Page MenuHomeFreeBSD

kevans (Kyle Evans)
User

Projects (7)

User Details

User Since
Oct 2 2015, 1:17 PM (292 w, 1 d)

Recent Activity

Wed, Apr 28

kevans added a comment to D30023: sysutils/debootstrap: add dependency on gnugrep.

Odd, looks like it just uses grep from PATH and I'm not seeing anywhere that it might use anything but that from your profile/login(1).

Wed, Apr 28, 12:46 PM

Thu, Apr 22

kevans added a comment to D29924: stats.7: Add dwatch(1) to the list.

Worth noting also that immediate MFC in base is a little unusual- typically there's a three day minimum soak time, and it should be denoted with a 'MFC after:' tag at the bottom of the commit.

Thu, Apr 22, 5:55 PM

Tue, Apr 20

kevans added a comment to D29848: config: create a 'requires' keyword for files file processing.
In D29848#670260, @imp wrote:
In D29848#669964, @manu wrote:

What's the difference between this and making an non-optional requirement for the file ?
Is it just to have a nicer error message than a linker error ?

One of my main annoyances with config(8) is that non-optional requirements require you to actually read and interpret files to understand how to get what you want, and the failure mode sucks- if you get it wrong, you don't even get a linker error IIRC because we simply don't build it unless all the requirements are there.

The difference between

a/b/c optional foo bar

and

a/b/c optional foo requires "bar"

is that if bar is not in the kernel in the first case you get no foo. No warning, no nothing, no foo. It is horribly abused to omit, for example, ethernet drivers that can't cope with inet not being in the kernel. The latter emits a nasty error and you don't make it past the config stage.

There's several concepts here that I'd like to explore in the future:
(1) needs X X must be configured, or you don't get this device even if it asked for
(2) requires X if X is not configured, that is an error
(3) implies X Turns X on when this device is present
(4) conflicts-with X Cannot be used at the same time X is present

config(8) only does (1) today. This adds (2) as a stop-gap before we rewrite config(8) entirely, likely in flua. I thought about doing (3) but didn't want to deal with detecting circular implies and such. It's not hard, I just didn't want to spend the time.

Tue, Apr 20, 3:43 PM
kevans added a comment to D29861: git hooks: Adjust hooks for the ports tree.

I may be misunderstanding the process you're describing, but it sounds to me like you're talking about a situation involving merging a pull request where the author is annotated. If the submission came from a BZ bug report, is the "Submitted by" field not still relevant?

My brain is definitely adapting slowly to the new git world order.

Tue, Apr 20, 2:36 PM
kevans added a comment to D29861: git hooks: Adjust hooks for the ports tree.

IMO if "Submitted by" gets restored, it should be noted that it's expected to be more of an anomaly than the norm -- reserved for situations where the original work was submitted by someone else, but changed substantially enough that the author should be updated.

Interesting. I always thought of that for "Reported by," and treated "Submitted by" as the norm.

Tue, Apr 20, 2:21 PM
kevans added a comment to D29861: git hooks: Adjust hooks for the ports tree.

Two thoughts here:

  1. Should we restore "Submitted by:" to this list?
Tue, Apr 20, 2:10 PM
kevans added a comment to D29848: config: create a 'requires' keyword for files file processing.
In D29848#669964, @manu wrote:

What's the difference between this and making an non-optional requirement for the file ?
Is it just to have a nicer error message than a linker error ?

Tue, Apr 20, 11:56 AM

Sat, Apr 17

kevans added inline comments to D29788: kern: ether_gen_addr: randomize on default hostuuid, too.
Sat, Apr 17, 2:39 PM

Fri, Apr 16

kevans updated the diff for D29787: man: document ether_gen_addr(9).

+SPDX

Fri, Apr 16, 1:31 AM
kevans requested review of D29788: kern: ether_gen_addr: randomize on default hostuuid, too.
Fri, Apr 16, 1:16 AM
kevans requested review of D29787: man: document ether_gen_addr(9).
Fri, Apr 16, 1:16 AM
kevans committed R10:77c89fa6f5cd: modules: remove stale if_wg reference (authored by kevans).
modules: remove stale if_wg reference
Fri, Apr 16, 1:00 AM

Thu, Apr 15

kevans added a comment to D29772: ipfw: do not use sleepable malloc in callout context..

Thanks!

Thu, Apr 15, 1:08 PM
kevans accepted D29772: ipfw: do not use sleepable malloc in callout context..
Thu, Apr 15, 1:08 PM

Wed, Apr 14

kevans committed R11:b02c01eca918: misc/freebsd-release-manifests: add 13.0-RELEASE, rm 13-RC* (authored by gjb).
misc/freebsd-release-manifests: add 13.0-RELEASE, rm 13-RC*
Wed, Apr 14, 4:13 AM

Tue, Apr 13

kevans committed R10:3016c5c2bf68: caroot: reroll the remaining certs (authored by kevans).
caroot: reroll the remaining certs
Tue, Apr 13, 10:01 PM
kevans committed R10:fac832b27105: caroot: remove certs distrusted for server auth (authored by kevans).
caroot: remove certs distrusted for server auth
Tue, Apr 13, 10:01 PM
kevans committed R10:c3510c941c0d: caroot: update CA bundle processor (authored by kevans).
caroot: update CA bundle processor
Tue, Apr 13, 10:01 PM
kevans committed R10:446169e0b6f0: caroot: routine cert update (authored by kevans).
caroot: routine cert update
Tue, Apr 13, 10:01 PM

Mon, Apr 12

kevans added inline comments to D29659: jail: convert several functions from int to bool.
Mon, Apr 12, 1:10 PM · Jails

Fri, Apr 9

kevans added a comment to D21892: freebsd-update: create a ZFS boot environment on install if the system supports it.

ping: Is there anything left for Dave to do here, or can we commit this?

Fri, Apr 9, 7:46 PM

Apr 1 2021

kevans accepted D29533: libfetch: Fix proxy authentication.

Sure

Apr 1 2021, 7:35 PM
kevans committed R10:4752dc88e64d: init: use explicit_bzero() for clearing passwords (authored by kevans).
init: use explicit_bzero() for clearing passwords
Apr 1 2021, 5:32 PM
kevans committed R10:e0ad785a5d29: libevent1: fix layout of duplicated RB_ENTRY() definition (authored by kevans).
libevent1: fix layout of duplicated RB_ENTRY() definition
Apr 1 2021, 5:31 PM
kevans committed R10:223d6caabd2e: init: use explicit_bzero() for clearing passwords (authored by kevans).
init: use explicit_bzero() for clearing passwords
Apr 1 2021, 5:30 PM
kevans added inline comments to D29512: Loader: support booting OS from memory disk (MD).
Apr 1 2021, 2:29 PM
kevans added inline comments to D29533: libfetch: Fix proxy authentication.
Apr 1 2021, 12:44 PM
kevans added a comment to D29533: libfetch: Fix proxy authentication.

re: commit message:

Apr 1 2021, 12:39 PM
kevans added a comment to D29526: add soshutdown() calls to the server side krpc whenever a socket is no longer usable.

The idea seems generally sane to me, but I would definitely wait for tuexen to pitch in. My understanding is that shutdown() will either progress us somewhat quickly to FIN_WAIT_1 or LAST_ACK depending on our prior state, and the tcp state machine will thereafter deliver the final blow modulo freeing local resources.

Apr 1 2021, 4:33 AM

Mar 31 2021

kevans added inline comments to D29512: Loader: support booting OS from memory disk (MD).
Mar 31 2021, 3:27 PM
kevans added inline comments to D29512: Loader: support booting OS from memory disk (MD).
Mar 31 2021, 3:22 PM
kevans added a comment to D29505: loader: create a generic vendor sub-menu place holder.

Actually, you can go ahead and tack this on there after the vendor example; I suspect the vendor example will be more useful to folks and the below is just for informational purposes.

Mar 31 2021, 2:11 AM
kevans added a comment to D29505: loader: create a generic vendor sub-menu place holder.

+1 in advance for the idea of creating a vendor entry and changing the example to just reference using the vendor entry. I'll likely come back through after and add an example of modifying a submenu or something (e.g., adding another boot option) just to have one of those covered.

Mar 31 2021, 1:28 AM

Mar 29 2021

kevans added a comment to D24570: Add support for jail.d.

Sorry for the delays on this; I dropped some inline comments based on the last bit of discussion that took place. I think we're in a good place to start driving this forward again. There are other things we can do to improve this, but IMO we can get the basics in place to match existing facilities then further enhance it once it's in-tree, e.g., @jamie's latest comment about an empty jail_list also indicating anything we can find in these directories.

Mar 29 2021, 7:08 PM

Mar 24 2021

kevans committed R10:64c01719e476: libevent1: fix layout of duplicated RB_ENTRY() definition (authored by kevans).
libevent1: fix layout of duplicated RB_ENTRY() definition
Mar 24 2021, 4:41 AM

Mar 18 2021

kevans accepted D29031: fusefs: fix two bugs regarding fcntl file locks.
Mar 18 2021, 8:02 PM

Mar 17 2021

kevans committed R10:f187d6dfbf63: base: remove if_wg(4) and associated utilities, manpage (authored by kevans).
base: remove if_wg(4) and associated utilities, manpage
Mar 17 2021, 2:16 PM
kevans committed R10:cb370b19715b: service(8): use an environment more consistent with init(8) (authored by andrew_tao173.riddles.org.uk).
service(8): use an environment more consistent with init(8)
Mar 17 2021, 12:12 AM

Mar 15 2021

kevans committed R10:ff92a03616c5: if_wg: fix build with DIAGNOSTICS (authored by kevans).
if_wg: fix build with DIAGNOSTICS
Mar 15 2021, 12:36 PM
kevans committed R10:25529f11cff8: if_wg: close the sockets if wg_socket_bind() failed (authored by kevans).
if_wg: close the sockets if wg_socket_bind() failed
Mar 15 2021, 6:29 AM
kevans committed R10:99158c7fbd56: if_wg: stop holding creds in wg_socket_init() (authored by kevans).
if_wg: stop holding creds in wg_socket_init()
Mar 15 2021, 6:29 AM
kevans committed R10:11704859477c: if_wg: fix the !INET6 support (authored by kevans).
if_wg: fix the !INET6 support
Mar 15 2021, 5:42 AM
kevans committed R10:e277a507d8f9: tests: netinet: add missing up from local change (authored by kevans).
tests: netinet: add missing up from local change
Mar 15 2021, 5:42 AM
kevans added a comment to D29221: Add a basic smoke test for if_wg.

Crap, sorry, I forgot to remove this before I merged our branch over. =(

Oh no problem. Was it worth committing at all in light of the port of netns.sh? I didn't spend much more time on this once I saw you were working on that.

Mar 15 2021, 5:28 AM
kevans added a comment to D29221: Add a basic smoke test for if_wg.

Crap, sorry, I forgot to remove this before I merged our branch over. =(

Mar 15 2021, 5:18 AM
kevans committed R10:74ae3f3e33b8: if_wg: import latest fixup work from the wireguard-freebsd project (authored by kevans).
if_wg: import latest fixup work from the wireguard-freebsd project
Mar 15 2021, 4:53 AM

Mar 14 2021

kevans added a reverting change for R10:84eaf2ccc6aa: x86: stop punishing VMs with low priority for TSC timecounter: R10:1a1592459393: x86: tsc: deprioritize TSC on VirtualBox.
Mar 14 2021, 7:46 PM
kevans committed R10:1a1592459393: x86: tsc: deprioritize TSC on VirtualBox (authored by kevans).
x86: tsc: deprioritize TSC on VirtualBox
Mar 14 2021, 7:45 PM

Mar 13 2021

kevans closed D29081: relnotes/13.0: add note about jail(8) cpuset behavior.
Mar 13 2021, 7:21 AM
kevans committed R9:fdef22474f7a: relnotes/13.0: add note about jail(8) cpuset behavior (authored by kevans).
relnotes/13.0: add note about jail(8) cpuset behavior
Mar 13 2021, 7:21 AM

Mar 12 2021

kevans added a reverting change for R10:84eaf2ccc6aa: x86: stop punishing VMs with low priority for TSC timecounter: R10:ec24f78e5b20: x86: tsc: deprioritize TSC on VirtualBox.
Mar 12 2021, 6:45 PM
kevans committed R10:ec24f78e5b20: x86: tsc: deprioritize TSC on VirtualBox (authored by kevans).
x86: tsc: deprioritize TSC on VirtualBox
Mar 12 2021, 6:44 PM
kevans added inline comments to D29221: Add a basic smoke test for if_wg.
Mar 12 2021, 5:12 PM
kevans added a comment to D29221: Add a basic smoke test for if_wg.

I saw the upstream commit to use the ifnet creator's creds for the socket, so one can keep the wg socket in the host after moving the interface into a child vnet. Working on a test for that too.

Mar 12 2021, 4:58 PM
kevans committed rP568179: MFH: r568178.
MFH: r568178
Mar 12 2021, 4:01 PM
kevans committed rP568178: emulators/qemu-user-static: update to fix copy_file_range() bug.
emulators/qemu-user-static: update to fix copy_file_range() bug
Mar 12 2021, 4:01 PM
kevans accepted D29221: Add a basic smoke test for if_wg.
Mar 12 2021, 5:43 AM

Mar 11 2021

kevans committed R10:7a1f6858dabf: jail(8): reset to root cpuset before attaching to run commands (authored by kevans).
jail(8): reset to root cpuset before attaching to run commands
Mar 11 2021, 7:26 PM

Mar 9 2021

kevans added inline comments to D29146: kern: pull pru_sockaddr call out into sosockaddr().
Mar 9 2021, 9:21 PM
kevans committed R10:ce53f92e6c81: wg(4): note the persistent-keepalive ifconfig(8) option (authored by kevans).
wg(4): note the persistent-keepalive ifconfig(8) option
Mar 9 2021, 8:22 PM
kevans closed D29143: if_wg: export tx_bytes, rx_bytes, and last_handshake.
Mar 9 2021, 7:51 PM
kevans committed R10:94dddbfd00b9: if_wg: export tx_bytes, rx_bytes, and last_handshake (authored by kevans).
if_wg: export tx_bytes, rx_bytes, and last_handshake
Mar 9 2021, 7:51 PM
kevans closed D29145: iflib: allow clone detach if not yet init.
Mar 9 2021, 7:51 PM
kevans committed R10:0dd691b41276: iflib: allow clone detach if not yet init (authored by kevans).
iflib: allow clone detach if not yet init
Mar 9 2021, 7:51 PM
kevans closed D29142: if_wg: wg_input: remove a couple locals (NFC).
Mar 9 2021, 7:50 PM
kevans committed R10:299f8977cef6: if_wg: wg_input: remove a couple locals (NFC) (authored by kevans).
if_wg: wg_input: remove a couple locals (NFC)
Mar 9 2021, 7:50 PM
kevans added inline comments to D29143: if_wg: export tx_bytes, rx_bytes, and last_handshake.
Mar 9 2021, 3:10 PM
kevans added a comment to D29146: kern: pull pru_sockaddr call out into sosockaddr().

This method is used in a number of other places - was it your intention to convert them at some point?

Mar 9 2021, 3:03 PM
kevans added a comment to D29145: iflib: allow clone detach if not yet init.

This happens after an error from iflib_pseudo_register()?

Mar 9 2021, 2:54 PM
kevans requested review of D29148: ifconfig: wg: stop requiring a listen-port.
Mar 9 2021, 1:12 PM
kevans requested review of D29147: if_wg: choose a port at random if zero/omitted.
Mar 9 2021, 1:11 PM
kevans requested review of D29146: kern: pull pru_sockaddr call out into sosockaddr().
Mar 9 2021, 1:11 PM
kevans requested review of D29145: iflib: allow clone detach if not yet init.
Mar 9 2021, 12:17 PM
kevans requested review of D29143: if_wg: export tx_bytes, rx_bytes, and last_handshake.
Mar 9 2021, 11:21 AM
kevans committed R10:b3dac3913dc9: ifconfig: allow displaying/setting persistent-keepalive (authored by kevans).
ifconfig: allow displaying/setting persistent-keepalive
Mar 9 2021, 11:19 AM
kevans committed R10:172a8241c9c9: ifconfig: wg: stop requiring peer endpoints (authored by kevans).
ifconfig: wg: stop requiring peer endpoints
Mar 9 2021, 11:19 AM
kevans closed D29125: kern: malloc: fix panic on M_WAITOK during THREAD_NO_SLEEPING().
Mar 9 2021, 11:18 AM
kevans committed R10:1ae20f7c70ea: kern: malloc: fix panic on M_WAITOK during THREAD_NO_SLEEPING() (authored by kevans).
kern: malloc: fix panic on M_WAITOK during THREAD_NO_SLEEPING()
Mar 9 2021, 11:18 AM
kevans closed D29124: if_wg: avoid sleeping under the net epoch.
Mar 9 2021, 11:18 AM
kevans committed R10:e80e371d79d8: if_wg: avoid sleeping under the net epoch (authored by kevans).
if_wg: avoid sleeping under the net epoch
Mar 9 2021, 11:18 AM
kevans requested review of D29142: if_wg: wg_input: remove a couple locals (NFC).
Mar 9 2021, 11:02 AM
kevans added a reverting change for R10:df554850858f: wg_input: avoid leaking due to an m_defrag failure: R10:bae59285f932: if_wg: return to m_defrag() of incoming mbuf, sans leak.
Mar 9 2021, 10:54 AM
kevans committed R10:bae59285f932: if_wg: return to m_defrag() of incoming mbuf, sans leak (authored by kevans).
if_wg: return to m_defrag() of incoming mbuf, sans leak
Mar 9 2021, 10:54 AM

Mar 8 2021

kevans added a reverting change for R10:84eaf2ccc6aa: x86: stop punishing VMs with low priority for TSC timecounter: R10:8cc15b0dfc2f: x86: tsc: deprioritize TSC on VirtualBox.
Mar 8 2021, 8:44 PM
kevans committed R10:8cc15b0dfc2f: x86: tsc: deprioritize TSC on VirtualBox (authored by kevans).
x86: tsc: deprioritize TSC on VirtualBox
Mar 8 2021, 8:44 PM
kevans closed D29132: x86: tsc: punish VirtualBox with low priority for TSC.
Mar 8 2021, 8:44 PM
kevans added a comment to D29132: x86: tsc: punish VirtualBox with low priority for TSC.
In D29132#652389, @kib wrote:

From what I understand in the timecounter selection logic, tc with negative tc_quality can be selected with sysctl, but not by newly added tunable. If yes, it might make sense to set quality to zero, not a negative.

Mar 8 2021, 8:15 PM
kevans updated the summary of D29132: x86: tsc: punish VirtualBox with low priority for TSC.
Mar 8 2021, 7:46 PM
kevans added a reverting change for R10:84eaf2ccc6aa: x86: stop punishing VMs with low priority for TSC timecounter: D29132: x86: tsc: punish VirtualBox with low priority for TSC.
Mar 8 2021, 7:46 PM
kevans requested review of D29132: x86: tsc: punish VirtualBox with low priority for TSC.
Mar 8 2021, 7:46 PM
kevans added inline comments to D29125: kern: malloc: fix panic on M_WAITOK during THREAD_NO_SLEEPING().
Mar 8 2021, 7:34 PM
kevans updated the diff for D29124: if_wg: avoid sleeping under the net epoch.

Address feedback from @markj

Mar 8 2021, 6:51 PM
kevans added inline comments to D29124: if_wg: avoid sleeping under the net epoch.
Mar 8 2021, 6:04 PM
kevans added inline comments to D29031: fusefs: fix two bugs regarding fcntl file locks.
Mar 8 2021, 3:55 PM
kevans committed R10:2b82c94acff6: if_wg: avoid null ptr deref (authored by kevans).
if_wg: avoid null ptr deref
Mar 8 2021, 6:27 AM
kevans committed R10:df554850858f: wg_input: avoid leaking due to an m_defrag failure (authored by kevans).
wg_input: avoid leaking due to an m_defrag failure
Mar 8 2021, 6:23 AM
kevans committed R10:d9a50109e238: if_wg: release correct lock in noise_remote_begin_session() (authored by kevans).
if_wg: release correct lock in noise_remote_begin_session()
Mar 8 2021, 6:23 AM
kevans requested review of D29125: kern: malloc: fix panic on M_WAITOK during THREAD_NO_SLEEPING().
Mar 8 2021, 6:19 AM
kevans requested review of D29124: if_wg: avoid sleeping under the net epoch.
Mar 8 2021, 6:19 AM
kevans accepted D29122: Make kern.timecounter.hardware tunable.

LGTM, thanks!

Mar 8 2021, 1:24 AM