Linux 5.4 rc1 is released

From LinuxReviews
Jump to navigationJump to search
Tux.png

Half the lines changed in this release candidate for the next upcoming kernel were AMD DRM definition headers. We leave it up to AMD fans to speculate on what fancy new graphics technology they plan to release with support for the indicated new features.

Kernel-questions-5.3rc8.jpg

Linus Torvalds had all of this to say about the new release:

"I didn't really extend the merge window by a day here, but I gave myself an extra day to merge my pending queue. Thus the Monday date for the rc1 rather than the usual Sunday afternoon.

And it wasn't all _that_ big or painful a merge window, for some reason I just didn't get to the end of the queue until fairly late in the second week, and continued to get a few more pull requests even then. Part of it was just other discussions too happening, so I didn't do _just_ merges all the time. But part of it was just that I also spent some of Sunday away from the computer, doing some welding instead.

Anyway, what I'm saying is that the Monday rc1 isn't really a sign of any real trouble or more issues than usual. More just random timing.

Size-wise, 5.4 looks to shape up very regular. It's almost exactly the same size as 5.3 was at the same stage, both in commits and in lines added (honestly in advertising: 5.3 had more lines removed mainly due to some isdn removal). Nothing major stands out, the most notable may be the long-pending lockdown patches that weren't all that big, but that now finally aren't tied to just EFI secure boot, so you can test them out other ways too.

But there's changes all over. The patch is (once again) dominated by some AMD DRM register definition header file churn, which is almost half of it just by lines. But that's the nature of those beasts, being largely autogenerated from hw descriptions, and you should ignore it.

And if you do ignore those (very boring) header file changes, you'll see mostly the usual distribution: about 60% driver updates (GPU, networking, sound, staging), with the rest being the usual mix of architecture updates (ia64 shows up thanks to removal of the odd sn platform, but it's otherwise mostly arm[64], powerpc and x86), tooling, filesystems, documentation and misc (core networking, code kernel and VM, include files etc etc).

Much too many changes to list individually, and over 1500 individual developers. The log below is just my usual "merge log" with who I merge from and a oneliner very rough description of the area it updates.

Go test, the merge window has now closed,

Linus"

Linus Torvalds on the Kernel Mailing List on Monday September 30, 2019

It's interesting to note that a large set of Kernel lockdown patches submitted for v5.4 did not make it into the release.

Linus had this to say about them earlier in the release-cycle:

" This is one of the pull requests that I have to go through commit by commit because of the history of this thing.

And I've yet to empty my queue of all the _regular_ things that came in this merge window, so I haven't had time."

Linus Torvalds on the Kernel Mailing List on Monday September 27, 2019

Too much to do, too little time. Sounds familiar.

The new kernel release candidate is, as always, available from kernel.org for those who want to test those new features that did make it into this release-candidate.


published 2019-10-01last edited 2019-10-01

0.00
(0 votes)


Add your comment
LinuxReviews welcomes all comments. If you do not want to be anonymous, register or log in. It is free.