View previous topic :: View next topic |
Author |
Message |
khumba n00b
Joined: 16 Jun 2011 Posts: 51
|
Posted: Thu Apr 23, 2015 12:37 am Post subject: Re: kdbuswreck |
|
|
miket wrote: | There you go. systemdamage and kdbuswreck. |
I like "dbust":
One Thousand Gnomes wrote: | If the userspace folks choose to continue to implement dbust over it but the kernel layer is clean and generic then all is good, because someone can replace dbust with something better. If its got dbust hard wired into it then its a complete mess. |
|
|
Back to top |
|
|
gwr Apprentice
Joined: 19 Nov 2014 Posts: 194
|
Posted: Thu Apr 23, 2015 12:42 pm Post subject: |
|
|
From lwn.net kdbuswreck article:
Quote: | attempt to derive a bit of light from the massive amounts of heat that have been generated so far |
That's not a biased statement at all, is it? |
|
Back to top |
|
|
Shamus397 Apprentice
Joined: 03 Apr 2005 Posts: 218 Location: Ur-th
|
Posted: Thu Apr 23, 2015 3:54 pm Post subject: |
|
|
It's funny to see LP whining about that article (still behind the paywall), and complaining about the title "The Kdbuswreck".
I didn't know that "mezcalero" was Poettering; it explains quite a bit from comments to articles I've read on LWN. |
|
Back to top |
|
|
roki942 Apprentice
Joined: 18 Apr 2005 Posts: 285 Location: Seattle
|
Posted: Thu Apr 23, 2015 4:31 pm Post subject: |
|
|
Boy he's really touchy. A bus wreck does not assume that there was something wrong with the bus. Though it can be the driver's fault .... and LP is driving the Kbus .... guess he's saying the light turned red too soon. |
|
Back to top |
|
|
depontius Advocate
Joined: 05 May 2004 Posts: 3509
|
Posted: Thu Apr 23, 2015 4:33 pm Post subject: |
|
|
Shamus397 wrote: | It's funny to see LP whining about that article (still behind the paywall), and complaining about the title "The Kdbuswreck".
I didn't know that "mezcalero" was Poettering; it explains quite a bit from comments to articles I've read on LWN. |
I rather like that "One Thousand Gnomes" seems to be Alan Cox. _________________ .sigs waste space and bandwidth |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Fri Apr 24, 2015 11:23 am Post subject: |
|
|
Shamus397 wrote: | I didn't know that "mezcalero" was Poettering; it explains quite a bit from comments to articles I've read on LWN. |
Could it be finally that he doesn't that much crazy fan, but a lot of nickname???
(don't read this if you are easy scare): "Could it be that ulenrich is... LP!!!" |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Fri Apr 24, 2015 7:25 pm Post subject: |
|
|
@krinn: Do you know that people with technical doubts in the whole *kit story are discredited with the "conspiracy theory" argument?
You make it too easy for some to fasttalk around the technical facts... |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Fri Apr 24, 2015 8:03 pm Post subject: |
|
|
mv wrote: | @krinn: Do you know that people with technical doubts in the whole *kit story are discredited with the "conspiracy theory" argument?
You make it too easy for some to fasttalk around the technical facts... |
I don't think, even the lamest one, would use a small taunt/funny quote to use it to backup any "conspiracy theory". Even someone not knowing ulenrich may get it easy. |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Apr 24, 2015 9:34 pm Post subject: |
|
|
Shamus397 wrote: | It's funny to see LP whining about that article (still behind the paywall), and complaining about the title "The Kdbuswreck". |
LP wrote: | Also, I don't think calling kdbus a "wreck" is appropriate at all. |
corbet wrote: | ...and I never did that. The title refers to the discussion, not the technology.
....
With regard to the title...perhaps it was a bad choice, but "buswreck" (or "trainwreck") is a fairly common English term for an unfortunate situation. I still believe that you have to stretch pretty hard to say that "The kdbuswreck" (note "the") somehow refers to the code. |
Germans associate "wreck" with "Schiffswrack" - a ship stranded on the shore. But later on jjmarin wrote: | I agree that "The kdbuswreck" is a misleading title, IMHO, I think the title should be more precise to convey the general meaning of the article, maybe something like "The kdbus debate wreck"... anyway, I'm sure there must be a much better suggestion for the title |
Perhaps following correction given by Linus himself solves the kdebuswreck debate:
http://lkml.iu.edu/hypermail/linux/kernel/1504.2/04915.html Quote: | No Greg.
Just remove the shit. Really. Take out the command line and the task
name. You already admitted that there is no actual valid use for it.
We don't add crap that then has to be disabled with secuirity rules
just because it was a bad interface. Just make the interface not do it
in the first place. It's that simple. |
|
|
Back to top |
|
|
truekaiser l33t
Joined: 05 Mar 2004 Posts: 801
|
Posted: Fri Apr 24, 2015 11:21 pm Post subject: |
|
|
ulenrich wrote: |
Perhaps following correction given by Linus himself solves the kdebuswreck debate:
http://lkml.iu.edu/hypermail/linux/kernel/1504.2/04915.html Quote: | No Greg.
Just remove the shit. Really. Take out the command line and the task
name. You already admitted that there is no actual valid use for it.
We don't add crap that then has to be disabled with secuirity rules
just because it was a bad interface. Just make the interface not do it
in the first place. It's that simple. |
|
At which point gkh cries, STOP OPPRESSING ME!!1!
and gets linus removed via the standard of conduct patch he rammed through earlier.. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Sat Apr 25, 2015 1:23 pm Post subject: |
|
|
Oh jeez; that's as about as useful as a claim from a manufacturer that some new product in the same range is better than their existing pile of turd.
A better comparison was the one between dbus, Gnome's ORB, and dcop. ie alternative technologies doing the same thing.
So in this case, a comparison between kdbus and TIPC is what's needed. Anything else is just propaganda^W advertising^W marketing^W "communicating" (nonsense.)
I see where krinn is coming from: you unfailingly repeat the same old gloss that we've heard before, and never address the substantive technical issues, when it comes to anything Poeterring-related.
I'm tired of repeating the same old URLs over and over. If you want to counter the claims, read them and just for a change, show us you've read them in the arguments you make. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Sat Apr 25, 2015 1:30 pm Post subject: |
|
|
truekaiser wrote: | At which point gkh cries, STOP OPPRESSING ME!!1!
and gets linus removed via the standard of conduct patch he rammed through earlier.. |
It's much more insidious than that; the whole idea is to set the tone for when Linus retires/dies.
Once he's gone, no-one else is going to have any sort of standing to simply talk straightforwardly, instead of couching everything in politically-(in)correct vague, which as we all know is exactly how corporates take over in the background.
Everyone ends up self-censoring, and their headspace is wasted on that, instead of correcting mistakes.
Multinationals play a long game, while forcing everyone else to focus on last quarter's results, or next week's shopping.
Meanwhile, "I can't tell Linus, but I sure as hell can tell you." |
|
Back to top |
|
|
roki942 Apprentice
Joined: 18 Apr 2005 Posts: 285 Location: Seattle
|
Posted: Sun Apr 26, 2015 10:45 am Post subject: |
|
|
oh I like this one http://lkml.iu.edu/hypermail/linux/kernel/1504.2/04649.html
Quote: | > Did I miss anything else here? Are there any technical reasons I'm
> forgetting about for why this can't be pulled in as-is for this merge
> window?
Like the outstanding NACKS ?
Greg - you are sounding like you have some kind of special entitlement to
ignore the way this works for everyone else. If you are feeling
frustrated, annoyed and led up several avenues at once then welcome to
the world of every other submitter who doesn't think have some kind of
magic stage door pass to get their crap in the kernel when there
are core maintainers asking hard and unanswerd questions and who have
nacked it.
There's no huge hurry. There are a bunch of things like the interactions
with cgroups, and the privilege and capability model which need careful
examination. Slipping it one release to get that right isn't a big deal -
it's not even as if you can't use hardware without it as with a driver
missing a merge - this is just a performance tweak.
Alan |
|
|
Back to top |
|
|
mrbassie l33t
Joined: 31 May 2013 Posts: 772 Location: over here
|
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6051 Location: Removed by Neddy
|
Posted: Sun Apr 26, 2015 11:54 am Post subject: |
|
|
Lol standard approach to build a system...
The number of times I have crafted a large analogue circuit in say Matlab, saber, simetrix ONLY for it to not run (damn Jacobian matrix...) Wasting days...
Only to start from scratch, build it up submodule by submodule with a suitable testbench strapped around to check it... Only for the complete model to work. 1st looks show there to be no obvious differences but there must be _________________
Quote: | Removed by Chiitoo |
|
|
Back to top |
|
|
arnvidr l33t
Joined: 19 Aug 2004 Posts: 629 Location: Oslo, Norway
|
Posted: Sun Apr 26, 2015 3:13 pm Post subject: |
|
|
mrbassie wrote: | http://lkml.iu.edu/hypermail/linux/kernel/1504.2/04853.html | And three days later, no reply. _________________
|
|
Back to top |
|
|
ct85711 Veteran
Joined: 27 Sep 2005 Posts: 1791
|
Posted: Sun Apr 26, 2015 3:20 pm Post subject: |
|
|
Actually the discussion is on the next archive page over. Here's the link to the start of the kdbus chain for this week. http://lkml.iu.edu/hypermail/linux/kernel/1504.3/00006.html (Haven't seen anything posted for yesterday, the 25th, though I am thinking it's just a delay before it gets updated on this archive. |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Sun Apr 26, 2015 6:20 pm Post subject: |
|
|
mrbassie wrote: | http://lkml.iu.edu/hypermail/linux/kernel/1504.2/04853.html | Very interesting read, especially this:
Quote: | This code has astonishingly complex interactions with all kinds of other
kernel subsystems and concerns. As a community we should understand
them and accept them before letting them in.
The only way I have seen anything make meaninful progress with those
kinds of interactions is for the pieces to be teased apart. And then
the code incrementally added to with all of the right people being
pulled into the discussion. |
Seems like he's saying they need to break their monolithic mess into small pieces that each...Oh, I don't know..."do one thing and do it well"? With those folks pigs will fly first I think. That philosophy somehow seems to be lost on everyone these days. I was reading an article this morning about Ubuntu 15.04 moving to systemd and the "controversy". The article totally white-washed it saying that systemd wasn't monolithic because it "creates 69 binaries"...what a flaming load of crap.
Tom |
|
Back to top |
|
|
Anon-E-moose Watchman
Joined: 23 May 2008 Posts: 6098 Location: Dallas area
|
Posted: Mon Apr 27, 2015 9:58 am Post subject: |
|
|
Quote: | This code has astonishingly complex interactions with all kinds of other
kernel subsystems and concerns. As a community we should understand
them and accept them before letting them in. |
This is the most frightening part of their push to have it in the kernel so quickly.
It touches all kinds of subsystems.
Which brings up the question "is that really needed" for dbus
or is this just a slight of hand to get more control over kernel pieces
and somehow filter them through sysd thereby cutting out any and all other init systems.
ie. a wrapper around the kernel...long live sysd.
I don't think they want people, especially the kernel devs to see what the true interactions
with all the subsystems is.
Unlike GKH's whine, I don't trust the ones who wrote kdbus, I don't trust their motives,
I don't trust GKH's push to have it in come hell or high water without any meaningful reviews.
Nack, Nack, Nack...ad infinitum. _________________ PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6051 Location: Removed by Neddy
|
Posted: Mon Apr 27, 2015 10:46 am Post subject: |
|
|
Anon-E-moose wrote: | Quote: | This code has astonishingly complex interactions with all kinds of other
kernel subsystems and concerns. As a community we should understand
them and accept them before letting them in. |
This is the most frightening part of their push to have it in the kernel so quickly.
It touches all kinds of subsystems.
Which brings up the question "is that really needed" for dbus
or is this just a slight of hand to get more control over kernel pieces
and somehow filter them through sysd thereby cutting out any and all other init systems.
ie. a wrapper around the kernel...long live sysd.
I don't think they want people, especially the kernel devs to see what the true interactions
with all the subsystems is.
Unlike GKH's whine, I don't trust the ones who wrote kdbus, I don't trust their motives,
I don't trust GKH's push to have it in come hell or high water without any meaningful reviews.
Nack, Nack, Nack...ad infinitum. |
Hanlon's razor...
They probably did that because they are userland coders and not use to atomic structure. The problem is if it went in as-is those tentrils would probably start getting abused. There is no need for it to start linking across everything... its a messagebus Userland <> Userland .... not a Userland --> deepKernel _________________
Quote: | Removed by Chiitoo |
|
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Mon Apr 27, 2015 3:23 pm Post subject: |
|
|
Naib wrote: | They probably did that because they are userland coders and not use to atomic structure. |
The real problem is they're not even good userland coders; they don't seem to know the POSIX base at all, let alone the rest of the UNIX world.
Consequently, they are blinkered in their approach; to put it mildly.
Quote: | The problem is if it went in as-is those tendrils would probably start getting abused. |
s/probably/definitely/ -- if there's a way to make money doing something, then someone will do it, even if most of us think it reprehensible.
If you think that's far-fetched, consider the milieu most of these nubs come from (Windoze) for the archetype of where RedHat is going.
Quote: | There is no need for it to start linking across everything... its a messagebus Userland <> Userland .... not a Userland --> deepKernel |
Follow it through, and you arrive at TIPC.
Unfortunately: "You won't get round us with that design nonsense.." ;-) |
|
Back to top |
|
|
Anon-E-moose Watchman
Joined: 23 May 2008 Posts: 6098 Location: Dallas area
|
Posted: Mon Apr 27, 2015 11:09 pm Post subject: |
|
|
Linus just posted this http://lkml.iu.edu/hypermail/linux/kernel/1504.3/02040.html
Quote: | >> That said, either you're running your test on a potato, or dbus is
>> seriously screwed up. No way should it take 4+ seconds to send a 1000b
>> message to back and forth 20k times. But as mentioned, I can't even
>> see what it's doing right now.
>
> Whee! I'm typing this email on a potato!
No, I think you're right, there's the other non-potato choice: "dbus
is seriously screwed up".
That thing has almost no kernel footprint. It's spending all it's time
in user space overhead.
Quite frankly, the whole "kdbus is important for performance" seems to
be *totally* invalidated by even a minimal look at profiles for that
thing. Here's the top-15 offender list:
2.62% gdbus libc-2.20.so [.] _int_malloc
2.43% gdbus libc-2.20.so [.] free
2.31% server libc-2.20.so [.] free
2.12% gdbus libc-2.20.so [.] malloc
1.77% gdbus libglib-2.0.so.0.4200.2 [.] g_utf8_validate
1.43% gdbus libglib-2.0.so.0.4200.2 [.] g_slice_alloc
1.41% gdbus libglib-2.0.so.0.4200.2 [.] g_hash_table_lookup
1.28% server libc-2.20.so [.] _int_malloc
1.27% gdbus libglib-2.0.so.0.4200.2 [.] g_mutex_lock
1.22% gdbus libglib-2.0.so.0.4200.2 [.] g_variant_unref
1.16% server libc-2.20.so [.] malloc
1.14% gdbus libglib-2.0.so.0.4200.2 [.] g_bit_lock
1.07% gdbus libglib-2.0.so.0.4200.2 [.] g_slice_free1
1.05% gdbus libglib-2.0.so.0.4200.2 [.] g_bit_unlock
1.01% gdbus libglib-2.0.so.0.4200.2 [.] g_mutex_unlock
there's not a kernel function in sight in the top-15, and it's all
just overhead. The above is from the server side, but the client looks
similar.
If somebody wants to speed up dbus, they should likely look at the
user-space code, not the kernel side.
My guess is that pretty much the entirely of the quoted kdbus
"speedup" isn't because it speeds up any kernel side thing, it's
because it avoids the user-space crap in the dbus server.
IOW, all the people who say that it's about avoiding context switches
are probably just full of shit. It's not about context switches, it's
about bad user-level code. |
LoL
and in the follow up post
Quote: | <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> IOW, all the people who say that it's about avoiding context switches
> are probably just full of shit. It's not about context switches, it's
> about bad user-level code.
Just to make sure, I did a system-wide profile (so that you can
actually see the overhead of context switching better), and that
didn't change the picture.
The scheduler overhead *might* be 1% or so.
So really. The people who talk about how kdbus improves performance
are just full of sh*t. Yes, it improves things, but the improvement
seems to be 100% "incidental", in that it avoids a few trips down the
user-space problems.
The real problems seem to be in dbus memory management (suggestion:
keep a small per-thread cache of those message allocations) and to a
smaller degree in the crazy utf8 validation (why the f*ck does it do
that anyway?), with some locking problems thrown in for good measure. |
_________________ PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
Back to top |
|
|
Anon-E-moose Watchman
Joined: 23 May 2008 Posts: 6098 Location: Dallas area
|
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6051 Location: Removed by Neddy
|
Posted: Tue Apr 28, 2015 6:53 am Post subject: |
|
|
Get ready for the sysd manoeuvre "kdbus is not about speed improvement its about, um, security... Wait no its about ... STOP THIS CAMPAIGN OF HATE"
Funny how they are having to review userland calls to get an idea what kdbus is suppose todo _________________
Quote: | Removed by Chiitoo |
|
|
Back to top |
|
|
|