| View previous topic :: View next topic |
| Author |
Message |
tld
Veteran
Joined: 09 Dec 2003
Posts: 1861
|
 Posted: Mon May 18, 2015 12:40 pm Post subject: |
 |
|
| krinn wrote: | Now if systemd depend on that patch (so systemd need a kernel with kdbus patch), that would force anyone to use a kernel with the patch (so a non vanilla kernel anymore), and write in stone redhat has fork the kernel.
And how distro will look at this: getting deeper redhat dependent or finally back to sanity and dropping systemd and all the shit. |
Exactly. Those distros let themselves get strong armed into using systemd by malicious dependencies by Gnome etc. They should have realized it wouldn't end there. It'll get very interesting if using systemd starts forcing them to use a kernel full of ugly hacks. |
|
| Back to top |
|
 |
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6370
Location: Dallas area
|
| Posted: Mon May 18, 2015 1:18 pm Post subject: |
|
|
| tld wrote: | | It'll get very interesting if using systemd starts forcing them to use a kernel full of ugly hacks. |
Which is part of the reason that RH is lobbying so hard for kdbus to get into the kernel.
They know that it's an uphill publicity battle to tell other distros not only do you need systemd but our patch set too.
Effectively killing everyone as a separate distro off, but with kdbus in the kernel,
they can easily convince themselves that RH still has their interests first and foremost.
_________________
UM780 xtx, 6.14 zen kernel, gcc 15, openrc, wayland
Got to love snowflakes, how does the world survive without them. |
|
| Back to top |
|
|
mrbassie
l33t
Joined: 31 May 2013
Posts: 855
Location: Go past the sign for cope, right at the sign for seethe. If you see the target you've missed it.
|
| Posted: Mon May 18, 2015 5:44 pm Post subject: |
|
|
| But, but, but we need to end the duplication of effort and have standardisation otherwise..umm..something bad! Oh yeah and we have to kill Microsoft because...something! We can't use desktops without systemd...oh wait! |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6090
Location: Removed by Neddy
|
| Posted: Thu Jun 04, 2015 6:28 am Post subject: |
|
|
Round ? I have lost count
http://lkml.iu.edu/hypermail/linux/kernel/1506.0/02614.html
_________________
#define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6370
Location: Dallas area
|
| Posted: Thu Jun 04, 2015 9:47 am Post subject: |
|
|
They keep on ignoring the fact that the kernel devs have told them (albeit not so plainly as I state it here)
that kdbus won't be allowed in the kernel the way it is written right now.
And that involves a lot more than just shuffling metadata.
This is what I see happening, the (k)dbus cabal will push, yet again, for kdbus to be included.
The kernel devs will say the same thing they've said for all the candidates so far "nack on it the way it is written"
Lather...rinse...repeat. 
_________________
UM780 xtx, 6.14 zen kernel, gcc 15, openrc, wayland
Got to love snowflakes, how does the world survive without them. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Thu Jun 04, 2015 3:18 pm Post subject: |
|
|
| Anon-E-moose wrote: | They keep on ignoring the fact that the kernel devs have told them (albeit not so plainly as I state it here)
that kdbus won't be allowed in the kernel the way it is written right now.
And that involves a lot more than just shuffling metadata.
This is what I see happening, the (k)dbus cabal will push, yet again, for kdbus to be included.
The kernel devs will say the same thing they've said for all the candidates so far "nack on it the way it is written"
Lather...rinse...repeat. |
It's odd, reading GregK-H's posts, it's as if those NAKs never happened.
I'm sure some social-anthropology student somewhere can describe it much better than any of us, but it strikes me that the use of language is quite interesting. By breezily carrying on, as if the thing hadn't been trashed into the ground, the impression is given that everything's hunky-dory in lala-land: "y'know, just those crazy kernel-devs carrying on as usual, heh what can you do.. roll up, roll up, business as usual." ("one born every second.")
From the pov of apparatchik-watching, it strikes me that anyone other than that shill would never even get away with it.
As such it represents a classic case of how you can coopt a "technical" community via social methods; geeks pretending that politics doesn't matter, are just easy victims -- precisely because they ignore the machinations, while being subject to the same old cognitive dissonance (a term originally from the propaganda^W advertising industry, as a way of describing what they do) as every other human being.
From a technical pov, "sd-bus coupled with GCC plugins to construct kdbus sandbox profiles during compilation thanks to sd-bus semantics" sounds like utter shite to me, especially the "thanks" part.
I dread to think what bs they'll inflict on gcc, which will then become a "required part of Lennux" so if you're unfortunate enough to be on a bindist tied into systemdbust, you'll be a botnet-vector waiting to be pwnt.
But still, the NSA can push patches for their malware to your machine, as can just about anyone else, so that's all good then. /s
Most of all it sounds like gibber: | Code: | "Maybe I'm an idiot, but I have no idea what anyone is talking about.
What is it? It's complete gibberish. It's insane.
When is this idiocy going to stop?" |
In any event, I fully expect the kdbust-crowd to keep pushing, and keep on acting as if the blatant problems can all be fixed in the wash.
Later it'll just be the usual "all complaints were responded to" becoming "all complaints were dealt with" even though everyone involved in the discussion saw the responses given as pitifully inadequate.
As Charlie Brooker puts it, "Welcome to the World of Bullshit." |
|
| Back to top |
|
|
sitquietly
Apprentice
Joined: 23 Oct 2010
Posts: 153
Location: On the Wolf River, Tennessee
|
| Posted: Thu Jun 04, 2015 9:11 pm Post subject: |
|
|
| steveL wrote: | .....
I'm sure some social-anthropology student somewhere can describe it much better than any of us, but it strikes me that the use of language is quite interesting. By breezily carrying on, as if the thing hadn't been trashed into the ground.....it represents a classic case of how you can coopt a "technical" community via social methods; geeks pretending that politics doesn't matter, are just easy victims..... |
You've helped me understand the situation in this thread. Thanks. it has caused me to take positive steps to reduce the future impact of the re-basing of *nix by the corporate and government boys. I've been able to run without avahi/consolekit/policykit/pulseaudio/systemd for a long while but getting rid of any dependence of my daily workflow on dbus has been more difficult. Even Gentoo ports pull in dbus for lots of packages where it is not actually needed for the function of the software. The prime example of this problem is gtk3, which I would like to have available as a gui toolkit but must avoid completely because it absolutely must have dbus in order to draw a user interface (??).
What I've found so far is that a dbus-free desktop is more responsive and presents fewer interruptions (less noise from popups telling me things I already know like "you just inserted a cdrom"). There have been few real losses. qpdfview is now a good substitute for okular for example. I had to learn to use handbrake's command line instead of the (gtk3-based) gui. I use a "simple" fluxbox/rox-filer desktop. It has the feel and workflow of a classic "direct manipulation", "drag and drop" interface.
I have actually found it more difficult to eliminate dbus from my FreeBSD system! Many of the FreeBSD ports do not provide "knobs" for disabling dbus, where the Gentoo package does provide for USE="-dbus". If dbus goes the way I think it will then FreeBSD desktop users will eventually have more trouble coping than Gentoo users. I'm planning to "git branch pure" in my FreeBSD ports clone and start modifying some of the ports to eliminate "linuxisms," such as dbus, where possible.
I was wondering if anyone has gone through gtk3 or kdelibs to patch out dbus? That would be helpful to my goal of avoiding messy tentacles. |
|
| Back to top |
|
|
depontius
Advocate
Joined: 05 May 2004
Posts: 3530
|
| Posted: Fri Jun 05, 2015 1:03 am Post subject: |
|
|
| steveL wrote: | | It's odd, reading GregK-H's posts, it's as if those NAKs never happened. |
I followed your link, looked elsewhere on the thread, and found this:
| Quote: | I also have a question is someone of you guys planning to submit a talk
to future Linux events about kdbus and sdbus ? I have a presentation not
finished about kdbus, sd-bus coupled with GCC plugins to construct kdbus
sandbox profiles during compilation thanks to sd-bus semantics, but
never had the time to finish it  |
What the heck is "sd-bus"? It looks obviously related to dbus in some way, but they seem to be shoving it sideways into gcc?!?
Google tells me that "sdbus" is something to do with Windows that causes a lot of BSODs, and "sd-bus" has people riding it.
_________________
.sigs waste space and bandwidth |
|
| Back to top |
|
|
augustin
Guru
Joined: 23 Feb 2015
Posts: 318
|
| Posted: Fri Jun 05, 2015 7:27 am Post subject: |
|
|
| sitquietly wrote: |
There have been few real losses. qpdfview is now a good substitute for okular for example. |
Sorry, what?
Okular depends on systemd? |
|
| Back to top |
|
|
arnvidr
l33t
Joined: 19 Aug 2004
Posts: 629
Location: Oslo, Norway
|
| Posted: Fri Jun 05, 2015 7:45 am Post subject: |
|
|
| augustin wrote: | | sitquietly wrote: |
There have been few real losses. qpdfview is now a good substitute for okular for example. |
Sorry, what?
Okular depends on systemd? |
From what I gathered, he removed gtk3, since it depends on dbus.
_________________
|
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6090
Location: Removed by Neddy
|
| Posted: Fri Jun 05, 2015 8:23 am Post subject: |
|
|
| depontius wrote: | | steveL wrote: | | It's odd, reading GregK-H's posts, it's as if those NAKs never happened. |
I followed your link, looked elsewhere on the thread, and found this:
| Quote: | I also have a question is someone of you guys planning to submit a talk
to future Linux events about kdbus and sdbus ? I have a presentation not
finished about kdbus, sd-bus coupled with GCC plugins to construct kdbus
sandbox profiles during compilation thanks to sd-bus semantics, but
never had the time to finish it |
What the heck is "sd-bus"? It looks obviously related to dbus in some way, but they seem to be shoving it sideways into gcc?!?
Google tells me that "sdbus" is something to do with Windows that causes a lot of BSODs, and "sd-bus" has people riding it. |
sd-bus is the backend hooks associated with sd-devices... systemd replacement for udev . Yes thats right kids udev is going away 
_________________
#define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
| Back to top |
|
|
augustin
Guru
Joined: 23 Feb 2015
Posts: 318
|
| Posted: Fri Jun 05, 2015 8:35 am Post subject: |
|
|
| arnvidr wrote: | | augustin wrote: | | sitquietly wrote: |
There have been few real losses. qpdfview is now a good substitute for okular for example. |
Sorry, what?
Okular depends on systemd? |
From what I gathered, he removed gtk3, since it depends on dbus. |
Thanks, but I must be getting something wrong.
Okular is a KDE application, no?
gtk is a GNOME library, right?
Doesn't okular use Qt / the KDE libs?
Also, what I still don't get is the technical relationship between dbus and systemd, beside the fact that they are being developed by the same people (?).
http://en.wikipedia.org/wiki/D-Bus |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6370
Location: Dallas area
|
| Posted: Fri Jun 05, 2015 9:38 am Post subject: |
|
|
| Naib wrote: | | Yes thats right kids udev is going away |
As predicted by many of us, well, except for the gentoo dev who kept on trying to lie to us about udev.
_________________
UM780 xtx, 6.14 zen kernel, gcc 15, openrc, wayland
Got to love snowflakes, how does the world survive without them. |
|
| Back to top |
|
|
bstaletic
Guru
Joined: 05 Apr 2014
Posts: 517
|
| Posted: Fri Jun 05, 2015 11:53 am Post subject: |
|
|
| So what happens when there's no more udev, only sdbus? Do we switch to devuan's vdev? Are there any viable alternatives to {e,}udev? |
|
| Back to top |
|
|
sitquietly
Apprentice
Joined: 23 Oct 2010
Posts: 153
Location: On the Wolf River, Tennessee
|
| Posted: Fri Jun 05, 2015 1:52 pm Post subject: |
|
|
| augustin wrote: | | arnvidr wrote: | | augustin wrote: | | sitquietly wrote: |
There have been few real losses. qpdfview is now a good substitute for okular for example. |
Sorry, what?
Okular depends on systemd? |
From what I gathered, he removed gtk3, since it depends on dbus. |
Thanks, but I must be getting something wrong.
Okular is a KDE application, no?
gtk is a GNOME library, right?
Doesn't okular use Qt / the KDE libs?
Also, what I still don't get is the technical relationship between dbus and systemd, beside the fact that they are being developed by the same people (?).
http://en.wikipedia.org/wiki/D-Bus |
I'm left utterly flabbergasted as to what any of this palaver about systemd has to do with the topic of this thread, which is "kdbus in the kernel".
You all were discussing the agenda to move dbus from user space into the kernel, "weaponized dbus". Is anyone here responding in a defensive way by preparing to remove dbus from your system. To eliminate dbus from your system requires eliminating all use of gtk3 AND kdelibs, amongst others. Okular of course is a kdelibs client. |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6370
Location: Dallas area
|
| Posted: Fri Jun 05, 2015 6:29 pm Post subject: |
|
|
| sitquietly wrote: | | I'm left utterly flabbergasted as to what any of this palaver about systemd has to do with the topic of this thread, which is "kdbus in the kernel". |
Possibly because the sysd people are some of the ones pushing for kdbus to be "included" in the kernel
and they've already said that once kdbus gets into the kernel, they will start using it.
| Quote: | | You all were discussing the agenda to move dbus from user space into the kernel, "weaponized dbus". Is anyone here responding in a defensive way by preparing to remove dbus from your system. To eliminate dbus from your system requires eliminating all use of gtk3 AND kdelibs, amongst others. Okular of course is a kdelibs client. |
I don't run dbus on my system and haven't for the last year or so.
That gtk3 requires dbus is an absurd dependency.
Note: I've blocked gtk3 and don't run kdelibs on my system.
I use openbox with the lxde panel (gtk2) and a few qt apps.
I have had to block some apps since they insist that dbus be available one way or another.
I find the next most suitable app to use in that case.
_________________
UM780 xtx, 6.14 zen kernel, gcc 15, openrc, wayland
Got to love snowflakes, how does the world survive without them. |
|
| Back to top |
|
|
patrix_neo
Guru
Joined: 08 Jan 2004
Posts: 520
Location: The Maldives
|
| Posted: Fri Jun 05, 2015 8:42 pm Post subject: |
|
|
I think I just read:
Resistance is futile, asimliate and propagate!
This is just the beginning of a very long journey between the good and the bad.
It's actually back.
The answer is twofold...first there is the heroes going against the oppression, and finally the truth will hitting everyone. That will happen. Tell me I am wrong...I dare you.
It might take a century in between...finally everyone is up to par. Muy Bien.
A thing like kdubus is just a way to cope with SystemD, if I am right. Which is, as I see it, a variable to destroy freedom as I have felt it to be.
Entanglement. I might be wrong, I might be foolish, but I will stand by it as long it is not.... |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sat Jun 06, 2015 5:32 pm Post subject: |
|
|
| bstaletic wrote: | | So what happens when there's no more udev, only sdbus? Do we switch to devuan's vdev? Are there any viable alternatives to {e,}udev? |
Just carry on using eudev, afaic. We don't need it to do any more than it already does; after all that was the point of udev in the first place, that it be a simple userland multiplexor of rt-netlink info from kernel notification-chains. (so the kernel doesn't have to worry about it, but the admin can, as desired by configuration.)
X is the classic example of a userland multiplexor.
Really, all desktop users really care about in the main, is being able to plug in removable devices, and have them show up.
We got past that point ages ago. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sat Jun 06, 2015 5:38 pm Post subject: |
|
|
| sitquietly wrote: | I've been able to run without avahi/consolekit/policykit/pulseaudio/systemd for a long while but getting rid of any dependence of my daily workflow on dbus has been more difficult. Even Gentoo ports pull in dbus for lots of packages where it is not actually needed for the function of the software. The prime example of this problem is gtk3, which I would like to have available as a gui toolkit but must avoid completely because it absolutely must have dbus in order to draw a user interface (??).
What I've found so far is that a dbus-free desktop is more responsive and presents fewer interruptions (less noise from popups telling me things I already know like "you just inserted a cdrom"). There have been few real losses. qpdfview is now a good substitute for okular for example. I had to learn to use handbrake's command line instead of the (gtk3-based) gui. I use a "simple" fluxbox/rox-filer desktop. It has the feel and workflow of a classic "direct manipulation", "drag and drop" interface. |
Yeah the RISC-OS filer was lovely. I must take a look at rox soon.
| Quote: | | I was wondering if anyone has gone through gtk3 or kdelibs to patch out dbus? That would be helpful to my goal of avoiding messy tentacles. |
Hmm I don't actually have an issue with dbus within the X session, though obviously I'd prefer dcop back again (or a better version, that is competitive with CORBA: either way both were much more performant than dbus.)
I don't see it happening soon, for kdelibs; it'd take a lot of work, and only really be feasible for the release after KF5.
I could be wrong; it might be simple enough to plug in another transport, and provide the qdbus interface.
Though really, one would need to address the GPL-functionality issue wrt licensing, to do it properly. |
|
| Back to top |
|
|
depontius
Advocate
Joined: 05 May 2004
Posts: 3530
|
| Posted: Sat Jun 06, 2015 6:01 pm Post subject: |
|
|
| steveL wrote: | | simple userland multiplexor of rt-netlink info from kernel notification-chains. (so the kernel doesn't have to worry about it, but the admin can, as desired by configuration.) |
Presuming kdbus does at some point make it into the kernel, it's going to be really interesting to see the fur flying when systemd developers try to remove rt-netlink in favor of kdbus.
_________________
.sigs waste space and bandwidth |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sun Jun 07, 2015 2:01 am Post subject: |
|
|
| steveL wrote: | | simple userland multiplexor of rt-netlink info from kernel notification-chains. (so the kernel doesn't have to worry about it, but the admin can, as desired by configuration.) |
| depontius wrote: | | Presuming kdbus does at some point make it into the kernel, it's going to be really interesting to see the fur flying when systemd developers try to remove rt-netlink in favor of kdbus. |
Yeah I saw you commented on that here, but forgot to respond. I don't think it's a good idea at all, since things like dhcpcd were the primary intended consumer of rt-netlink, not udev. (the clue's in the name, ofc.)
Notification-chains just flowed naturally into passing off device information to userland.
Anyhow, another follow-up discussion is here (technical.) |
|
| Back to top |
|
|
tclover
Guru
Joined: 10 Apr 2011
Posts: 516
|
| Posted: Sun Jun 07, 2015 7:38 am Post subject: |
|
|
| steveL wrote: | [...]
I'm sure some social-anthropology student somewhere can describe it much better than any of us, but it strikes me that the use of language is quite interesting. By breezily carrying on, as if the thing hadn't been trashed into the ground, the impression is given that everything's hunky-dory in lala-land: "y'know, just those crazy kernel-devs carrying on as usual, heh what can you do.. roll up, roll up, business as usual." ("one born every second.")
From the pov of apparatchik-watching, it strikes me that anyone other than that shill would never even get away with it.
As such it represents a classic case of how you can coopt a "technical" community via social methods; geeks pretending that politics doesn't matter, are just easy victims -- precisely because they ignore the machinations, while being subject to the same old cognitive dissonance (a term originally from the propaganda^W advertising industry, as a way of describing what they do) as every other human being.
[...] |
Nice post! Just quoting the tasty part of it... What a good summary this "Empire of risng scum" is. Iit couldn't better summarized. Thanks for this, I did not really know (the origin of) apparatchik, though, the word doesn't really matter--I am not one of those who are in love with words/language if you'd like. Only the underlying,--not so underlain anyway like any PR, ads or "empty words, just void" as I like to call them,--the dynamics, the discrete subversion of "the aim" does.
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/ |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6370
Location: Dallas area
|
| Posted: Sun Jun 07, 2015 9:35 am Post subject: |
|
|
| steveL wrote: | | bstaletic wrote: | | So what happens when there's no more udev, only sdbus? Do we switch to devuan's vdev? Are there any viable alternatives to {e,}udev? |
Just carry on using eudev, afaic. We don't need it to do any more than it already does; |
Indeed, for a long time I was using udev 171-r6 before switching to eudev.
I didn't switch because udev wasn't working, I just got tired of having to modify ebuilds
because the gentoo devs didn't (and still don't) have a clue as to how to write an ebuild
using generic dependencies.
_________________
UM780 xtx, 6.14 zen kernel, gcc 15, openrc, wayland
Got to love snowflakes, how does the world survive without them. |
|
| Back to top |
|
|
asturm
Developer
Joined: 05 Apr 2007
Posts: 9420
|
| Posted: Sun Jun 07, 2015 10:11 am Post subject: |
|
|
| Anon-E-moose wrote: | because the gentoo devs didn't (and still don't) have a clue as to how to write an ebuild
using generic dependencies. |
Pardon? |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6370
Location: Dallas area
|
| Posted: Sun Jun 07, 2015 12:45 pm Post subject: |
|
|
Used to a requirement for a dependency (pre the devs from the last few years) was simple, ie "udev"
In their infinite wisdom they've now gone to ">=udev-[whatever version they use]" as a dependency requirement.
Many of the ebuilds that I was changing was simply from things like this.
It's not that it didn't work in the more generic sense, they worked perfectly fine.
Like many of the changes that have been foisted upon the users by gentoo devs
ie libav vs ffmpeg as default for example, they seem to be ill thought out.
_________________
UM780 xtx, 6.14 zen kernel, gcc 15, openrc, wayland
Got to love snowflakes, how does the world survive without them. |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Kernel & Hardware
