On 08/04/2016 12:41 PM, Tobias Klausmann wrote:

On 04.08.2016 18:55, Larry Finger wrote:

...

If you are running TW in a VirtualBox Virtual Machine, DO NOT install this snapshot!!!!! There is a bug that breaks xboxvideo for both KDE and Gnome. Not only does the new 4.7.0-1 kernel fail to boot to X, but something has broken X even when using older kernels.

The specifics of the problem are unknown, but you have been warned.

Larry

There has been an xorg-server update lately (1.18.3 -> 1.18.4). Maybe the updated server is the problem, you could check /var/log/xorg.0.log to see what is wrong.

That may be part of the problem, but a kernel BUG is triggered when xboxvideo loads. I need to find the reason for it first. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/05/2016 02:20 AM, Herbert Graeber wrote:

Am 04.08.2016 um 20:29 schrieb Larry Finger:

...

On 08/04/2016 12:41 PM, Tobias Klausmann wrote:

...

On 04.08.2016 18:55, Larry Finger wrote:

...

If you are running TW in a VirtualBox Virtual Machine, DO NOT install this snapshot!!!!! There is a bug that breaks xboxvideo for both KDE and Gnome. Not only does the new 4.7.0-1 kernel fail to boot to X, but something has broken X even when using older kernels.

The specifics of the problem are unknown, but you have been warned.

Larry

There has been an xorg-server update lately (1.18.3 -> 1.18.4). Maybe the updated server is the problem, you could check /var/log/xorg.0.log to see what is wrong.

That may be part of the problem, but a kernel BUG is triggered when xboxvideo loads. I need to find the reason for it first. It's not a kernel bug. The really old xboxvideo module (5.0.18) is not compatible with the new kernel. Maybe an update of virtualbox (5.0.26 or better 5.1.2) will fix that.

If you have installed 20160830, you can boot wth the previous kernel and everything is well.

I have been trying to get later versions of VirtualBox to build in OBS for some time; however, they fail because the i586 build runs out of memory even though it has been allowed to have as much as 20 GB. Other times, both flavors run out of disk space, and yesterday, a build failed because of networking problems. I have finally disabled 32-bit builds in the hope that something will make it through. Unfortunately, that has not yet happened. Given the long delays in OBS, each time one of these builds has to be resubmitted, it takes about 2 weeks until the next such failure. I have no idea how to get an updated version into TW. One other thing. I installed a locally-built version of 5.0.26 onto TW. It also failed. Perhaps there is a problem with xorg-server. Version 5.0.26 is supposed to be building at the moment, but I have no idea if it will succeed, or if it will fix the problems. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/05/2016 11:48 AM, Richard Brown wrote:

Top posting because this is your obligatory reminder that a reliance of 3rd party kernel drivers is not recommended for Tumbleweed users

This is clearly documented on the Tumbleweed wiki - https://en.opensuse.org/Portal:Tumbleweed

The reason for this recommendation is precisely because of issues such as these - they happen and will continue to happen as long as the 3rd parties in question (Nvidia and Oracle/VirtualBox) do not keep up with the pace of Kernel development

Now I've said 'We told you so', I do hope this thread finds suitable workarounds for users who are affected, and if not, I'd strongly encourage the use of nouveau Nvidia drivers instead of the proprietary drivers, and the use of KVM instead of Virtualbox.

Good advice, but what do users that have Windows hosts do about virtual machines? Is there a KVM implementation for them? One other question: Why is the Factory build request for VB 5.1.2 only reached the "accepted" stage even though it was sent to OBS 10 days ago? Yes, I know that 5.0.18 is old, but 5.0.22 was sent to OBS on June 20, and 5.0.26 was sent on July 18. How can we ever get an updated version when nothing ever comes out the back end? Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 5 August 2016 at 19:04, Larry Finger <Larry.Finger@lwfinger.net> wrote:

On 08/05/2016 11:48 AM, Richard Brown wrote:

...

Top posting because this is your obligatory reminder that a reliance of 3rd party kernel drivers is not recommended for Tumbleweed users

This is clearly documented on the Tumbleweed wiki - https://en.opensuse.org/Portal:Tumbleweed

The reason for this recommendation is precisely because of issues such as these - they happen and will continue to happen as long as the 3rd parties in question (Nvidia and Oracle/VirtualBox) do not keep up with the pace of Kernel development

Now I've said 'We told you so', I do hope this thread finds suitable workarounds for users who are affected, and if not, I'd strongly encourage the use of nouveau Nvidia drivers instead of the proprietary drivers, and the use of KVM instead of Virtualbox.

Good advice, but what do users that have Windows hosts do about virtual machines? Is there a KVM implementation for them?

While I would normally recommend the inverse, having Windows as a VM in a Linux host, Windows ships with the Hyper-V hypervisor, which has it's GPL licensed drivers baked into the Linux Kernel So I would recommend Hyper-V long before I would recommend Virtualbox

One other question: Why is the Factory build request for VB 5.1.2 only reached the "accepted" stage even though it was sent to OBS 10 days ago? Yes, I know that 5.0.18 is old, but 5.0.22 was sent to OBS on June 20, and 5.0.26 was sent on July 18. How can we ever get an updated version when nothing ever comes out the back end?

I believe Dominique and Jan are answering this in some detail, but from my casual glance it would appear that you neglected to ensure that the dependencies your requests required were fufillable in Tumbleweed. That makes them somewhat hard to accept. Regards, -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 5 August 2016 at 19:33, Larry Finger <Larry.Finger@lwfinger.net> wrote:

Actually, someone else submitted the build for 5.1.2. That said, the request for updating kbuild to a version >= svn 2808 has been in the accepted state for 14 days. I suppose that I could determine what is holding it up, but I am growing tired of this whole process. I guess this is what I get for trying to help.

Where? https://build.opensuse.org/package/requests/openSUSE:Factory/kbuild quite clearly shows no requests in any state for 12 months -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 05.08.2016 um 19:50 schrieb Larry Finger:

On 08/05/2016 12:37 PM, Richard Brown wrote:

...

On 5 August 2016 at 19:33, Larry Finger <Larry.Finger@lwfinger.net> wrote:

...

Actually, someone else submitted the build for 5.1.2. That said, the request for updating kbuild to a version >= svn 2808 has been in the accepted state for 14 days. I suppose that I could determine what is holding it up, but I am growing tired of this whole process. I guess this is what I get for trying to help.

Where?

https://build.opensuse.org/package/requests/openSUSE:Factory/kbuild quite clearly shows no requests in any state for 12 months

https://build.opensuse.org/request/show/412899

This is a request targeting the package kbuild in the project Virtualization. Someone should forward it to the real development project of kbuild: devel:tools:building After that, it can be submitted to Factory. Then, vbox can be accepted for Factory. Thomas -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 05.08.2016 um 20:53 schrieb Larry Finger:

On 08/05/2016 12:53 PM, Thomas Leineweber wrote:

...

Am 05.08.2016 um 19:50 schrieb Larry Finger:

...

On 08/05/2016 12:37 PM, Richard Brown wrote:

...

On 5 August 2016 at 19:33, Larry Finger <Larry.Finger@lwfinger.net> wrote:

...

Actually, someone else submitted the build for 5.1.2. That said, the request for updating kbuild to a version >= svn 2808 has been in the accepted state for 14 days. I suppose that I could determine what is holding it up, but I am growing tired of this whole process. I guess this is what I get for trying to help.

Where?

https://build.opensuse.org/package/requests/openSUSE:Factory/kbuild quite clearly shows no requests in any state for 12 months

https://build.opensuse.org/request/show/412899

This is a request targeting the package kbuild in the project Virtualization. Someone should forward it to the real development project of kbuild: devel:tools:building

After that, it can be submitted to Factory. Then, vbox can be accepted for Factory.

I do not have enough privilege to forward it to devel:tools:building; however, I have added a comment in that project. I hope it gets through to the maintainer.

did you really try? Let me do that for you (I have no special rights in OBS!): https://build.opensuse.org/package/show/devel:tools:building/kbuild and press "Submit package". Fill in the form and press "OK". Et voila: https://build.opensuse.org/request/show/417161 I thought you know the drill..... Thomas -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Fr, Aug 05 2016, Larry Finger <Larry.Finger@lwfinger.net> wrote:

No, I do not. I have "learned" a few steps needed to branch a package, build it locally, commit it, and issue a service request. Anything more than that is a black box.

When you want to update a package in Factory you should start with branching from oS:Factory, which will redirect you to the associated devel project. Virtualization/kbuild is just a link to it, and request 412899 should really have been rejected in the first place. Andreas. -- Andreas Schwab, SUSE Labs, schwab@suse.de GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 "And now for something completely different." -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 5 August 2016 at 19:50, Larry Finger <Larry.Finger@lwfinger.net> wrote:

On 08/05/2016 12:37 PM, Richard Brown wrote:

...

On 5 August 2016 at 19:33, Larry Finger <Larry.Finger@lwfinger.net> wrote:

...

Actually, someone else submitted the build for 5.1.2. That said, the request for updating kbuild to a version >= svn 2808 has been in the accepted state for 14 days. I suppose that I could determine what is holding it up, but I am growing tired of this whole process. I guess this is what I get for trying to help.

Where?

https://build.opensuse.org/package/requests/openSUSE:Factory/kbuild quite clearly shows no requests in any state for 12 months

https://build.opensuse.org/request/show/412899

Virtualization:Kbuild is not the Devel Project for kbuild in Factory https://build.opensuse.org/package/show/openSUSE:Factory/kbuild Developed at devel:tools:building https://en.opensuse.org/openSUSE:Factory_development_model#Devel_Projects Any attempt to submit kbuild from Virtualization to openSUSE:Factory would be automatically rejected: https://en.opensuse.org/openSUSE:Factory_submissions -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

05.08.2016 20:25, Richard Brown пишет:

So I would recommend Hyper-V long before I would recommend Virtualbox

Please explain how I am supposed to use Hyper-V on Windows 7 host. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Richard Brown composed on 2016-08-05 18:48 (UTC+0200):

Now I've said 'We told you so', I do hope this thread finds suitable workarounds for users who are affected, and if not, I'd strongly encourage the use of nouveau Nvidia drivers instead of the proprietary drivers, and the use of KVM instead of Virtualbox.

Do note that nouveau is not the only NVidia driver alternative. FOSS video driver development focus has been transitioning from chip-specific to generic[1]. That means instead of much of the effort that has in the past been going into intel, nouveau or ati/amd/radeon drivers, has been redirected into the modesetting driver. To use modesetting, the easiest way is simply to uninstall the chip-specific xf86- or proprietary driver. In Tumbleweed back in March Egbert Eich provided us with a new driver config technique using /etc/X11/xorg_pci_ids/*.ids. Docs about it I've yet to locate, but here are a few references: https://bugzilla.opensuse.org/show_bug.cgi?id=972126 https://build.opensuse.org/request/show/377691 https://build.opensuse.org/request/show/379539 [1] https://www.reddit.com/r/archlinux/comments/4cojj9/it_is_probably_time_to_di... -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Tobias Klausmann composed on 2016-08-06 01:18 (UTC+0200):

Felix Miata wrote:

...

Do note that nouveau is not the only NVidia driver alternative. FOSS video driver development focus has been transitioning from chip-specific to generic[1]. That means instead of much of the effort that has in the past been going into intel, nouveau or ati/amd/radeon drivers, has been redirected into the modesetting driver.

...

To use modesetting, the easiest way is simply to uninstall the chip-specific xf86- or proprietary driver. In Tumbleweed back in March Egbert Eich provided us with a new driver config technique using /etc/X11/xorg_pci_ids/*.ids. Docs about it I've yet to locate, but here are a few references:

...

https://bugzilla.opensuse.org/show_bug.cgi?id=972126 https://build.opensuse.org/request/show/377691 https://build.opensuse.org/request/show/379539

...

[1] https://www.reddit.com/r/archlinux/comments/4cojj9/it_is_probably_time_to_di...

Sorry to say but you took this a bit wrong. You have several parts within the graphics stack, one part is the X servers driver(s) *xf86-* which can mostly be replaced by the generic xf86-video-modesetting. But that is only a small part of the picture, you have libdrm (intel, nouveau ,...), Mesa with its OpenGL implementation for the particular hardware you use and finally the kernel module. If you are going to use the modesetting driver, you will still end up with the Mesa implementation and the kernel module for your hardware (if you are not using the nvidia closed source driver or some other closed source driver).

The modesetting driver heavily depends on the OpenGL implementation for your hardware to accelerated 2D content (glamor), so if you have a bug there, you will face problems even by only using a desktop.

I realize there are several components to video for Xorg, but non-developer types can't be expected to keep track of which component does what or depends on what. From where I sit, X either starts or not, and crashes or not, and produces corruption or not. If all are nots, I'm good. If something isn't a not, and not a not is a problem for me, I boot something else, to see whether either a different installation on the same hardware, or on different hardware, matches the problem observed, or not. If I think my observation can be useful to other users or X-related devs, and don't find it to be a reported problem anywhere, I consider to either mention it in an appropriate forum, and/or file a bug. When I find a problem in X, usually I find it's not an openSUSE exclusive, so no one here hears from me about it before upstream does, if ever. :-)

On the long run the generic driver is the way to go, but personally i'm not certain its already the time to throw away the xf86-video-* drivers,

I don't see any material inconsistency between your last paragraph and what I wrote. The main thrust of my response was that NVidia gfxchip TW users are not stuck with nouveau as their only fallback when a kernel upgrade breaks their proprietary video driver and they have good reason to want to use that freshly installed kernel. Everything else was tangential. I have a bunch of installations using the server-built-in modesetting driver instead of the chip-specific drivers, and not demonstrating any observable negative impact in comparison. In fact, I have nearly every installation with a gfxchip supported by the modesetting driver configured to use the modesetting driver. That this works may be because I don't like bling and do everything I know how to avoid its use. I have no idea whether or when I am or could be looking at "3D" something or other. All _I_ see on _my_ screens are two dimensional objects. :-) -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andrei Borzenkov composed on 2016-08-06 12:18 (UTC+0300):

Felix Miata composed:

...

I don't see any material inconsistency between your last paragraph and what I wrote. The main thrust of my response was that NVidia gfxchip TW users are not stuck with nouveau

modesetting X11 driver requires nouveau, as nouveau is the only KMS driver available.

My customary procedure to enable use of the modesetting driver in 42.1 systems with NVidia gfxcards is 'zypper rm xf86-video-nouveau'. In TW, sometimes I do it the same way, and other times, use the /etc/X11/xorg_pci_ids/*.ids method provided by Egbert, avoiding its use via specification in xorg.con*. The result is Xorg.0.logs loaded with lines including the string "modeset(0)", and no lines that include the string "nouveau(0)". -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/

On vendredi, 5 août 2016 18.48:11 h CEST Richard Brown wrote:

Top posting because this is your obligatory reminder that a reliance of 3rd party kernel drivers is not recommended for Tumbleweed users

At least, not for those who can't handle those breaking deal. Once you know that you firmly depend on one or the other, you can play with zypper lock until the dust settle down.

This is clearly documented on the Tumbleweed wiki - https://en.opensuse.org/Portal:Tumbleweed

The reason for this recommendation is precisely because of issues such as these - they happen and will continue to happen as long as the 3rd parties in question (Nvidia and Oracle/VirtualBox) do not keep up with the pace of Kernel development

Now I've said 'We told you so', I do hope this thread finds suitable workarounds for users who are affected, and if not, I'd strongly encourage the use of nouveau Nvidia drivers instead of the proprietary drivers, and the use of KVM instead of Virtualbox.

Cheers,

Rich

It is a good advise, but when you have a non "mass saled low end gpu" or need for professional reason truly 3D acceleration, nouveau is not enough. It was working great enough with my 4.5 year old quadro K2000 gpu (it has took time to get a good support, but it come), but nouveau is absolutely not usable on my new 6 month old quadro M2000M beside having 7 consoles. It need resources and time to run nouveau debug, improving the stack and so, the most annoying being the fact you have to debug on the hardware you would like to use for work. So be it «finger middle» nvidia driver ;-) About KVM/libvirt (which I'm a big fan from the beginning 7 years ago) I also try to use it most of the time (even running Windows from XP to 10 and Server edition with the virtio drivers, if you don't need 3D acceleartion). But there's also some reason to use virtualbox. It's when you need to prepare Virtual Machines and have to deliver them on various plateforms, where virtualbox is available. Perhaps I've miss it, but as I know, there's no way to run a qemu/kvm libvirt machine easily under Windows or OSX, isn't it ? As I actually dependent for VB I will have to check how to help Larry :-) -- Bruno Friedmann Ioda-Net Sàrl www.ioda-net.ch Bareos Partner, openSUSE Member, fsfe fellowship GPG KEY : D5C9B751C4653227 irc: tigerfoot -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 05.08.2016 um 18:48 schrieb Richard Brown:

Top posting because this is your obligatory reminder that a reliance of 3rd party kernel drivers is not recommended for Tumbleweed users

Oh, Richard being helpful as always...

Now I've said 'We told you so', I do hope this thread finds suitable workarounds for users who are affected, and if not, I'd strongly encourage the use of nouveau Nvidia drivers instead of the proprietary drivers, and the use of KVM instead of Virtualbox.

On the next event where we meet I'll watch with astonishment how you start KVM on my Windows Notebook. Larry: thanks for your work on keeping virtualbox updated. It's really appreciated. Best regards, Stefan -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Hi Dominique, On 08.08.2016 09:09, Dominique Leuenberger / DimStar wrote:

On Sun, 2016-08-07 at 22:08 +0200, Stefan Seyfried wrote:

...

On the next event where we meet I'll watch with astonishment how you start KVM on my Windows Notebook. [...] https://qemu.weilnetz.de/w64/

He might not have the worst chances winning that challenge - his bigger problem is likely walking around with a Windows host to start with

"on my Windows Notebook", see above. But of course i should have stated the obvious: a working, usable solution is required. Which Virtualbox on Windows hosts definitely is. And qemu (when I tried it last time, maybe a year ago), certainly is not. It's fun, but not if you need it to do your work. And -- because of missing acceleration -- it's terribly slow, it's not KVM. But you probably knew all that. What I'm really objecting to is Richard always^H^H^H^H^H^Hoften trying to scare away contributors and users by making it look like they are doing "evil" or "bad" things, and that they are stupid even assuming that anything will work at all if they are not doing "right" -- as defined by him. This was the case with his "addon repositories are bad" crusade and I see this repeatedly with "nvidia -- don't use it!" -- "virtualbox -- you are doing it wrong!". (Fortunately for the "addon repos" crusade, it was shown quite clearly at OSC2016 that Richard's opinion is the opinion of a small minority.) While I 100% agree about the technical problems that externally maintained kernel drivers bring into the equation, it is almost never helpful to tell people "just buy other hardware" (the NVidia case, alternatively "just don't use any features of your hardware") or "just don't use your software of choice" (without telling them what to use instead! or without even considering if there is an alternative). I know I have violated this rule in the past, because of the fun a well crafted flamewar simply is, but today I mostly try to abide to the motto "If you don't have anything helpful to say, then stay quiet". Yes, new users should be warned about the problems that 3rd party drivers can bring. But in this thread, where Larry warns about a currently present problem, an answer like "I told you so that you were going to suffer for the blasphemy you are committing" -- sorry, the only result of this i can imagine (and I hope it will not happen) is, that Larry sooner or later will choose to no longer do the -- obviously frivolous -- task of maintaining free software (virtualbox is GPL after all) for openSUSE. Best regards, Stefan -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On Mon, 2016-08-08 at 09:37 +0200, Stefan Seyfried wrote:

Hi Dominique,

On 08.08.2016 09:09, Dominique Leuenberger / DimStar wrote:

...

On Sun, 2016-08-07 at 22:08 +0200, Stefan Seyfried wrote:

...

On the next event where we meet I'll watch with astonishment how you start KVM on my Windows Notebook. [...]

https://qemu.weilnetz.de/w64/

He might not have the worst chances winning that challenge - his bigger problem is likely walking around with a Windows host to start with

"on my Windows Notebook", see above. But of course i should have stated the obvious: a working, usable solution is required.

Which Virtualbox on Windows hosts definitely is. And qemu (when I tried it last time, maybe a year ago), certainly is not. It's fun, but not if you need it to do your work. And -- because of missing acceleration -- it's terribly slow, it's not KVM. But you probably knew all that.

What I'm really objecting to is Richard always^H^H^H^H^H^Hoften trying to scare away contributors and users by making it look like they are doing "evil" or "bad" things, and that they are stupid even assuming that anything will work at all if they are not doing "right" -- as defined by him.

This was the case with his "addon repositories are bad" crusade and I see this repeatedly with "nvidia -- don't use it!" -- "virtualbox -- you are doing it wrong!". (Fortunately for the "addon repos" crusade, it was shown quite clearly at OSC2016 that Richard's opinion is the opinion of a small minority.) "shown quite clearly" - care to elaborate?

M -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBCAAGBQJXqGoBAAoJEFskXWf93mhXw9QQAIKdikp/AnNVEosvSJRUuxKi tD2J31nTgrPNWYULyKbeUtTZZ1GLxEpG+Zia6bWtjINrc0qaiR2BwDobGlWUPh6O e8ZHbIAvRfTxowBD/ti2nSMtLHsqGCDKtzPN7I5tQIgL+ZGGA7lPfJXgVIpzPN1p iLp/JptCLCbtflNLWpgSLaLF8bwqjKFPaVtCoa3oN1NJtgr0+i+XTtdFmMIDX0n0 BgHnMPiIwMmKv3c/Z0zWbFKos+vLE0SZvMzX3c/pr0mifu2iaBo7injUewVqj3K6 1Gk7NzwH8nlxfxI6JjOuFj7cq9XNuosv7hsQsB69JvVNFTs/j3gs5Je9xCOMHMbx FvuF89yqJRPwlX8oqSxjRcUf2v55RnyduYySrlU8uCqrV390NfBrCgIf5TkgGktj 1GGpSrS08iveN4HXV3B5YzRPDD9AFEtF1GVldlhfqX5Cq7QHLwO8BlaZBiwMyMiW XONEEwNjQlJRsi4ZQQK7vggHHppCI6wKt7DOT8pfDO9LF/ec2eMtRcrUL/GEJ1y5 ZkEIX3OqE1YnZExvk9Bd+bT9OyHgYRFS8abOcKHFEy8QiCMtBGnaoKd7ps6BxjlH Q7oBY0vvQerFd0YjN8rOyPJIWBlSdiTZ9+m76ggC/fjrhgRWZm0H8R2E6Q7mFh3d n1KwbifjyZZ9LJhnnA5e =zFAh -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On Mon, 2016-08-08 at 14:45 +0200, Stefan Seyfried wrote:

The discussion in the large room was quite lively and -- at least that's what I have taken from it -- the "addon repos are the root of all evil" opinion was not shared by many, while the more liberal "rather have some

I recall some people objecting to need to limit usage of addon repositores, however far from majority (of both present people and contributors).

well maintained addon repository with maintainers doing their job but not willing to bother submitting everything to factory than not having those packages at all" was widely accepted. Again, few voices - here I would ask you if you were in same room but I actually recall you stating something about "humilliation associated with review on submission to Factory" - so you were there, perhaps you just decided to hear what you wanted to hear. (At least by the people I talked to, but then it is of course possible that I surrounded myself only with the people that share my opinion :-)) There seems to be certain potential for misunderstanding, someone speaking to you and politely nodding to your arguments does not necessary share your views - maybe he is just being polite and/or is not interested in having argument with you.

Cheers Martin P.S I would like to state my full support for elimination of addon repositories and their supporters -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBCAAGBQJXqIPAAAoJEFskXWf93mhXotoP/1Y3ubhhW0WfYY3Gvk2S71GV aMEy7PI1MJAicCkNoLJKoFo86hkeC0xFywh8+/TKlyHGL9UPbB343UAAAJjc+bxz lwf6+6iLZG6J99ZMU36uwauxrtJpQhlYWUwe6P6khvTd2BRwxYZDyky/0o5BhHEr pUmJ87aIvvHcjZT1Vr9Qh4rH4CKuC75e0IZwzBdE3/PaKbBZEPOSLKSR/pcy8/1K ZS/AXfEfaAWisfoCLB0JOuVsx7IAaQp/aFP+UJAIIkANzd0wSdI/Ix24pZ6V/Awn c/Nkoutq4yNi6uNB/7N5BuWVgJdPP16iQQdexcOPIoZ05ix+cPpR4h4pi18wMj/G 7axQgmrVw7VYKaixUi5HF63Qtq9ahVYXFDUa59TlXvtcgRcGyNQ7zJmHjuLLXoDo piqcVN2VRaU+cQ3TqI5MbCJjEhQy5oL4tC2elmhl5z43DQmV03oG6JrF/OX/6bHZ mKu5Fjg/bK1YbLs0pWN6jr3txt4J6U/ni9NQu3GzHL1cOfWoRoE/z8v/EdodV00l MEn00aFk0mFcoG4JdizUB///rNRttzlpEgfrKR5/ZodQKHltaE+/0Fx5pxiESRMo 7kZHdzdbr+vYQ4VjhJ+yNw+el0KuHy6kWPE0pUQKlEHgPzSZHOI38/M5beai8Fgn rz5QvbauxXjhTvz0xCoE =1sJ+ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On Mon, 2016-08-08 at 09:10 -0400, Patrick Shanahan wrote:

* Martin Pluskal <martin@pluskal.org> [08-08-16 09:07]:

...

Cheers

Martin

P.S I would like to state my full support for elimination of addon repositories and their supporters

ie: you then advocate the demise of openSUSE ???

Of course not - I love openSUSE M -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBCAAGBQJXqIuoAAoJEFskXWf93mhXt3MP/iEAK0iT0oEDjsuqDzivMcZb QdqPZkOX4CK4QjY2aoZI2Rf0wiyepkmOMmiXCjXKHTEydmpZ7/cJ/nGIPOe4cJl4 XYh1VCgabl/Typ36rkPsUDJn1ElEFvzpD+jyT9sNK5RGYprCIqvBEvItWMtAeIt6 7FmVmLurPEvLqCfUdXxgj0gN7e1BSR/Z6ZnivKYd9RL+EGJr55cSHXquG5R6Iy9V FBp6SJzU8eMY0CXZOk63go6YIBmUtnfD9gvcGKpKq0Xm6oNAniJig2DbfANbmDdU nS3RPlbUmAg/Qw8hGX1aL4KTb38PyaRUXbsctjKwijxBTC3CxZqrOxuYTWENsyhb /CPwZ1NobmiUxnTH5He+CfoxsXOsbJyMKEQcTnXBE+Wnt3U05Svb5MKQvbkVoOYV JRCtXYxaQeWwDHPd8fGOdCHvrPVLNf4a0LryQ3AeXXPeMXjNE/Q2GxnNuk4hy6lK OwTO416ep3X/CahG6ZwFFwlshltaODH7kDI+MyK/G573CuYC3QbcGw6UrSLklQF3 ZwIpMrZqympZgL9DQ4PxPJkAidCpKc6TTiYh19Uk60nfcxF45X5IJ1FPqepqhdCu pw2CQL4kxN28ynM9VFqlqTheOuSSkmzepNfGuwpPXqX+2IxmU/UkmLomX8Vazk4l vTmWhIBIFQcz6D9PIkoG =bg+4 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On Mon, 2016-08-08 at 15:10 +0200, Stephan Kulow wrote:

On 08.08.2016 15:06, Martin Pluskal wrote:

...

On Mon, 2016-08-08 at 14:45 +0200, Stefan Seyfried wrote:

P.S I would like to state my full support for elimination of addon repositories and their supporters

You want to eliminate people? I think you're violating quite some openSUSE rules here.

1) I just support it 2) It was of course joke of ill taste (as most of my jokes are) M -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBCAAGBQJXqIsrAAoJEFskXWf93mhX2h4P/0m+WXD/fpWVF0GDNthzyFHG hQET04hdUdY/ECX/tdZT27Xc52WFb2HBTOPDcKSlb7qY4lIbIBFnRALSGJ66SI0k Qq9dBGeb8T/isbVGtzm4qRXnlLQuPmssxJPV1bsMbJmCnYGbrrSTMTIQNBbtdV3O CXumagmG6sxzfFUpXebCgfytv5teHneo19moOTgST5pg+nUG3apKo9C23EKV2TXf uPe2bSmtyMXW9TDhV1h0G0BSr1oHH45fBWaQ3ZKkSEqok8PioxSd/P/bKWUBzd5c c4iqy5d1NmKgeiBetjQZmCW9Bb7A/J3j9bULol7efpGOlRR6vLRGhuIY2S5ZOW0y Y9Ck0n5fL3IMrixMFL8mK7d533YgkB4bvRd6oQDvmSxX+FcF16RrjCu91mZNb5Kt pozbzfbdKaSHmhyeIrMva9OPH9v/xG+rbwNSz/WanD4+7HjOdJh6RqYWncbMiu0c qThaZ7MHX5BrBL2Sbqvz4a1VWZMkkeOdcoYBrXw4mRjMfWSaNBUGBJ/JTFama+5q qulaS9+WbxtUgVxL1pVkdT9NqcRGAYFN+zfpFcJyvCAr2YoaPgADGYmILvzZARWN kmXLyC5c3oUY2OtqpCFAowJcLnaAkangIaAABf8l4Co4uRLka/+C3HwC09GfNeqF dyUfuxNZ9N0Yb0kWvQyF =noXJ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

2016-08-05 18:48 GMT+02:00 Richard Brown:

Top posting because this is your obligatory reminder that a reliance of 3rd party kernel drivers is not recommended for Tumbleweed users

I've considered a long time whether to comment this because this half an off-topic. At the end I decided express my support for Larry. Excuses for this spam to people helping to solve the real problems here.

This is clearly documented on the Tumbleweed wiki - https://en.opensuse.org/Portal:Tumbleweed

True if you want to use OpenSUSE in a student laboratory. Misleading in the real world. This is a typical viewpoint of a support department - We told you, so we'll never support you.

The reason for this recommendation is precisely because of issues such as these - they happen and will continue to happen as long as the 3rd parties in question (Nvidia and Oracle/VirtualBox) do not keep up with the pace of Kernel development

Now I've said 'We told you so', I do hope this thread finds suitable workarounds for users who are affected, and if not, I'd strongly encourage the use of nouveau Nvidia drivers instead of the proprietary drivers, and the use of KVM instead of Virtualbox.

Especially organizations have internal policies how to distribute images to their members. You often can't change this, because there are reasons beyond your personal view (multi-platform, usability). The support departments I got in touch with formerly distributed VMWare and now VirtualBox images (Windows and SLES as guests). I can immediately use them in Tumbleweed, and to be honest, VirtualBox for Tumbleweed made by Larry has never disappointed me. the emulation of all OS provides was running well and performantly, including graphics and sound. --> Virtualbox is a great standalone virtualization solution, immediately usable and intuitive, even without a manual. Regarding the sidekick to NVidia - it seems you haven't been a long-term NVidia graphics user. I still don't consider performance, just the wish not to crash after minutes or hours of work. --> NVidia proprietary drivers do not optimally integrate with the kernel and Xorg, but they don't hang up or crash in most cases. The "nouveau" driver is a neverending construction area introducing breaking code changes to each new kernel mainline, otherwise I'd never use the proprietary one. Thumbs up to Larry and all maintainers. This time especially to those delivering packages with a different license than GPL :-) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/05/2016 12:06 PM, Jan Engelhardt wrote:

On Friday 2016-08-05 18:04, Larry Finger wrote:

...

I have been trying to get later versions of VirtualBox to build in OBS for some time; however, they fail because the i586 build runs out of memory even though it has been allowed to have as much as 20 GB.

20 GB don't help if a single process is limited by 4G. Code authors are also not inculpable. Gazillions of C++ headers, huge TUs, and passing 1000 .o files into the linker at once. And then LTO maybe. :(

The part I do not understand is that the local build is fine. Only the OBS build running under KVM crashes. Something seems to be wrong there. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/06/2016 08:55 AM, Bruno Friedmann wrote:

On vendredi, 5 août 2016 12.17:05 h CEST Larry Finger wrote:

...

On 08/05/2016 12:06 PM, Jan Engelhardt wrote:

...

On Friday 2016-08-05 18:04, Larry Finger wrote:

...

I have been trying to get later versions of VirtualBox to build in OBS for some time; however, they fail because the i586 build runs out of memory even though it has been allowed to have as much as 20 GB.

20 GB don't help if a single process is limited by 4G. Code authors are also not inculpable. Gazillions of C++ headers, huge TUs, and passing 1000 .o files into the linker at once. And then LTO maybe. :(

The part I do not understand is that the local build is fine. Only the OBS build running under KVM crashes. Something seems to be wrong there.

Larry it should be quite easy to build with kvm locally too.

I would like to try to run them, as my osc has double setup (kvm and chroot) Now I would like to make it useful, and I confess I've been lost during the thread about which version we're talking about.

Would you mind to confirm me the link of ?

To set up a local build, do the following: osc branch Virtualization virtualbox That will create a project for you and will suggest the command to check out that project. Use that command to get the sources. Then cd to the directory with those sources, and then 'osc build' to build the x86_64 version. To try the 32-bit version, you will need to edit file virtualbox.spec. At line 163, you will find a line containing "ExclusiveArch" that has been commented out. Remove the # and add it to the next line. You can then use the command "osc build i586" to build the 32-bit version. I have just switched my local builds to use KVM and I will see how that goes. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On samedi, 6 août 2016 19.24:47 h CEST Bruno Friedmann wrote:

On samedi, 6 août 2016 12.10:17 h CEST Larry Finger wrote:

...

On 08/06/2016 08:55 AM, Bruno Friedmann wrote:

...

On vendredi, 5 août 2016 12.17:05 h CEST Larry Finger wrote:

...

On 08/05/2016 12:06 PM, Jan Engelhardt wrote:

...

On Friday 2016-08-05 18:04, Larry Finger wrote:

...

I have been trying to get later versions of VirtualBox to build in OBS for some time; however, they fail because the i586 build runs out of memory even though it has been allowed to have as much as 20 GB.

20 GB don't help if a single process is limited by 4G. Code authors are also not inculpable. Gazillions of C++ headers, huge TUs, and passing 1000 .o files into the linker at once. And then LTO maybe.

:(

The part I do not understand is that the local build is fine. Only the OBS build running under KVM crashes. Something seems to be wrong there.

Larry it should be quite easy to build with kvm locally too.

I would like to try to run them, as my osc has double setup (kvm and chroot) Now I would like to make it useful, and I confess I've been lost during the thread about which version we're talking about.

Would you mind to confirm me the link of ?

To set up a local build, do the following:

osc branch Virtualization virtualbox

That will create a project for you and will suggest the command to check out that project. Use that command to get the sources. Then cd to the directory with those sources, and then 'osc build' to build the x86_64 version.

To try the 32-bit version, you will need to edit file virtualbox.spec. At line 163, you will find a line containing "ExclusiveArch" that has been commented out. Remove the # and add it to the next line. You can then use the command "osc build i586" to build the 32-bit version.

I have just switched my local builds to use KVM and I will see how that goes.

Larry

I've co Virtualization virtualbox and hack a bit to unexclude i586 and then yeap make the changes in .spec to be able to build it.

First try on 8 cores has crashed with errors ... so to have a trace readable I'm rebuilding with -j1

To not overload the mailing list, did you have already a boo# for that failure ?

Ok so from what I read a trouble with pae and bad code :-) http://susepaste.org/11865719 We need patches or fixes from upstream :-o -- Bruno Friedmann Ioda-Net Sàrl www.ioda-net.ch Bareos Partner, openSUSE Member, fsfe fellowship GPG KEY : D5C9B751C4653227 irc: tigerfoot -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On samedi, 6 août 2016 16.11:52 h CEST Larry Finger wrote:

On 08/06/2016 02:59 PM, Bruno Friedmann wrote:

...

Ok so from what I read a trouble with pae and bad code :-)

http://susepaste.org/11865719

We need patches or fixes from upstream :-o

I'm still trying to get the memory and disk requirements correct, and have not yet reached that point.

We will likely need to patch the VB source to fix this problem.

Larry

The non building is related to the removed cpu_has_pge in 4.7.0 Commit https://lkml.org/lkml/2016/3/29/582 And the upstream fix is https://www.virtualbox.org/changeset/63028/vbox Was certainly there in the previous 4.7 kernel path removed a bit too quickly in the 5.1.2 submission. So i686 has build locally and seems to have build correctly. I'm making a branch and see how we can get it build everytime. -- Bruno Friedmann Ioda-Net Sàrl www.ioda-net.ch Bareos Partner, openSUSE Member, fsfe fellowship GPG KEY : D5C9B751C4653227 irc: tigerfoot

07.08.2016 20:22, Larry Finger пишет:

On 08/07/2016 05:06 AM, Bruno Friedmann wrote:

...

The non building is related to the removed cpu_has_pge in 4.7.0

Commit https://lkml.org/lkml/2016/3/29/582

And the upstream fix is https://www.virtualbox.org/changeset/63028/vbox Was certainly there in the previous 4.7 kernel path removed a bit too quickly in the 5.1.2 submission.

So i686 has build locally and seems to have build correctly. I'm making a branch and see how we can get it build everytime.

Thanks for the patch. With it, the i586 build works correctly, and x86_64 also builds using chmod. With the KVM builds, I'm still trying to get builds of both flavors to succeed. I just got a failure message for Virtualization/virtualbox in Kernel_stable_standard/x86_64 that says:

[ 906s] [ 906s] cc1plus: out of memory allocating 7785984 bytes after a total of 160129024 bytes [ 908s] kmk: *** [/home/abuild/rpmbuild/BUILD/VirtualBox-5.1.2/out/linux.amd64/release/obj/vboxsoap/gen/webservice/soapC-3.o] Error 1

Why should a build need more than 160 GB?

MB

Can any OBS experts answer that?

Larry

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Thu, 2016-08-04 at 16:16 -0400, Patrick Shanahan wrote:

* Larry Finger <Larry.Finger@lwfinger.net> [08-04-16 12:56]:

...

If you are running TW in a VirtualBox Virtual Machine, DO NOT install this snapshot!!!!! There is a bug that breaks xboxvideo for both KDE and Gnome. Not only does the new 4.7.0-1 kernel fail to boot to X, but something has broken X even when using older kernels.

The specifics of the problem are unknown, but you have been warned.

NVidia prop installer also fails with kernel 4.7.xxx

Please people: can you chose a SUBJECT in your emails that makes ANY sense in order to browse archives at any moment in time? Thank you Dominique