Page MenuHomeFreeBSD

j-nitrology.com (Jason Wolfe)
User

Projects

User Details

User Since
Mar 21 2015, 1:39 AM (194 w, 5 d)

Recent Activity

Aug 20 2018

j-nitrology.com added a comment to D16471: Some TCP reassembly improvements.

D16626 expanded on these changes and was recently committed. I believe we can close this one out.

Aug 20 2018, 8:44 PM

Jul 21 2018

j-nitrology.com accepted D16282: NULL out cc_data in pluggable TCP {cc}_cb_destroy.
Jul 21 2018, 8:23 PM

Jun 7 2018

j-nitrology.com added a comment to D15686: convert inpcbhash rlock to epoch.

This has been tested under moderate load in our lab @ LLNW over a few days with no stability issues. We haven't gotten to more in depth profiling at higher loads, but I also saw no performance concerns.

Jun 7 2018, 7:17 AM

Jan 25 2018

j-nitrology.com added a comment to D4154: Clean up unused bandwidth entry in the TCP hostcache.

If it could be useful to others, might you commit it?

Jan 25 2018, 6:24 PM

Oct 3 2017

j-nitrology.com added reviewers for D12575: Check so_error early in sendfile() call: glebius, jtl.
Oct 3 2017, 4:35 PM

Sep 26 2017

j-nitrology.com added a comment to D11003: Load balance sockets with new SO_REUSEPORT_LB option.

I believe we need to update cddl/lib/libdtrace/tcp.d with this change also:

Sep 26 2017, 3:12 AM

Jun 21 2017

j-nitrology.com accepted D9478: Add quirks for Samsung 750 EVO SSD.

Assuming this has been used locally with NCQ TRIM for a while or it's been confirmed with imp's suggestion, we should get this in.

Jun 21 2017, 6:10 PM

Jun 1 2017

j-nitrology.com added a reviewer for D11016: Add a sysctl to toggle the use of the sockets LOWAT when calculating auto window growth: gnn.
Jun 1 2017, 9:29 PM

May 31 2017

j-nitrology.com added a comment to D11016: Add a sysctl to toggle the use of the sockets LOWAT when calculating auto window growth.

It may be more proper for this to go into tflag, but the downsides there is then it won't apply to existing sessions and we consume more space (which is low with the current padding). Open to any opinions, as locally we just have it hard coded on in any case.

May 31 2017, 9:55 PM
j-nitrology.com created D11016: Add a sysctl to toggle the use of the sockets LOWAT when calculating auto window growth.
May 31 2017, 9:52 PM

Feb 14 2017

j-nitrology.com added a reviewer for D9465: Remove unnecessary ifdef soup from struct tcpcb: pkelsey.
Feb 14 2017, 7:24 PM

Feb 2 2017

j-nitrology.com abandoned D7413: Fix possible memory leak on error return in krpc_portmap.

This will be reworked in the future to be proper, abandoning for now. Thank you for the follow up.

Feb 2 2017, 10:47 PM
j-nitrology.com removed a reviewer for D7413: Fix possible memory leak on error return in krpc_portmap: sbruno.
Feb 2 2017, 10:46 PM

Jan 17 2017

j-nitrology.com retitled D9209: Add 4k quirk for Micron 5100 and Intel S3610 SSDs from to Add 4k quirk for Micron 5100 and Intel S3610 SSDs.
Jan 17 2017, 7:52 AM

Oct 11 2016

j-nitrology.com added a comment to D7413: Fix possible memory leak on error return in krpc_portmap.

Sorry for the delay, this had dropped off my radar until sbruno prodded me about it. Thanks for taking the time to review, looking over the goto outs in krpc_call() reveals the possible double-frees as you've mentioned.

Oct 11 2016, 8:14 PM
j-nitrology.com updated the diff for D7413: Fix possible memory leak on error return in krpc_portmap.
Oct 11 2016, 8:10 PM
j-nitrology.com set the repository for D7413: Fix possible memory leak on error return in krpc_portmap to rS FreeBSD src repository.
Oct 11 2016, 8:02 PM
j-nitrology.com updated the diff for D7413: Fix possible memory leak on error return in krpc_portmap.
Oct 11 2016, 8:01 PM
j-nitrology.com updated the diff for D7413: Fix possible memory leak on error return in krpc_portmap.

Allow krpc_call() to do all freeing in out: to avoid possible double-frees

Oct 11 2016, 7:59 PM

Aug 4 2016

j-nitrology.com retitled D7413: Fix possible memory leak on error return in krpc_portmap from to Fix possible memory leak on error return in krpc_portmap.
Aug 4 2016, 6:41 AM
j-nitrology.com retitled D7412: Fix possible memory leaks in ixl VLAN addition/removal from to Fix possible memory leaks in ixl VLAN addition/removal.
Aug 4 2016, 6:37 AM

Jun 1 2016

j-nitrology.com closed D6683: Add Micron Enterprise SSD 4k quirks.
Jun 1 2016, 10:39 PM
j-nitrology.com reclaimed D6683: Add Micron Enterprise SSD 4k quirks.
Jun 1 2016, 10:39 PM
j-nitrology.com abandoned D6683: Add Micron Enterprise SSD 4k quirks.
Jun 1 2016, 10:38 PM
j-nitrology.com added a comment to D6683: Add Micron Enterprise SSD 4k quirks.

The 4096 physical sector size reported by the hardware is properly being keyed off of (r228846), and we are properly setting a stripesize of 4096 based on this. The quirk is not needed for this drive.

Jun 1 2016, 10:37 PM
j-nitrology.com added a comment to D6683: Add Micron Enterprise SSD 4k quirks.

Relevant output from drives:

Jun 1 2016, 9:30 PM
j-nitrology.com retitled D6683: Add Micron Enterprise SSD 4k quirks from to Add Micron Enterprise SSD 4k quirks.
Jun 1 2016, 9:19 PM

May 17 2016

j-nitrology.com added a comment to D6085: Collapse/vm_fault races and small fixes for radix insertion failures.

We have been running the stable/10 iteration of this patch at Limelight on a number of boxes over the past few weeks. We have had no ill effects, and these machines have not hit the process stuck in vodead bug. We don't have a way to reliably reproduce, and we did hit it fairly rarely so this isn't conclusive. Since this review has been fairly quiet I did want to give some positive feedback though. Thanks!

May 17 2016, 3:05 AM

May 16 2016

j-nitrology.com added a comment to D6198: Add sysctl to disable the TCP hostcache.

When someone has a moment, might this look good after the updates?

May 16 2016, 2:28 AM

May 4 2016

j-nitrology.com updated D6198: Add sysctl to disable the TCP hostcache.
May 4 2016, 6:23 PM
j-nitrology.com updated the diff for D6198: Add sysctl to disable the TCP hostcache.

I've moved all functionality to post declaration and updated the comment about the return on tcp_hc_getmtu().
@rwatson yes I just figure it's a near free feature should you be using jails to do various testing. It has been tested with both v4 and v6 traffic.

May 4 2016, 6:03 PM
j-nitrology.com updated the diff for D6198: Add sysctl to disable the TCP hostcache.

The default needs to maintain the current behavior

May 4 2016, 4:29 AM
j-nitrology.com added a comment to D6198: Add sysctl to disable the TCP hostcache.
May 4 2016, 4:23 AM
j-nitrology.com retitled D6198: Add sysctl to disable the TCP hostcache from to Add sysctl to disable the TCP hostcache.
May 4 2016, 4:19 AM

Mar 29 2016

j-nitrology.com accepted D5753: CAM: Add 4k quirks for Samsung PM851 SSD.

We don't have official word from Samsung, but through light research all MZ7 drives are recent SSDs.

Mar 29 2016, 4:57 AM

Mar 23 2016

j-nitrology.com updated the diff for D5711: Add 4k enabled cam quirks for Samsung SM863 Series SSDs.

Drop extraneous E

Mar 23 2016, 5:06 AM

Mar 22 2016

j-nitrology.com retitled D5711: Add 4k enabled cam quirks for Samsung SM863 Series SSDs from to Add 4k enabled cam quirks for Samsung SM863 Series SSDs.
Mar 22 2016, 10:12 PM

Jan 26 2016

j-nitrology.com added a comment to D5024: Implement sysctl for the min/max values of the TCP persist timer.

So all of the current VNET tooling expects compile time constants, it appears this is what may have been holding back the simple transition of everything in tcp_timer.h based on hz. Talking with @bz, this commit is good to go, and we'll circle back and virtualize this and other timers with a revised method.

Jan 26 2016, 6:54 AM

Jan 22 2016

j-nitrology.com added a comment to D5024: Implement sysctl for the min/max values of the TCP persist timer.
In D5024#106674, @bz wrote:

These should be per-VNET changeable really; I see no harm at least why they shouldn't be?

Jan 22 2016, 1:38 AM

Jan 21 2016

j-nitrology.com retitled D5024: Implement sysctl for the min/max values of the TCP persist timer from to Implement sysctl for the min/max values of the TCP persist timer.
Jan 21 2016, 9:14 PM

Dec 11 2015

j-nitrology.com updated subscribers of D2254: Fix a number of issues with in6 ifaddr list locking..
Dec 11 2015, 4:16 AM

Dec 9 2015

j-nitrology.com updated subscribers of D4364: inpcb L2/L3 caching.
Dec 9 2015, 8:07 PM · transport
j-nitrology.com updated subscribers of D4306: Re-add route caching for TCP.
Dec 9 2015, 8:01 PM · transport

Nov 20 2015

j-nitrology.com updated subscribers of D4225: Fix dupack processing to detect loss correctly..
Nov 20 2015, 3:20 AM

Nov 19 2015

j-nitrology.com updated subscribers of D4205: Add an option to use rfc6675 based pipe/inflight bytes calculation in cubic..
Nov 19 2015, 9:38 AM

Nov 13 2015

j-nitrology.com retitled D4154: Clean up unused bandwidth entry in the TCP hostcache from to Clean up unused bandwidth entry in the TCP hostcache.
Nov 13 2015, 9:54 PM

Oct 31 2015

j-nitrology.com updated subscribers of D3521: Implement callout_drain_async().
Oct 31 2015, 12:08 AM

Oct 30 2015

j-nitrology.com added a comment to D4039: Add support for sysctl knobs to live tune the tx packet processing limits in igb and fix a wrap-around bug..
In D4039#84487, @hiren wrote:

Is there a way we can unify *_process_limit and *_processing_limit? I see both being used and that is confusing. I am not sure if its just the code and if user is presented with only one but in any case we should choose one and stick with it. BTW, this can be done as a followup/separate commit.

This change looks okay to me.

Oct 30 2015, 8:44 PM
j-nitrology.com updated D4039: Add support for sysctl knobs to live tune the tx packet processing limits in igb and fix a wrap-around bug..
Oct 30 2015, 5:33 PM

Oct 22 2015

j-nitrology.com updated subscribers of D3971: Follow rfc6675 and maintain sacked bytes to calculate correct inflight/pipe calculation..
Oct 22 2015, 12:51 AM

Oct 15 2015

j-nitrology.com added a comment to D3858: Add an ability to specify initial congestion window..

It's worth noting that you could set the initcwnd via the net.inet.tcp.slowstart_flightsize sysctl from r50673 16 years ago until it was removed in r226447. 9.2 was the first release without it, and we used it heavily at Limelight during its tenure. Its long lived existence and removal not receiving much notice should at least ease concerns a bit about users blowing things up.

Oct 15 2015, 6:29 AM

Oct 9 2015

j-nitrology.com updated subscribers of D3858: Add an ability to specify initial congestion window..
Oct 9 2015, 10:32 PM

Sep 24 2015

j-nitrology.com updated the test plan for D3719: Add support for sysctl knobs to live tune the per interrupt rx/tx packet processing limits in ixgbe(4).
Sep 24 2015, 5:49 PM

Sep 23 2015

j-nitrology.com updated the diff for D3719: Add support for sysctl knobs to live tune the per interrupt rx/tx packet processing limits in ixgbe(4).

Clean up now unused old process_limits, we use the ones in the adapter struct for all limits in the rx/tx task queues.

Sep 23 2015, 9:01 AM
j-nitrology.com retitled D3719: Add support for sysctl knobs to live tune the per interrupt rx/tx packet processing limits in ixgbe(4) from to Add support for sysctl knobs to live tune the per interrupt rx/tx packet processing limits in ixgbe(4).
Sep 23 2015, 7:54 AM

Aug 31 2015

j-nitrology.com added a comment to D3301: mac address does not update when removing the primary iface from a lagg .

This tests well and resolves the proposed issue for me on head.

Aug 31 2015, 10:16 PM · network

Jun 30 2015

j-nitrology.com added a project to D2950: Add missing DTrace scripts that exist in the source tree to the Makefile for install: DTrace.
Jun 30 2015, 6:07 PM · DTrace
j-nitrology.com updated the diff for D2950: Add missing DTrace scripts that exist in the source tree to the Makefile for install.

Order SCRIPTS alphabetically.

Jun 30 2015, 2:27 AM · DTrace
j-nitrology.com retitled D2950: Add missing DTrace scripts that exist in the source tree to the Makefile for install from to Add missing DTrace scripts that exist in the source tree to the Makefile for install.
Jun 30 2015, 2:03 AM · DTrace

Jun 16 2015

j-nitrology.com updated subscribers of D2599: Decompose TCP INP_INFO lock to increase short-lived connections scalability.
Jun 16 2015, 11:39 PM

Jun 10 2015

j-nitrology.com added a comment to D2763: Fix a callout race condition introduced in TCP timers callouts with r281599..

So far this is looking solid for us. Both with defaults and lowered keep alives on the same traffic patterns that caused the cores prior. Running with net.inet.tcp.per_cpu_timers = 1

Jun 10 2015, 5:22 PM

Jun 6 2015

j-nitrology.com added a comment to D2079: Fix TCP timers use-after-free old race conditions.
Jun 6 2015, 7:54 PM
j-nitrology.com added a comment to D2079: Fix TCP timers use-after-free old race conditions.

It sounds like you guys are well into this, but FWIW we at Limelight are seeing the same and would be happy to sacrifice some machines to the debugging altar should it be needed. Everything has been TIME_WAIT so far, but I've only analyzed a handful of cores. The rest of the CB is still populated.

Jun 6 2015, 7:03 PM

Apr 14 2015

j-nitrology.com updated subscribers of D2296: Full 32bit flowid for igb(4).
Apr 14 2015, 10:17 PM