Mailing List Archive

On Tue, Sep 3, 2013 at 10:24 AM, Frank Peters <frank.peters@comcast.net> wrote:
> A new linux kernel, 3.11, is now released and, as expected, the
> nvidia-drivers fail to emerge.
>
> For the impatient, a patch is available:
>
> http://pastebin.com/qV30u23p
>
> Just drop the patch file into /etc/portage/patches/x11-drivers/nvidia-drivers
> and you'll be good to go.
>
> Frank Peters


Nice. Thanks Frank!

- Mark

On 3 Sep 2013 20:25, "Frank Peters" <frank.peters@comcast.net> wrote:
>
> A new linux kernel, 3.11, is now released and, as expected, the
> nvidia-drivers fail to emerge.
>
> For the impatient, a patch is available:
>
> http://pastebin.com/qV30u23p
>
> Just drop the patch file into
/etc/portage/patches/x11-drivers/nvidia-drivers
> and you'll be good to go.
>
> Frank Peters

Thanks Frank :)

Alex

On 03/09/13 20:24, Frank Peters wrote:
> A new linux kernel, 3.11, is now released and, as expected, the
> nvidia-drivers fail to emerge.
>
> For the impatient, a patch is available:
>
> http://pastebin.com/qV30u23p
>
> Just drop the patch file into /etc/portage/patches/x11-drivers/nvidia-drivers
> and you'll be good to go.

Note that 3.10 is the new "long term support" kernel by upstream, so if
you decide not to upgrade to the latest and greatest, 3.10 will be
receiving updates for a long time to come.

http://kroah.com/log/blog/2013/08/04/longterm-kernel-3-dot-10

On Tue, Sep 3, 2013 at 10:53 AM, Nikos Chantziaras <realnc@gmail.com> wrote:
<SNIP>
>
> Note that 3.10 is the new "long term support" kernel by upstream, so if you
> decide not to upgrade to the latest and greatest, 3.10 will be receiving
> updates for a long time to come.
>
> http://kroah.com/log/blog/2013/08/04/longterm-kernel-3-dot-10


Great info Nikos. Thanks. I'm a bit thrown by a new Gentoo-Linux
option in the kernel but it looks like a good move long term for the
distro. Just have to test that I didn't break anything basic.

My machine here is getting into an Nvidia pickle that so far I've been
unable to get out of. Problem is I'm not sure where to go but I
haven't really asked anyone as I wasn't even quite sure how to
approach the problem. I'm not necessarily looking for a technical
solution today. Really I'm just interested in how others would
approach the problem.

Currently I'm running gentoo-sources-3.8.13 with
nvidia-drivers-313.30. For whatever reason my long running xorg.conf
file will no longer run with any nvidia driver newer than 313.30. X
just doesn't run. Based on your post I just built 3.10.7 and it
appears that 313.30 isn't supported with this newer kernel so pretty
soon I'm gonna have problems...

Possible directions to go:

1) Fix xorg.conf. Best solution but so far I haven't been able to do
it even with help from the Gentoo forums. Where to go? Don't know.
Nvidia forums maybe? Here? Dunno.

2) Drop nvidia-drivers and try nouveau. May be a reasonable video
direction - my screen-based video/graphics needs are not high -
however I have long terms needs of accessing the GPU as a compute
engine in R so I _think_ I have to stick with nvidia-drivers to do
that. Maybe not but I don't know at this time and I could deal with
that later. (I know Duncan's vote, if he's reading, goes here.) :-)

3) Buy one or two new VGA cards. I currently run 3 screens, 2 on a
GTX465 and 1 on a 8400 GS. My trading partner uses 2 8400's without
issues (so far) so I could dump the GTX465, buy 1 8400 and I'd be OK
but would lose a lot of compute power. None the less it would likely
work.

4) Something else...

All considered responses welcomed.

Cheers,
Mark

Mark Knecht <markknecht@gmail.com> skribis:
> 2) Drop nvidia-drivers and try nouveau. May be a reasonable video
> direction - my screen-based video/graphics needs are not high -
> however I have long terms needs of accessing the GPU as a compute
> engine in R so I _think_ I have to stick with nvidia-drivers to do
> that. Maybe not but I don't know at this time and I could deal with
> that later. (I know Duncan's vote, if he's reading, goes here.) :-)

I’ve had real trouble with nouveau graphics, not considering cuda
computing issues, and so have resigned myself to nvidia’s fetters for
the while. I’d switch graphics cards to be freer but, like fictional
Jerry Seinfeld’s nana, I’m ‘on a very fixed income’. :)

On Tue, Sep 3, 2013 at 12:51 PM, Barry Schwartz
<chemoelectric@chemoelectric.org> wrote:
> Mark Knecht <markknecht@gmail.com> skribis:
>> 2) Drop nvidia-drivers and try nouveau. May be a reasonable video
>> direction - my screen-based video/graphics needs are not high -
>> however I have long terms needs of accessing the GPU as a compute
>> engine in R so I _think_ I have to stick with nvidia-drivers to do
>> that. Maybe not but I don't know at this time and I could deal with
>> that later. (I know Duncan's vote, if he's reading, goes here.) :-)
>
> I’ve had real trouble with nouveau graphics, not considering cuda
> computing issues, and so have resigned myself to nvidia’s fetters for
> the while. I’d switch graphics cards to be freer but, like fictional
> Jerry Seinfeld’s nana, I’m ‘on a very fixed income’. :)


She's a great lady. Gotta luv her, and you're advice also. :-)

One thing I hate about my current setup is that with two different
nvidia cards OpenGL only works on the 2 screens attached to the
GTX465. No video apps work on the 3rd display, and the KDE OpenGL
features like (from memory - it's been so loon) Alt-F8, or whatever it
was that allows me to see all 6 desktops at once, don't work at all.
My friends machine which has 2 8400's work perfectly in that respect.

Finding a new card isn't out of the question. I'd prefer to find a
single card that drove 4 monitors, like maybe one or two of the
Quadro's do, but it's money and risk as I don't know of any Gentoo
User types using those cards.

I suspect I can disassociate the CUDA stuff if I have to. I don't have
to tell X11 that the 465 is in the machine, and I have 3 PCI Express
slots so I could potentially have 2 VGAs and 1 CUDA. (I think...)

Thanks for the inputs.

Cheers,
Mark

Mark Knecht wrote:
> Great info Nikos. Thanks. I'm a bit thrown by a new Gentoo-Linux
> option in the kernel but it looks like a good move long term for the
> distro. Just have to test that I didn't break anything basic. My
> machine here is getting into an Nvidia pickle that so far I've been
> unable to get out of. Problem is I'm not sure where to go but I
> haven't really asked anyone as I wasn't even quite sure how to
> approach the problem. I'm not necessarily looking for a technical
> solution today. Really I'm just interested in how others would
> approach the problem. Currently I'm running gentoo-sources-3.8.13 with
> nvidia-drivers-313.30. For whatever reason my long running xorg.conf
> file will no longer run with any nvidia driver newer than 313.30. X
> just doesn't run. Based on your post I just built 3.10.7 and it
> appears that 313.30 isn't supported with this newer kernel so pretty
> soon I'm gonna have problems... Possible directions to go: 1) Fix
> xorg.conf. Best solution but so far I haven't been able to do it even
> with help from the Gentoo forums. Where to go? Don't know. Nvidia
> forums maybe? Here? Dunno. 2) Drop nvidia-drivers and try nouveau. May
> be a reasonable video direction - my screen-based video/graphics needs
> are not high - however I have long terms needs of accessing the GPU as
> a compute engine in R so I _think_ I have to stick with nvidia-drivers
> to do that. Maybe not but I don't know at this time and I could deal
> with that later. (I know Duncan's vote, if he's reading, goes here.)
> :-) 3) Buy one or two new VGA cards. I currently run 3 screens, 2 on a
> GTX465 and 1 on a 8400 GS. My trading partner uses 2 8400's without
> issues (so far) so I could dump the GTX465, buy 1 8400 and I'd be OK
> but would lose a lot of compute power. None the less it would likely
> work. 4) Something else... All considered responses welcomed. Cheers,
> Mark


I don't know if it is related or not but I to have issues with Nvidia
drivers and the little KDE panel/kicker thingy at the bottom. After
being logged in for a while, the panel thing would lock up tight. The
clock wouldn't update, I couldn't use it to switch desktops, click the K
menu or anything. The best version I found for a while was this:

x11-drivers/nvidia-drivers-313.30

and kernel:

3.9.5-gentoo

Right now I have that kernel version working with:

nvidia-drivers-319.49

I have a Nvidia GeForce GT 220 video card. I think it has 1GB of ram or
some amount likely larger than I need. At this point, I can't tell if
it is Nvidia, KDE, or some other mismatch causing this. Right now, I'm
just sticking with what I know works.

May not be related but it may help you to compare notes maybe?

Dale

:-) :-)

--
I am only responsible for what I said ... Not for what you understood or how you interpreted my words!

Mark Knecht posted on Tue, 03 Sep 2013 13:25:15 -0700 as excerpted:

> I suspect I can disassociate the CUDA stuff if I have to. I don't have
> to tell X11 that the 465 is in the machine, and I have 3 PCI Express
> slots so I could potentially have 2 VGAs and 1 CUDA. (I think...)

There's actually some brand new developments in that regard -- hit the
Linux news in the last week, and scheduled for 3.12.

So what's the big deal? They're splitting the formerly single-device KMS/
DRM into two separate devices, one of which will be render/compute-only,
and thus require lower privs -- classic Unix user/group file permissions,
not the root or CAP_SYS_ADMIN that the current device requires, and a
second that will be mode-setting/display-controller only, accessible via
clone from the current DRM-control node (which is currently unused).
Because no new resources are created -- they're simply cloned from the
existing control node, required privs here can be reduced as well. The
practical effect of the latter will be another step toward allowing user-
priv-only X (and wayland).

But the former is bigger news in srovider. For just $5 per month, you can
get yourself a Cloud server with 512 MB of RAM, 20 GB super-fast SSD,
free snapshots, plus backups for a minimal fee. All via a simple
graphical interface.

And by signing up with this referral link, you can help support this
website.

If you are reading this, your ad could also be occupying this space.
Contact us to make it happen.
2 Comments

ome ways, as it means any app running as a user/group with suitable file
permissions will be able to open compute nodes, no root or CAP_SYS_ADMIN
required. That will in turn dramatically open up the possibilities for
"ordinary application" use of compute-nodes, thus allowing pretty much
any app to use those resources, instead of forcing the severe privilege
restrictions currently needed to avoid huge security issues, currently.

So why is all this happening now? It turns out that while on x86/amd64,
it has historically been the same hardware device handling both rendering
and mode-setting, on arm, which is where the big money is now since
that's what all the mobile devices seem to run on, these functions are
often split into two different pieces of hardware. The existing unified
setup could be made to work, but it wasn't a particularly good fit, while
the new setup is a natural fit.

Meanwhile, while they could have solved their own little hardware problem
in their corner of the computing world in just their own drivers, the
bigger solution was to split the functionality up in general, thus
allowing solutions to a bunch of thorny problems with the existing
unified hardware as well. =:^) In addition to the security and compute
resource availability issues mentioned above that this solves, it also
makes solving problems such as the newer embedded/dedicated gpu pairing
on laptops, with the display only attached to the embedded, thus making
the dedicated card that's supposed to activate for games and stuff
useless, under the current unified setup. Similarly, it'll now be
possible to run different user X sessions on different displays driven by
the same GPU hardware -- multi-seat-on- single-gpu-multi-display, which
currently isn't possible, because only one app can be DRM-master on a gpu
at once, and that app must be run by one user, it can't be shared.

So the separate render/modeset hardware on arm is driving better and far
more flexible solutions for (normally) unified render/modeset hardware on
x86/amd64, too! =:^)

Original blog announcement:

https://dvdhrm.wordpress.com/2013/09/01/splitting-drm-and-kms-device-
nodes/

Initial LWN coverage. (Not much there besides the single paragraph
pointer to the above at present, but kernel devs often comment there, and
LWN will very likely have a feature article covering it in some depth in
one of its coming weekly editions, too.)

http://lwn.net/Articles/565464

Phoronix (pre-blog announcement coverage):

http://www.phoronix.com/scan.php?page=news_item&px=MTQ0OTI

(earlier...)

http://www.phoronix.com/scan.php?page=news_item&px=MTQ0MzQ

(earlier still, July...)

http://www.phoronix.com/scan.php?page=news_item&px=MTQwNTA

--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman

Am Wed, 4 Sep 2013 06:01:48 +0000 (UTC)
schrieb Duncan <1i5t5.duncan@cox.net>:

[...]
> So what's the big deal? They're splitting the formerly single-device KMS/
> DRM into two separate devices, one of which will be render/compute-only,
> and thus require lower privs -- classic Unix user/group file permissions,
> not the root or CAP_SYS_ADMIN that the current device requires, and a
> second that will be mode-setting/display-controller only, accessible via
> clone from the current DRM-control node (which is currently unused).
> Because no new resources are created -- they're simply cloned from the
> existing control node, required privs here can be reduced as well. The
> practical effect of the latter will be another step toward allowing user-
> priv-only X (and wayland).
>
> But the former is bigger news in srovider. For just $5 per month, you can
> get yourself a Cloud server with 512 MB of RAM, 20 GB super-fast SSD,
> free snapshots, plus backups for a minimal fee. All via a simple
> graphical interface.
>
> And by signing up with this referral link, you can help support this
> website.
>
> If you are reading this, your ad could also be occupying this space.
> Contact us to make it happen.
> 2 Comments

Woah! Is it normal for ads to be placed smack in the middle of your emails?
I've heard of ads inserted at the top or bottom of HTML-Emails, but this is
just plain ridiculous.

And if that wasn't bad enough, it appears to have taken part of the next
paragraph, too ("ome ways...").

> ome ways, as it means any app running as a user/group with suitable file
> permissions will be able to open compute nodes, no root or CAP_SYS_ADMIN
> required. That will in turn dramatically open up the possibilities for
> "ordinary application" use of compute-nodes, thus allowing pretty much
> any app to use those resources, instead of forcing the severe privilege
> restrictions currently needed to avoid huge security issues, currently.
>
[...]

--
Marc Joliet
--
"People who think they know everything really annoy those of us who know we
don't" - Bjarne Stroustrup

Hi All,
I've reported two problems recently to the list:

1) Problems with X11 configuration on my 3 monitor, 2 card setup
2) A >2 minute login problem

This morning, based on Nikos' comments yesterday about
gentoo-sources-3.10 being a long term support kernel, I brought up
3.10.7. As with earlier recent attempts I didn't get X at all.

As an experiment I tried removing everything from my xorg.conf having
to do with the second card & monitor and tried the newest
nvidia-driver package listed on their site as "Long Lived Branch"
which is 319.49. Amazingly, _both_ problems were solved. X came up
nicely (on 2 monitors only) and login is immediate.

I think for my needs I'll remove the unused card, live with 2
monitors for now and consider how to proceed while maintaining a long
term supported nature to the kernel and video drivers. I run 'mostly
stable' so as long as long term kernels get security updates and I
don't add new hardware to this 3 year old machine I think I'll be
happy.

Cheers,
Mark

On Tue, Sep 3, 2013 at 1:25 PM, Mark Knecht <markknecht@gmail.com> wrote:
> On Tue, Sep 3, 2013 at 12:51 PM, Barry Schwartz
> <chemoelectric@chemoelectric.org> wrote:
>> Mark Knecht <markknecht@gmail.com> skribis:
>>> 2) Drop nvidia-drivers and try nouveau. May be a reasonable video
>>> direction - my screen-based video/graphics needs are not high -
>>> however I have long terms needs of accessing the GPU as a compute
>>> engine in R so I _think_ I have to stick with nvidia-drivers to do
>>> that. Maybe not but I don't know at this time and I could deal with
>>> that later. (I know Duncan's vote, if he's reading, goes here.) :-)
>>
>> I’ve had real trouble with nouveau graphics, not considering cuda
>> computing issues, and so have resigned myself to nvidia’s fetters for
>> the while. I’d switch graphics cards to be freer but, like fictional
>> Jerry Seinfeld’s nana, I’m ‘on a very fixed income’. :)
>
>
> She's a great lady. Gotta luv her, and you're advice also. :-)
>
> One thing I hate about my current setup is that with two different
> nvidia cards OpenGL only works on the 2 screens attached to the
> GTX465. No video apps work on the 3rd display, and the KDE OpenGL
> features like (from memory - it's been so loon) Alt-F8, or whatever it
> was that allows me to see all 6 desktops at once, don't work at all.
> My friends machine which has 2 8400's work perfectly in that respect.
>
> Finding a new card isn't out of the question. I'd prefer to find a
> single card that drove 4 monitors, like maybe one or two of the
> Quadro's do, but it's money and risk as I don't know of any Gentoo
> User types using those cards.
>
> I suspect I can disassociate the CUDA stuff if I have to. I don't have
> to tell X11 that the 465 is in the machine, and I have 3 PCI Express
> slots so I could potentially have 2 VGAs and 1 CUDA. (I think...)
>
> Thanks for the inputs.
>
> Cheers,
> Mark

Mark Knecht posted on Wed, 04 Sep 2013 12:09:51 -0700 as excerpted:

> Amazingly, _both_ problems were solved. X came up nicely (on 2 monitors
> only) and login is immediate.
>
> I think for my needs I'll remove the unused card, live with 2
> monitors for now and consider how to proceed while maintaining a long
> term supported nature to the kernel and video drivers. I run 'mostly
> stable' so as long as long term kernels get security updates and I don't
> add new hardware to this 3 year old machine I think I'll be happy.

Currently, things really do seem to be moving toward the single-card/
multi-output solution, for multi-monitor, and they've been moving that
way for awhile, as multiple outputs on a single card become more common.
There's a lot of very nasty graphics coordination issues that simply
disappear when it's a single card driving everything.

At my last graphics update, a big one as it was part of a full system
update, finally moving from an old AGP bus mobo (on an 8-year-old system)
to modern PCIE, I was happy to find not just double-output, but triple-
output, was now reasonably common in a mainline card.

Couple that with the trends towards larger monitors and digital-TV
finally allowing the merge of TVs and monitors at larger sizes and
reasonable (TV) HD-resolutions (TV resolutions are reasonable for
monitors now too; for years they were simply too low), thus bringing down
"big-screen" TV/monitor prices into the "normal people's budget" realm,
and the now commonly available 40-inch-plus TV/monitor solutions really
do drive down the need for multiple monitors as well.

Stated elsewise, it's an unusual use-case indeed that really NEEDS more
than triple-monitor at 1920x1080 each, at 42-72 inches per monitor, and
that's now within reach of a reasonably mainline budget, on just a single
very solidly mainline single graphics card.

FWIW, I'm a bit physically size-constrained here and that's the effective
cap on total graphics real-estate available to me now. I'm running
triple monitor, two 42-inch full-HD 1920x1080 in stacked config for
1920x2160 as mains, with an old 21-inch full-HD logically stacked on top
of that for a 1920x3240 desktop, but the third one is physically located
to the side, displaying real-time system monitor information at double
the font size (actually triple, I think) since it's half the physical
size at the same resolution, thus still making it still readable with my
now aging (mid to late 40s, reading/computer glasses!) eyes.

I /think/ it was this list where that came up in a thread and I posted
screenshots, a few months ago.

But... with dual full-HD 42-inch-TV monitors as my mains, I really don't
/need/ the third one and only threw it into the mix when I realized I had
both a still unused output and an old monitor laying around from before
the upgrade to 42-inch. It's nice to have as it allows me to get the
real-time system monitor stuff entirely out of the normal working area
and off my main monitors while still keeping it visible at all times, but
that's it, nice-to-have, NOT don't-know-what-I'd-do-without-it.

OTOH, if I wasn't so space constrained and with a somewhat bigger budget
to work with, I'd probably have tried to find a 4-output card at a semi-
reasonable price-point, and would have gone with a 4 by 52-60-inch
monitors config in classic 2x2 layout. Or even a (still definitely *NOT*
mainline and thus *QUITE* expensive, AFAIK) 6-output setup and a classic
3x2 layout, all say 60-inch monitors, double-stacked, left/center/right.

But of course the 6X in 3x2 layout thing is entirely dreamland and would
take probably double my income level (at least) to put it in anything
like reasonable territory, as there's a lot of other things (including a
bigger house/room to put it in!) that would take priority before I got to
that. I'm really quite happy indeed with my current setup, tho I might
well upgrade that current 21-inch to a 32-inch at some point, about the
biggest I could fit where I have it, and if I upgrade to quad-output
graphics at some point, in theory I could add a second 32-inch beside
that before I'd really have my space maxed out, but at this point I
really can't say what I'd actually use it for.

--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman

Marc Joliet posted on Wed, 04 Sep 2013 10:57:16 +0200 as excerpted:

> Woah! Is it normal for ads to be placed smack in the middle of your
> emails?

Woops! Uncaught middle-click-paste while scrolling!

Sorry!

<sheepish> I suppose I'm lucky it was an ad and that I hadn't just
checked my bank account online or something. </sheepish>

--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman