[opensuse] prime-select: unable to query GPU information
Hi, I just upgraded my asus laptop with intel/nvidia graphics from 42.1 to 42.3, all went well so far. But when entering prime-select nvidia I get: ERROR unable to query GPU information The script runs anyway, I guess, because when I reboot at some point the screen with the boot messages begins to flicker heavily (I could not read the last message...). I had to restart and boot to runlevel 3, then entered "prime-sectect intel" and now I can boot normally - but of course only using the intel card. I don't know which information you need to help me: please ask and I'll try to provide it. Thanks for your help! Daniel -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
12.01.2018 21:58, Daniel Bauer пишет:
Hi,
I just upgraded my asus laptop with intel/nvidia graphics from 42.1 to 42.3, all went well so far.
But when entering
prime-select nvidia
I get: ERROR unable to query GPU information
The message comes from nvidia-xconfig --query-gpu-info. As far as I can tell, it actually expects nvidia kernel module to be loaded (it tries to load it itself). So you may try to debug this utility until it works. Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
The script runs anyway, I guess, because when I reboot at some point the screen with the boot messages begins to flicker heavily (I could not read the last message...). I had to restart and boot to runlevel 3, then entered "prime-sectect intel" and now I can boot normally - but of course only using the intel card.
prime-select is supposed to generate xorg.conf and enable nVidia specific GL libraries. You can check resulting configuration file and manually adjust it until it works.
I don't know which information you need to help me: please ask and I'll try to provide it.
Thanks for your help!
Daniel
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 13.01.2018 um 06:58 schrieb Andrei Borzenkov:
12.01.2018 21:58, Daniel Bauer пишет:
Hi,
I just upgraded my asus laptop with intel/nvidia graphics from 42.1 to 42.3, all went well so far.
But when entering
prime-select nvidia
I get: ERROR unable to query GPU information
The message comes from nvidia-xconfig --query-gpu-info. As far as I can tell, it actually expects nvidia kernel module to be loaded (it tries to load it itself). So you may try to debug this utility until it works.
Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
The script runs anyway, I guess, because when I reboot at some point the screen with the boot messages begins to flicker heavily (I could not read the last message...). I had to restart and boot to runlevel 3, then entered "prime-sectect intel" and now I can boot normally - but of course only using the intel card.
prime-select is supposed to generate xorg.conf and enable nVidia specific GL libraries. You can check resulting configuration file and manually adjust it until it works.
I don't know which information you need to help me: please ask and I'll try to provide it.
Thanks for your help!
Daniel
Thank you for the explanations. I will see if my minimal knowledge leads to some results this weekend using your information... Daniel -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
I had no time to play with my laptop, but here I am again... Am 13.01.2018 um 06:58 schrieb Andrei Borzenkov:
12.01.2018 21:58, Daniel Bauer пишет:
... when entering
prime-select nvidia
I get: ERROR unable to query GPU information
The message comes from nvidia-xconfig --query-gpu-info. As far as I can tell, it actually expects nvidia kernel module to be loaded (it tries to load it itself). So you may try to debug this utility until it works.
That's far away of what my knowledge allows :-)
Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
I see, but I have no idea how to get that PCI-ID Hwinfo tells me so many things that I have no idea what to use :-) I uploaded a screenshot of the hardware infor screen: https://daniel-bauer.com/test/hwinfo_nvidia.jpg
prime-select is supposed to generate xorg.conf and enable nVidia specific GL libraries. You can check resulting configuration file and manually adjust it until it works.
So, I guess, next step is to find out the PCI ID, then somehow put it into that script, and then look if it works... Can you help me: - how to find that PCI ID - how to put put it in that script (I mean which code to remove and which to add instead) ? Or, if meanwhile there are some solutions on how to run prime-select on 42.3, I be happy of course to hear about. Googling didn't help me. Thanks for further help! Daniel -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday, 2018-03-30 at 13:04 +0200, Daniel Bauer wrote:
Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
I see, but I have no idea how to get that PCI-ID Hwinfo tells me so many things that I have no idea what to use :-) I uploaded a screenshot of the hardware infor screen: https://daniel-bauer.com/test/hwinfo_nvidia.jpg
If you manage that to generate in English, it would be easier for us ;-) You could create this script: /usr/local/bin/ingles: #!/bin/sh LANG=en_US.UTF-8 \ LC_ALL=en_US.UTF-8 \ DICTIONARY=english \ KDE_LANG=en_US.UTF-8 \ exec "$@" then call: ingles hwinfo --gfxcard - -- Cheers, Carlos E. R. (from openSUSE 42.3 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlq+LxUACgkQtTMYHG2NR9W5UACcDgaMcG7UrkyJCWDqVjDy3ifG ztkAn0HbAynmbpDUqz6xazrPUu2c28Mh =boPN -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 30.03.2018 um 14:35 schrieb Carlos E. R.:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Friday, 2018-03-30 at 13:04 +0200, Daniel Bauer wrote:
Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
I see, but I have no idea how to get that PCI-ID Hwinfo tells me so many things that I have no idea what to use :-) I uploaded a screenshot of the hardware infor screen: https://daniel-bauer.com/test/hwinfo_nvidia.jpg
If you manage that to generate in English, it would be easier for us ;-)
here it is: meitli:~ # ingles hwinfo --gfxcard 09: PCI 02.0: 0300 VGA compatible controller (VGA) [Created at pci.378] Unique ID: _Znp.RAVeT1Md3J0 SysFS ID: /devices/pci0000:00/0000:00:02.0 SysFS BusID: 0000:00:02.0 Hardware Class: graphics card Model: "Intel HD Graphics 530" Vendor: pci 0x8086 "Intel Corporation" Device: pci 0x191b "HD Graphics 530" SubVendor: pci 0x1043 "ASUSTeK Computer Inc." SubDevice: pci 0x1c5d Revision: 0x06 Driver: "i915" Driver Modules: "i915" Memory Range: 0xdd000000-0xddffffff (rw,non-prefetchable) Memory Range: 0xb0000000-0xbfffffff (ro,non-prefetchable) I/O Ports: 0xf000-0xf03f (rw) IRQ: 128 (484136 events) Module Alias: "pci:v00008086d0000191Bsv00001043sd00001C5Dbc03sc00i00" Driver Info #0: Driver Status: i915 is active Driver Activation Cmd: "modprobe i915" Config Status: cfg=no, avail=yes, need=no, active=unknown 23: PCI 100.0: 0302 3D controller [Created at pci.378] Unique ID: VCu0.hDjcjYACgT6 Parent ID: vSkL.fWlq6741dtC SysFS ID: /devices/pci0000:00/0000:00:01.0/0000:01:00.0 SysFS BusID: 0000:01:00.0 Hardware Class: graphics card Model: "nVidia GM107M [GeForce GTX 960M]" Vendor: pci 0x10de "nVidia Corporation" Device: pci 0x139b "GM107M [GeForce GTX 960M]" SubVendor: pci 0x1043 "ASUSTeK Computer Inc." SubDevice: pci 0x1c5d Revision: 0xa2 Memory Range: 0xde000000-0xdeffffff (rw,non-prefetchable) Memory Range: 0xc0000000-0xcfffffff (ro,non-prefetchable) Memory Range: 0xd0000000-0xd1ffffff (ro,non-prefetchable) I/O Ports: 0xe000-0xefff (rw) Memory Range: 0xdf000000-0xdf07ffff (ro,non-prefetchable,disabled) IRQ: 255 (no events) Module Alias: "pci:v000010DEd0000139Bsv00001043sd00001C5Dbc03sc02i00" Driver Info #0: Driver Status: nouveau is not active Driver Activation Cmd: "modprobe nouveau" Config Status: cfg=no, avail=yes, need=no, active=unknown Attached to: #8 (PCI bridge) Primary display adapter: #9 -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-03-30 15:43, Daniel Bauer wrote:
Am 30.03.2018 um 14:35 schrieb Carlos E. R.:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Friday, 2018-03-30 at 13:04 +0200, Daniel Bauer wrote:
Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
I see, but I have no idea how to get that PCI-ID Hwinfo tells me so many things that I have no idea what to use :-) I uploaded a screenshot of the hardware infor screen: https://daniel-bauer.com/test/hwinfo_nvidia.jpg
If you manage that to generate in English, it would be easier for us ;-)
here it is:
meitli:~ # ingles hwinfo --gfxcard 09: PCI 02.0: 0300 VGA compatible controller (VGA) [Created at pci.378] Unique ID: _Znp.RAVeT1Md3J0 SysFS ID: /devices/pci0000:00/0000:00:02.0 SysFS BusID: 0000:00:02.0 Hardware Class: graphics card Model: "Intel HD Graphics 530" Vendor: pci 0x8086 "Intel Corporation" Device: pci 0x191b "HD Graphics 530" SubVendor: pci 0x1043 "ASUSTeK Computer Inc." SubDevice: pci 0x1c5d Revision: 0x06 Driver: "i915" Driver Modules: "i915" Memory Range: 0xdd000000-0xddffffff (rw,non-prefetchable) Memory Range: 0xb0000000-0xbfffffff (ro,non-prefetchable) I/O Ports: 0xf000-0xf03f (rw) IRQ: 128 (484136 events) Module Alias: "pci:v00008086d0000191Bsv00001043sd00001C5Dbc03sc00i00" Driver Info #0: Driver Status: i915 is active Driver Activation Cmd: "modprobe i915" Config Status: cfg=no, avail=yes, need=no, active=unknown
23: PCI 100.0: 0302 3D controller [Created at pci.378] Unique ID: VCu0.hDjcjYACgT6 Parent ID: vSkL.fWlq6741dtC SysFS ID: /devices/pci0000:00/0000:00:01.0/0000:01:00.0 SysFS BusID: 0000:01:00.0 Hardware Class: graphics card Model: "nVidia GM107M [GeForce GTX 960M]" Vendor: pci 0x10de "nVidia Corporation" Device: pci 0x139b "GM107M [GeForce GTX 960M]" SubVendor: pci 0x1043 "ASUSTeK Computer Inc." SubDevice: pci 0x1c5d Revision: 0xa2 Memory Range: 0xde000000-0xdeffffff (rw,non-prefetchable) Memory Range: 0xc0000000-0xcfffffff (ro,non-prefetchable) Memory Range: 0xd0000000-0xd1ffffff (ro,non-prefetchable) I/O Ports: 0xe000-0xefff (rw) Memory Range: 0xdf000000-0xdf07ffff (ro,non-prefetchable,disabled) IRQ: 255 (no events) Module Alias: "pci:v000010DEd0000139Bsv00001043sd00001C5Dbc03sc02i00" Driver Info #0: Driver Status: nouveau is not active Driver Activation Cmd: "modprobe nouveau" Config Status: cfg=no, avail=yes, need=no, active=unknown Attached to: #8 (PCI bridge)
Primary display adapter: #9
Well, the info is there, but like you I don't know which of the "pci" strings there is the one. My guess would be "0000:01:00.0" -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
Am 30.03.2018 um 18:46 schrieb Carlos E. R.:
On 2018-03-30 15:43, Daniel Bauer wrote:
Am 30.03.2018 um 14:35 schrieb Carlos E. R.:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Friday, 2018-03-30 at 13:04 +0200, Daniel Bauer wrote:
Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
I see, but I have no idea how to get that PCI-ID Hwinfo tells me so many things that I have no idea what to use :-)
23: PCI 100.0: 0302 3D controller [Created at pci.378] Unique ID: VCu0.hDjcjYACgT6 Parent ID: vSkL.fWlq6741dtC SysFS ID: /devices/pci0000:00/0000:00:01.0/0000:01:00.0 SysFS BusID: 0000:01:00.0 Hardware Class: graphics card Model: "nVidia GM107M [GeForce GTX 960M]" Vendor: pci 0x10de "nVidia Corporation" Device: pci 0x139b "GM107M [GeForce GTX 960M]" SubVendor: pci 0x1043 "ASUSTeK Computer Inc." SubDevice: pci 0x1c5d Revision: 0xa2 Memory Range: 0xde000000-0xdeffffff (rw,non-prefetchable) Memory Range: 0xc0000000-0xcfffffff (ro,non-prefetchable) Memory Range: 0xd0000000-0xd1ffffff (ro,non-prefetchable) I/O Ports: 0xe000-0xefff (rw) Memory Range: 0xdf000000-0xdf07ffff (ro,non-prefetchable,disabled) IRQ: 255 (no events) Module Alias: "pci:v000010DEd0000139Bsv00001043sd00001C5Dbc03sc02i00" Driver Info #0: Driver Status: nouveau is not active Driver Activation Cmd: "modprobe nouveau" Config Status: cfg=no, avail=yes, need=no, active=unknown Attached to: #8 (PCI bridge)
Primary display adapter: #9
Well, the info is there, but like you I don't know which of the "pci" strings there is the one.
My guess would be "0000:01:00.0"
So I tried to replace nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'` by nvidia_busid='0000:01:00.0' in the script, saved it with another name and run it. Then I couldn't log out anymore, I had to shutdown via konsole. After reboot the screen flickered heavily, and I had to boot into level 3 and run the script to put intel again. So, either that was not the bus-id, or I wrote it wrong, or... Problem is that I have no idea about these grep and sed things and if I used the correct kind of quote sign... What else can I try? -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 30.03.2018 um 20:33 schrieb Daniel Bauer:
So I tried to replace
nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'`
by
nvidia_busid='0000:01:00.0'
Now I tried to use nvidia_busid='01:00.0' This again leads to the error message "unable to query GPU information" Googling just shows me that this prime-thing doesn't work anywhere. Is it possible that I bought an expensive laptop and with opensuse can only have the graphics performance of a cheapest laptop? That would be a REAL showstopper... -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Daniel Bauer composed on 2018-03-30 21:11 (UTC+0200):
Now I tried to use nvidia_busid='01:00.0' This again leads to the error message "unable to query GPU information"
Googling just shows me that this prime-thing doesn't work anywhere. Is it possible that I bought an expensive laptop and with opensuse can only have the graphics performance of a cheapest laptop? That would be a REAL showstopper...
Did you try ArchLinux? https://wiki.archlinux.org/index.php/PRIME Took me long time to find: Asus GL552VW-CN113T https://www.asus.com/us/ROG-Republic-Of-Gamers/ROG-GL552VW/specifications/ -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Carlos E. R. composed on 2018-03-30 14:35 (UTC+0200):
On Friday, 2018-03-30 at 13:04 +0200, Daniel Bauer wrote:
I see, but I have no idea how to get that PCI-ID Hwinfo tells me so many things that I have no idea what to use :-) I uploaded a screenshot of the hardware infor screen: https://daniel-bauer.com/test/hwinfo_nvidia.jpg
If you manage that to generate in English, it would be easier for us ;-)
You could create this script:
/usr/local/bin/ingles:
#!/bin/sh LANG=en_US.UTF-8 \ LC_ALL=en_US.UTF-8 \ DICTIONARY=english \ KDE_LANG=en_US.UTF-8 \ exec "$@"
then call:
ingles hwinfo --gfxcard
When all you want is a PCI ID, that produces needlessly verbose output. Next time try: # lspci -nn | grep VGA 00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0402] (rev 06) or if you want driver information too: # lspci -nnk | grep -A3 VGA 00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0402] (rev 06) Subsystem: Micro-Star International Co., Ltd. [MSI] Device [1462:7850] Kernel driver in use: i915 Kernel modules: i915 Probably wouldn't need conversion to English this way. -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 30.03.2018 um 16:42 schrieb Felix Miata:
When all you want is a PCI ID, that produces needlessly verbose output. Next time try:
# lspci -nn | grep VGA 00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0402] (rev 06)
or if you want driver information too:
# lspci -nnk | grep -A3 VGA 00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0402] (rev 06) Subsystem: Micro-Star International Co., Ltd. [MSI] Device [1462:7850] Kernel driver in use: i915 Kernel modules: i915
Probably wouldn't need conversion to English this way.
This only gives me output for the intel graphics, not for the nvidida. I guess I'd have to replace "VGA" by "3D" or something. Anyway, I already posted the output according to Carlos' suggestion - and still don't know which is that PCI ID that the script needs and where and how to put it :-) -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Daniel Bauer composed on 2018-03-30 16:55 (UTC+0200):
Felix Miata composed:
When all you want is a PCI ID, that produces needlessly verbose output. Next time try:
# lspci -nn | grep VGA 00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0402] (rev 06)
or if you want driver information too:
# lspci -nnk | grep -A3 VGA 00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0402] (rev 06) Subsystem: Micro-Star International Co., Ltd. [MSI] Device [1462:7850] Kernel driver in use: i915 Kernel modules: i915
Probably wouldn't need conversion to English this way.
This only gives me output for the intel graphics, not for the nvidida. I guess I'd have to replace "VGA" by "3D" or something. Anyway, I already posted the output according to Carlos' suggestion - and still don't know which is that PCI ID that the script needs and where and how to put it :-)
The first characters of output are the PCI ID. [####:####] is the device ID, first part manufacturer code, second part device-specific. Change -A3 to something larger, e.g. -A6 or -A8, and the info for the NVidia should show up. man lspci /-A<Enter> -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
30.03.2018 14:04, Daniel Bauer пишет:
I had no time to play with my laptop, but here I am again...
Am 13.01.2018 um 06:58 schrieb Andrei Borzenkov:
12.01.2018 21:58, Daniel Bauer пишет:
... when entering
prime-select nvidia
I get: ERROR unable to query GPU information
The message comes from nvidia-xconfig --query-gpu-info. As far as I can tell, it actually expects nvidia kernel module to be loaded (it tries to load it itself). So you may try to debug this utility until it works.
That's far away of what my knowledge allows :-)
Looking at prime-select, it is using nvidia-xonfig only to determine PCI ID of nVidia card, so you ma actual hardcode them for debugging to see if it works otherwise.
I see, but I have no idea how to get that PCI-ID Hwinfo tells me so many things that I have no idea what to use :-) I uploaded a screenshot of the hardware infor screen: https://daniel-bauer.com/test/hwinfo_nvidia.jpg
prime-select is supposed to generate xorg.conf and enable nVidia specific GL libraries. You can check resulting configuration file and manually adjust it until it works.
So, I guess, next step is to find out the PCI ID, then somehow put it into that script, and then look if it works...
Can you help me: - how to find that PCI ID
Your screenshot shows 01:00:0 for PCI ID, which translates into PCI:1:0:0 xorg.conf BusID.
- how to put put it in that script (I mean which code to remove and which to add instead) ?
In prime-select.sh: gpu_info=`nvidia-xconfig --query-gpu-info` nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'` Just comment out both lines and put literal nvidia_busid="PCI:1:0:0" instead.
Or, if meanwhile there are some solutions on how to run prime-select on 42.3, I be happy of course to hear about. Googling didn't help me.
Thanks for further help!
Daniel
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 31.03.2018 um 06:57 schrieb Andrei Borzenkov:
In prime-select.sh:
gpu_info=`nvidia-xconfig --query-gpu-info` nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'`
Just comment out both lines and put literal
nvidia_busid="PCI:1:0:0"
instead.
Thank you. Unfortunately it doesn't work :-( Whether I put nvidia_busid="PCI:1:0:0" or nvidia_busid='01:00.0' after running the script I cannot logout anymore (message something like: KSM-LOGOUT-GREETER aborted), and after a reboot the screen flickers. I then reboot into init 3 and switch back to intel.. Any more things I can try? -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 31/03/2018 à 09:34, Daniel Bauer a écrit :
Any more things I can try?
I have a problem probably similar to yours (and I couldn't solve it and had no time to investigate more), may be this can be of some help? http://dodin.org/wiki/pmwiki.php?n=Doc.OptimusVideo jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
31.03.2018 10:34, Daniel Bauer пишет:
Am 31.03.2018 um 06:57 schrieb Andrei Borzenkov:
In prime-select.sh:
gpu_info=`nvidia-xconfig --query-gpu-info` nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'`
Just comment out both lines and put literal
nvidia_busid="PCI:1:0:0"
instead.
Thank you. Unfortunately it doesn't work :-(
Whether I put
nvidia_busid="PCI:1:0:0" or nvidia_busid='01:00.0'
after running the script I cannot logout anymore (message something like: KSM-LOGOUT-GREETER aborted), and after a reboot the screen flickers.
I then reboot into init 3 and switch back to intel..
Any more things I can try?
Well, showing Xorg.log would be a start I guess. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 31.03.2018 um 09:56 schrieb Andrei Borzenkov:
31.03.2018 10:34, Daniel Bauer пишет:
Am 31.03.2018 um 06:57 schrieb Andrei Borzenkov:
In prime-select.sh:
gpu_info=`nvidia-xconfig --query-gpu-info` nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'`
Just comment out both lines and put literal
nvidia_busid="PCI:1:0:0"
instead.
Thank you. Unfortunately it doesn't work :-(
Whether I put
nvidia_busid="PCI:1:0:0" or nvidia_busid='01:00.0'
after running the script I cannot logout anymore (message something like: KSM-LOGOUT-GREETER aborted), and after a reboot the screen flickers.
I then reboot into init 3 and switch back to intel..
Any more things I can try?
Well, showing Xorg.log would be a start I guess.
Here it is: https://www.daniel-bauer.com/test/Xorg.0.log -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
31.03.2018 11:41, Daniel Bauer пишет:
Am 31.03.2018 um 09:56 schrieb Andrei Borzenkov:
31.03.2018 10:34, Daniel Bauer пишет:
Am 31.03.2018 um 06:57 schrieb Andrei Borzenkov:
In prime-select.sh:
gpu_info=`nvidia-xconfig --query-gpu-info` nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'`
Just comment out both lines and put literal
nvidia_busid="PCI:1:0:0"
instead.
Thank you. Unfortunately it doesn't work :-(
Whether I put
nvidia_busid="PCI:1:0:0" or nvidia_busid='01:00.0'
after running the script I cannot logout anymore (message something like: KSM-LOGOUT-GREETER aborted), and after a reboot the screen flickers.
I then reboot into init 3 and switch back to intel..
Any more things I can try?
Well, showing Xorg.log would be a start I guess.
Here it is: https://www.daniel-bauer.com/test/Xorg.0.log
It does not show any trace of nVidia drivers being loaded at all. It also is using modesetting driver while suse-prime should configure Xorg to use Intel. It does not look like this log was collected after switch to nVidia. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 31.03.2018 um 10:50 schrieb Andrei Borzenkov:
31.03.2018 11:41, Daniel Bauer пишет:
Am 31.03.2018 um 09:56 schrieb Andrei Borzenkov:
31.03.2018 10:34, Daniel Bauer пишет:
Am 31.03.2018 um 06:57 schrieb Andrei Borzenkov:
In prime-select.sh:
gpu_info=`nvidia-xconfig --query-gpu-info` nvidia_busid=`echo "$gpu_info" |grep -i "PCI BusID"|sed 's/PCI BusID ://'|sed 's/ //g'`
Just comment out both lines and put literal
nvidia_busid="PCI:1:0:0"
instead.
Thank you. Unfortunately it doesn't work :-(
Whether I put
nvidia_busid="PCI:1:0:0" or nvidia_busid='01:00.0'
after running the script I cannot logout anymore (message something like: KSM-LOGOUT-GREETER aborted), and after a reboot the screen flickers.
I then reboot into init 3 and switch back to intel..
Any more things I can try?
Well, showing Xorg.log would be a start I guess.
Here it is: https://www.daniel-bauer.com/test/Xorg.0.log
It does not show any trace of nVidia drivers being loaded at all. It also is using modesetting driver while suse-prime should configure Xorg to use Intel. It does not look like this log was collected after switch to nVidia.
That's what I thought. Problem is that when I try to switch to nvidia (run the script), this does not trigger any new lines in the xorg log. But after running the script to switch to nvidia I cannot boot anymore. The screen flickers before any login is possible. So I tried the following: - run the script to change to nvidia - shutdown, tried to boot and waited for the flickering screen - reset, booted into level 3 - copied Xorg.0.log, so that it does not get overwritten the next boot This log yes talks about nvidia: https://www.daniel-bauer.com/test/Xorg.save It says I should look at system's kernel log for additional error messages, but I don't know how... Maybe this log helps a bit further? -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-03-31 11:48, Daniel Bauer wrote:
It does not show any trace of nVidia drivers being loaded at all. It also is using modesetting driver while suse-prime should configure Xorg to use Intel. It does not look like this log was collected after switch to nVidia.
That's what I thought.
Problem is that when I try to switch to nvidia (run the script), this does not trigger any new lines in the xorg log.
But after running the script to switch to nvidia I cannot boot anymore. The screen flickers before any login is possible.
So I tried the following: - run the script to change to nvidia - shutdown, tried to boot and waited for the flickering screen - reset, booted into level 3 - copied Xorg.0.log, so that it does not get overwritten the next boot
Or boot normally and get /var/log/Xorg.0.log.old :-)
This log yes talks about nvidia: https://www.daniel-bauer.com/test/Xorg.save
It says I should look at system's kernel log for additional error messages, but I don't know how...
Maybe this log helps a bit further?
[ 32.077] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the [ 32.077] (EE) NVIDIA: system's kernel log for additional error messages and [ 32.077] (EE) NVIDIA: consult the NVIDIA README for details. So you should look at /var/log/messages if it exists, or journalctl if it doesn't. Time stamp around this time: [ 32.038] (==) Log file: "/var/log/Xorg.0.log", Time: Sat Mar 31 11:34:15 2018 -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
Am 31.03.2018 um 14:49 schrieb Carlos E. R.:
On 2018-03-31 11:48, Daniel Bauer wrote:
It does not show any trace of nVidia drivers being loaded at all. It also is using modesetting driver while suse-prime should configure Xorg to use Intel. It does not look like this log was collected after switch to nVidia.
That's what I thought.
Problem is that when I try to switch to nvidia (run the script), this does not trigger any new lines in the xorg log.
But after running the script to switch to nvidia I cannot boot anymore. The screen flickers before any login is possible.
So I tried the following: - run the script to change to nvidia - shutdown, tried to boot and waited for the flickering screen - reset, booted into level 3 - copied Xorg.0.log, so that it does not get overwritten the next boot
Or boot normally and get /var/log/Xorg.0.log.old :-)
This log yes talks about nvidia: https://www.daniel-bauer.com/test/Xorg.save
It says I should look at system's kernel log for additional error messages, but I don't know how...
Maybe this log helps a bit further?
[ 32.077] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the [ 32.077] (EE) NVIDIA: system's kernel log for additional error messages and [ 32.077] (EE) NVIDIA: consult the NVIDIA README for details.
So you should look at /var/log/messages if it exists, or journalctl if it doesn't. Time stamp around this time:
[ 32.038] (==) Log file: "/var/log/Xorg.0.log", Time: Sat Mar 31 11:34:15 2018
So I hope I extracted from var/log/messages that session that tried to boot with nvidia: https://www.daniel-bauer.com/test/messages.txt I guess, the problem reports start here: Mar 31 11:33:58 meitli systemd[1]: Started X Display Manager. But looking at these lines for me they could be written in Chinese, too, they just don't say anything to me. Maybe to somebody here? -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
31.03.2018 17:16, Daniel Bauer пишет:
Am 31.03.2018 um 14:49 schrieb Carlos E. R.:
On 2018-03-31 11:48, Daniel Bauer wrote:
It does not show any trace of nVidia drivers being loaded at all. It also is using modesetting driver while suse-prime should configure Xorg to use Intel. It does not look like this log was collected after switch to nVidia.
That's what I thought.
Problem is that when I try to switch to nvidia (run the script), this does not trigger any new lines in the xorg log.
But after running the script to switch to nvidia I cannot boot anymore. The screen flickers before any login is possible.
So I tried the following: - run the script to change to nvidia - shutdown, tried to boot and waited for the flickering screen - reset, booted into level 3 - copied Xorg.0.log, so that it does not get overwritten the next boot
Or boot normally and get /var/log/Xorg.0.log.old :-)
This log yes talks about nvidia: https://www.daniel-bauer.com/test/Xorg.save
It says I should look at system's kernel log for additional error messages, but I don't know how...
Maybe this log helps a bit further?
[ 32.077] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the [ 32.077] (EE) NVIDIA: system's kernel log for additional error messages and [ 32.077] (EE) NVIDIA: consult the NVIDIA README for details.
So you should look at /var/log/messages if it exists, or journalctl if it doesn't. Time stamp around this time:
[ 32.038] (==) Log file: "/var/log/Xorg.0.log", Time: Sat Mar 31 11:34:15 2018
So I hope I extracted from var/log/messages that session that tried to boot with nvidia: https://www.daniel-bauer.com/test/messages.txt
It does not show anything related to nvidia kernel module. Before you can do anything with nvidia PRIME you must have necessary kernel modules up and running.
I guess, the problem reports start here: Mar 31 11:33:58 meitli systemd[1]: Started X Display Manager.
But looking at these lines for me they could be written in Chinese, too, they just don't say anything to me. Maybe to somebody here?
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 31.03.2018 um 18:31 schrieb Andrei Borzenkov:
It does not show anything related to nvidia kernel module. Before you can do anything with nvidia PRIME you must have necessary kernel modules up and running.
Yes, I've wondered about that, too. But shouldn't the script make me have the modules up and running? Before I updated to 42.3 I just had to run that script and it changed either to nvidia or to intel. I had nothing more to do. To be sure I have the recent drivers, this morning I even re-installed all nvidia things with Yast... I've uploaded a new messages (already mentioned in other post) -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
31.03.2018 20:58, Daniel Bauer пишет:
Am 31.03.2018 um 18:31 schrieb Andrei Borzenkov:
It does not show anything related to nvidia kernel module. Before you can do anything with nvidia PRIME you must have necessary kernel modules up and running.
Yes, I've wondered about that, too. But shouldn't the script make me have the modules up and running?
If "script" means prime-select.sh - of course not. The only thing script does is to configure user space (Xorg and libraries). It does not touch kernel driver at all.
Before I updated to 42.3 I just had to run that script and it changed either to nvidia or to intel. I had nothing more to do. To be sure I have the recent drivers, this morning I even re-installed all nvidia things with Yast...
I've uploaded a new messages (already mentioned in other post)
There is still no trace of nvidia kernel module there. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 31.03.2018 um 21:24 schrieb Andrei Borzenkov:
31.03.2018 20:58, Daniel Bauer пишет:
Am 31.03.2018 um 18:31 schrieb Andrei Borzenkov:
It does not show anything related to nvidia kernel module. Before you can do anything with nvidia PRIME you must have necessary kernel modules up and running.
Yes, I've wondered about that, too. But shouldn't the script make me have the modules up and running?
If "script" means prime-select.sh - of course not. The only thing script does is to configure user space (Xorg and libraries). It does not touch kernel driver at all.
Ok. And what can I do then? And why were the modules there before and now not anymore?
Before I updated to 42.3 I just had to run that script and it changed either to nvidia or to intel. I had nothing more to do. To be sure I have the recent drivers, this morning I even re-installed all nvidia things with Yast...
I've uploaded a new messages (already mentioned in other post)
There is still no trace of nvidia kernel module there.
I know. What can i do that nvidia kernel module leaves trace? -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
31.03.2018 22:34, Daniel Bauer пишет:
Am 31.03.2018 um 21:24 schrieb Andrei Borzenkov:
31.03.2018 20:58, Daniel Bauer пишет:
Am 31.03.2018 um 18:31 schrieb Andrei Borzenkov:
It does not show anything related to nvidia kernel module. Before you can do anything with nvidia PRIME you must have necessary kernel modules up and running.
Yes, I've wondered about that, too. But shouldn't the script make me have the modules up and running?
If "script" means prime-select.sh - of course not. The only thing script does is to configure user space (Xorg and libraries). It does not touch kernel driver at all.
Ok. And what can I do then? And why were the modules there before and now not anymore?
Before I updated to 42.3 I just had to run that script and it changed either to nvidia or to intel. I had nothing more to do. To be sure I have the recent drivers, this morning I even re-installed all nvidia things with Yast...
I've uploaded a new messages (already mentioned in other post)
There is still no trace of nvidia kernel module there.
I know. What can i do that nvidia kernel module leaves trace?
Well, start troubleshooting it. Those are standard steps unrelated to sue-prime. Do you have the correct version matching your hardware? Do you have module at all (modinfo nvidia)? Is it loaded (lsmod)? Can you load it manually (modprobe nvidia)? If no, what is error message and messages in dmesg immediately after this command? If yes, is it probably blacklisted (/etc/modprobe.conf, /etc/modprobe.d)? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 01.04.2018 um 08:07 schrieb Andrei Borzenkov:
Well, start troubleshooting it. Those are standard steps unrelated to sue-prime.
Do you have the correct version matching your hardware? Do you have module at all (modinfo nvidia)? Is it loaded (lsmod)? Can you load it manually (modprobe nvidia)? If no, what is error message and messages in dmesg immediately after this command? If yes, is it probably blacklisted (/etc/modprobe.conf, /etc/modprobe.d)?
Ok, thank you. Let's see, if I got one step further... First: When I upgraded to 42.3 there was no nvidia repo for 42.3, so I left it with 42.2. No there is this repo, I changed to it, and updated all nvidia. It still doesn't work, but some things chanmged , at least.. Because of the long output of the commands I've put the in textfiles on my server: the module is here: modinfo nvidia: https://www.daniel-bauer.com/test/modinfo.txt it seems loaded: lsmod: https://www.daniel-bauer.com/test/lsmod.txt (as it seems loaded I did not try "modprobe nvidia") There is no /etc/modprobe.conf file or directory here. In the folder /etc/modprobe.d there is /etc/modprobe.d/50-blacklist.conf that contains "blacklist nvidiafb" /etc/modprobe.d/50-nvidia-default.conf that contains "options nvidia NVreg_DeviceFileUID=0 NVreg_DeviceFileGID=33 NVreg_DeviceFileMode=0660 install nvidia PATH=$PATH:/bin:/usr/bin; if /sbin/modprobe --ignore-install nvidia; then if /sbin/modprobe nvidia_uvm; then if [ ! -c /dev/nvidia-uvm ]; then mknod -m 660 /dev/nvidia-uvm c $(cat /proc/devices | while read major device; do if [ "$device" == "nvidia-uvm" ]; then echo $major; break; fi ; done) 0; chown :video /dev/nvidia-uvm; fi; fi; if [ ! -c /dev/nvidiactl ]; then mknod -m 660 /dev/nvidiactl c 195 255; chown :video /dev/nvidiactl; fi; devid=-1; for dev in $(ls -d /sys/bus/pci/devices/*); do vendorid=$(cat $dev/vendor); if [ "$vendorid" == "0x10de" ]; then class=$(cat $dev/class); classid=${class%%00}; if [ "$classid" == "0x0300" -o "$classid" == "0x0302" ]; then devid=$((devid+1)); if [ ! -c /dev/nvidia${devid} ]; then mknod -m 660 /dev/nvidia${devid} c 195 ${devid}; chown :video /dev/nvidia${devid}; fi; fi; fi; done; /sbin/modprobe nvidia_drm; if [ ! -c /dev/nvidia-modeset ]; then mknod -m 660 /dev/nvidia-modeset c 195 254; chown :video /dev/nvidia-modeset; fi; fi" /etc/modprobe.d/nvidia-default.conf that contains only "blacklist nouveau" ------------ So, once again, now with the new drivers, I run the script to switch to nvidia, and again could not log out. I "shutdown -r now" and when booting the system "freezed" at some point, i.e. just nothing happened anymore: the screen with the boot messages upt to this point just staid there and nothing more. I booted with init 3 and copied the Xorg-files: Xorg.0.log: https://www.daniel-bauer.com/test/Xorg.save1.txt (which I guess is the one of the failed boot?) Xorg.0.log.old: https://www.daniel-bauer.com/test/Xorg.save.old1.txt (which I guess is the one of the former "intel session" in which I run the switch-to-nvidia script) /var/log/messages: https://www.daniel-bauer.com/test/messages3.txt - begins after the "shutdown"-command - failed boot, pressed power button - boot into level 3, run script to switch to intel, shutdown - normal boot Now there's a lot of Nvidia-messages, but none that sais something useful /to me/... What shall I do/try now? -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-04-02 11:06, Daniel Bauer wrote:
------------ So, once again, now with the new drivers, I run the script to switch to nvidia, and again could not log out. I "shutdown -r now" and when booting the system "freezed" at some point, i.e. just nothing happened anymore: the screen with the boot messages upt to this point just staid there and nothing more.
I booted with init 3 and copied the Xorg-files:
Xorg.0.log: https://www.daniel-bauer.com/test/Xorg.save1.txt (which I guess is the one of the failed boot?)
According to this one, there is no problem. Normal closing.
Xorg.0.log.old: https://www.daniel-bauer.com/test/Xorg.save.old1.txt (which I guess is the one of the former "intel session" in which I run the switch-to-nvidia script)
Both logs seems to be the same thing. This is mine - stand alone nvidia card: [ 1397.901] (--) Depth 24 pixmap format is 32 bpp [ 1397.901] (II) NVIDIA: Using 768.00 MB of virtual memory for indirect memory access. [ 1397.904] (II) NVIDIA(0): ACPI: failed to connect to the ACPI event daemon; the daemon [ 1397.904] (II) NVIDIA(0): may not be running or the "AcpidSocketPath" X [ 1397.904] (II) NVIDIA(0): configuration option may not be set correctly. When the [ 1397.904] (II) NVIDIA(0): ACPI event daemon is available, the NVIDIA X driver will [ 1397.904] (II) NVIDIA(0): try to use it to receive ACPI event notifications. For [ 1397.904] (II) NVIDIA(0): details, please see the "ConnectToAcpid" and [ 1397.904] (II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X [ 1397.904] (II) NVIDIA(0): Config Options in the README. [ 1397.907] (II) NVIDIA(0): Setting mode "DFP-0:nvidia-auto-select" [ 1397.966] (==) NVIDIA(0): Disabling shared memory pixmaps [ 1397.966] (==) NVIDIA(0): Backing store enabled [ 1397.966] (==) NVIDIA(0): Silken mouse enabled [ 1397.966] (**) NVIDIA(0): DPMS enabled [ 1397.966] (II) Loading sub module "dri2" [ 1397.966] (II) LoadModule: "dri2" [ 1397.966] (II) Module "dri2" already built-in [ 1397.966] (II) NVIDIA(0): [DRI2] Setup complete [ 1397.966] (II) NVIDIA(0): [DRI2] VDPAU driver: nvidia [ 1397.966] (--) RandR disabled [ 1397.970] (II) Initializing extension GLX [ 1397.970] (II) Indirect GLX disabled.(II) config/udev: Adding input device Power Button (/dev/input/event1) This is your Xorg.0.log: [ 19.908] (II) NVIDIA: Using 6144.00 MB of virtual memory for indirect memory [ 19.908] (II) NVIDIA: access. [ 19.913] (II) NVIDIA(0): ACPI: failed to connect to the ACPI event daemon; the daemon [ 19.913] (II) NVIDIA(0): may not be running or the "AcpidSocketPath" X [ 19.913] (II) NVIDIA(0): configuration option may not be set correctly. When the [ 19.913] (II) NVIDIA(0): ACPI event daemon is available, the NVIDIA X driver will [ 19.913] (II) NVIDIA(0): try to use it to receive ACPI event notifications. For [ 19.913] (II) NVIDIA(0): details, please see the "ConnectToAcpid" and [ 19.913] (II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X [ 19.913] (II) NVIDIA(0): Config Options in the README. [ 19.937] (II) NVIDIA(0): Setting mode "NULL" [ 19.939] (==) NVIDIA(0): Disabling shared memory pixmaps [ 19.940] (==) NVIDIA(0): Backing store enabled [ 19.940] (==) NVIDIA(0): Silken mouse enabled [ 19.940] (==) NVIDIA(0): DPMS enabled [ 19.940] (II) Loading sub module "dri2" [ 19.940] (II) LoadModule: "dri2" [ 19.940] (II) Module "dri2" already built-in [ 19.940] (II) NVIDIA(0): [DRI2] Setup complete [ 19.940] (II) NVIDIA(0): [DRI2] VDPAU driver: nvidia [ 19.940] (--) RandR disabled [ 19.943] (II) Initializing extension GLX [ 19.943] (II) Indirect GLX disabled. This is your Xorg.0.log.old: [ 1351.989] (II) NVIDIA: Using 6144.00 MB of virtual memory for indirect memory [ 1351.989] (II) NVIDIA: access. [ 1351.992] (II) NVIDIA(0): ACPI: failed to connect to the ACPI event daemon; the daemon [ 1351.992] (II) NVIDIA(0): may not be running or the "AcpidSocketPath" X [ 1351.992] (II) NVIDIA(0): configuration option may not be set correctly. When the [ 1351.992] (II) NVIDIA(0): ACPI event daemon is available, the NVIDIA X driver will [ 1351.992] (II) NVIDIA(0): try to use it to receive ACPI event notifications. For [ 1351.992] (II) NVIDIA(0): details, please see the "ConnectToAcpid" and [ 1351.992] (II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X [ 1351.992] (II) NVIDIA(0): Config Options in the README. [ 1352.010] (II) NVIDIA(0): Setting mode "NULL" [ 1352.012] (==) NVIDIA(0): Disabling shared memory pixmaps [ 1352.012] (==) NVIDIA(0): Backing store enabled [ 1352.012] (==) NVIDIA(0): Silken mouse enabled [ 1352.013] (==) NVIDIA(0): DPMS enabled [ 1352.013] (II) Loading sub module "dri2" [ 1352.013] (II) LoadModule: "dri2" [ 1352.013] (II) Module "dri2" already built-in [ 1352.013] (II) NVIDIA(0): [DRI2] Setup complete [ 1352.013] (II) NVIDIA(0): [DRI2] VDPAU driver: nvidia [ 1352.013] (--) RandR disabled [ 1352.015] (II) Initializing extension GLX [ 1352.015] (II) Indirect GLX disabled.
What shall I do/try now?
Isn't there a wiki page at openSUSE that explains how to use these hybrids? -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
On 2018-04-02 11:06, Daniel Bauer wrote:
------------ So, once again, now with the new drivers, I run the script to switch to nvidia, and again could not log out. I "shutdown -r now" and when booting the system "freezed" at some point, i.e. just nothing happened anymore: the screen with the boot messages upt to this point just staid there and nothing more.
I booted with init 3 and copied the Xorg-files:
Xorg.0.log: https://www.daniel-bauer.com/test/Xorg.save1.txt (which I guess is the one of the failed boot?)
According to this one, there is no problem. Normal closing.
Xorg.0.log.old: https://www.daniel-bauer.com/test/Xorg.save.old1.txt (which I guess is the one of the former "intel session" in which I run the switch-to-nvidia script)
Both logs seems to be the same thing.
This is mine - stand alone nvidia card:
[ 1397.901] (--) Depth 24 pixmap format is 32 bpp [ 1397.901] (II) NVIDIA: Using 768.00 MB of virtual memory for indirect memory access. [ 1397.904] (II) NVIDIA(0): ACPI: failed to connect to the ACPI event daemon; the daemon [ 1397.904] (II) NVIDIA(0): may not be running or the "AcpidSocketPath" X [ 1397.904] (II) NVIDIA(0): configuration option may not be set correctly. When the [ 1397.904] (II) NVIDIA(0): ACPI event daemon is available, the NVIDIA X driver will [ 1397.904] (II) NVIDIA(0): try to use it to receive ACPI event notifications. For [ 1397.904] (II) NVIDIA(0): details, please see the "ConnectToAcpid" and [ 1397.904] (II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X [ 1397.904] (II) NVIDIA(0): Config Options in the README. [ 1397.907] (II) NVIDIA(0): Setting mode "DFP-0:nvidia-auto-select" [ 1397.966] (==) NVIDIA(0): Disabling shared memory pixmaps [ 1397.966] (==) NVIDIA(0): Backing store enabled [ 1397.966] (==) NVIDIA(0): Silken mouse enabled [ 1397.966] (**) NVIDIA(0): DPMS enabled [ 1397.966] (II) Loading sub module "dri2" [ 1397.966] (II) LoadModule: "dri2" [ 1397.966] (II) Module "dri2" already built-in [ 1397.966] (II) NVIDIA(0): [DRI2] Setup complete [ 1397.966] (II) NVIDIA(0): [DRI2] VDPAU driver: nvidia [ 1397.966] (--) RandR disabled [ 1397.970] (II) Initializing extension GLX [ 1397.970] (II) Indirect GLX disabled.(II) config/udev: Adding input device Power Button (/dev/input/event1)
This is your Xorg.0.log:
[ 19.908] (II) NVIDIA: Using 6144.00 MB of virtual memory for indirect memory [ 19.908] (II) NVIDIA: access. [ 19.913] (II) NVIDIA(0): ACPI: failed to connect to the ACPI event daemon; the daemon [ 19.913] (II) NVIDIA(0): may not be running or the "AcpidSocketPath" X [ 19.913] (II) NVIDIA(0): configuration option may not be set correctly. When the [ 19.913] (II) NVIDIA(0): ACPI event daemon is available, the NVIDIA X driver will [ 19.913] (II) NVIDIA(0): try to use it to receive ACPI event notifications. For [ 19.913] (II) NVIDIA(0): details, please see the "ConnectToAcpid" and [ 19.913] (II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X [ 19.913] (II) NVIDIA(0): Config Options in the README. [ 19.937] (II) NVIDIA(0): Setting mode "NULL" [ 19.939] (==) NVIDIA(0): Disabling shared memory pixmaps [ 19.940] (==) NVIDIA(0): Backing store enabled [ 19.940] (==) NVIDIA(0): Silken mouse enabled [ 19.940] (==) NVIDIA(0): DPMS enabled [ 19.940] (II) Loading sub module "dri2" [ 19.940] (II) LoadModule: "dri2" [ 19.940] (II) Module "dri2" already built-in [ 19.940] (II) NVIDIA(0): [DRI2] Setup complete [ 19.940] (II) NVIDIA(0): [DRI2] VDPAU driver: nvidia [ 19.940] (--) RandR disabled [ 19.943] (II) Initializing extension GLX [ 19.943] (II) Indirect GLX disabled.
This is your Xorg.0.log.old:
[ 1351.989] (II) NVIDIA: Using 6144.00 MB of virtual memory for indirect memory [ 1351.989] (II) NVIDIA: access. [ 1351.992] (II) NVIDIA(0): ACPI: failed to connect to the ACPI event daemon; the daemon [ 1351.992] (II) NVIDIA(0): may not be running or the "AcpidSocketPath" X [ 1351.992] (II) NVIDIA(0): configuration option may not be set correctly. When the [ 1351.992] (II) NVIDIA(0): ACPI event daemon is available, the NVIDIA X driver will [ 1351.992] (II) NVIDIA(0): try to use it to receive ACPI event notifications. For [ 1351.992] (II) NVIDIA(0): details, please see the "ConnectToAcpid" and [ 1351.992] (II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X [ 1351.992] (II) NVIDIA(0): Config Options in the README. [ 1352.010] (II) NVIDIA(0): Setting mode "NULL" [ 1352.012] (==) NVIDIA(0): Disabling shared memory pixmaps [ 1352.012] (==) NVIDIA(0): Backing store enabled [ 1352.012] (==) NVIDIA(0): Silken mouse enabled [ 1352.013] (==) NVIDIA(0): DPMS enabled [ 1352.013] (II) Loading sub module "dri2" [ 1352.013] (II) LoadModule: "dri2" [ 1352.013] (II) Module "dri2" already built-in [ 1352.013] (II) NVIDIA(0): [DRI2] Setup complete [ 1352.013] (II) NVIDIA(0): [DRI2] VDPAU driver: nvidia [ 1352.013] (--) RandR disabled [ 1352.015] (II) Initializing extension GLX [ 1352.015] (II) Indirect GLX disabled.
What shall I do/try now?
Isn't there a wiki page at openSUSE that explains how to use these hybrids? There is on Bumblebee, not on Prime. The packages were introduced by Bo Simonsen, but he seems no longer around, otherwise he would have shown up in
Op maandag 2 april 2018 12:44:45 CEST schreef Carlos E. R.: this thread. -- Gertjan Lettink a.k.a. Knurpht openSUSE Board Member openSUSE Forums Team -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 02.04.2018 um 15:30 schrieb Knurpht @ openSUSE:
Op maandag 2 april 2018 12:44:45 CEST schreef Carlos E. R.:
Isn't there a wiki page at openSUSE that explains how to use these hybrids?
There is on Bumblebee, not on Prime. The packages were introduced by Bo Simonsen, but he seems no longer around, otherwise he would have shown up in this thread.
Yes, I am googling all the time, but either the information is outdated (there was no problem before 42.3...) or it is contradictory and I don't know what to "believe" - despite of the fact, that most of the things I find I simply don't understand.... The Bumblebee thing seems to me too complicated, also considering future updates. Furthermore I read that it's quite slow, and above all, google also find problems with 42.3... Prime-select was wonderful: simply switch between one and the other. If I only knew why it doesn't work anymore now... -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-04-02 15:30, Knurpht @ openSUSE wrote:
Op maandag 2 april 2018 12:44:45 CEST schreef Carlos E. R.:
On 2018-04-02 11:06, Daniel Bauer wrote:
Isn't there a wiki page at openSUSE that explains how to use these hybrids? There is on Bumblebee, not on Prime. The packages were introduced by Bo Simonsen, but he seems no longer around, otherwise he would have shown up in this thread.
I'm a bit confused, but as I'm on the search for a small laptop for travel, I'm interested. Bumblebee and Prime are different solutions to handle laptops with hibrid nvidia/intel video hardware, right? But the hardware is the same, I understand. What is the best current software solution? Methinks, I will search for a laptop without enhanced graphics. Less trouble, if I can find such a machine. But is is a problem for Daniel... he needs good graphics. -- Cheers/Saludos Carlos E. R. (testing openSUSE Leap 15.0, at Minas-Anor) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
04.04.2018 22:57, Carlos E. R. пишет:
Methinks, I will search for a laptop without enhanced graphics. Less trouble, if I can find such a machine.
In 3+ years of having new notebook I never used nVidia dGPU. I settled on nouveau simply to keep it powered off without using out of tree kernel module. I would probably use it for video, but a) does not work with bumblebee b) extremely hard to setup with nouveau, needs extra firmware and drivers c) does not work with nVidia binary drivers using DRI_PRIME and I do not want full switch to dGPU (that is main topic of this thread) So it is just one more piece of hardware that is more or less useless but does not harm. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 04.04.2018 um 21:57 schrieb Carlos E. R.:
On 2018-04-02 15:30, Knurpht @ openSUSE wrote:
Op maandag 2 april 2018 12:44:45 CEST schreef Carlos E. R.:
On 2018-04-02 11:06, Daniel Bauer wrote:
Isn't there a wiki page at openSUSE that explains how to use these hybrids? There is on Bumblebee, not on Prime. The packages were introduced by Bo Simonsen, but he seems no longer around, otherwise he would have shown up in this thread.
I'm a bit confused, but as I'm on the search for a small laptop for travel, I'm interested.
Bumblebee and Prime are different solutions to handle laptops with hibrid nvidia/intel video hardware, right?
As much as I understand, Prime switches completely to the enhanced graphics (what I want), while bumblebee adds like a "virtual second X session running additionally" (excuse the very poor amateur explanation) for just a single program you start with bumblebee, thus reducing efficiency. Furthermore it is quite complicated to start a program using bumblebee with a click on a desktop symbol or a dolphin entry. You more or less must start everything bumblebee from console.
But the hardware is the same, I understand. What is the best current software solution?
From googling I understand that bumblebee (at least) on opensuse 42.3 now also has severe problems. So the best software solution is a windows or mac computer - I think under opensuse there is simply none. At least now.
Methinks, I will search for a laptop without enhanced graphics. Less trouble, if I can find such a machine.
Yes, the enhanced graphics is absolutely useless for now. It doesn't harm your computer, it is just there doing nothing but enhancing the price of the laptop. You can just ignore it as a feature and look at the rest of the machine.
But is is a problem for Daniel... he needs good graphics.
It is /really/ a problem. It was working before. It does not work anymore. I bought an expensive computer for nothing. I can use it for email and internet - all that I can do with my mobile phone, too. The regression made my laptop worthless for what I actually bought it. It's like flushing money into the toilet. (If you think I am frustrated by this: you got it :-) ) -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Thu, Apr 5, 2018 at 12:26 PM, Daniel Bauer <linux@daniel-bauer.com> wrote:
Bumblebee and Prime are different solutions to handle laptops with hibrid nvidia/intel video hardware, right?
As much as I understand, Prime switches completely to the enhanced graphics (what I want), while bumblebee adds like a "virtual second X session running additionally" (excuse the very poor amateur explanation) for just a single program you start with bumblebee, thus reducing efficiency.
That's correct for nVidia binary drivers. For KMS drivers (modesetting/nouveau/intel) there is now bumblebee-like support for offloading single program to discrete GPU. It does not require completely separate X server and is more efficient than bumblebee, but it won't work with nVidia drivers for several reasons. ...
It is /really/ a problem. It was working before. It does not work anymore. I bought an expensive computer for nothing. I can use it for email and internet - all that I can do with my mobile phone, too. The regression made my laptop worthless for what I actually bought it. It's like flushing money into the toilet.
I sort of confirm it for Leap 42.3 even without nVidia drivers. I do not have physical system to play with so I configured PCI pass through into KVM VM with Leap 42.3. Attempting to configure PRIME offload from nVidia into virgl results in black screen. X server is started and apparently does everything (I can check with xrandr that it sees both adapters and sets up source offload) but there is no pictures on screen. As soon as I try to run "xrandr --auto" to make it re-detect displays kernel crashes ... Now my case may be complicated by the fact that I'm using VM and virgl adapter and I have no idea if it even supposed to work. OTOH in Internet you can find various hints that doing it with DRI2 (that I have) may be problematic and may require some application support like special composer. I'm not intimately familiar with these topics. I probably will test with Tumbleweed/GNOME next. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 05/04/2018 à 11:47, Andrei Borzenkov a écrit :
That's correct for nVidia binary drivers. For KMS drivers (modesetting/nouveau/intel) there is now bumblebee-like support for offloading single program to discrete GPU. It does not require completely separate X server and is more efficient than bumblebee, but it won't work with nVidia drivers for several reasons.
do you have a link? I'm very interested thanks jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Thu, Apr 5, 2018 at 2:06 PM, jdd@dodin.org <jdd@dodin.org> wrote:
Le 05/04/2018 à 11:47, Andrei Borzenkov a écrit :
That's correct for nVidia binary drivers. For KMS drivers (modesetting/nouveau/intel) there is now bumblebee-like support for offloading single program to discrete GPU. It does not require completely separate X server and is more efficient than bumblebee, but it won't work with nVidia drivers for several reasons.
do you have a link? I'm very interested
I'm feeling like google expert ... https://nouveau.freedesktop.org/wiki/Optimus/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 05/04/2018 à 13:12, Andrei Borzenkov a écrit :
On Thu, Apr 5, 2018 at 2:06 PM, jdd@dodin.org <jdd@dodin.org> wrote:
Le 05/04/2018 à 11:47, Andrei Borzenkov a écrit :
That's correct for nVidia binary drivers. For KMS drivers (modesetting/nouveau/intel) there is now bumblebee-like support for offloading single program to discrete GPU. It does not require completely separate X server and is more efficient than bumblebee, but it won't work with nVidia drivers for several reasons.
do you have a link? I'm very interested
I'm feeling like google expert ...
I see no difference with https://en.opensuse.org/SDB:NVIDIA_Bumblebee http://dodin.org/wiki/pmwiki.php?n=Doc.OptimusVideo#toc8 and it work (or seems to) on my computer, with no change since several month except regular openSUSE updates. Bumblebee is said to run, optirun kdenlive start kdenlive, no immediate way to see if it's really better linux-owxt:/home/jdd # optirun --status Bumblebee status: Ready (3.2.1). X inactive. Discrete video card is off. linux-owxt:/home/jdd # optirun glxgears Running synchronized to the vertical refresh. The framerate should be approximately the same as the monitor refresh rate. 6240 frames in 5.0 seconds = 1247.935 FPS what I notice is that I can change # cat /proc/acpi/bbswitch 0000:01:00.0 OFF not even with vi (it's immediately reversed) best guess is that some bug was fixed in newer kernel (if it works) my asus computer is of the second kind as said here https://01.org/linuxgraphics/gfx-docs/drm/gpu/vga-switcheroo.html "Dual GPUs but only one of them is connected to outputs" thanks jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 05/04/2018 à 11:26, Daniel Bauer a écrit :
(If you think I am frustrated by this: you got it :-) )
not sure linux application can really take use of cuda processors, they already miss part of the multi processor features. this makes seeing effects on kdenlive much difficult no idea what Windows do with it jdd -- http://dodin.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Mon, Apr 2, 2018 at 12:06 PM, Daniel Bauer <linux@daniel-bauer.com> wrote:
Xorg.0.log: https://www.daniel-bauer.com/test/Xorg.save1.txt (which I guess is the one of the failed boot?)
This looks rather different from all previous logs. It shows that both Intel and nVidia Xorg drivers are correctly loaded. Now for PRIME you need to enable KMS support in nVidia with modeset=1 nvidia-drm module parameter. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 02.04.2018 um 13:07 schrieb Andrei Borzenkov:
On Mon, Apr 2, 2018 at 12:06 PM, Daniel Bauer <linux@daniel-bauer.com> wrote:
Xorg.0.log: https://www.daniel-bauer.com/test/Xorg.save1.txt (which I guess is the one of the failed boot?)
This looks rather different from all previous logs. It shows that both Intel and nVidia Xorg drivers are correctly loaded. Now for PRIME you need to enable KMS support in nVidia with modeset=1 nvidia-drm module parameter.
So, I booted with adding "nvidia-drm.modeset=1" (without quotes) to the end of the boot command line, then run the script to switch to nvidia. But still the same result :-( -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-03-31 16:16, Daniel Bauer wrote:
Am 31.03.2018 um 14:49 schrieb Carlos E. R.:
So you should look at /var/log/messages if it exists, or journalctl if it doesn't. Time stamp around this time:
[ 32.038] (==) Log file: "/var/log/Xorg.0.log", Time: Sat Mar 31 11:34:15 2018
So I hope I extracted from var/log/messages that session that tried to boot with nvidia: https://www.daniel-bauer.com/test/messages.txt
But your log ends precisely "Mar 31 11:34:14", should be just after that time. And does not contain the word "nvidia". -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
Am 31.03.2018 um 19:06 schrieb Carlos E. R.:
But your log ends precisely "Mar 31 11:34:14", should be just after that time. And does not contain the word "nvidia".
It ends there because that's when I pressed the rest button. So, I made a new test to have a more complete /var/log/messages: https://www.daniel-bauer.com/test/messages2.txt But: the word nvidia does not appear neither... What I did (where the log begins): - opened a console, logged in as root - run the script do switch to nvidia - "shutdown -r now" - waited until during the boot the screen began to flicker - pressed reset (power) button - booted with init 3 - logged in as root - run the script to switch to intel - "shutdown -r now" - booted normal and logged in as user A lot of lines :-) -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (6)
-
Andrei Borzenkov -
Carlos E. R. -
Daniel Bauer -
Felix Miata -
jdd＠dodin.org -
Knurpht ＠ openSUSE