Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Nouveau crashes xserver - something is broken but what ?
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Desktop Environments
View previous topic :: View next topic  
Author Message
jfo
n00b
n00b


Joined: 07 Sep 2015
Posts: 37

PostPosted: Sun Apr 16, 2017 5:37 am    Post subject: Nouveau crashes xserver - something is broken but what ? Reply with quote

This is a new thread to summarize the issues I have experienced, in the hope that this might be useful
to someone in a similar situation.

Following some recent updates, I found myself no longer able to get kde to start.
My machine is relatively new and for more than a year now, I have been using-- with success--
the nouveau driver. After much tinkering I believe that some recent change possibly in the
nouveau driver itself, in mesa, libdrm or in a related package causes the Xserver to crash.
The crash is immediate and leaves my card (GeForce GT 730) in a strange state - the only remedy is to reboot.
(the kernel does not crash - the machine remains accessible over the network).
Unfortunately, I found out that the cause of the crash is very diffficult to isolate. The Xorg log contains no useful
information. There is an indication of an error, but no traceback information gets captured.

To verify my hypothesis I tried the proprietary nvidia driver. Low and behold kde starts normally. However, the situation with the proprietary driver is less than ideal because unbelievably the framebuffer support is completely broken on UEFI systems even though UEFI has been around for a few years already.

some informations:

AsRock X99 WS/ i7 5820K/ GeForce Gt 730 - booting in UEFI mode

xf86-video-nouveau-1.0.13 [edit: xf86-video-nouveau-1.0.14]
mesa-17.03
libdrm-2.4.79
kernel-4.9.9 ( I tried a few in the 4.10 series as well)
xorg-server-1.19.3

I would file a bug - but I would need to narrow down the problem. I am running out of ideas (and time).


Last edited by jfo on Sun Apr 16, 2017 4:22 pm; edited 1 time in total
Back to top
View user's profile Send private message
Roman_Gruber
Advocate
Advocate


Joined: 03 Oct 2006
Posts: 3806
Location: Austro Bavaria

PostPosted: Sun Apr 16, 2017 1:10 pm    Post subject: Re: Nouveau crashes xserver - something is broken but what Reply with quote

jfo wrote:
To verify my hypothesis I tried the proprietary nvidia driver. Low and behold kde starts normally. However, the situation with the proprietary driver is less than ideal because unbelievably the framebuffer support is completely broken on UEFI systems even though UEFI has been around for a few years already.



May I ask which functionality is broken with framebuffer with the binary driver?

Are you talking about these settings?

Code:
ASUS-G75VW roman # grep BUFFER /usr/src/linux/.config
CONFIG_DMA_SHARED_BUFFER=y
CONFIG_CDROM_PKTCDVD_BUFFERS=64
CONFIG_FRAMEBUFFER_CONSOLE=y
CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY=y
# CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
# CONFIG_IIO_BUFFER is not set


I have set my text console via grub with

Code:
GRUB_GFXMODE=1920x1080x32
GRUB_GFXPAYLOAD_LINUX=keep


as you can see 1920x1080 resolution with 32bit

--

UEFI is broken by design.

missing functionality because microsoft / intel designed it very badly. open kernel bugs since years here regarding this and special keys of my hardware

--
https://en.wikipedia.org/wiki/Linux_framebuffer
Quote:
The now defunct DirectFB is another project aimed at providing a framework for hardware acceleration of the Linux framebuffer
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 42592
Location: 56N 3W

PostPosted: Sun Apr 16, 2017 1:19 pm    Post subject: Reply with quote

jfo,

Are there any messages in dmesg when your graphics card locks up?

The binary blob nvidia driver does not provide a framebuffer console and leaves dire warnings in dmesg if you try to add one.
However, framebuffer drivers that do not try to configure the hardware may work, ignore the warning.
These are EFI Framebuffer, Simple Framebuffer and VESA Framebuffer.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
jfo
n00b
n00b


Joined: 07 Sep 2015
Posts: 37

PostPosted: Sun Apr 16, 2017 4:11 pm    Post subject: Reply with quote

NeddySeagoon,

There is *no* message in dmesg when the card locks up.
All I can tell is that the Xserver has crashed ( there is an error in the Xorg log - but no traceback is captured).

It seems that whether EFI Framebuffer, Simple Framebuffer and VESA Framebuffer work at all
depends on (poorly documented) factors including the version of the EFI firmware itself, and/or the
graphics card firmware. So far my attempts with the above fbs have resulted
in a blank console screen.

See https://devtalk.nvidia.com/default/topic/827139/uefi-nvidia-vga-console-complaints-/
Quote:

The message is a little misleading in UEFI mode. What it means it that the GPU was initialized to a graphical mode using the legacy VGA BIOS, regardless of whether the system was booted in UEFI mode or not. Typically this happens if the Compatibility Support Module (CSM) is enabled in the system BIOS. If you have an option to disable CSM in the SBIOS, please try that.

Aaron Plattner
NVIDIA Linux Graphics

#8
Posted 04/22/2015 02:31 PM


I have an option to disable CSM (I have not tried yet) however - it is all or nothing - and many extension cards still rely on CSM mode for initialization.
The whole thing is a horrible mess. I have nothing but respect and admiration for the free software developers that somehow manage
to get things sort of working.
Back to top
View user's profile Send private message
Jaglover
Watchman
Watchman


Joined: 29 May 2005
Posts: 6971
Location: Saint Amant, Acadiana

PostPosted: Sun Apr 16, 2017 4:44 pm    Post subject: Reply with quote

You may want to see if there is a bug filed in bugs.freedesktop.org. You are not the only one having those nouveau freezes.
_________________
Please learn how to denote units correctly!
Back to top
View user's profile Send private message
jfo
n00b
n00b


Joined: 07 Sep 2015
Posts: 37

PostPosted: Sun Apr 16, 2017 6:05 pm    Post subject: Reply with quote

Jaglover

I did take look at the FreeDesktop bugzilla. This is not a "freeze" i.e. the server does not freeze .. it completely crashes
on kde/plasma startup. It is not obvious at all whether the issue is nouveau, Xorg or something else.
The closest reports I found are on the Arch linux forums and a prime suspect seems to be Xorg.
I feel there is not enough information to file a useful bug report.
Back to top
View user's profile Send private message
jfo
n00b
n00b


Joined: 07 Sep 2015
Posts: 37

PostPosted: Sun Apr 16, 2017 7:35 pm    Post subject: Reply with quote

Just for completeness: I succeeded in getting a working console by enabling
EFI Framebuffer *and* disabling the BIOS compatibility module for the video
only. It turns out that my board (AsRock x99 WS) allows one to selectively
enable/disable EFI mode for different classes of devices. Since the
video card is initialized in EFI mode, the ominous warnings are gone.
BTW all my other attempts with other fb devices failed (EFI mode or compatibility mode).
In particular; I could never get Simple Framebuffer to display anything.
One quirk: I have a 2 monitor setup. With nouveau and the vesafb both displays
show the console output. With EFI fb, only the "principal" display is enabled.
May be there is a module parameter to enable this, I do not know. Anyway,
I am happy enough to have a functional console.

My only comment at this point is that the documentation for the procedure to
deal with the nvidia drivers in EFI mode is inadequate.
Back to top
View user's profile Send private message
davidm
Guru
Guru


Joined: 26 Apr 2009
Posts: 557
Location: US

PostPosted: Tue Apr 18, 2017 6:21 pm    Post subject: Reply with quote

jfo,

I am thinking there is a great possibility that our issues are the same or at least related. I am also using nouveau and had Plasma/KDE suddenly crash upon startup without making any other config changes other than recent updates.

I am also using what seems to be the same GPU as you or near it:

Code:

01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 730] (rev a1) (prog-if 00 [VGA controller])
        Subsystem: eVga.com. Corp. GF108 [GeForce GT 730]
        Flags: bus master, fast devsel, latency 0, IRQ 28
        Memory at fc000000 (32-bit, non-prefetchable) [size=16M]
        Memory at f0000000 (64-bit, prefetchable) [size=128M]
        Memory at f8000000 (64-bit, prefetchable) [size=32M]
        I/O ports at dc80 [size=128]
        Expansion ROM at 000c0000 [disabled] [size=128K]
        Capabilities: [60] Power Management version 3
        Capabilities: [68] MSI: Enable+ Count=1/1 Maskable- 64bit+
        Capabilities: [78] Express Endpoint, MSI 00
        Capabilities: [b4] Vendor Specific Information: Len=14 <?>
        Capabilities: [100] Virtual Channel
        Capabilities: [128] Power Budgeting <?>
        Capabilities: [600] Vendor Specific Information: ID=0001 Rev=1 Len=024 <?>
        Kernel driver in use: nouveau


I am able to use fluxbox from 'startx /usr/bin/startfluxbox' and it is fully functional other than a crash I noticed once when going full screen in smplayer.

Here is the topic where I reported on my issues:
https://forums.gentoo.org/viewtopic-t-1062162.html
Back to top
View user's profile Send private message
jfo
n00b
n00b


Joined: 07 Sep 2015
Posts: 37

PostPosted: Wed Apr 19, 2017 1:28 am    Post subject: Reply with quote

davidm -

Good work ! and thank you ! I will give a libdrm downgrade a try when I have a moment, but it is very likely that libdrm is the culprit.
That would explain why the proprietary nvidia driver works ... afaik it has no libdrm dependencies.

Here is what I see on my system

Code:

qlop -l | grep libdrm

Mon Dec  5 19:45:30 2016 >>> x11-libs/libdrm-2.4.74
Sat Jan 28 09:40:53 2017 >>> x11-libs/libdrm-2.4.75
Wed Mar 29 19:14:50 2017 >>> x11-libs/libdrm-2.4.76
Sat Apr   8 10:03:47 2017 >>> x11-libs/libdrm-2.4.77
Sun Apr  9 09:54:42 2017 >>> x11-libs/libdrm-2.4.79


So there were 3 successive updates about a week ago.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Desktop Environments All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum