kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

justm3 · 2020-03-15 07:40:45

Hi
Box with old nvidia GeForce GT 630 and nvidia-390xx-lts-1:390.132-26 drivers (now in aur)
Upgraded today kernel from linux-lts-5.4.24-1 to linux-lts-5.4.25-2.
It will not boot. Changed to tty2 and downgraded kernel. All ok.
I have not tried the nouveau drivers.

Is it a bug(known?) or something I missed?

V1del · 2020-03-15 10:14:09

Since the package is now in the AUR you need to remember to rebuild the nvidia-390xx-lts kernel module each time the kernel is updated.

If you don't want to do that yourself you could also install linux-lts-headers and install nvidia-390xx-dkms that way the kernel module will be automatically rebuilt on kernel updates.

TheSgtBilko · 2020-03-15 15:13:41

Why was nvidia-390xx pushed to AUR? According to NVidia it's supported throughout 2022 !

justm3 · 2020-03-15 16:00:35

Thanks for the reply.
I had tried the rebuild (which is why I mentioned the aur). It does not work. The system hangs at boot.
I will give it a try with dkms.

Thanks again.

seth · 2020-03-15 16:33:23

Also notice that the AUR package is not an all-in-one one, you'll need to install nvidia-390xx-utils from AUR as well (downloading the driver blob twice, running nvidias installer twice … sound like this should be a split package)

TheSgtBilko wrote:

Why was nvidia-390xx pushed to AUR?

Probably orphaned by the former maintainer, but "+1" - if there're no strong reasons this should probably return to the repos afap.
The GPU set is probably still prevalent and the package is interlinked with kernel updates and will just silently fail, failing its users GUI.

There's not even a warning on the homepage. Sound as if we'll get a bunch of related threads… sigh :-(

For the time being, I'd strongly recommend to use the dkms package (assuming the upstream update frequency isn't too high this will spare you re-installing the kernel package w/ every kernel and in this run a risk to break compatibility with the utils package (unless you update that every time as well), still breaking the GUI…

eschwartz · 2020-03-15 16:40:24

seth wrote:

Also notice that the AUR package is not an all-in-one one, you'll need to install nvidia-390xx-utils from AUR as well (downloading the driver blob twice, running nvidias installer twice … sound like this should be a split package)
TheSgtBilko wrote:
Why was nvidia-390xx pushed to AUR?
Probably orphaned by the former maintainer, but "+1" - if there're no strong reasons this should probably return to the repos afap.
The GPU set is probably still prevalent and the package is interlinked with kernel updates and will just silently fail, failing its users GUI.

Does this count as a strong reason? https://lists.archlinux.org/pipermail/a … 29895.html

(Note this thread started in July 2018!)

seth wrote:

There's not even a warning on the homepage. Sound as if we'll get a bunch of related threads… sigh :-(

Why should there be a news post that a package was dropped to the AUR?

seth wrote:

For the time being, I'd strongly recommend to use the dkms package (assuming the upstream update frequency isn't too high this will spare you re-installing the kernel package w/ every kernel and in this run a risk to break compatibility with the utils package (unless you update that every time as well), still breaking the GUI…

I would *strongly* recommend users never install AUR modules except as a dkms module. This one is not unusual in that regard.

Last edited by eschwartz (2020-03-15 16:41:32)

seth · 2020-03-15 16:52:04

Well, the package was (as assumed) orphaned. If that's the only reason and it sees a maintainer, there's no reason left to not have it in the repos, right?
(Maybe I should have stressed "other, strong, reason…")

Why should there be a news post that a package was dropped to the AUR?

As indicated, the stalled update is easily missed and will leave users with a broken GUI stack.
No, that should totally not be a problem for arch users, but you've been around for a while so let's please not pretend the forum reality would not exist. (Yes, maybe those users wouldn't read the new either)

One could argue that this requires manual intervention on kernel updates for nvidia-390xx users and manual intervention requirements on even fringe packages usually get a news post, don't they?

eschwartz · 2020-03-15 16:58:45

Why does it need manual intervention? The AUR maintainer of nvidia-390xx has bumped it for the new kernel version, no one has adopted the -lts package but someone should bump its pkgrel too. Then users would automatically rebuild it via an AUR helper, and there would be no problem. Alternatively, users can remove packages which are marked as foreign.

And, sure, if a maintainer is interested in maintaining it, it can and should be added back to the repos. But the reason we are in this situation in the first place is because we have been looking for a maintainer for 20 months and no one has stepped up... this is possibly the longest period of a maintainer actively looking and begging for people to take over an orphaned package, in the history of archlinux (or at least my memory of it).

seth · 2020-03-15 17:12:42

If I've installed the (non-dkms) repo package for the past couple of years, will I automatically install the AUR package with the next "pacman -Syu"? (that's a rethorical question…)
So I've to pick up the AUR package - manually - or face a "broken" system with the next kernel update (and then figure what's wrong, what to do, how to do it…)

To me, that's akin to https://www.archlinux.org/news/sshd-nee … nssh-82p1/ - just that rebooting won't fix it either.

justm3 · 2020-03-15 18:13:38

DKMS obviously worked. In my case, it took about 2 hours to build and get it to a finished state (this is not a personal computer-my wife uses it as well)

I mean no offense to all views and statements. But I would like to ask this question.

If I am a "regular" user, possibly updating once a week or month and need to do my work after the update, how on earth am I going to do that?

Thanks.

V1del · 2020-03-15 20:32:05

2 hours to build a dkms package? What did it do exactly? What are the rest of your system specs? On my admittedly beefy for the time skylake with 8 threads and 4.6Ghz it takes around five minutes per installed kernel.

FWIW, as long as the actual version of the nvidia-390xx-utils doesn't change you can continue to work on the old kernel as normal, with of course the usual caveat that new modules won't be loaded but that has always been a limitation.

paulkerry · 2020-03-15 21:03:01

justm3 wrote:

Box with old nvidia GeForce GT 630 and nvidia-390xx-lts-1:390.132-26 drivers (now in aur)

In case you are not aware, the GeForce GT 630 is supported by nvidia 440.64, so you should just be able to use the nvidia-lts package in extra.

See "Supported products" on...
https://www.nvidia.com/Download/driverR … 7462/en-us

Cheers
Paul.

V1del · 2020-03-15 21:16:21

Depends, I have seen many OEM provided 630 being plain rebranded 550s or so. An lspci/xorg log/dmesg would provide clarity here.

liewkj · 2020-03-15 23:55:05

Somehow, I found 390xx more stable than the latest drivers if you are still on GFxxx/GKxxx generation of GPUs.

justm3 · 2020-03-16 04:19:51

V1del wrote:

2 hours to build a dkms package? What did it do exactly? What are the rest of your system specs? On my admittedly beefy for the time skylake with 8 threads and 4.6Ghz it takes around five minutes per installed kernel...

Core2 Duo E4600 @ 2.40 & dimms at @667 with 3 spinners at raid10 assembly (or raid 1E if you like to call it)
That's way away from beefy, would you not say so?

paulkerry wrote:

In case you are not aware, the GeForce GT 630 is supported by nvidia 440.64, so you should just be able to use the nvidia-lts package in extra...

Thank you, I am aware. It does NOT work.

Thanks everyone for your contribution. I really appreciate it.
Be well

bwidlar · 2020-03-17 08:05:47

Same problem here, downgrading to kernel 5.5.8-arch1-1 works.

V1del · 2020-03-17 08:19:57

You didn't really read the thread if that's your conclusion... It's an AUR package, for the foreseeable future you need to rebuild the kernel package for kernel updates. Otherwise you are frozen on this kernel forever.

bwidlar · 2020-03-17 08:34:45

V1del wrote:

You didn't really read the thread if that's your conclusion... It's an AUR package, for the foreseeable future you need to rebuild the kernel package for kernel updates. Otherwise you are frozen on this kernel forever.

Dont worry, I read the thread. I just prefer this solution than the aur package.
Thanks for the advice.

loqs · 2020-03-17 17:58:30

justm3 wrote:

paulkerry wrote:
In case you are not aware, the GeForce GT 630 is supported by nvidia 440.64, so you should just be able to use the nvidia-lts package in extra...
Thank you, I am aware. It does NOT work.

What is the PID of the chip?

justm3 · 2020-03-18 05:59:21

loqs wrote:

What is the PID of the chip?

As much info as I can get

lshw -class display

01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GF108 [GeForce GT 630] [10de:0f00] (rev a1)

and

lspci -vv

01:00.0 VGA compatible controller: NVIDIA Corporation GF108 [GeForce GT 630] (rev a1) (prog-if 00 [VGA controller])
        Subsystem: Micro-Star International Co., Ltd. [MSI] GF108 [GeForce GT 630]
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0
        Interrupt: pin A routed to IRQ 31
        Region 0: Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
        Region 1: Memory at d0000000 (64-bit, prefetchable) [size=128M]
        Region 3: Memory at dc000000 (64-bit, prefetchable) [size=32M]
        Region 5: I/O ports at 9c00 [size=128]
        Expansion ROM at 000c0000 [virtual] [disabled] [size=128K]
        Capabilities: [60] Power Management version 3
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-
        Capabilities: [68] MSI: Enable+ Count=1/1 Maskable- 64bit+
                Address: 00000000fee02004  Data: 4029
        Capabilities: [78] Express (v2) Endpoint, MSI 00
                DevCap: MaxPayload 128 bytes, PhantFunc 0, Latency L0s unlimited, L1 <64us
                        ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset- SlotPowerLimit 0.000W
                DevCtl: CorrErr- NonFatalErr- FatalErr- UnsupReq-
                        RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+
                        MaxPayload 128 bytes, MaxReadReq 512 bytes
                DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-
                LnkCap: Port #0, Speed 2.5GT/s, Width x16, ASPM L0s L1, Exit Latency L0s <256ns, L1 <4us
                        ClockPM+ Surprise- LLActRep- BwNot- ASPMOptComp-
                LnkCtl: ASPM Disabled; RCB 64 bytes Disabled- CommClk+
                        ExtSynch- ClockPM+ AutWidDis- BWInt- AutBWInt-
                LnkSta: Speed 2.5GT/s (ok), Width x16 (ok)
                        TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
                DevCap2: Completion Timeout: Not Supported, TimeoutDis+, NROPrPrP-, LTR-
                         10BitTagComp-, 10BitTagReq-, OBFF Not Supported, ExtFmt-, EETLPPrefix-
                         EmergencyPowerReduction Not Supported, EmergencyPowerReductionInit-
                         FRS-, TPHComp-, ExtTPHComp-
                         AtomicOpsCap: 32bit- 64bit- 128bitCAS-
                DevCtl2: Completion Timeout: 50us to 50ms, TimeoutDis-, LTR-, OBFF Disabled
                         AtomicOpsCtl: ReqEn-
                LnkCtl2: Target Link Speed: 2.5GT/s, EnterCompliance- SpeedDis-
                         Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
                         Compliance De-emphasis: -6dB
                LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete-, EqualizationPhase1-
                         EqualizationPhase2-, EqualizationPhase3-, LinkEqualizationRequest-
        Capabilities: [b4] Vendor Specific Information: Len=14 <?>
        Capabilities: [100 v1] Virtual Channel
                Caps:   LPEVC=0 RefClk=100ns PATEntryBits=1
                Arb:    Fixed- WRR32- WRR64- WRR128-
                Ctrl:   ArbSelect=Fixed
                Status: InProgress-
                VC0:    Caps:   PATOffset=00 MaxTimeSlots=1 RejSnoopTrans-
                        Arb:    Fixed- WRR32- WRR64- WRR128- TWRR128- WRR256-
                        Ctrl:   Enable+ ID=0 ArbSelect=Fixed TC/VC=01
                        Status: NegoPending- InProgress-
        Capabilities: [128 v1] Power Budgeting <?>
        Capabilities: [600 v1] Vendor Specific Information: ID=0001 Rev=1 Len=024 <?>

Last edited by justm3 (2020-03-18 08:08:46)

seth · 2020-03-18 06:47:39

rebranded GeForce GT 440…
https://en.wikipedia.org/wiki/GeForce_6 … s#Products

Please edit yourr post and wrap the output in code tags.

justm3 · 2020-03-18 08:54:54

Thank you for the link.
It clearly explains that 390 driver is the only way to go.
Code tags added.

Thank you again.

Arch Linux

#1 2020-03-15 07:40:45

kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#2 2020-03-15 10:14:09

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#3 2020-03-15 15:13:41

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#4 2020-03-15 16:00:35

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#5 2020-03-15 16:33:23

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#6 2020-03-15 16:40:24

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#7 2020-03-15 16:52:04

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#8 2020-03-15 16:58:45

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#9 2020-03-15 17:12:42

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#10 2020-03-15 18:13:38

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#11 2020-03-15 20:32:05

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#12 2020-03-15 21:03:01

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#13 2020-03-15 21:16:21

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#14 2020-03-15 23:55:05

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#15 2020-03-16 04:19:51

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#16 2020-03-17 08:05:47

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#17 2020-03-17 08:19:57

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#18 2020-03-17 08:34:45

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#19 2020-03-17 17:58:30

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#20 2020-03-18 05:59:21

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#21 2020-03-18 06:47:39

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

#22 2020-03-18 08:54:54

Re: kernel linux-lts-5.4.25-2 and nvidia-390xx-lts problems

Board footer