udev fork
Hi,
I know that many people do not like the way udev development went recently, therefore we have a systemd-free fork of udev. At the moment most of the changes (except for systemd integration) have been ported, and no stability issues are noticed. I hope it will be useful for those who don't like systemd being pushed in their systems.
P.S. This maillist seems to be dead for quite a while and most of the discussion moved to systemd-related lists, so I hope nobody would mind if we use it as our primary maillist.
Quoted 8 lines · On 17:49 Wed 12 Sep , [email protected] wrote:
Hi,
I know that many people do not like the way udev development went recently, therefore we have a systemd-free fork of udev. At the moment most of the changes (except for systemd integration) have been ported, and no stability issues are noticed. I hope it will be useful for those who don't like systemd being pushed in their systems.
Quoted 8 lines
P.S. This maillist seems to be dead for quite a while and most of the discussion moved to systemd-related lists, so I hope nobody would mind if we use it as our primary maillist.--
To unsubscribe from this list: send the line "unsubscribe linux-hotplug" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
BTW, I totally forgot to include a link to the repo. Here it is:
https://bitbucket.org/braindamaged/udev
Quoted 4 lines · On Wed, Sep 12, 2012 at 05:49:52PM +0000, [email protected] wrote:
Hi,
I know that many people do not like the way udev development went recently,
What specifically are you objecting to?
Quoted 1 line
therefore we have a systemd-free fork of udev.
udev works without systemd running on the system just fine for me, doesn't it for you?
Quoted 8 lines
At the moment most of the changes (except for systemd integration) have been ported, and no stability issues are noticed. I hope it will be useful for those who don't like systemd being pushed in their systems.
P.S. This maillist seems to be dead for quite a while and most of the discussion moved to systemd-related lists, so I hope nobody would mind if we use it as our primary maillist.
Who is "we"?
And note, if you are forking udev (which is fine by me, no objections), be careful with the release numbers and naming, to keep others from getting confused.
thanks,
greg k-h
Quoted 26 lines · Bruce Dubbs
Quoted 13 lines · [email protected] wrote:
Quoted 12 lines · On 17:49 Wed 12 Sep , [email protected] wrote:
Hi,
I know that many people do not like the way udev development went recently, therefore we have a systemd-free fork of udev. At the moment most of the changes (except for systemd integration) have been ported, and no stability issues are noticed. I hope it will be useful for those who don't like systemd being pushed in their systems.
P.S. This maillist seems to be dead for quite a while and most of the discussion moved to systemd-related lists, so I hope nobody would mind if we use it as our primary maillist.Quoted 7 lines
Quoted 4 lines
--
To unsubscribe from this list: send the line "unsubscribe linux-hotplug" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.htmlBTW, I totally forgot to include a link to the repo. Here it is:
https://bitbucket.org/braindamaged/udevA fork is way too much effort. The only thing you need is a different
Makefile. For a package that is Linux only, using autotools seems to be a tremendous amount of overkill.
Quoted 6 lines
http://www.linuxfromscratch.org/lfs/view/development/chapter06/udev.html
http://anduin.linuxfromscratch.org/sources/other/udev-lfs-189.tar.bz2Works for us.
-- Bruce
This is not entirely true. A different makefile might be a temporary solution, but udev is already dependent on libsystemd, and I'm afraid this process may go much much further.
Quoted 30 lines · On Wed, Sep 12, 2012 at 06:09:09PM +0000, [email protected] wrote:
Quoted 29 lines · Bruce Dubbs
Quoted 13 lines · [email protected] wrote:
Quoted 12 lines · On 17:49 Wed 12 Sep , [email protected] wrote:
Hi,
I know that many people do not like the way udev development went recently, therefore we have a systemd-free fork of udev. At the moment most of the changes (except for systemd integration) have been ported, and no stability issues are noticed. I hope it will be useful for those who don't like systemd being pushed in their systems.
P.S. This maillist seems to be dead for quite a while and most of the discussion moved to systemd-related lists, so I hope nobody would mind if we use it as our primary maillist.Quoted 7 lines
Quoted 4 lines
--
To unsubscribe from this list: send the line "unsubscribe linux-hotplug" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.htmlBTW, I totally forgot to include a link to the repo. Here it is:
https://bitbucket.org/braindamaged/udevA fork is way too much effort. The only thing you need is a different
Makefile. For a package that is Linux only, using autotools seems to be a tremendous amount of overkill.http://www.linuxfromscratch.org/lfs/view/development/chapter06/udev.html
http://anduin.linuxfromscratch.org/sources/other/udev-lfs-189.tar.bz2
Quoted 7 lines
Quoted 3 lines
Works for us.
-- Bruce
This is not entirely true. A different makefile might be a temporary solution, but udev is already dependent on libsystemd, and I'm afraid this process may go much much further.
I don't understand, what specifically is your objection with something that might happen in the future? And, since it hasn't happened, how can you object to it?
confused,
greg k-h
Quoted 7 lines · Greg KH
Quoted 4 lines · On Wed, Sep 12, 2012 at 05:49:52PM +0000, [email protected] wrote:
Hi,
I know that many people do not like the way udev development went recently,
What specifically are you objecting to?
Extra dependencies and L. Poettering's position that running udev without systemd is a dead end and will not be supported in the future (this seem to be already discussed here before).
Quoted 4 lines
Quoted 1 line
therefore we have a systemd-free fork of udev.
udev works without systemd running on the system just fine for me, doesn't it for you?
Call me paranoid, but I'm afraid that won't last long.
Quoted 9 lines
Quoted 8 lines
At the moment most of the changes (except for systemd integration) have been ported, and no stability issues are noticed. I hope it will be useful for those who don't like systemd being pushed in their systems.
P.S. This maillist seems to be dead for quite a while and most of the discussion moved to systemd-related lists, so I hope nobody would mind if we use it as our primary maillist.
Quoted 5 lines
Who is "we"?
And note, if you are forking udev (which is fine by me, no objections), be careful with the release numbers and naming, to keep others from getting confused.
Me, a bunch of other guys commiting code, and possible users.
Thanks for the advice.
Quoted 4 lines · On Wed, Sep 12, 2012 at 06:15:43PM +0000, [email protected] wrote:
Quoted 3 lines · Greg KH
Quoted 2 lines · On Wed, Sep 12, 2012 at 05:49:52PM +0000, [email protected] wrote:
Hi,
Quoted 8 lines
Quoted 4 lines
Quoted 2 lines
I know that many people do not like the way udev development went recently,
What specifically are you objecting to?
Extra dependencies and L. Poettering's position that running udev without systemd is a dead end and will not be supported in the future (this seem to be already discussed here before).
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
Personally, I think running a Linux system without systemd is a deadend, but hey, what do I know about these things? :)
Quoted 6 lines
Quoted 4 lines
Quoted 1 line
therefore we have a systemd-free fork of udev.
udev works without systemd running on the system just fine for me, doesn't it for you?
Call me paranoid, but I'm afraid that won't last long.
Fear as a development motivator? That's odd, really?
greg k-h
Quoted 20 lines · Greg KH
Quoted 4 lines · On Wed, Sep 12, 2012 at 06:15:43PM +0000, [email protected] wrote:
Quoted 3 lines · Greg KH
Quoted 2 lines · On Wed, Sep 12, 2012 at 05:49:52PM +0000, [email protected] wrote:
Hi,
Quoted 8 lines
Quoted 4 lines
Quoted 2 lines
I know that many people do not like the way udev development went recently,
What specifically are you objecting to?
Extra dependencies and L. Poettering's position that running udev without systemd is a dead end and will not be supported in the future (this seem to be already discussed here before).
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
Personally, I think running a Linux system without systemd is a deadend, but hey, what do I know about these things? :)
Quoted 8 lines
Quoted 6 lines
Quoted 4 lines
Quoted 1 line
therefore we have a systemd-free fork of udev.
udev works without systemd running on the system just fine for me, doesn't it for you?
Call me paranoid, but I'm afraid that won't last long.
Fear as a development motivator? That's odd, really?
brain damaged?
Lucas De Marchi
Quoted 2 lines · Greg KH
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
The configure scripts require packages that are not in LFS. We do not want to add them just to satisfy a systemd build that we don't want. However creating a custom Makefile to build udev from the current systemd sources was not particularly hard.
Quoted 2 lines
Personally, I think running a Linux system without systemd is a deadend, but hey, what do I know about these things? :)
I understand that big distros only want to support one methodology, but in my opinion systemd is a solution only needed by a very small percentage of users. It is quite opaque for new users trying to understand the boot process. We don't use an initrd for the same reason.
The nice thing about Linux is that one size does not have to fit all.
-- Bruce
linuxfromscratch.org - Your distro, your rules.
Quoted 6 lines · Bruce Dubbs
Quoted 2 lines · Greg KH
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
The configure scripts require packages that are not in LFS.
Like what? Can't you add them?
Quoted 3 lines
We do not want to add them just to satisfy a systemd build that we don't want. However creating a custom
Makefileto build udev from the current systemd sources was not particularly hard.
So you don't offer systemd to any LFS user either?
Quoted 4 lines
Quoted 2 lines
Personally, I think running a Linux system without systemd is a deadend, but hey, what do I know about these things? :)
I understand that big distros only want to support one methodology,
No, that's not why they are switching to systemd.
Quoted 2 lines
but in my opinion systemd is a solution only needed by a very small percentage of users.
I don't think you really understand what systemd offers. I don't know anyone who has used it that has wanted to switch back. Also, it solves numerous problems that people have been having for years. And further, it's becoming a requirement for large industry groups that use Linux, because they too want what it offers.
That's not to say you don't want it, that's fine, I understand, but to deride it by saying only a small number of users want it is disingenuous.
Quoted 2 lines
It is quite opaque for new users trying to understand the boot process.
The 100+ man pages are not descriptive enough? :)
Quoted 1 line
We don't use an initrd for the same reason.
That's fine, but then how do you support a separate /usr partition? And handle kernels built for a wide range of systems?
Quoted 1 line
The nice thing about Linux is that one size does not have to fit all.
Sure, and that's fine. But I think you are shortchanging your users here. Again, just my opinion. Best of luck,
greg k-h
Quoted 1 line · Greg KH
That's fine, but then how do you support a separate /usr partition?
With fstab. Believe me, it still works unless you move everything to /usr. But thats like blowing off your foot and then saying that everyone needs a kludge from now on.
Quoted 9 lines · Greg KH
Quoted 6 lines · Bruce Dubbs
Quoted 2 lines · Greg KH
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
The configure scripts require packages that are not in LFS.
Like what? Can't you add them?
intltool, glib, gperf, gobject-introspection.
intl needs XML::Parser. glib needs libffi and python and can use pcre, attr, d-bus, gamin, and gtk-doc. gobject-introspection also needs glib and can use cairo and gtk-doc. cairo needs libpng, glib, and pixman and can use fontconfig, gtk+, xorg libraries (and on and on).
They are all in BLFS but they are not needed for all users. For instance, if you want to build a system and only want to add a web server, they are not needed.
Quoted 5 lines
Quoted 3 lines
We do not want to add them just to satisfy a systemd build that we don't want. However creating a custom
Makefileto build udev from the current systemd sources was not particularly hard.So you don't offer systemd to any LFS user either?
We could add it to BLFS, but I think it would require us to redo all our boot scripts. Users do have the systemd sources and are free to use it if they desire.
Quoted 11 lines
Quoted 4 lines
Quoted 2 lines
Personally, I think running a Linux system without systemd is a deadend, but hey, what do I know about these things? :)
I understand that big distros only want to support one methodology,
No, that's not why they are switching to systemd.
Quoted 2 lines
but in my opinion systemd is a solution only needed by a very small percentage of users.
I don't think you really understand what systemd offers.
Perhaps not, but we also have not had any requests for systemd. I've been programming since 1965 and using Unix like systems since about 1988 and have not run into the problems that systemd solves. We all have different perspectives and I'm sure you have many instances where it is a good solution.
Quoted 9 lines
I don't know
anyone who has used it that has wanted to switch back. Also, it solves numerous problems that people have been having for years. And further, it's becoming a requirement for large industry groups that use Linux, because they too want what it offers.That's not to say you don't want it, that's fine, I understand, but to deride it by saying only a small number of users want it is disingenuous.
I didn't say that only a small number of users want it. The vast majority of users don't know or care. One major reason users want to build from source is because they think the major distros are bloated.
The major reason for us to even publish LFS/BLFS is to help users understand how things fit together.
Quoted 4 lines
Quoted 2 lines
It is quite opaque for new users trying to understand the boot process.
The 100+ man pages are not descriptive enough? :)
The fact that there are 100+ pages needed is the point.
Quoted 4 lines
Quoted 1 line
We don't use an initrd for the same reason.
That's fine, but then how do you support a separate /usr partition? And handle kernels built for a wide range of systems?
We don't support every combination directly. If /usr is on /dev/sda7, there is no problem. Just put it in fstab. We don't support encrypted partitions or other less common setups. In our next version, we may merge /bin, /lib, and /sbin into /usr though.
Quoted 4 lines
Quoted 1 line
The nice thing about Linux is that one size does not have to fit all.
Sure, and that's fine. But I think you are shortchanging your users here. Again, just my opinion.
Our users are free to do what they think is right. We even try to help when users do things outside of LFS/BLFS. There is no shortchanging.
-- Bruce
Quoted 9 lines · Bruce Dubbs
Quoted 8 lines · Greg KH
Quoted 6 lines · Bruce Dubbs
Quoted 2 lines · Greg KH
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
The configure scripts require packages that are not in LFS.
Quoted 8 lines
Quoted 1 line
Like what? Can't you add them?
intltool, glib, gperf, gobject-introspection.
intl needs
XML::Parser. glib needs libffi and python and can use pcre, attr, d-bus, gamin, and gtk-doc. gobject-introspection also needs glib and can use cairo and gtk-doc. cairo needs libpng, glib, and pixman and can use fontconfig, gtk+, xorg libraries (and on and on).
Pkg X "can use" pkg Y (where Y is something that one might or might not want to install) is not an argument against requiring pkg X.
I'm one who thinks (on the basis of experience with home-rolled systems), that systemd really is a smarter, faster, more comprehensible, and more user-manageable way to get a Linux system up and running than sysvinit plus a big mess of shell scripts.
However, I take your point about some of the systemd dependencies, direct and indirect (even though systemd's configure script has a fair number of useful --disable-whatever options).
Why intltool, for instance? Systemd has a --disable-nls option in its configure script. But this is in fact just automake fraud; there's really no way to disable nls (and everything it brings in, including intltool), so far as I can tell.
--
Allin Cottrell
Department of Economics
Wake Forest University
Quoted 15 lines · Allin Cottrell
Quoted 13 lines · Bruce Dubbs
Quoted 9 lines · Greg KH
Quoted 6 lines · Bruce Dubbs
Quoted 2 lines · Greg KH
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
The configure scripts require packages that are not in LFS.
Like what? Can't you add them?
intltool, glib, gperf, gobject-introspection.
Quoted 7 lines
Quoted 4 lines
intl needs
XML::Parser. glib needs libffi and python and can use pcre, attr, d-bus, gamin, and gtk-doc. gobject-introspection also needs glib and can use cairo and gtk-doc. cairo needs libpng, glib, and pixman and can use fontconfig, gtk+, xorg libraries (and on and on).Pkg X "can use" pkg Y (where Y is something that one might or might not want to install) is not an argument against requiring pkg X.
It is for LFS. Every user builds every package from source. That's the purpose of LFS.
Quoted 4 lines
I'm one who thinks (on the basis of experience with home-rolled systems), that systemd really is a smarter, faster, more comprehensible, and more user-manageable way to get a Linux system up and running than sysvinit plus a big mess of shell scripts.
After dealing with LFS users for 10 years, my experience is different. If we were building a binary distro to distribute to users, I might agree with you, but we try to make things easy to understand. The base LFS system has about 2000 lines of shell scripts. Compare to about 150K of C code in systemd. If a script has a problem, there are typically about 5 lines in a start or stop. Plowing through all the C code is a lot more difficult.
Quoted 3 lines
However, I take your point about some of the systemd dependencies, direct and indirect (even though systemd's configure script has a fair number of useful --disable-whatever options).
They have rejected patches that fix the problem.
Quoted 4 lines
Why intltool, for instance? Systemd has a --disable-nls option in its configure script. But this is in fact just automake fraud; there's really no way to disable nls (and everything it brings in, including intltool), so far as I can tell.
That's why we have a hand crafted Makefile. I don't understand why autotools are needed for a package that only has one target architecture.
Our Makefile (udev, gudev, keymap, and gir) is only 674 lines. Systemd's configure.ac is 812 lines and is a lot more difficult to understand if you are not an autotools wizard.
I'll also note that a complete LFS build and install for (base) udev takes about 10 seconds. Boot time for a typical base LFS system is 8 seconds.
-- Bruce
Quoted 21 lines · Bruce Dubbs
Quoted 20 lines · Allin Cottrell
Quoted 17 lines · Bruce Dubbs
Quoted 9 lines · Greg KH
Quoted 6 lines · Bruce Dubbs
Quoted 2 lines · Greg KH
What dependencies? Run time? Build time? And why are dependencies bad? Do you have no ram in your system for them?
The configure scripts require packages that are not in LFS.
Like what? Can't you add them?
intltool, glib, gperf, gobject-introspection.
intl needs
XML::Parser. glib needs libffi and python and can use pcre, attr, d-bus, gamin, and gtk-doc. gobject-introspection also needs glib and can use cairo and gtk-doc. cairo needs libpng, glib, and pixman and can use fontconfig, gtk+, xorg libraries (and on and on).
Quoted 5 lines
Quoted 2 lines
Pkg X "can use" pkg Y (where Y is something that one might or might not want to install) is not an argument against requiring pkg X.
It is for LFS. Every user builds every package from source. That's the purpose of LFS.
I don't see it. "Can use" Y means it doesn't have to use Y: if you're building X from scratch and don't have a (pressing) use for whatever pkg Y provides, then say --disable-Y when configuring X (or maybe that's not even needed, if Y is auto-detected). Not a problem.
Quoted 11 lines
Quoted 4 lines
I'm one who thinks (on the basis of experience with home-rolled systems), that systemd really is a smarter, faster, more comprehensible, and more user-manageable way to get a Linux system up and running than sysvinit plus a big mess of shell scripts.
After dealing with LFS users for 10 years, my experience is different. If we were building a binary distro to distribute to users, I might agree with you, but we try to make things easy to understand. The base LFS system has about 2000 lines of shell scripts. Compare to about 150K of C code in systemd. If a script has a problem, there are typically about 5 lines in a start or stop. Plowing through all the C code is a lot more difficult.
OK, opinions may differ on this. But I'm not talking about making a distro either, just running a DIY Linux system (not strictly LFS, but making grateful use of LFS from time to time).
I've found that the C code of systemd "just works"; the only thing I have to worry about is the *.service files, which are easier to manage than shell scripts plus the menagerie of shell functions they call. I'm no more required to concern myself with systemd's C code that I was with sysvinit's C code.
Quoted 5 lines
Quoted 3 lines
However, I take your point about some of the systemd dependencies, direct and indirect (even though systemd's configure script has a fair number of useful --disable-whatever options).
They have rejected patches that fix the problem.
That's relevant. Any specifics? Did you have a patch to really disable nls?
Quoted 8 lines
Quoted 4 lines
Why intltool, for instance? Systemd has a --disable-nls option in its configure script. But this is in fact just automake fraud; there's really no way to disable nls (and everything it brings in, including intltool), so far as I can tell.
That's why we have a hand crafted
Makefile. I don't understand why autotools are needed for a package that only has one target architecture.
For my part I like autoconf but consider automake the spawn of Satan, so I'm part-way with you on that. Makefiles that can be read by human beings -- and don't contain 95% irrelevant repetitive numbskull boilerplate -- are certainly my preference.
--
Allin Cottrell
Department of Economics
Wake Forest University
Quoted 10 lines · Allin Cottrell
Quoted 5 lines
Quoted 2 lines
Pkg X "can use" pkg Y (where Y is something that one might or might not want to install) is not an argument against requiring pkg X.
It is for LFS. Every user builds every package from source. That's the purpose of LFS.
I don't see it. "Can use" Y means it doesn't have to use Y: if you're building X from scratch and don't have a (pressing) use for whatever pkg Y provides, then say --disable-Y when configuring X (or maybe that's not even needed, if Y is auto-detected). Not a problem.
The dependency links are a problem. Theoretically, you can build X without Y and then build Y and go back and rebuild X. One or two times is not a big deal, but when the chain gets long, most users don't want to repeat the process. As an extreme example, libreoffice takes almost 10 hours to build on some systems. If you leave out something, you don't want to do that again.
Quoted 17 lines
Quoted 5 lines
Quoted 4 lines
I'm one who thinks (on the basis of experience with home-rolled systems), that systemd really is a smarter, faster, more comprehensible, and more user-manageable way to get a Linux system up and running than sysvinit plus a big mess of shell scripts.
Quoted 7 lines
After dealing with LFS users for 10 years, my experience is different. If we were building a binary distro to distribute to users, I might agree with you, but we try to make things easy to understand. The base LFS system has about 2000 lines of shell scripts. Compare to about 150K of C code in systemd. If a script has a problem, there are typically about 5 lines in a start or stop. Plowing through all the C code is a lot more difficult.
OK, opinions may differ on this. But I'm not talking about making a distro either, just running a DIY Linux system (not strictly LFS, but making grateful use of LFS from time to time).
Quoted 5 lines
I've found that the C code of systemd "just works"; the only thing I have to worry about is the *.service files, which are easier to manage than shell scripts plus the menagerie of shell functions they call. I'm no more required to concern myself with systemd's C code that I was with sysvinit's C code.
I can see your point, but sysv's init is only one small program. The functions are all in one 800 line file of scripts (with substantial comments).
Everything gets installed with one 'make install', but you can look at any script with any text tool to see exactly what is happening. It's really quite transparent.
Quoted 7 lines
Quoted 5 lines
Quoted 3 lines
However, I take your point about some of the systemd dependencies, direct and indirect (even though systemd's configure script has a fair number of useful --disable-whatever options).
They have rejected patches that fix the problem.
That's relevant. Any specifics? Did you have a patch to really disable nls?
http://lists.freedesktop.org/archives/systemd-devel/2012-June/005407.html
-- Bruce
Quoted 7 lines · On 9/12/2012 10:49 AM, [email protected] wrote:
Hi,
I know that many people do not like the way udev development went recently, therefore we have a systemd-free fork of udev. At the moment most of the changes (except for systemd integration) have been ported, and no stability issues are noticed. I hope it will be useful for those who don't like systemd being pushed in their systems.
First, I believe that the overhead of maintaining a systemd+udevd configuration that allows disabling of all systemd dependencies is relatively trivial. Therefore, not doing so (especially when others are submitting patches to help it along) is the height of seppuku arrogance/laziness on the part of the systemd+udevd development team.
Having said this, as the maintainer of MiniMyth, I have found the disjoint nature of sysvinit and udev to be cumbersome in a world of hotplug hardware that requires starting/stopping userspace daemons or changing userspace configuration. I (and I imagine others) have found hotplug requires combining of init that starts/stops services and udev that can trigger what services need to be started and stopped. In order to solve this problem for MiniMyth some years ago, I wrote my own custom sysvinit and udev scripts that make it work for MiniMyth. However, this has been a kludge. If the tighter integration of init and hotplug makes this more simple over time (and this is a big if and often change makes things more complicated), then I believe there value.
However, for systems that do not need hotplug (of which there are many), requiring the overhead of systemd at either build or run time is a bad idea.
This is just my opinion based on my experience. Take it or leave it.