WARNING - OLD (2018-KIT) OZSOUTH KERNELS OUTDATED

[ozsouth]

WARNING:-
2018-KIT KERNELS & OLDER FDRVS SHOULD NOT BE USED WITH WOOF-CE PUPPIES CREATED AFTER 2022.
STRUCTURAL CHANGES WILL LIKELY CAUSE CORRUPTION.
EXCEPTION: Slackware derivatives.

Well, my methods for kernels & fdrvs just hit 'the wall'. Some Pups in 2022 & most Pups from 2023 on, that are produced via
woof-ce, have a different structure to that of many years prior, & it is necessary to use kernels & fdrvs produced via woof-ce.
Failure to do so will likely result in corruption. The new structure is in line with Ubuntu's structure & will become universal.
This will standardise Puppy offerings. Trying to hold back the tide of progress is not worthwhile. I used 2018 Kernel-Kit Master,
with only my essential-for-operation fixes.

Puppies created prior to 2023 (with the exception of some 2022 woof-ce offerings) can still use my previous kernels & fdrvs.
S15Pup64-22.12+1 is unaffected by this, and I will continue to use it, but mostly use my Chromebook.

[ally]

Oz

I've added this warning to the huge kernels mirror

https://archive.org/details/Puppy_Linux_Huge-Kernels

[Marv]

@ozsouth, Is the change discussed here: https://forum.puppylinux.com/viewtopic. ... 920#p78920 the cause? Just trying to swim with the tide...

[rockedge]

It's possible to convert the zdrv and fdrv structure to be able to run on both the old and new. I have experimented with having the fdrv have both /lib and /usr/lib directory structure and the kernel can be interchangeable without recompiling anything

[Marv]

I've converted the 6.1.5 library I use in all my pups at this point and it boots and runs correctly in the 64b jammypup with the 'new' library structure.

@dimkr

I also have converted my FP64ce_9.6_radky6 ydrives for testing. So far the JWM flavour runs correctly with one lib needed by PCManFM present in FP64 but not in jammy64 added. I haven't tested the LXDE flavour as yet. While I was unsquashing and resquashing, I pulled Palemoon and Abiword from the main jammypup64ce_10.0 SFS. Posting in portable Brave from a non-savefile non-fdrv or bdrv, personalized 8.1 MB ydrv, 6.1.5 64b kernel jammypup64ce_10.0 install. So, The bridge while I'm running both pre and post change puppies is a bit of a PITA, but this old dog can make the leap.

@rockedge, we need a jammypup location/discussion thread....

Update: Changing the directory structure in my LXDE ydrv broke LXDE itself pretty seriously. Boot is OK, Openbox and PCManFM are OK, but lxpanel doesn't start and there seems to be a general mimetypes failure. I may pick away at it but it remains very much a side-excursion for now.

[dimkr]

@ozsouth No need to be so dramatic! woof-CE still defaults to the old layout, without the /usr symlinks (set USR_SYMLINKS=yes in _00build.conf to opt-in to the "new" layout). AFAIK Slackware refuses to switch, so woof-CE must support the old layout. Anyone using woof-CE is free to build a Puppy without those symlinks, even an Ubuntu based one. You can build jammy64 with the old layout: it will work mostly OK, but you lose (some?) Ubuntu compatibility to increase compatibility with old .pet packages and SFSs.

[ozsouth]

Sorry if that sounded overly dramatic, folks. With the direction linux in general is moving, I felt it necessary to draw a 'line in the sand' & not push against the tide. Whilst there is some freedom to build old styles, that is not where the majority of effort is headed, so I now see this as the time to change. I don't want to see average folk trying my older style kernels/fdrvs on new Pups & coming unstuck. No-one needs that unnecessary problem. Also, many 'old pup' users simply stick to the original issued kernel, so I'm seeing single-digit downloads. Most folk simply want a Puppy iso to download & run. Peebee has made some produced under the new system readily available. The new S15Pup64-22.12+2 is polished & nicely automated, as I expect the new Jammypup64 (& others) will be.

[Jasper]

@ozsouth

I do enjoy trying out your new kernel updates. I have only jumped from the original FossaPup64-95 kernel to 6.0.x and then onto 6.1.x.

With each incremental update, I have seen an improvement in my OS especially with memory/hardware management.

When I do try an update I always provide feedback and when successful, screenshots as evidence. Hopefully, to encourage other members to attempt the same.

I took those steps only as to me it looked like huge jumps a bit like moving from up stages in the MS OS releases.

How I feel about it is that it may have taken you 3 hours to do the work, but for me your are increasing the lifeline of my PC with the many hours that I am still able to use a PC which I might have consider replacing. For me that's a "win"

Personally, I have learned a lot in the space of a few months and this OS is incredibly versatile in it's deployment, usage and the updates provided by you and other members on-board. The mantra I always hear is the freedom of choice.

Keep up the amazing work, it is appreciated

[Jasper]

@ozsouth

Sorry, I forgot to add that I use your Microcode security update alongside your kernels.

I was unaware of the Spectre problem and your update ensures security as do the kernels I use.

You provide a better UIX.

Thank you

[ozsouth]

@Jasper - your unsuccessful attempt to use one of my kernels in a 2023 Pup is what I'm trying to avoid. In your case no serious damage was done, but I can't guarantee that. I've had some suggestions that perhaps an occasional kernel with the appropriate 'Not for use in post-2022 Pups' warning is desirable, so maybe for now I will produce one early in each of the four seasons - just an aufs/overlayfs/voluntary preempt one, as they are the most popular. That should be frequent enough to cover security concerns.

[Jasper]

@ozsouth

It is all part and parcel of the learning process. When I have time, I can extract the fdrv and see if I need to add an additional driver to get it working. Apart from that I would be using it right now.

However, you decide to roll out these kernels and the time in-between will sufficient. I was surprised how often you managed to do all the work.

It's a huge benefit if t works straight out the box, but if it involves a little input from users that is up to them to decide what they want to do. So it's obvious that you cannot guarantee it will work for each individual who decides to use it.

Hardware changes and I understand the kernel deprecates older hardware.

I'm happy with the status quo and your decision. I would have been sad if you had decided to call it a day

Many thank again!