| View previous topic :: View next topic |
| Author |
Message |
Naib
Watchman
Joined: 21 May 2004
Posts: 6065
Location: Removed by Neddy
|
|
| Back to top |
|
 |
ct85711
Veteran
Joined: 27 Sep 2005
Posts: 1791
|
 Posted: Fri Oct 30, 2015 3:54 pm Post subject: |
 |
|
Now, hopefully the devs will scrap kdbus as a whole and throw it away....
It's been said, that most of the reason to make kdbus is because of dbus inefficiencies. Maybe they'll figure it out and remake dbus and be done with everything... |
|
| Back to top |
|
|
davidm
Guru
Joined: 26 Apr 2009
Posts: 557
Location: US
|
| Posted: Fri Oct 30, 2015 4:28 pm Post subject: |
|
|
I'm using systemd on Gentoo myself but OUCH. I do think they need to slow down, focus more on quality and be far less pushy. Make people want to switch by killing them with the features and the stability and not by forcing them to do so with underhanded tricks and power plays. |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6148
Location: Dallas area
|
| Posted: Fri Oct 30, 2015 5:03 pm Post subject: |
|
|
From the Naib link:
| Quote: | | The upstream developers asked me to remove the module from Fedora while they rethink some of the approach they are taking with kdbus. |
Maybe the NAKS from the kernel devs, everytime they have asked for it to be included in the kernel, is starting to sink in.
Either that or they have found out that the "vast gains" of using kdbus aren't that vast after all.
_________________
UM780, 6.1 zen kernel, gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6065
Location: Removed by Neddy
|
| Posted: Fri Oct 30, 2015 5:10 pm Post subject: |
|
|
what's just as interesting was Fedora failed a few internal reviews and F23 wasn't going to be released. Today... it now passes and F23 is being released.
related or co-incidental?
_________________
| Quote: | | Removed by Chiitoo |
|
|
| Back to top |
|
|
depontius
Advocate
Joined: 05 May 2004
Posts: 3522
|
| Posted: Fri Oct 30, 2015 7:29 pm Post subject: |
|
|
| Anon-E-moose wrote: | | Either that or they have found out that the "vast gains" of using kdbus aren't that vast after all. |
The gains may be measured in lock-in, not performance. The latter may be minor compared to the former.
The real question here is how fast after kdbus goes into the kernel, presuming it does, will someone (GKH?) call for removal of netlink as a method of reporting udev events? You know, the lock-in of systemd as the one true whatchamacallit. (I use that term because it's fearfully more than an init system.)
_________________
.sigs waste space and bandwidth |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Fri Oct 30, 2015 7:40 pm Post subject: |
|
|
| davidm wrote: | | I'm using systemd on Gentoo myself but OUCH. I do think they need to slow down, focus more on quality and be far less pushy. Make people want to switch by killing them with the features and the stability and not by forcing them to do so with underhanded tricks and power plays. |
davidm ... but those are the "features" ...
"systemd is poorly designed? Its claims, wrt the "brokeness" of the function systemd intended to replace, or its own "features", are spurious? No, it has innovative underhanded tricks and power plays ... haytre".
The problem with systemd is that its now like the boy who cried wolf, no one should give it, or its adherents, any credence.
best ... khay |
|
| Back to top |
|
|
roki942
Apprentice
Joined: 18 Apr 2005
Posts: 285
Location: Seattle
|
| Posted: Fri Oct 30, 2015 7:49 pm Post subject: |
|
|
Can I be the first to say it?
"LP this is your wake up call!"
edit:
Oops, my bad .... that sounds personal, but I never met the guy.
Should have been: "Systemd this is your wake up call!"
Last edited by roki942 on Fri Oct 30, 2015 9:33 pm; edited 1 time in total |
|
| Back to top |
|
|
GFCCAE6xF
Apprentice
Joined: 06 Aug 2012
Posts: 295
|
| Posted: Fri Oct 30, 2015 8:22 pm Post subject: |
|
|
| Naib wrote: | what's just as interesting was Fedora failed a few internal reviews and F23 wasn't going to be released. Today... it now passes and F23 is being released.
related or co-incidental? |
Leave it for the reader to research 
http://pkgs.fedoraproject.org/cgit/kernel.git/ |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Fri Oct 30, 2015 10:17 pm Post subject: |
|
|
| GFCCAE6xF wrote: | | http://pkgs.fedoraproject.org/cgit/kernel.git/ |
Lol: | Code: | | Drop kdbus Lines -52081/+3 |
The whole of the "UNIX Operating System Source Code Level Six", as presented in Lions[1], is 9073 lines.
[1] Lions' Commentary on UNIX 6th Edition with Source Code (John Lions, 1977, published: 1996 Peer-to-Peer.)
Sure, it's a "specially-edited selection" but the point stands. According to ESR in the foreword, it's the "entire source of the UNIX Version 6 kernel." |
|
| Back to top |
|
|
gwr
Apprentice
Joined: 19 Nov 2014
Posts: 194
|
| Posted: Sat Oct 31, 2015 11:08 am Post subject: |
|
|
| steveL wrote: | | GFCCAE6xF wrote: | | http://pkgs.fedoraproject.org/cgit/kernel.git/ |
Lol: | Code: | | Drop kdbus Lines -52081/+3 |
The whole of the "UNIX Operating System Source Code Level Six", as presented in Lions[1], is 9073 lines.
[1] Lions' Commentary on UNIX 6th Edition with Source Code (John Lions, 1977, published: 1996 Peer-to-Peer.)
Sure, it's a "specially-edited selection" but the point stands. According to ESR in the foreword, it's the "entire source of the UNIX Version 6 kernel." |
The vast majority of lines of code appears to be the endlessly verbose XML documention that is needlessly intermixed with the source code. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sat Oct 31, 2015 5:30 pm Post subject: |
|
|
| gwr wrote: | | The vast majority of lines of code appears to be the endlessly verbose XML documentation that is needlessly intermixed with the source code. |
*shudder* ;) |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54578
Location: 56N 3W
|
| Posted: Sat Oct 31, 2015 5:33 pm Post subject: |
|
|
There has to be a halloween reference to kdbus there somewhere but I can't spot it.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6065
Location: Removed by Neddy
|
| Posted: Sat Oct 31, 2015 5:48 pm Post subject: |
|
|
| NeddySeagoon wrote: | | There has to be a halloween reference to kdbus there somewhere but I can't spot it. |
the dead will rise again?
_________________
| Quote: | | Removed by Chiitoo |
|
|
| Back to top |
|
|
depontius
Advocate
Joined: 05 May 2004
Posts: 3522
|
| Posted: Sat Oct 31, 2015 6:07 pm Post subject: |
|
|
| Naib wrote: | | NeddySeagoon wrote: | | There has to be a halloween reference to kdbus there somewhere but I can't spot it. |
the dead will rise again? |
Oh, come on! It's obvious... "Trick or Treat!" (with a higher probability of the former, most here would say)
_________________
.sigs waste space and bandwidth |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Sat Oct 31, 2015 7:08 pm Post subject: |
|
|
| depontius wrote: | | Naib wrote: | | NeddySeagoon wrote: | | There has to be a halloween reference to kdbus there somewhere but I can't spot it. |
the dead will rise again? |
Oh, come on! It's obvious... "Trick or Treat!" (with a higher probability of the former, most here would say) |
no, no, "all your base are belong to us" ... as the new improved halloween is 365 days a year.
best ... khay |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sun Nov 01, 2015 2:15 pm Post subject: |
|
|
Enough with the quote trees already: flatten 'em out, people! ;x
| NeddySeagoon wrote: | | There has to be a halloween reference to kdbus there somewhere but I can't spot it. |
| Naib wrote: | | the dead will rise again? |
| depontius wrote: | | Oh, come on! It's obvious... "Trick or Treat!" (with a higher probability of the former, most here would say) |
"Trick or Trick" ;) (story)
Heh: "All your Data are belong to US" seems apposite.
Though "All your data belong to us" is a different matter, relating to the freedom of information implicit in the "digital age", and the rights of everyone to knowledge, since all knowledge we have comes from our society.
That's why the sociopathic fantasy of the "Nietschzian Superman" is so useless; as if no-one ever had to change his dirty nappy.. ;-) |
|
| Back to top |
|
|
gwr
Apprentice
Joined: 19 Nov 2014
Posts: 194
|
| Posted: Fri Nov 06, 2015 12:18 pm Post subject: |
|
|
Article: kdbus not dead - from the systemd conference
http://www.golem.de/news/systemd-konferenz-kdbus-ist-nicht-tot-1511-117313.html
From Google Translate:
| Quote: |
The systemd maintainer Lennart Poettering reaffirmed a developer conference that kdbus will continue hand and "not dead". The implementation in the kernel and userspace will however rebuilt. How long that will take, is not yet clear.
For more than two years working Linux hackers and the systemd team with kdbus on implementing an interprocess communication (IPC) at the kernel level, which is a kind of replica of the previous D-Bus functionality. Systemd-founder Lennart Poettering has "not exactly a success story," described the development of this technology at the Conference as systemd, which is primarily due to the criticism from the kernel community well. But Poettering said just as clearly: "kdbus is not dead".
The presence of the IPC in the kernel is necessary, for example to already be able to communicate early in the boot process it. The need to use tools like NetworkD. But despite these repeated statements on the need for clear kdbus, the systemd team asked two weeks ago about that kdbus module will be removed from the kernel of Fedora Rawhide development branch. This seems like a harsh step backwards. Had the team but only this summer recommended a large-scale test using the module.
Kdbus is rebuilt
Actually, however, this is only a strategic "step back". Because According to Poettering of the team is just starting with major reconstruction work on kdbus. So currently would be implemented to the classic D-Bus in the kernel still a lot of compatibility. In future this will probably be outsourced, so should about the semantics are implemented in userspace. But yet it is not clear how much of the code will ultimately changed because the work on it are still running.
In an interview with Golem.de confirmed Poettering that will be discussed with the renovation work on some of the criticism from the kernel community. The work could lead to a more generic IPC. Presumably, the technology has thus also a greater chance of being taken up in the medium term but still in the Linux kernel.
|
|
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6148
Location: Dallas area
|
| Posted: Fri Nov 06, 2015 2:33 pm Post subject: |
|
|
Even if it never gets included in the kernel (like bfs, bfq and reiser4 to name a few) it doesn't mean it's dead.
They can carry the patches to the kernel and all the distros can apply it along with many others that they apply anyway.
I'm only really concerned that kdbus doesn't make it into the kernel with the (crappy) code that they've had in the past.
I'm sure there will be lots of commentary on lkml over it when the rewrite gets released.
_________________
UM780, 6.1 zen kernel, gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
|
gwr
Apprentice
Joined: 19 Nov 2014
Posts: 194
|
| Posted: Fri Nov 06, 2015 2:40 pm Post subject: |
|
|
| Anon-E-moose wrote: | | lots of commentary on lkml over it when the rewrite gets released. |
I find it interesting that they wish to re-architect it, given that they just finished trying to push it into the kernel and fought bitterly that it was ready. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Fri Nov 06, 2015 3:41 pm Post subject: |
|
|
| Quote: | | The presence of the IPC in the kernel is necessary, for example to already be able to communicate early in the boot process it. The need to use tools like NetworkD. But despite these repeated statements on the clear need for kdbus |
Nah, these statements are simply bullshit, that's why no-one's playing with your square ball..
IPC is necessary in the kernel, sure, but that's why we have at least 5 variants of it already.
| Quote: | | Actually, however, this is only a strategic "step back". |
Well, how about a "strategic" decision: NO. You do not know what you are doing, and everything you are trying to do is better done using existing facilities, while forgetting about some of the idiot design "constraints" inflicted by your wilful ignorance and lack of basic research, plus a large dose of intransigence (ie: pig-headed behaviour.)
Poettering, Sievers and Kroah-Hartman are simply not qualified to architect such solutions, and the pain they have inflicted on admins and users over the last 7 or 8 years must stop.
Go away. Perhaps move into management, with the rest of the bulshytters, and let the Peter principle continue.
| Quote: | | In an interview with Golem.de confirmed Poettering that will be discussed with the renovation work on some of the criticism from the kernel community. The work could lead to a more generic IPC. Presumably, the technology has thus also a greater chance of being taken up in the medium term but still in the Linux kernel. |
It's called TIPC, and it's already in the kernel, and has been available for Linux for 15 years. It also happens to be multi-platform.
"Catch-up with the modern-world, numpty; we sorted this stuff out in the 1990s, while you guys were gawking at your first view of a WIMP display (done badly.)"
~ Unix-greybeard. |
|
| Back to top |
|
|
gwr
Apprentice
Joined: 19 Nov 2014
Posts: 194
|
| Posted: Fri Nov 06, 2015 7:36 pm Post subject: |
|
|
I can't give conference talks and impress my fans with old technology. |
|
| Back to top |
|
|
krinn
Watchman
Joined: 02 May 2003
Posts: 7470
|
| Posted: Sat Nov 07, 2015 10:46 am Post subject: |
|
|
| steveL wrote: | | It's called TIPC |
From the features list on the wikipedia link, i see it really lack the feature
- Made by Pottering.
As such, i don't think they would use it, as this missing is a huge stopper. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sat Nov 07, 2015 12:55 pm Post subject: |
|
|
| gwr wrote: | | I can't give conference talks and impress my fans with old technology. |
Lul.. the saps could just run dcop over TIPC, and it would all be about the uses they put the technology to: that would get them loads of new conference material, about how efficient they were being, on the one hand, and about how the desktop apps were run over domain-specific, tuned mechanisms, rather than clumped through one badly-designed bottleneck ("now in 60 moving parts!")
They could even use capnrpc, and compare against a forked X, running under the logged-in user's uid, since it has an inter-desktop app-protocol already (or used to.)
There's plenty of fun stuff to be done (or brought-back), for people who actually have the time and inclination to work on it.
Just follow the advice of the kernel network hackers, and focus on the fun part in userland, instead of trying to hijack Linux.
Do this by understanding all the options available, from POSIX first, so that your apps run on other systems too, and select the ones which are most suited, rather than just what some nub told you to use, for those rare, but important, parts where you're not using a library.
| krinn wrote: | From the features list on the wikipedia link, i see it really lack the feature
- Made by Pottering.
As such, i don't think they would use it, as this missing is a huge stopper. |
Hah; well you called it: | krinn wrote: | does anyone think we're close to have a new challenger for the "software fiasco of the year" with kdbus?
What will be interesting is what redhat will do, sure they can have a patch kernel for themselves, calling it a fork or not, but it's personal use. 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. |
and here we are, 5 months later:
Hopefully, the Linux kernel bods will reconsider the lazy-thinking that says "let's hand over all control of cgroups to systemd" as well.
(discussed here and further below.) |
|
| Back to top |
|
|
krinn
Watchman
Joined: 02 May 2003
Posts: 7470
|
| Posted: Sat Nov 07, 2015 2:06 pm Post subject: |
|
|
| steveL wrote: | Hah; well you called it: | krinn wrote: | | does anyone think we're close to have a new challenger for the "software fiasco of the year" with kdbus? |
|
By that time, i didn't want to push my seer capacity as most unbelievers are always shocked when i do ; i think now everyone can just admit it: krinn is a seer!
My only regret, is that it looks like Redhat may not be pushing systemd (like i suppose earlier).
But i'm just a seer, never said i was God (i could, but i will wait until everyone admit my seer status first). |
|
| Back to top |
|
|
|
Kernel & Hardware
