The Ultimate "When Will The Next Slackware Release Arrive" MegaThread
SlackwareThis Forum is for the discussion of Slackware Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Originally posted by win32sux
real slackers compile their own kernels anyways, so it really doesn't matter to them what patrick does with the distro's kernel...
Amen!
And they also update their system with Patricks's apps in -current, or compile from source; they are not still on the bottle waiting for Patrick to release a new Slackware version so that they can use new software.
True.
I love the consistant reliability of Slackware and prefer to wait for the next release to upgrade. Whilst a lot of apps have been compiled from source, I don't have a lot of time to play around with my system, and so depend upon Slackware's stability for trouble free computing (It's consistency is a joy). I am looking forward to the next release and am perfectly happy to wait.
I prefer the true and tried, rather than the bleeding edge that many distros impose on their unwary users. Time spent learning Slackware/Linux is far more productive than time wasted hacking around some unstable alternative.
In my experience "Slackware makes (is) good sense!"
Yeah, I'm finding that what you guys told me when I first started is so true, it's not like windows where you need to get your incremental patches, fixes, updates, every day.
Maybe for something like openoffice, ok, but I dont need bleeding xorg 7.0 that gives DRI for maybe .0002% of the linux user population.
I eagerly awaiting the release of 11, I only have 10.0 and currently making my own distro from it. Installed only the packages I want to make a Desktop only OS and tarred and gzipped the drive. At present I'm working on a graphical installer, like Redhat's anaconda, except it will only install a single file instead of packages. If all goes to plan and I can iron out some problems i'm having, I should have it finished when 11 is released, then I can use 11 instead.
There's one thing I don't quite understand - if someone could post a pointer to where I can read up on this I'd greatly appreciate it:
It seems like Pat will support both 2.4 and 2.6 in the upcoming release. At the same time, it's generally frowned upon replacing the default system kernel headers even though one upgrades from 2.4.x to 2.6.x kernels. (I'm presently on 2.6.16x but keep the 10.2 2.4.x kernel headers). To complicate further, I get the impression from LKML that 2.6 kernel headers are preferred for glibc etc.
So - for the upcoming Slackware release - how can he support both 2.4 and 2.6 glibc/etc kernel headers, and what are the consequences both compatibility (both future and historic) and performance wise?
Is there a negative impact for those of us running 2.6 kernels if he still bases the next release on 2.4 from a header and library point of view?
The thing is that you must use the kernel headers that glibc is compiled against.
You can have 2.2, 2.4, 2.5, 2.6, and 2.17.6 kernels all compiled against the version of glibc that Slackware-10.2 uses.
You cannot use kernel headers compiled with glibc-2.3.5 and then update to glibc-2.3.6 and use the same kernel. You will have to recompile your kernel against that new glibc, or software won't compile. So long as you leave the glibc alone, and the headers are compiled against it, you'll have no problems with new kernels.
Notice in the -current ChangeLog Pat states:
Quote:
a/glibc-solibs-2.3.6-i486-3.tgz: Recompiled against 2.4.32 and 2.6.15.6
kernel headers. Yes, I have seen that shiny-looking glibc-2.4 release on
ftp.gnu.org, but glibc-2.4 completely drops support for linuxthreads, and
therefore will not support vanilla Linux 2.4.x kernels. I don't think
we're quite ready for that yet around here.
When he recompiled glibc, he updated the kernel headers. So if you updated your system to that glibc and were using an older Slackware generic kernel, you would have needed to update to one of those 2 kernels Pat recompiled against.
Also notice in the ChangeLog:
Quote:
kernels/huge26.s/*: Upgraded huge26.s kernel to 2.6.16.24.
The name of the big kernel with many built-in options has been changed from
test26.s to huge26.s to reflect that Slackware 11.0 will consider the
2.6.16.x kernel series to be a supported kernel series. However, I'm
probably going to leave the bare.i 2.4.32 kernel as the default kernel (or
perhaps sata.i?) as it has very good performance and probably better security
due to the simpler and longer-tested design. I might apply or at least make
available in the kernel-source package for 2.4.32 a patch to fix direct
rendering with 2.4.x kernels and X.Org 6.9.0 or newer. Since anyone using
Slackware for server use isn't likely to be loading the DRI modules, it's
untouched code on those machines and won't affect server stability (well,
depending on what, if anything, outside of the module is changed in the
kernel). It is probably a safe enough patch to apply. I'd rather ship 100%
vanilla kernels (and might, with the patch "on the side"), but DRI does not
work without the patch past X.Org 6.8.2. Is this enough text here?
Perhaps I should rename this my "ChangeBlog".
Just as it is right now, you will install all the packages with the kernel you select. If you're running Pat's generic 2.6.17.4 kernel in -current/testing/ now you would have needed to install:
Code:
kernel-generic-2.6.17.4-i486-1.tgz (Pat's generic kernel)
kernel-headers-2.6.17.4-i386-1.tgz (the headers for that kernel, compiled against Slackware -current's glibc-2.3.6-i486-3)
kernel-modules-2.6.17.4-i486-1.tgz (all the modules for that kernel's config)
kernel-source-2.6.17.4-noarch-1.tgz (the source, but you don't *have* to install it I don't think)
As for compatibility issues, when you keep your glibc from the initial install, and keep those kernel headers intact; you can compile new kernels all day long. I went through about 20 different kernel compiles on Slackware-10.1 -- trying new ones hoping for some ACPI support. I never changed glibc, and never touched my /usr/src/ area, and never experienced one single problem. You might want to read my little kernel compile guide and rant here. I still think building under /home as Linus says, and not touching /usr/src/ is the best idea.
Quote:
Originally posted by Yalla-One
I get the impression from LKML that 2.6 kernel headers are preferred for glibc etc.
Which version of glibc would they be referring to?
IMO they are preferred, but that's also debatable. There are more apps that we must take into consideration, which is where Pat's present 2.4 vs. 2.6 dillema lies. Most people are using a much later version of udev, HAL, NTPL. etc. with later versions of glibc and 2.6 kernels. I think this is the water into which Pat will wade with -current after Slackware-11.0 is released.
Quote:
Originally posted by Yalla-One
So - for the upcoming Slackware release - how can he support both 2.4 and 2.6 glibc/etc kernel headers, and what are the consequences both compatibility (both future and historic) and performance wise?
Is there a negative impact for those of us running 2.6 kernels if he still bases the next release on 2.4 from a header and library point of view?
I think what guys are missing is that Pat will have, as outlined above from -current, two complete sets of kernels and headers compiled against whatever glibc Slackware-11.0 comes with that you can choose when you do the install.
As for performance, there is no doubt that 2.6.x kernels perform better than 2.4 kernels. I'll leave the stability issue to those who want to argue that, but I've had no kernel oops or other problems with anything from 2.6.11 onward.
Tue Jul 18 22:37:26 CDT 2006
a/lilo-22.7.2-i486-1.tgz: Upgraded to lilo-22.7.2.
kde/koffice-1.5.2-i486-1.tgz: Upgraded to koffice-1.5.2.
Thanks to the KOffice team who did incredible work on this.
kdei/koffice-l10n-*-noarch-1.tgz:
Upgraded to l10n packages for koffice-1.5.2.
n/samba-3.0.23-i486-2.tgz: Patched a problem in nsswitch/wins.c that
caused crashes in the wins and/or winbind libraries. Thanks to
Mikhail Kshevetskiy for pointing out the issue and offering a
reference to the patch in Samba's source repository.
Thanks again to Andrea for this batch of kernel packages, and also thanks
for compiling all those intermediate kernels that were replaced upstream
and went unreleased in Slackware -current...
Ah, the things that go on here behind the scenes. ;-)
extra/linux-2.6.16.27/alsa-driver-1.0.11_2.6.16.27-i486-1.tgz:
Upgraded to alsa-driver-1.0.11 compiled for Linux 2.6.16.27.
extra/linux-2.6.16.27/kernel-generic-2.6.16.27-i486-1.tgz:
Upgraded to Linux 2.6.16.27 generic kernel.
extra/linux-2.6.16.27/kernel-headers-2.6.16.27-i386-1.tgz:
Upgraded to Linux 2.6.16.27 kernel headers.
extra/linux-2.6.16.27/kernel-modules-2.6.16.27-i486-1.tgz
Upgraded to Linux 2.6.16.27 kernel modules.
extra/linux-2.6.16.27/kernel-source-2.6.16.27-noarch-1.tgz
Upgraded to Linux 2.6.16.27 kernel source.
kernels/huge26.s/*: Upgraded huge26.s kernel to 2.6.16.27.
kernels/test26.s/*: Upgraded test26.s kernel to 2.6.17.6.
testing/packages/linux-2.6.17.6/kernel-generic-2.6.17.6-i486-1.tgz:
Upgraded to Linux 2.6.17.6 generic kernel.
testing/packages/linux-2.6.17.6/kernel-headers-2.6.17.6-i386-1.tgz:
Upgraded to Linux 2.6.17.6 kernel headers.
testing/packages/linux-2.6.17.6/kernel-modules-2.6.17.6-i486-1.tgz
Upgraded to Linux 2.6.17.6 kernel modules.
testing/packages/linux-2.6.17.6/kernel-source-2.6.17.6-noarch-1.tgz
Upgraded to Linux 2.6.17.6 kernel source.
I wonder how long to go now. I have a machine that I just upgraded the CPU and motherboard in and am going to reformat my Linux partition and install 11 when it comes out. For the time being, Windows will have to do :/.
hmm, you're right about something here, but I still think (or atleast hope) Pat will make it the default. atleast it's a start using the 2.6 kernel headers. Anyway, it hasn't been released yet, who knows !?
hmm, you're right about something here, but I still think (or atleast hope) Pat will make it the default. atleast it's a start using the 2.6 kernel headers. Anyway, it hasn't been released yet, who knows !?
Yes, it is interesting indeed to see what 11 will be like. It isn't certain yet what Pat will use as the default kernel, he says in Friday's changelog that he will "probably" use the 2.4x kernel as the default.
I'll be happy with what ever Pat decides to use. If he releases 11 with the 2.4x kernel then that'll motivate me to finally compile a 2.6x kernel.
11 is almost here:-)
well, we know that both 2.4.x and 2.6.16.y will be provided and supported... so what difference does it make which one is default?? geez, you're gonna use the one you want either way... slackware 11.0 will be ready when it's ready... i think it would be cool if patrick's TODO list would be public so we could know how far along we really are without all the speculation... it might even help speed things along by letting testers know what to focus on...
Because glibc compiled against 2.6.x kernels performs MUCH better than against 2.4.x, and glibc 2.4 is lightyears better than 2.3, yet we are still on 2.3 because kernel 2.4 is the default.
THere are many, many great things that we cannot have if 2.6 isn't the default. Even if it is offered, the 2.4 as default policy is preventing us from accessing those other benefits of 2.6.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.