Moved pid unlink. Added Reasons for script. Added signals #299
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
pid unlink was called before all processes were complete. This could leave dhcp6c still runing sending and waiting for release but any check for process by PID returned false, moved to just before exit(0); Env var REASONS only returned NBI, now returns various reasons for script call, these can be: REASON=INFO REASON=REPLY REASON=RENEW REASON=RELEASE REASON=REBIND REASON=EXIT REASON=OTHER EXIT will be signalled when the client exits. OTHER is a catchall and should never happen. Signals SIGUSR1 and SIGUSR2 are added. These are used to force an exit ignoring the state of the no-release flag. SIGUSR1 forces a no-release exit. SGUSR2 forces a release exit. Logging info changes to case and text; logging change to indicate PD or IA in info mode rather than debug.
|
I would like to hear from @hrs-allbsd about hrs-allbsd/wide-dhcpv6#5 before proceed with this change |
|
So would I, it's been there a week. |
|
I took the liberty of contacting @hrs-allbsd. He will try to look at the PR and merge it on Friday. |
|
Friday came and went and the PR has still not been merged. I sent another email to @hrs-allbsd, but I haven't heard back yet. Considering that the WIDE / KAME projects shut down in 2005, it's not surprising that there have been virtually no changes to WIDE-DHCPv6 (other than another recent fix for pfsense a few months ago) for 10 or 11 years. However, there have been many revisions to the RFCs that were in place in 2005 when WIDE-DHCPv6 was released, as well as new RFCs. WIDE-DHCPv6 is a dead project comprising one person (@hrs-allbsd) and it's clearly not his priority. On the other hand, KEA DHCP is being actively developed and will replace ISC DHCP. Maybe pfsense should move to KEA DHCP. |
|
Kea is a v4 \ v6 server, i cannot see any mention of a dhcpv6 client; it also appears to be c++ which means using extra libraries in the same way as Dibbler when i got that to work for Kev. At present there is not a really good v6 client. Dibbler is not bad, but has not been fully ported to *BSD and is missing some things that make it hard to use with pfSense, or any BSD system.
Its not pfSense in this case that's the problem, its FreeBSD.
Sent from: YOGA Tablet 2-1050F
On 26 Feb 2017 18:29, bimmerdriver <notifications@github.com> wrote:
Friday came and went and the PR has still not been merged. I sent another email to @hrs-allbsd, but I haven't heard back yet. Considering that the WIDE / KAME projects shut down in 2005, it's not surprising that there have been virtually no changes to WIDE-DHCPv6 (other than another recent fix for pfsense a few months ago) for 10 or 11 years. However, there have been many revisions to the RFCs that were in place in 2005 when WIDE-DHCPv6 was released, as well as new RFCs. WIDE-DHCPv6 is a dead project comprising one person (@hrs-allbsd) and it's clearly not his priority. On the other hand, KEA DHCP is being actively developed and will replace ISC DHCP. Maybe pfsense should move to KEA DHCP.
…--
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub:
#299 (comment)
|
|
I didn’t notice that. If so, it makes sense.
|
|
I've cherry-picked freebsd/freebsd-ports@3a3ac4a to RELENG_2_3 and devel, it will be available on next round of snapshots. |
netgate-git-updates
pushed a commit
that referenced
this pull request
Jun 26, 2019
[ Robert Edmonds ] * Release 1.3.2. * Use protobuf 3.7.1 in the Travis-CI environment (#368). * Fix test suite build failure on newer versions of protobuf (#369). [ Ilya Lipnitskiy ] * Fix proto3 repeated scalar field default packing behavior (#330, #377). [ Adam Cozzette ] * Fix out-of-bounds read in scan_length_prefixed_data() (#375, #376). [ Jurriaan Bremer ] * Fix -Wdeclaration-after-statement warning in parse_oneof_member() (#360). [ Hayri Ugur Koltuk ] * Fix SIGSEGV in protobuf_c_message_check() on messages with unpopulated oneof members (#358). [ Italo Guerrieri ] * Do not allow tag values of 0 in protobuf messages, as these are not allowed by proto2 or proto3 (#299). The patch for version 1.3.1 is no longer required. Sort Makefile statements to pacify portlint. Reported by: portscout Sponsored by: Farsight Security, Inc.
netgate-git-updates
pushed a commit
that referenced
this pull request
May 29, 2020
- Updates and documents a patch. Changes since 2.1.1: Library: * A library view showing recent books and reading progress (#119) * Search books by metadata * E-book discovery with OPDS (#6, #253) New supported formats: * FictionBook (.fb2, .fb2.zip) (#128) * Comic book archive (.cbr, .cbz, .cbt, .cb7) (#254) * Plain text (.txt) (#226) * Unpacked EPUB files Layout: * Option to set maximum page width (#192) * The "Automatic" layout will now show four columns when the page width is wide (#257) Image viewer: * Improved image viewer, with new shortcuts and the ability to rotate (#299) and invert images * Option to disable image viewer, or to open images on double, middle, or right click (#315) Text-to-speech: * Improved text-to-speech configuration UI * Ability to set text-to-speech voice based on book language (#134) Annotations * Import annotations from JSON * Search in annotations * Annotations are now sorted in the same order as they appear in the book (#318) * "Allow Unsafe Content" now only enables JavaScript; external content will no longer be loaded. This is an important security fix. If you are running an old version of Foliate and cannot upgrade, you should disable "Allow Unsafe Content". * WebKit processes are now sandboxed Other featues and fixes: * Improved support for vertical and right-to-left books * Support for StarDict dictionaries (#302) (requires new optional dependency sdcv) * Fixed current location not preserved when resizing (#152, #204) * Fixed auto-hiding header bar under some themes (#316) * Command-line options for showing version and adding books to library * Option to clear or disable cache * When an HTML page contains invalid self-closing anchor tags, it is now parsed as XHTML (#275, #288) Technical changes: * Foliate now stores file locations in ~/.local/share/com.github.johnfactotum.Foliate/library/uri-store.json, and saves book covers to ~/.cache/com.github.johnfactotum.Foliate. Both can be disabled. * New optional dependency tracker (gir1.2-tracker-2.0 on Debian based distributions), which can be used to get file locations. * New optional dependency libhandy (gir1.2-handy-0.0). It's not absolutely required, but installing it would make the experience better on larger as well as smaller screens. * To open remote files, Foliate now optionally depends on libsoup (gir1.2-soup-2.4). Foliate now bundles two additional libraries: * libarchivejs, which is MIT licensed. It is a WASM port of the popular libarchive C library. * crypto-js, which is MIT licensed. The MD5 module is used to generate identifiers for files that don't have unique identifiers.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
pid unlink was called before all processes were complete. This could leave dhcp6c still running sending and waiting for release but any check for process by PID returned false, moved to just before exit(0); This is probably the reason for cases of multiple dhcp6c instances. The original method of killing dhcp6c using SIGTERM, SIGTERM, SIGKILL would fail as there was no pid by the time it got to SIGKILL.
Env var REASONS only returned NBI, now returns various reasons for script call, these can be:
REASON=INFO
REASON=REPLY
REASON=RENEW
REASON=RELEASE
REASON=REBIND
REASON=EXIT
REASON=OTHER
EXIT will be signalled when the client exits.
OTHER is a catchall and should never happen.
Signals SIGUSR1 and SIGUSR2 are added. These are used to force an exit ignoring the state of the no-release flag. SIGUSR1 forces a no-release exit. SGUSR2 forces a release exit. SIGUSER1 is very useful as
it kills dhcp6c quickly without sending and waiting for release on a WAN down event.
Logging info changes to text case and text; logging change to indicate PD or IA in info mode rather than debug.
This client will work without script updates being required.