The availability of a PAE kernel might encourage some of them to install manjaro32 to help with testing. (Not me, I will compile linux-pae from AUR it you won’t make it available and stay with manjaro32 anyway.)
4.14 introduced new features compared to 4.9, some users with very old machines might still want to use these features.
Another point is that a kernel series which gets additional testers (PAE capable machines) whould be of benefit to all users even if they don’t need all the additional features because of this additional testing.
More important: If linux414 is to be shipped with the next ISO then it should be nonPAE on the ISO!
But if you want not to increase the number of supported kernels or even to reduce it linux41 and linux49 could be dropped. linux41 will become EOL anyway soon https://www.kernel.org/releases.html
linux49 was Manjaro default kernel for a long time it might have some more users. But does it have many features which linux44 hasn’t got?
For anyone interested in what PAE is …
As long as PAE kernels don’t adversely effect system without PAE support, this seems like a no brainer.
Indeed, I’m not sure what would even be the point of a 32 bit kernel once you don’t need to crave for even some hundred of MBs of ram?
And I’m on the broadly sure side that CPUs that don’t even support it are quite… behind 4GB in maximum supported memory.
Actually… I guess like people with [not that old] 32-bit only cpus actually would kinda bother about this (even if it’s just about going from 3GBs to 4)
As said, it won’t boot on non-PAE cpus.
Which are quite old, thus few people.
But on the other hand I’m skeptical about how many people would benefit of this too.
Doesn’t matter. It’s my memory, I paid for it, and damnit, I want to use it!
Curiously, I just installed manjaro32 on an old eeepc 1025c with 4GB of RAM and noticed that it only uses 3. So, as it doesn’t support 64 bits, I’d be interested in a PAE kernel, yeah (I used to use one on it with Mint/Xubuntu). Though this one is a really weird animal, every kernel above 3.16 mess with fan control for some reason.
We overlooked it back them.
Plus I didn’t really know what I was doing at the time…
(I still don’t, but don’t tell anyone )
So - the question now becomes, which kernel package do I have to create for this? I really don’t want to have a -pae version for each kernel as that doubles the effort every time the kernels get an update.
Can I assume anyone with a PAE-enabled CPU will be fine with 4.14?
Well as I said only 3.16 works 100% well on my machine, but it would be too much to ask for a rebuild only for my special case.
If you offer a -pae version of each kernel you can justify dropping some of them.
linux414 and linux414-pae
linux44 and linux44-pae
linux316 and linux316-pae
Won’t -pae kernels benefit massively from ccache if the regular kernel has been just built?
Let’s wait for more responses to get a better picture.
Mayybeee… In fact, yes, they should. I’d have to try it to see what effect it would have in practice.
A PAE option just for the newest LTS kernel series would be a good compromise (well, for almost everyone.) The older LTS kernels support a larger variety of old hardware. Newer kernels occasionally drop support for “ancient” devices. Most ancient 32 bit devices won’t have that much memory.
Oddly, my atom baytrail device doesn’t need PAE as it only has 2 GB RAM soldered, no sockets for more. The 32 bit OS is a little stingier with RAM so it’s a better choice for my 64 bit capable system. But a PAE kernel would boot fine on it.
A bit few votes so far, but thanks for those who voted and wrote. Few is much more than zero.
I hope for some more feedback!
I’m using (more dabbling occasionally) manjaro32 xfce on an old AMD system that has a 64-bit processor and only 2GB of DDR2-800 RAM at present, some of which is shared by the onboard GPU. so it’s ideal as a test machine for manjaro32 but could equally be upgraded to 4GB RAM and happily run 64-bit instead.
Personally I have no need for PAE but if it means someone with an older machine gets to enjoy the distribution more then why not make the kernel available if you have the time to maintain the extra package(s) involved as well.
I’ll give this a +1 my pure test machine for manjaro32 is 64bit and has 4GB ram. I’ve also recently acquired an ancient mobile centrino (I think) embedded system that’ll I’ll spin into a test system shortly.
You haven’t checked the box “I’m running manjaro32”.
The Centrino laptop will probaly have 1GB at most.
Fixed It’s actually a Celeron mobile CPU not a centrino (I was wrong), clocked at a blazing 1.3Ghz and a whopping 1GB of DDR266 RAM…
It’s having some issues installing Manjaro32 right now with lightdm.service coredumping on installer boot. I don’t have much time to really diagnose but it’s booting and installed Debian 9.4 fine so I can rule out hardware issues. I’ll make it my weekend project.
This is the old beast, if interested. It’s a rev1.5 so slightly better CPU that mentioned on website.
17.1.7-rc1 image fails to boot
Is there any news about when will the pae kernel packages be available?
No news, but you are invited to participate in the vote and tell why you would like to have a PAE kernel!