> There are wins reported for Cassandra, Hadooop, MySQL/MariaDB, Memcached, MongoDB, PostgreSQL, Redis, and improving the behavior in general for systems with limited RAM capacities or a lot of memory intense activities.
> MGLRU is already shipping with Google Android devices and also patched into Google's Chrome OS kernel.
Why would they? Linus changes version numbers on a whim, he stated that multiple times. Partly to make it obvious that all kernel releases are equal, partly for no reason at all.
Please don't complain about tangential annoyances—things like article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
Yes, ultimately as there is no perfect and never break system. But pushing that is pointless as we need to have a safeguard now. What should it be? We can't wait for the God or Godet to come. We need a human solution.
And that might be faulty. Still, 5k lines are better than ...
It was working well before. And the Linux kernel has a history of highly pozzed, insecure RNG's that were given to the community by cryptographic "experts" as well.
Wireguard is probably the part of the kernel I trust _the most_ since it's < 5,000 lines of code and easily audited. I don't think the alphabet boys are involved in this one.
It's one of the most carefully reviewed pieces of cryptographic software on the Internet. You could just use Google, or, if you like, Google Scholar, to verify that for yourself. Please don't write troll comments on HN.
31 comments
[ 3.2 ms ] story [ 17.3 ms ] threadhttps://www.phoronix.com/scan.php?page=news_item&px=Linux-5....
https://www.phoronix.com/scan.php?page=news_item&px=Linux-5....
https://www.phoronix.com/scan.php?page=news_item&px=Linux-5....
https://www.phoronix.com/scan.php?page=news_item&px=Linux-5....
https://www.phoronix.com/scan.php?page=news_item&px=Btrfs-Li...
https://www.phoronix.com/scan.php?page=news_item&px=Intel-IF...
But MGLRU alone is impactful enough to finally fix a multi-decade desktop limitation, so I don't even care about other changes.
https://www.phoronix.com/scan.php?page=news_item&px=MGLRU-v1...
> There are wins reported for Cassandra, Hadooop, MySQL/MariaDB, Memcached, MongoDB, PostgreSQL, Redis, and improving the behavior in general for systems with limited RAM capacities or a lot of memory intense activities.
> MGLRU is already shipping with Google Android devices and also patched into Google's Chrome OS kernel.
https://lwn.net/Articles/856931/
https://lwn.net/Articles/894859/
An aspect I found interesting was that, in addition to the LRU aspect, it includes a PID controller to try to learn or compensate for its mistakes.
- Unrolled tweets: https://threadreaderapp.com/thread/1528494394604761094.html
- The merge commit: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...
- Some discussion: https://lore.kernel.org/lkml/YmlMGx6+uigkGiZ0@zx2c4.com/
- The commit from that discussion: https://git.kernel.org/pub/scm/linux/kernel/git/crng/random....
- Source code: https://git.kernel.org/pub/scm/linux/kernel/git/crng/random....
- More commits: https://git.kernel.org/pub/scm/linux/kernel/git/crng/random....
It is.
Please don't complain about tangential annoyances—things like article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
https://news.ycombinator.com/newsguidelines.html
That was very well written for sure. If only all projects had merge commits with so much useful info!
And that might be faulty. Still, 5k lines are better than ...
You are welcome.