp: Blackbox logging and tcp accounting together can cause a crash. (authored by rrs).
tcp: Blackbox logging and tcp accounting together can cause a crash.
Apr 18 2023, 12:40 PM
rrs closed D39622: tcp: Blackbox logging and tcp accounting together can cause a crash..
Apr 18 2023, 12:40 PM

Apr 17 2023

rrs updated the diff for D39622: tcp: Blackbox logging and tcp accounting together can cause a crash..

Opps lets get the sets of the log union in the right place.

Apr 17 2023, 7:11 PM
rrs requested review of D39622: tcp: Blackbox logging and tcp accounting together can cause a crash..
Apr 17 2023, 7:06 PM

Apr 14 2023

rrs closed D39576: tcp: stack unloading crash in rack and bbr.
Apr 14 2023, 7:52 PM
rrs committed rG3cc7b6673280: tcp: stack unloading crash in rack and bbr (authored by rrs).
tcp: stack unloading crash in rack and bbr
Apr 14 2023, 7:44 PM
rrs requested review of D39576: tcp: stack unloading crash in rack and bbr.
Apr 14 2023, 6:55 PM
rrs accepted D39574: tcp_hpts: use queue(9) STAILQ for the input queue.
Apr 14 2023, 6:03 PM
rrs closed D39565: tcp: rack pacing has some caveats that need to be obeyed when LRO is missing..
Apr 14 2023, 1:34 PM
rrs committed rG9903bf34f01f: tcp: rack pacing has some caveats that need to be obeyed when LRO is missing (authored by rrs).
tcp: rack pacing has some caveats that need to be obeyed when LRO is missing
Apr 14 2023, 1:34 PM