Nvidia binary driver FTBS due to DKMS layer violation

[Stefan Bader]

Filing this against the 340-updates version but possibly the same applies to older versions, too. The nvidia source package produces two individual dkms packages: nvidia-340-updates, nvidia-340-updates-uvm. The problem is that the DKMS build of the nvidia-uvm module runs compile steps inside the nvidia modules build directory. This is violating the DKMS assumption that each module can be build independently (there is no way of describing cross-modules dependencies and even more important, the autoinstall step after a new kernel is installed will run the modules build in parallel).

Since nvidia and nvidia-uvm are very dependent on each other the right course of action seems to be to combine both sources in one DKMS module that produces two kernel modules (this is supported by DKMS). For the transition this resulting dkms package needs to have a breaks/replaces for the nvidia-uvm package.

[dino99 (9d9)]

That's all the mess met by ignoring that issue
https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-346-updates/+bug/1268257

[Mateusz Stachowski (stachowski-mateusz)]

This bug most likely affects the 331 drivers too because they also have the separate uvm module.

Also isn't this bug the source of Bug #1268257 which is simply enormus and currently affects 2376 users and I'm also affected by it. In Vivid I had multiple failed installs of nvidia modules for new kernels and the bug report always led me to Bug #1268257.

The drivers that make problems have or in the case of 346 had a separate -uvm package.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Launchpad Janitor (janitor)]

This bug was fixed in the package nvidia-graphics-drivers-340 - 340.76-0ubuntu2

---------------
nvidia-graphics-drivers-340 (340.76-0ubuntu2) vivid; urgency=medium

  * debian/rules,
    debian/templates/control.in,
    debian/templates/dkms_nvidia_uvm.conf.in,
    debian/templates/nvidia-graphics-drivers-uvm.dirs.in,
    debian/templates/nvidia-graphics-drivers-uvm.install.in,
    debian/templates/nvidia-graphics-drivers-uvm.postinst.in,
    debian/templates/nvidia-graphics-drivers-uvm.prerm.in:
    - Build nvidia and nvidia-uvm together, as the latter
      build depends on the former (LP: #1431753).
    - Merge the -uvm package into the main nvidia driver
      package, and provide a transitional package for -uvm.
    - Make sure to remove all upstream dkms.conf files.
 -- Alberto Milone <email address hidden> Mon, 30 Mar 2015 17:09:32 +0200

[Launchpad Janitor (janitor)]

This bug was fixed in the package nvidia-graphics-drivers-340-updates - 340.76-0ubuntu2

---------------
nvidia-graphics-drivers-340-updates (340.76-0ubuntu2) vivid; urgency=medium

  * debian/rules,
    debian/templates/control.in,
    debian/templates/dkms_nvidia_uvm.conf.in,
    debian/templates/nvidia-graphics-drivers-uvm.dirs.in,
    debian/templates/nvidia-graphics-drivers-uvm.install.in,
    debian/templates/nvidia-graphics-drivers-uvm.postinst.in,
    debian/templates/nvidia-graphics-drivers-uvm.prerm.in:
    - Build nvidia and nvidia-uvm together, as the latter
      build depends on the former (LP: #1431753).
    - Merge the -uvm package into the main nvidia driver
      package, and provide a transitional package for -uvm.
    - Make sure to remove all upstream dkms.conf files.
 -- Alberto Milone <email address hidden> Mon, 30 Mar 2015 17:14:52 +0200

[KennoVO (kenno-xs4all)]

Sorry, but 331 is still the most recent driver supported by the LTS release. While "LTS" doesn't seem to mean much to Canonical anymore, at least try to keep up the appearance. Especially for a bug that affects thousands of users, not all of whom will be willing/skilled to upgrade their kernels (and then upgrade them again and again every time short-term support runs out).

[KennoVO (kenno-xs4all)]

Dear Alberto and others, Could you please please backport this fix to the 331 driver, which is the most recent one available for 14.04 LTS (Trusty)? Or at the very least explain why this might not be as trivial as it sounds? This bug impacts a massive amount of Trusty users, and failing to get a graphical interface every time the kernel is upgraded is quite high-profile; a lot of people are abandoning Ubuntu altogether for it. Our whole workplace may do so soon because one of the programs we use professionally requires the proprietary Nvidia driver and our support people cannot afford the time commitment to have everyone use the non-LTS branches.

I also feel I need to point out that the "users affected" and bug heat of this report ( lp:1431753 ) are misleadingly low. Most of the users who trigger this bug land at lp:1268257 . If you look at the stats of *that* bug and all the recent comments, the whole affair starts to look like a trainwreck. We tried to make lp:1268257 a duplicate of this bug ( lp:1431753 ), which would presumably bump the latter to the top of the priority list, but were prevented from doing so by a bug in Launchpad itself ( lp:1450251 ).

[Alberto Milone (albertomilone)]

I'm going to backport the fix together with the 346 release and with the 340 release in 14.04. Sorry for the inconvenience.

[KennoVO (kenno-xs4all)]

Alberto, thank you sooo much, that is great news. Especially since there seems to be another wave of bitter comments and confusion on lp:1268257 . Two questions:
- Am I understanding correctly that you'll fix the problem for the 331 driver AND that we'll have the option of using the 340 or 346 driver with the 3.13 kernel in 14.04 ? (For me personally, _either_ solution would work. Having newer drivers is always nice, but there might hypothetically be people with really old hardware who require the old driver.)
- Do you have any idea when either solution will be available to end-users?

[Alberto Milone (albertomilone)]

331 is the previous long term release. I plan to introduce 340 (the legacy driver) to migrate users from 331 (so that hardware support is the same), whereas 346 (the current long term release) will have to be installed manually (if your hardware is supported).

[KennoVO (kenno-xs4all)]

Nice! Will this work with the 3.13 kernel in 14.04 ? Also, do you have any idea when it will be available to end-users?

[KennoVO (kenno-xs4all)]

*bump*

[Nailer1887 (barry-titterton)]

This bug affects me for driver-331.

[teo1978 (teo8976)]

> We tried to make lp:1268257 a duplicate of this bug ( lp:1431753 ), which would presumably bump the latter to the top of the priority list, but were prevented from doing so by a bug in Launchpad itself ( lp:1450251 ).

I tried to edit duplicates of lp:1268257 to make duplicate of this bug ( lp:1431753 ), but after I succesfully managed to do it with a handful of them, it started to systematically time out.

It seems that lp:1450251 not only happens when the duplicate has a lot of duplicates, but also when the "target" has (and "a lot" is like around a dozen, it's really pathetic)

[teo1978 (teo8976)]

Ok I think I've found the way:
- pick each dupe of lp:1268257
- remove its duplicate status
- AFTER that, mark it as dupe of this one (1431753)

This way, the "fire" in this one will increase. I'll do a few until I get tired.

[Alberto Milone (albertomilone)]

I am already working on it (I've been busy with other work). There's no need to make this any more complicated than it needs to be.

[Pablo (pabloa98)]

Still not fixed drivers version 331? Sorry I am so picky. I guess now I understand LTS meaning: bug fixing is long term, very long. In this case: 2 years.

[Todd (dudemansomebody)]

hahah couldnt agree more @pabloa98

[[e-mail] (rbscycle)]

SO, since I don't know that much .....
On one of my computers (HP s3720f) my display disappeared completely. I
tried a bunch of things. The last thing I tried seem to work:

$ sudo apt-get --purge remove xserver-xorg-video-nouveau
$ sudo ubuntu-drivers devices
A list of drivers will appear. Pick a driver to install and...
sudo apt-get install nvidia-172 (for example)

Interestingly enough, this computer (HP s3720f) has not complained about
any updates since ... at least not yet!!!

On 15-06-17 09:49 AM, Todd wrote:
> hahah couldnt agree more @pabloa98
>

--

[Heath (heathnaylor)]

This seems to me like a trivial bug fix, is this just nVidia not wanting to devote time to a fix?

[xtsbdu3reyrbrmroezob (xtsbdu3reyrbrmroezob)]

No vendor cares about linux
On Thu, Jun 18, 2015 at 11:46 AM Heath <email address hidden> wrote:

> This seems to me like a trivial bug fix, is this just nVidia not wanting
> to devote time to a fix?
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1405694).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

[paulwilson05 (paulwilson05)]

I want zero downtime, top tier speed, and zero cost.

On Thu, Jun 18, 2015 at 2:09 PM, Kristian Erik Hermansen <
<email address hidden>> wrote:

> No vendor cares about linux
> On Thu, Jun 18, 2015 at 11:46 AM Heath <email address hidden> wrote:
>
> > This seems to me like a trivial bug fix, is this just nVidia not wanting
> > to devote time to a fix?
> >
> > --
> > You received this bug notification because you are subscribed to a
> > duplicate bug report (1405694).
> > https://bugs.launchpad.net/bugs/1431753
> >
> > Title:
> > Nvidia binary driver FTBS due to DKMS layer violation
> >
> > To manage notifications about this bug go to:
> >
> >
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
> >
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1399042).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

--
Thanks,
-Paul

[[e-mail] (rbscycle)]

That may be so. However, I care. I use Linux as my everyday computer
at home for everything. I use the "LTS" version because I expect less
problems, and don't really know a hell-of-a-lot about the "underneath"
workings of Linux. If this means I have to go to another distribution,
so be it.

On 15-06-18 03:09 PM, Kristian Erik Hermansen wrote:
> No vendor cares about linux
> On Thu, Jun 18, 2015 at 11:46 AM Heath <email address hidden> wrote:
>
>> This seems to me like a trivial bug fix, is this just nVidia not wanting
>> to devote time to a fix?
>>
>> --
>> You received this bug notification because you are subscribed to a
>> duplicate bug report (1405694).
>> https://bugs.launchpad.net/bugs/1431753
>>
>> Title:
>> Nvidia binary driver FTBS due to DKMS layer violation
>>
>> To manage notifications about this bug go to:
>>
>> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>>

--

[xtsbdu3reyrbrmroezob (xtsbdu3reyrbrmroezob)]

Just use Debian. No one considers Ubuntu secure any more after the privacy
violations embedded by Canonical.

On Thu, Jun 18, 2015 at 12:41 PM <email address hidden> <email address hidden>
wrote:

> That may be so. However, I care. I use Linux as my everyday computer
> at home for everything. I use the "LTS" version because I expect less
> problems, and don't really know a hell-of-a-lot about the "underneath"
> workings of Linux. If this means I have to go to another distribution,
> so be it.
>
>
> On 15-06-18 03:09 PM, Kristian Erik Hermansen wrote:
> > No vendor cares about linux
> > On Thu, Jun 18, 2015 at 11:46 AM Heath <email address hidden> wrote:
> >
> >> This seems to me like a trivial bug fix, is this just nVidia not wanting
> >> to devote time to a fix?
> >>
> >> --
> >> You received this bug notification because you are subscribed to a
> >> duplicate bug report (1405694).
> >> https://bugs.launchpad.net/bugs/1431753
> >>
> >> Title:
> >> Nvidia binary driver FTBS due to DKMS layer violation
> >>
> >> To manage notifications about this bug go to:
> >>
> >>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
> >>
>
> --
>
>
> ** Attachment added: "ronnie6.png"
>
> https://bugs.launchpad.net/bugs/1431753/+attachment/4416893/+files/ronnie6.png
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1405694).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

[Vadim Peretokin (vperetokin)]

Can you guys not litter the bug report?

On Fri, Jun 19, 2015 at 6:18 AM, Kristian Erik Hermansen <
<email address hidden>> wrote:

> Just use Debian. No one considers Ubuntu secure any more after the privacy
> violations embedded by Canonical.
>
> On Thu, Jun 18, 2015 at 12:41 PM <email address hidden> <email address hidden>
> wrote:
>
> > That may be so. However, I care. I use Linux as my everyday computer
> > at home for everything. I use the "LTS" version because I expect less
> > problems, and don't really know a hell-of-a-lot about the "underneath"
> > workings of Linux. If this means I have to go to another distribution,
> > so be it.
> >
> >
> > On 15-06-18 03:09 PM, Kristian Erik Hermansen wrote:
> > > No vendor cares about linux
> > > On Thu, Jun 18, 2015 at 11:46 AM Heath <email address hidden> wrote:
> > >
> > >> This seems to me like a trivial bug fix, is this just nVidia not
> wanting
> > >> to devote time to a fix?
> > >>
> > >> --
> > >> You received this bug notification because you are subscribed to a
> > >> duplicate bug report (1405694).
> > >> https://bugs.launchpad.net/bugs/1431753
> > >>
> > >> Title:
> > >> Nvidia binary driver FTBS due to DKMS layer violation
> > >>
> > >> To manage notifications about this bug go to:
> > >>
> > >>
> >
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
> > >>
> >
> > --
> >
> >
> > ** Attachment added: "ronnie6.png"
> >
> >
> https://bugs.launchpad.net/bugs/1431753/+attachment/4416893/+files/ronnie6.png
> >
> > --
> > You received this bug notification because you are subscribed to a
> > duplicate bug report (1405694).
> > https://bugs.launchpad.net/bugs/1431753
> >
> > Title:
> > Nvidia binary driver FTBS due to DKMS layer violation
> >
> > To manage notifications about this bug go to:
> >
> >
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
> >
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1412105).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

[Pablo (pabloa98)]

Could you fix it?

[xtsbdu3reyrbrmroezob (xtsbdu3reyrbrmroezob)]

Can't fix binaries without patching
On Thu, Jun 18, 2015 at 4:35 PM Pablo <email address hidden> wrote:

> Could you fix it?
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1405694).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

[Stefan Bader (smb)]

Alberto is rolling this back into older releases over time. To be fair it is an annoying issue but not one that is so fatal one would not be able to recover a system. In the majority of cases this ends up being just a false system issue notification. Because on kernel update the parallel builds are actually started 3 times. By now I just made it a habit of doing the following whenever kernel updates come in:

- run "dkms status" to see whether both nvidia modules appear for the latest kernel (bbswitch is always there)
- if one(or both) is missing, do "sudo dkms install -m <module> -v <module version> -k <kernel version>
   (the name and versions taken from the dkms status report)
- "sudo rm /var/crash/*nvidia*" to ignore the build failures before

When the fixed version of the driver hits this can be noticed by having only one module reported by dkms for nvidia. The *-uvm one will be gone and the module it produced will be part of the remaining dkms module which then produces two kernel modules.

[Simon Reed (xubuntu-o)]

"To be fair it is an annoying issue but not one that is so fatal one would not be able to recover a system."

It leaves the system not working. That's is more than just annoying. Also, I actually did re-install the operating system at least once because of this bug to try to get this PC working again. I actually spent about three days struggling with driver versions and trying to work out what the real problem was. I was also swopping out video cards to see if that helped. Any sane person would have gone out and bought a Windows licence and installed that instead.

"By now I just made it a habit of doing the following…"

Sadly, those instructions are the sort of reason people say Linux is not ready for the mainstream. How do I tell Auntie Margaret over the phone she just needs to "sudo dkms install -m <module> -v <module version> -k <kernel version>" ?

It might be trivial to an operating systems programmer with a computer science degree. But some of us want to USE Ubuntu-based operating systems for work purposes, not for working on.

I am sorry to add a whinge to this saga, but to imply it's just a nuisance and there's an easy workaround is not how it feels to some of the user community.

[josh (joshuaglennmurphy)]

Nima this really isn't the place for that talk. Will EVERYONE please keep
this thread on topic about this bug and only this bug please
On 19 Jun 2015 21:51, "Nima" <email address hidden> wrote:

> I want to go to another linux distribution ,
> Can any body give me a hint about it?
> CentOS or Opensuse? Which one is better?
> which one has a full resource of software like ,for example "ubuntu
> software center"?
> Thanks for any help in advance
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1412043).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> Status in nvidia-graphics-drivers-331 package in Ubuntu:
> Confirmed
> Status in nvidia-graphics-drivers-331-updates package in Ubuntu:
> Confirmed
> Status in nvidia-graphics-drivers-340 package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-340-updates package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-346 package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-346-updates package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-331 source package in Trusty:
> Triaged
> Status in nvidia-graphics-drivers-331-updates source package in Trusty:
> Triaged
> Status in nvidia-graphics-drivers-340 source package in Trusty:
> Confirmed
> Status in nvidia-graphics-drivers-340-updates source package in Trusty:
> Confirmed
> Status in nvidia-graphics-drivers-346 source package in Trusty:
> Confirmed
> Status in nvidia-graphics-drivers-346-updates source package in Trusty:
> Confirmed
>
> Bug description:
> Filing this against the 340-updates version but possibly the same
> applies to older versions, too. The nvidia source package produces two
> individual dkms packages: nvidia-340-updates, nvidia-340-updates-uvm.
> The problem is that the DKMS build of the nvidia-uvm module runs
> compile steps inside the nvidia modules build directory. This is
> violating the DKMS assumption that each module can be build
> independently (there is no way of describing cross-modules
> dependencies and even more important, the autoinstall step after a new
> kernel is installed will run the modules build in parallel).
>
> Since nvidia and nvidia-uvm are very dependent on each other the right
> course of action seems to be to combine both sources in one DKMS
> module that produces two kernel modules (this is supported by DKMS).
> For the transition this resulting dkms package needs to have a
> breaks/replaces for the nvidia-uvm package.
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

[Zewbie (zewbie)]

Try fedora 21 not 22

Sent from my Sprint Samsung Galaxy S® 6.-------- Original message --------
From: Nima <email address hidden>
Date: 06/19/2015 3:43 AM (GMT-07:00)
To: <email address hidden>
Subject: [Bug 1431753] Re: Nvidia binary driver FTBS due to DKMS layer
  violation

I want to go to another linux distribution ,
Can any body give me a hint about it?
CentOS or Opensuse? Which one is better?
which one has a full resource of software like ,for example "ubuntu software center"?
Thanks for any help in advance

--
You received this bug notification because you are subscribed to a
duplicate bug report (1382608).
https://bugs.launchpad.net/bugs/1431753

Title:
  Nvidia binary driver FTBS due to DKMS layer violation

Status in nvidia-graphics-drivers-331 package in Ubuntu:
  Confirmed
Status in nvidia-graphics-drivers-331-updates package in Ubuntu:
  Confirmed
Status in nvidia-graphics-drivers-340 package in Ubuntu:
  Fix Released
Status in nvidia-graphics-drivers-340-updates package in Ubuntu:
  Fix Released
Status in nvidia-graphics-drivers-346 package in Ubuntu:
  Fix Released
Status in nvidia-graphics-drivers-346-updates package in Ubuntu:
  Fix Released
Status in nvidia-graphics-drivers-331 source package in Trusty:
  Triaged
Status in nvidia-graphics-drivers-331-updates source package in Trusty:
  Triaged
Status in nvidia-graphics-drivers-340 source package in Trusty:
  Confirmed
Status in nvidia-graphics-drivers-340-updates source package in Trusty:
  Confirmed
Status in nvidia-graphics-drivers-346 source package in Trusty:
  Confirmed
Status in nvidia-graphics-drivers-346-updates source package in Trusty:
  Confirmed

Bug description:
  Filing this against the 340-updates version but possibly the same
  applies to older versions, too. The nvidia source package produces two
  individual dkms packages: nvidia-340-updates, nvidia-340-updates-uvm.
  The problem is that the DKMS build of the nvidia-uvm module runs
  compile steps inside the nvidia modules build directory. This is
  violating the DKMS assumption that each module can be build
  independently (there is no way of describing cross-modules
  dependencies and even more important, the autoinstall step after a new
  kernel is installed will run the modules build in parallel).

  Since nvidia and nvidia-uvm are very dependent on each other the right
  course of action seems to be to combine both sources in one DKMS
  module that produces two kernel modules (this is supported by DKMS).
  For the transition this resulting dkms package needs to have a
  breaks/replaces for the nvidia-uvm package.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions

[teo1978 (teo8976)]

> Nima this really isn't the place for that talk.

It has become the place since nobody has given a fuck about fixing this huge system-screwing bug for years (and even now, they insist in treating it as a minor annoyance).

[Camil Staps (f7-info)]

Zewbie, what is wrong with you? Did you not read #29 and #35? Do you have any idea how many people are getting an email when you comment on this bug? Have a little responsibility.

[xtsbdu3reyrbrmroezob (xtsbdu3reyrbrmroezob)]

OSX is the way to go for sure...sorry for the annoying email alert

On Fri, Jun 19, 2015 at 4:32 AM Camil Staps <email address hidden> wrote:

> Zewbie, what is wrong with you? Did you not read #29 and #35? Do you
> have any idea how many people are getting an email when you comment on
> this bug? Have a little responsibility.
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1405694).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

[teo1978 (teo8976)]

And I'm pretty scared of uninstalling the drivers without first disabling them, because I don't trust at all the package uninstallation process to handle things well.

[teo1978 (teo8976)]

Indeed, I think I installed Bumblebee, because I thought that it would handle automatic switching between intel and nvidia as needed, but no way! it turns out it just allows you (or is supposed to allow, because I never managed to do even that) to manually run stuff on NVidia. So, in practice I have (or used to have until this update) bumblebee, nvidia drivers, and the intel GPU actually used all the time with the nvidia card just sleeping and useless. And I don't care, as long as my system doesn't start freezing every time I unplug the external screen.
All was fine (except for the errors showing up at kernel updates), and now everything is broken again (i.e. nvidia drivers being used, and causing random freezes because they don't support Optimus - or at least that's the diagnosis I was given when I first had the problem a couple of years ago)

[Anna (jaruzanka)]

I have a smilar case as teo1978. - I have a dell laptop with Optimus technology with both Intel and Nvidia GPU.

Before the upgrade from 331 to 340 I never managed to get the Nvidia GPU running, so I just used Intel instead. After the upgrade, the Nvidia GPU was turned on by default but something was wrong with the configuration in the xorg.conf file. Because of this I got a black screen after rebooting.

The solution that I linked a few posts up actually has a thorough explanation of what happened. It shows how to fix the xorg.conf file and how to prevent ubuntu from changing it back after reboot (it was crucial for me).

After all that my laptop started to actually use the Nvidia GPU for both daily use and manually running simulation on GPU.

[teo1978 (teo8976)]

So, the issue I pointed out is definitely confirmed.

Please, whoever can, change back the status at least for 340-Tursty, because the fix is partially broken.

[teo1978 (teo8976)]

Or should we file a separate bug?

[dino99 (9d9)]

hey all users still complaining:

- be sure to have a clean nvidia driver installation: not a system where the driver receive upgrades but still with borked settings behind the scene
- so "sudo apt-get purge nvidia*" is mandatory first, then install the required driver

- as the '331' driver is now deprecated, and replaced by the '340' one, that means the still opened '331' are useless

So this report is closed now; meaning you need to open a new report if you really need

[Robin (robingape)]

Teo,

the Optimus issue is definitely a *separate bug*, requiring its own
resolution.

R.

On 13/08/15 14:48, teo1978 wrote:
> Or should we file a separate bug?
>

[Alberto Milone (albertomilone)]

Just to clarify:

1) Do not install the driver unless you want to use it. Uninstalling the driver will not break your system. By design a driver will be enabled if you install it. We keep an option to disable the driver but that's only to be used by gpu-manager (which is part of ubuntu-drivers-common), in order to support hybrid graphics. That is not meant to be disabled manually.

2) Optimus is supported by the NVIDIA driver. @teo1978 feel free to file a separate bug report about the external screen issue.

3) Optimus seems to fail with 340 on some systems, and even with 346 on others. We moved from the "modesetting" driver (which NVIDIA recommend but it's not the only driver that works) to "intel" because some systems didn't work (either at all or correctly). On such systems using the modesetting driver is the only solution. Please file a separate bug report about this (and file it against "ubuntu-drivers-common")

[teo1978 (teo8976)]

> Teo,

> the Optimus issue is definitely a *separate bug*, requiring its own
> resolution.

I was not talking about the optimus issue.
I was talking about the fact that the update of the NVidia drivers overwrote some system settings that shouldn't have been touched. Before the update, my system was set up to not use nVidia drivers; after the update, it was set to use them. That's a bug in the update that was released.

I'm not quite sure whether or not that should be considered a separate issue, because it is an issue directly caused by the fix to this issue. That is, the fix to this issue caused a bug of its own.

[teo1978 (teo8976)]

> hey all users still complaining:
>
> - be sure to

No man, if one had a system that was working and properly set up, and after getting an update it is broken (or something that was correctly configured gets misconfigured, or any configuration unnecessarily changed at all), there is a bug.

We aren't supposed to "be sure to" anything. The updates should make sure to not screw up a working system.

The bug should NOT be closed, unless a new one is opened for the issue caused by the fix to this bug.

I am not going to waste my time reconfiguring my system every time a new update comes out and screws up the settings that I already had to manually fix.

[teo1978 (teo8976)]

> 2) Optimus is supported by the NVIDIA driver.

That's not what NVidia told me. I asked for support and they told me it's officially unsupported on Linux.
And what I can tell for sure is that it doesn't work. The screen switching issue is well known.

[teo1978 (teo8976)]

> By design a driver will be enabled if you install it.

However, if it is disable when you UPGRADE the driver, it shouldn't be reenabled.

> We keep an option to disable the driver but [...] That is not meant to be disabled manually.

I don't know about now, but at the time I installed the driver, there was a nice graphical interface, accessible somewhere from the System Settings menu (back when Ubuntu had a menu), where you could choose whether to use NVidia proprietary drivers. And I chose not to.
It wasn't some arcane command line that I found out by googling around, it was something that was made accessible to a "normal user". If things have changed since, you must take into account that there are machines still around that have been upgraded since those times. I can't remember exactly which Ubuntu version that was. It may have been 13.04, or 12.something

[teo1978 (teo8976)]

> Uninstalling the driver will not break your system

On another machine, merely *installing* it did break the system. So forgive me if I'm skeptical that uninstalling it "will not" break my system. I take that as a "should not".

[teo1978 (teo8976)]

@albertomilone please see https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-340/+bug/1483388 and mark its importance as critical (prevents boot) as soon as possible.

Also feel free to correct anything that may be wrong in the report.

[Alberto Milone (albertomilone)]

On 13-08-15 16:50:40, teo1978 wrote:
> > By design a driver will be enabled if you install it.
>
> However, if it is disable when you UPGRADE the driver, it shouldn't be
> reenabled.

Right, that usually doesn't happen when you upgrade to a new version of
the same driver (for example upgrading nvidia-331 from 331.19 to
331.25). However, when you are migrated to a different driver package,
e.g. from nvidia-331 to nvidia-340, you will lose your settings, because
they are different packages and they provide different alternatives
(note: the alternatives are what allows you to enable or disable a
driver).

This, while unfortunate, is not a bug, and we cannot change it, due to
the way alternatives work.

[Alberto Milone (albertomilone)]

On 13-08-15 16:45:47, teo1978 wrote:
> > 2) Optimus is supported by the NVIDIA driver.
>
> That's not what NVidia told me. I asked for support and they told me it's officially unsupported on Linux.
> And what I can tell for sure is that it doesn't work. The screen switching issue is well known.

That only means that, if something goes wrong, you're on your own (at
least as far as NVIDIA are concerned). Support for Optimus is something
that we provide using our own system. What the NVIDIA driver supports is
part of the RandR 1.4, which we use for Optimus.

I won't comment on the screen switching issue, unless I see the relevant
logs (in a separate bug report), as it could depend on a number of
things (including your hardware configuration).

[teo1978 (teo8976)]

> Right, that usually doesn't happen when you upgrade to a new version of
> the same driver (for example upgrading nvidia-331 from 331.19 to
> 331.25). However, when you are migrated to a different driver package,
> e.g. from nvidia-331 to nvidia-340, you will lose your settings, because
> they are different packages and they provide different alternatives

Yes, I know. But since in this case installing a different driver packages was forced by an update (without even a warning or asking for confirmation, btw), whatever process was responsible for that should take that into account and take care of "migrating" the settings.

> This, while unfortunate, is not a bug,

Yes it is. You are only giving a diagnosis of the exact cause of the bug (which by the way I already knew), but that doesn't make it less of a bug.

> and we cannot change it, due to
> the way alternatives work.

I seriously doubt that's IMPOSSIBLE to fix. It may not be easy, but if you can do it manually, it can be done automatically. You only need to produce a script that looks at the settings (alternatives, or whatever) before uninstalling the old package, and then applies them (with all the necessary remappings and modifications) after installing the new one.
It may be a lot of work but it's NECESSARY and it was part of fixing this issue. A part that hasn't been done.

[teo1978 (teo8976)]

> Support for Optimus is something that we provide using our own system

Cool! (except it doesn't work, but I'll file a separate bug, though I'm surprised there's none already if it's supposed to work, as it's widely known that it doesn't at all)

So this page (among a few that I remember having read) must be out of date?
https://wiki.ubuntu.com/Ubuntu/Wishlist?highlight=%28nvidia%29%7C%28optimus%29
"""
Whish list
- NVIDIA Optimus integration (though this may be up to NVIDIA, not Canonical)
"""

[teo1978 (teo8976)]

I knew it must have already been reported.
Here it is:
https://bugs.launchpad.net/xorg-server/+bug/1220426

Importance undecided and unassigned sinse 2013

[Alberto Milone (albertomilone)]

On 14-08-15 10:52:23, teo1978 wrote:
> > This, while unfortunate, is not a bug,
>
> Yes it is. You are only giving a diagnosis of the exact cause of the bug
> (which by the way I already knew), but that doesn't make it less of a
> bug.
>

The alternatives switched back to automatic mode, and in automatic mode
the nvidia alternative has a higher priority than mesa's (which is what
you want).

You can either select the mesa alternative, call ldconfig, and update
the initramfs, or uninstall the nvidia package (as I recommended).

> > and we cannot change it, due to
> > the way alternatives work.
>
> I seriously doubt that's IMPOSSIBLE to fix. It may not be easy, but if you can do it manually, it can be done automatically. You only need to produce a script that looks at the settings (alternatives, or whatever) before uninstalling the old package, and then applies them (with all the necessary remappings and modifications) after installing the new one.
> It may be a lot of work but it's NECESSARY and it was part of fixing
> this issue. A part that hasn't been done.

Such work is not "necessary", as yours is really an unsupported corner
case. If you insist on keeping the nvidia driver installed and disabled,
against what we recommend, that is your problem.

[teo1978 (teo8976)]

> Such work is not "necessary", as yours is really an unsupported corner
> case. If you insist on keeping the nvidia driver installed and disabled,
> against what we recommend, that is your problem.

And how was I supposed to know what you "recommend"?

Back in the days when I installed Ubuntu on my machine, there was a friendly interface somewhere like "Software & Updates" > "Additional Drivers" or something, where I could choose to install and use the NVidia drivers.
Then, when I saw they wouldn't work, I went to the same interface and chose not to use them. No warning appeared telling me that was "not recommended" or that I should uninstall the drivers.

Should I have read the whole Ubuntu documentation, like "Let's see if what I just did through the standard interface is not recommended, and if they recommend doing something else like uninstalling the driver".

It's not that I "insist" in keeping the driver installed and disabled. I did that ages ago and forgot about it. And I very much doubt it's that much of a "corner case". Right in this thread you have another user to whom the exact same thing happened.

And even if we are few, we got in our situation by just using what then was the user interface (I mean, we didn't tinkle and mess around with stuff), so if we got to a situation that was not supported, something was wrong in Ubuntu in the first place, and if we got into a situation which was then supported and then stopped being, then it's Ubuntu responsibility to make sure that automatic updates take that into account and keep things working.

And also, if a piece of software is designed in such a way that if you stop using it yoy have to uninstall it otherwise it may be harmful, it is badly designed.

[teo1978 (teo8976)]

If you have a machine that works, and you run automatic updates, and your machine stops working, there's a bug. If there's anything wrong in the machine configuration when you are updating, it must be detected, a warning given, and the option to continue at your own risk or skip that update.

[Jerry (jerrywone)]

Unfortunately, to get newer versions and hopefully out of Nvidia kernel
update hell, I went from LTS to the next regular unsupported, hoping to go
beyond that one to contemporary, but all went astray into unstable and
unbootable, probably in the filing to upgrade multiple times, repository
problems and bugs in same...

Possibly waiting a bit on 15.10alpha to mature and upgrade nonbootable disk
from cdr/dvdr/usbstick/netboot to get back to a working system.

Waiting for godot was too frustrating.

On Fri, Aug 14, 2015, 09:36 teo1978 <email address hidden> wrote:

> If you have a machine that works, and you run automatic updates, and
> your machine stops working, there's a bug. If there's anything wrong in
> the machine configuration when you are updating, it must be detected, a
> warning given, and the option to continue at your own risk or skip that
> update.
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1405597).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> Status in nvidia-graphics-drivers-331 package in Ubuntu:
> Invalid
> Status in nvidia-graphics-drivers-331-updates package in Ubuntu:
> Invalid
> Status in nvidia-graphics-drivers-340 package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-340-updates package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-346 package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-346-updates package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-331 source package in Trusty:
> Invalid
> Status in nvidia-graphics-drivers-331-updates source package in Trusty:
> Invalid
> Status in nvidia-graphics-drivers-340 source package in Trusty:
> Fix Released
> Status in nvidia-graphics-drivers-340-updates source package in Trusty:
> Fix Released
> Status in nvidia-graphics-drivers-346 source package in Trusty:
> Fix Released
> Status in nvidia-graphics-drivers-346-updates source package in Trusty:
> Fix Released
>
> Bug description:
> Filing this against the 340-updates version but possibly the same
> applies to older versions, too. The nvidia source package produces two
> individual dkms packages: nvidia-340-updates, nvidia-340-updates-uvm.
> The problem is that the DKMS build of the nvidia-uvm module runs
> compile steps inside the nvidia modules build directory. This is
> violating the DKMS assumption that each module can be build
> independently (there is no way of describing cross-modules
> dependencies and even more important, the autoinstall step after a new
> kernel is installed will run the modules build in parallel).
>
> Since nvidia and nvidia-uvm are very dependent on each other the right
> course of action seems to be to combine both sources in one DKMS
> module that produces two kernel modules (this is supported by DKMS).
> For the transition this resulting dkms package needs to have a
> breaks/replaces for the nvidia-uvm package.
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscript...

[Alberto Milone (albertomilone)]

On 14-08-15 13:24:50, teo1978 wrote:
> > Such work is not "necessary", as yours is really an unsupported corner
> > case. If you insist on keeping the nvidia driver installed and disabled,
> > against what we recommend, that is your problem.
>
> And how was I supposed to know what you "recommend"?
>
> Back in the days when I installed Ubuntu on my machine, there was a friendly interface somewhere like "Software & Updates" > "Additional Drivers" or something, where I could choose to install and use the NVidia drivers.
> Then, when I saw they wouldn't work, I went to the same interface and chose not to use them. No warning appeared telling me that was "not recommended" or that I should uninstall the drivers.
>

Different use cases were supported in 12.04. For example, you could have
multiple nvidia drivers installed at the same time. That changed after
12.04.

> And also, if a piece of software is designed in such a way that if you
> stop using it yoy have to uninstall it otherwise it may be harmful, it
> is badly designed.

Ok, that implies that all drivers are badly designed. It's your opinion.
This, however, is a bug report.

Now please stop posting, as this bug report is closed.

[teo1978 (teo8976)]

> Different use cases were supported in 12.04

That's my point: you can't just drop supported usecases and do nothing about it. If I'm doing something supported and I upgrade, and that's not supported any more, the upgrade MUST guide me through a way to fix whatever is wrong.

> Now please stop posting, as this bug report is closed.

I'm posting because it must be reopened.

[teo1978 (teo8976)]

> > And also, if a piece of software is designed in such a way that if you
> > stop using it yoy have to uninstall it otherwise it may be harmful, it
> > is badly designed.
>
> Ok, that implies that all drivers are badly designed. It's your opinion.

Don't pretend you are stupid when you aren't. This driver used to have an option (nicely accessible through user interface) which allowed to just not use it. It was *designed* to be either used or not used. And when not used, it would do nothing and interfere with nothing.
So, the newer version doesn't that anymore? Then you can't replace one version with another without performing some check. The update is incomplete and buggy if it does that.

[graingert (tagrain)]

Guys it's a closed source driver. I'm being a hypocrite here, but chillax.
On 14 Aug 2015 21:16, "teo1978" <email address hidden> wrote:

> > > And also, if a piece of software is designed in such a way that if you
> > > stop using it yoy have to uninstall it otherwise it may be harmful, it
> > > is badly designed.
> >
> > Ok, that implies that all drivers are badly designed. It's your opinion.
>
> Don't pretend you are stupid when you aren't. This driver used to have an
> option (nicely accessible through user interface) which allowed to just not
> use it. It was *designed* to be either used or not used. And when not used,
> it would do nothing and interfere with nothing.
> So, the newer version doesn't that anymore? Then you can't replace one
> version with another without performing some check. The update is
> incomplete and buggy if it does that.
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1411195).
> https://bugs.launchpad.net/bugs/1431753
>
> Title:
> Nvidia binary driver FTBS due to DKMS layer violation
>
> Status in nvidia-graphics-drivers-331 package in Ubuntu:
> Invalid
> Status in nvidia-graphics-drivers-331-updates package in Ubuntu:
> Invalid
> Status in nvidia-graphics-drivers-340 package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-340-updates package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-346 package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-346-updates package in Ubuntu:
> Fix Released
> Status in nvidia-graphics-drivers-331 source package in Trusty:
> Invalid
> Status in nvidia-graphics-drivers-331-updates source package in Trusty:
> Invalid
> Status in nvidia-graphics-drivers-340 source package in Trusty:
> Fix Released
> Status in nvidia-graphics-drivers-340-updates source package in Trusty:
> Fix Released
> Status in nvidia-graphics-drivers-346 source package in Trusty:
> Fix Released
> Status in nvidia-graphics-drivers-346-updates source package in Trusty:
> Fix Released
>
> Bug description:
> Filing this against the 340-updates version but possibly the same
> applies to older versions, too. The nvidia source package produces two
> individual dkms packages: nvidia-340-updates, nvidia-340-updates-uvm.
> The problem is that the DKMS build of the nvidia-uvm module runs
> compile steps inside the nvidia modules build directory. This is
> violating the DKMS assumption that each module can be build
> independently (there is no way of describing cross-modules
> dependencies and even more important, the autoinstall step after a new
> kernel is installed will run the modules build in parallel).
>
> Since nvidia and nvidia-uvm are very dependent on each other the right
> course of action seems to be to combine both sources in one DKMS
> module that produces two kernel modules (this is supported by DKMS).
> For the transition this resulting dkms package needs to have a
> breaks/replaces for the nvidia-uvm package.
>
> To manage notifications about this bug go to:
>
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-331/+bug/1431753/+subscriptions
>

[Gianfranco Costamagna (costamagnagianfranco)]

Are you aware that each comment triggers THOUSAND of emails to other users?

Please respect them, and use private messages with your exchanges with Alberto.

This particular bug is FIXED. it might have triggered regressions, or something new I do not know, and opening new bugs is the correct way to handle it.

You can't close the same bug twice, and you can't reopen a fixed bug.

Leave with the design of LP, and go in a new bug.

thanks.

[teo1978 (teo8976)]

> Are you aware that each comment triggers THOUSAND of emails to other users?

Are you aware that you can unsubscribe if you want to?

> You can't close the same bug twice, and you can't reopen a fixed bug.

No I can't but there are definitely people who can.

> This particular bug is FIXED. it might have triggered regressions,
> and opening new bugs is the correct way to handle it.

Well that's a way of seeing it (I'd say a bugfix that fixes half of the bug or does it in the wrong way is not a complete fix and hence leaves the bug not fixed).

Ok, I'll go open a new issue.

[Gianfranco Costamagna (costamagnagianfranco)]

>Are you aware that you can unsubscribe if you want to?

yes, but this doesn't change the point.

>No I can't but there are definitely people who can.

it would be useless.

>Well that's a way of seeing it (I'd say a bugfix that fixes half of the bug or does it in the wrong way is not a complete fix and hence >leaves the bug not fixed).
>
>Ok, I'll go open a new issue.

this is how debhelper works, and how the syntax of closing bugs works on LP and Debian.
if you want to fix this bug go on the new open bug and help Alberto with the stuff he requested.

Unfortunately dealing with closed and proprietary software is not so easy, and I'm telling this with my Debian Developer hat on.

Sorry for that, but posting here will annoy people, and for sure not help you.

And since I'm a DD, I'll follow the bug you created, and try to help there, because I'm not bothered by your mails, but other people are.
(thinking about freedom is also caring about other people's issues)

[KennoVO (kenno-xs4all)]

The Optimus issue most definitely must be filed as a separate bug (and comments on it go there). In order for a bug tracking system like this to work well, the technical scope of a bug must be narrowly defined; that's software development good practice 101. If the scope of a bug is vague or allowed to creep, no developer will touch the "assign to me" button with a 10-foot pole. In fact, this is exactly the reason why lp:1268257 has been sitting there for 18 months without any progress; the error description consisted of a *symptom* that could have a very large number of causes, so nobody was willing to go look into all of them. It was only after it was realized that lp:1431753 was the cause in the vast majority of cases that things started moving. In this context, kudos go to teo1978 for forcing the breakthrough by finding and applying a workaround for lp:1450251 . You guys have to forgive him for being a bit testy. Even the most patient person would throw a fit after waiting 18 moths for a trivial-looking fix to a bug that made thousands of users lose their GUI every few weeks, and then getting another show-stopping problem when applying that fix!

Anyhow, the scope of *this* bug is the DKMS violation, which is fully fixed. From a *technical* perspective, the Optimus issue is an entirely unrelated bug. The fact that it's triggered by the present fix does not matter a this stage.

<slightly offtopic>Just to set the record straight (you might want to use part of this in the description of the new bug): there are (at least) 2 separate ways to get Optimus support working in Ubuntu:
(1) nvidia-prime : allows you to switch between running the whole desktop on Intel or Nvidia. The choice is made in the nvidia control panel and takes effect when logging out and back in. It's far from a perfect solution but at least it's easy to use.
(2) bumblebee : allows you to specify on which card to run a given program at the command-line (or by editing the application launcher). I personally prefer this route because it gives me more granular control and doesn't force you to log out, but novice users may find it cumbersome.
It appears that the configuration files of the packages that are involved in both (1) and (2) did not anticipate drivers newer than 331 to be used with 14.04, resulting in a blank screen upon login. The fix for (1) is in Anna's link: http://vxlabs.com/2015/02/05/solving-the-ubuntu-14-04-nvidia-346-nvidia-prime-black-screen-issue/ . As for (2), see my post: https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-340-updates/+bug/1431753/comments/62 .
Now, if canonical was really serious about LTS, they would have a sufficiently large and professional team working on this. The person responsible for the nvidia driver would have remembered to contacted the people responsible for the gpu-manager and bumblebee packages, or at the very least have been reminded of it by my post, which I made when the fix was still in the "proposed" stage. The actual release of the fix would have been delayed while those other developers would be updating the configuration files of their respective packages (that's, like, less than an hour of work), and th...

[teo1978 (teo8976)]

> From a *technical* perspective, the Optimus issue is an entirely unrelated bug.
> The fact that it's triggered by the present fix does not matter a this stage.

I was not talking about the Optimus issue (I never argued that should be part of this issue, though I admit I did make some OT comment about that).
I was talking about the fact that the update that fixes THIS very issue causes the reenabling of NVidia drivers which were previously disabled, hence changing a system configuration that should not be touched.

But anyway, since I've been told that the fact that the fix to this bug is broken is a separate bug, I've filed a separate bug. That wouldn't have generated such a long exchange if Alberto didn't insist in claiming that the fix is not broken at all, that everything is ok, and that there is no additional problem to be fixed.

[Alexander (sturmlocke86)]

Sup guys,

quick update: Seems Alberto was right, my family and friends switched back to Nouveau via the driver utility and were then able to install the new "tested" and "recommended" Nvidia drivers 346.82 via the stable repo provided by Ubuntu. No need to use -updates. System stability and speed has been restored! I can't verify this myself, since I have given up on Nvidia (running on AMD and Intel + Open Source drivers now), thus: I have added a screenshot that a friend sent me to this post, showing the new drivers available to Nvidia users.

Thanks @whoever fixed this!

Cheers,
Alex

[J K Cunningham (jeffrey-jkcunningham)]

On 08/14/2015 01:46 PM, teo1978 wrote:
>> Are you aware that each comment triggers THOUSAND of emails to other
> users?
>
> Are you aware that you can unsubscribe if you want to?
>
>
I unsubscribed from this bug report days ago and it's still sending me
your silly tirades.
Would you please STFU?

[Robin (robingape)]

JK,

the link you need to fully unsubscribe is well hidden:
https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-340-updates/+bug/1431753/+subscriptions,
and then the bottom link, assuming that you are signed in to launchpad.

HTH

R.

On 15/08/15 02:33, J K Cunningham wrote:
> On 08/14/2015 01:46 PM, teo1978 wrote:
>>> Are you aware that each comment triggers THOUSAND of emails to other
>> users?
>>
>> Are you aware that you can unsubscribe if you want to?
>>
>>
> I unsubscribed from this bug report days ago and it's still sending me
> your silly tirades.
> Would you please STFU?
>

[J K Cunningham (jeffrey-jkcunningham)]

Much obliged.
--Jeff

On 08/15/2015 03:03 AM, Robin wrote:
> JK,
>
> the link you need to fully unsubscribe is well hidden:
> https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-340-updates/+bug/1431753/+subscriptions,
> and then the bottom link, assuming that you are signed in to launchpad.
>
> HTH
>
> R.
>
> On 15/08/15 02:33, J K Cunningham wrote:
>> On 08/14/2015 01:46 PM, teo1978 wrote:
>>>> Are you aware that each comment triggers THOUSAND of emails to other
>>> users?
>>>
>>> Are you aware that you can unsubscribe if you want to?
>>>
>>>
>> I unsubscribed from this bug report days ago and it's still sending me
>> your silly tirades.
>> Would you please STFU?
>>