legal: Replace and relocate Apache-2.0 LICENSE to comply reuse #17
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
Currently, if lowpan_adaptation_queue_size() returns > 2, the tun interface is not read. However, poll() still wake up the main loop. This behavior lead to a useless wake up of the main loop and a waste of CPU resources. This issue typically happens during throughput tests when the Linux stack send frames faster than the RCP is able to send. See WI_SUN-3494. Signed-off-by: Jérôme Pouiller <[email protected]>
Only the local copy of the variable was incremented.
tcflush() does not work for this purpose as it discards unsent bytes.
This prevents frame cancelation, which creates annoyances regarding framing detection on the RCP.
APIv2 has an additional field to allow for a different payload size between the request and the confirmation. This oversight went unnoticed because the way the payload is contructed starts with 00 01 02..., which was interpreted as "a 1 byte buffer with contents 0x02". Testing the TX mode makes the issue apparent as the RCP expects to read payload length of 0 but instead finds that there are not enough bytes.
With --no-daemon, dnsmasq outputs logs to both stderr and syslog, which results in line duplication in journalctl. Using --keep-in-foreground fixes this issue.
Since PAN Advertisements transmission can be canceled by trickle, it is best to also include JM-IEs in data frames.
WI_SUN-3496
Aparently this is how the API is supposed to work, but it makes the DELETE endpoint very confusing and/or useless.
Aparently this is how the API is supposed to work, but it makes the DELETE endpoint very confusing and/or useless.
Reviewed-by: Kylian Balan <[email protected]>
Reviewed-by: Kylian Balan <[email protected]>
The hop limit selected at a tunnel entry is a local configuration of the
ingress, and has no relation with the received hop limit of the packet
to encapsulate. RFC 2473 section 6.3 describes the use of hop limit in
the case of tunnels in more details.
IP tunnels are normally invisible to tools like traceroute since an
increment of the probe hop limit covers the whole tunnel. However with
this quirk, traceroute needs to increase to hop limit for every
intermediate Wi-SUN router in the tunnel, but it also does not receive
ICMP error messages since they are destined to the ingress.
Here is an example to demonstrate the difference in behavior:
+ 2001:db8::1
Backhaul--| | Ethernet link
+ 2001:db8::2 (eth0)
| Linux + wsbrd
+ fd12:3456::200:0:80:3 (tun0)
| | Radio
| fd12:3456::200:0:80:0
Wi-SUN----| | Radio
| fd12:3456::200:0:80:1
| | Radio
+ fd12:3456::200:0:80:2
Executing traceroute from the backhaul device:
Before:
$ traceroute fd12:3456::200:0:80:2
traceroute to fd12:3456::200:0:80:2 (fd12:3456::200:0:80:2), 30 hops max, 80 byte packets
1 2001:db8::1 (2001:db8::1) 0.323 ms 0.368 ms 0.280 ms
2 * * *
3 * * *
4 fd12:3456::200:0:80:2 (fd12:3456::200:0:80:2) 40.184 ms 43.054 ms 45.506 ms
After:
$ traceroute fd12:3456::200:0:80:2
traceroute to fd12:3456::200:0:80:2 (fd12:3456::200:0:80:2), 30 hops max, 80 byte packets
1 2001:db8::1 (2001:db8::1) 0.353 ms 0.393 ms 0.318 ms
2 fd12:3456::200:0:80:2 (fd12:3456::200:0:80:2) 19.228 ms 25.906 ms *
WI_SUN-3547
The section title was modified twice by commits:
- 7501fa2e57e ("doc: fix other typos")
- f8105d4d8a6 ("doc: fix typos in README")
Before, TR_DROP traces would be printed when a confirmation is received without a frame (ie. for broadcast frames or TX failures), and pcap_file is active.
For an empty iobuf, this could cause a segfault. Not that this sentinel check was not necessary: if strnlen() returns n, it must have not found a terminator. Reported by SonarQube.
Delete the neighbor cache entry in case the node did not remove itself. Otherwise routing will choose an "ARO route" instead of a "DAO route", which will fail until ARO expiration. WI_SUN-3594
The new RPL implementation accidentally removed the neighbor insersion
on DAO reception. Consequently, only neighbors registered uding NS(ARO)
were present in the kernel tables.
See:
- d479a00b4c ("stack: instantiate new RPL implementation 3/3")
- b1e9b9cd26 ("stack: introduce new RPL implementation 2/3")
- 41424c9a65 ("stack: drop legacy RPL implementation 1/3")
WI_SUN-3608
WI_SUN-3614
This FATAL was proven necessary considering we could miss confirmation frames and end up in a instable state. WI_SUN-3652
This should never happen outside of the init phase. WI_SUN-3652
wsbrd docker container is no longer maintained and its usage is not recommended. Origin: SiliconLabsInternal/wisun-br-linux#2 WI_SUN-3385
For the record reusability has been improved a bit, using reuse lint tool: Files with license information: 272 / 307 The following files have no licensing information: * common/spinel.h * tools/tbu/api.yaml Origin: SiliconLabs#16 Signed-off-by: Philippe Coval <[email protected]>
For the record the file has been snapshoted this way: mkdir -p LICENSES ; lynx -dump https://www.silabs.com/about-us/legal/master-software-license-agreement \ | tail +161 | head -n 842 | tee LICENSES/LicenseRef-MSL | sha256sum # 1cfe8601a36fbff16686cf84d36dbd20c4c691faa9c4672231ba2eefa6f37e1b - Origin: SiliconLabs#16 Signed-off-by: Philippe Coval <[email protected]>
Note content of file has been changed,
It is a good practice to align to Original files
they are validated and trusted by SBOM tools that verify checksums
(bitbake for instance),
For the record:
Checksums-Sha256: eb0aad9b316fb87f01ec635a13ad59d63bf5cdbb90419ea4e62c11c19c5edeea LICENSE-Apache-2.0
Checksums-Sha256: 074e6e32c86a4c0ef8b3ed25b721ca23aca83df277cd88106ef7177c354615ff LICENSES/Apache-2.0.txt
Note file has been downloaded using:
reuse download Apache-2.0
Successfully downloaded LICENSES/Apache-2.0.txt.
Note that Pelion copyright can not apply to the context
of text of Apache work, (i believe it is a mistake)
the copyright is tracked on sources codes.
Current Reuse status is:
* Missing licenses: GPL-2.0
* Used licenses: Apache-2.0, GPL-2.0, LicenseRef-MSLA
* Files with copyright information: 272 / 309
* Files with license information: 272 / 309
Origin: SiliconLabs#17
Signed-off-by: Philippe Coval <[email protected]>
rzr
force-pushed
the
phcoval/review/main
branch
from
6014818 to
9a79519
Compare
|
This version is not formatted in 80 columns. Could we use this one: https://www.apache.org/licenses/LICENSE-2.0.txt |
|
I used the one provided by reuse , I believe this checksum is known as reference for related sdpx tools so I aligned to it: $ curl https://raw.githubusercontent.com/spdx/license-list-data/main/text/Apache-2.0.txt | sha256sum | grep 074e6e32c86a4c0ef8b3ed25b721ca23aca83df277cd88106ef7177c354615ff |
|
It looks like spdx uniformized style of licences to make it more consistent (and more semantic?). Style is very subjective, is it blocker for this PR ? |
As suggested in review, use the 80col formated version of doc Signed-off-by: Philippe Coval <[email protected]>
jerome-pouiller
approved these changes
Nov 19, 2024
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.
Note content of file has been changed,
It is a good practice to align to Original files
they are validated and trusted by SBOM tools that verify checksums (bitbake for instance),
For the record:
Checksums-Sha256: eb0aad9b316fb87f01ec635a13ad59d63bf5cdbb90419ea4e62c11c19c5edeea LICENSE-Apache-2.0
Checksums-Sha256: 074e6e32c86a4c0ef8b3ed25b721ca23aca83df277cd88106ef7177c354615ff LICENSES/Apache-2.0.txt
Note file has been downloaded using:
reuse download Apache-2.0
Successfully downloaded LICENSES/Apache-2.0.txt.
Note that Pelion copyright can not apply to the context of text of Apache work, (i believe it is a mistake) the copyright is tracked on sources codes.
Current Reuse status is:
Origin: https://github.com/SiliconLabs/wisun-br-linux/pulls?q=author%3Arzr