You are not logged in.
Hi there, officially this is my first Arch Linux forum post I'v been using the linux-ck-ivybridge and linux-ck-ivybridge-headers and nvidia-ck-ivybridge for quite some time, last night i reinstalled my system and now I can't install the nvidia driver ... I get the following output
kronikpillow@susanoo ~ % p -S nvidia-ck-ivybridge
resolving dependencies...
looking for conflicting packages...
Packages (1) nvidia-ck-ivybridge-2:435.21-9
Total Download Size: 13.21 MiB
Total Installed Size: 13.34 MiB
:: Proceed with installation? [Y/n] Y
:: Retrieving packages...
error: failed retrieving file 'nvidia-ck-ivybridge-2:435.21-9-x86_64.pkg.tar.zst' from repo-ck.com : The requested URL returned error: 404
warning: failed to retrieve some files
error: failed to commit transaction (unexpected error)
Errors occurred, no packages were upgraded.
Offline
@kronikpillow:
https://wiki.archlinux.org/index.php/Pa … stallation
"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman
Offline
@kronikp - https://bbs.archlinux.org/viewtopic.php … 1#p1868051
Doesn't explain why your sync-db is showing the packages though... try rerfeshing.
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
Help no Nvidia-ck
pacman -S nvidia-ck-zen
error: target not found: nvidia-ck-zen
Offline
Help no Nvidia-ck
pacman -S nvidia-ck-zen
error: target not found: nvidia-ck-zen
nvidia-ck now rests in AUR.
Offline
Help no Nvidia-ck
pacman -S nvidia-ck-zen
error: target not found: nvidia-ck-zen
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
With the recent update removing extramodules, I don't have the time/resources to keep providing the broadcom and nvidia* packages. With the release of 5.2.21 today, I will point users of broadcom-wl-ck, nvidia-ck, nvidia-340xx-ck, and nvidia-390xx-ck to the respective dkms packages which are now live in the AUR and will build these for you automatically.
If you're using a linux-ck optimized package, you will need to change the respective dkms PKGBUILD to reflect that package's name. For example, linux-ck-zen. Look in the dkms PKGBUILD for the line defining the _kernelname variable. Change it from 'linux-ck' to 'linux-ck-zen' (or to match your kernel package).
Hi, as discussed on the AUR, I do not think you need an nvidia-xxx-dkms package specific to each custom linux-xxx kernel (unless you need a particular patch to the nvidia driver so that it can work with your custom kernel) - if I am not mistaken this is the purpose of dkms. On my system, nvidia-dkms reinstalls (and recompiles the non-binary part of) nvidia for every kernel (stock and custom ones) every time one of them is upgraded. This is because pacman hooks trigger nvidia reinstall by dkms. This is valid for other dkms packages (such as wifi drivers). When not using a dkms package though, you need to maintain one for every kernel (nvidia-ck, nvidia-zen etc).
Anyone to confirm or correct me?
Offline
Just want to say thanks to graysky for the effort and diligence.
Last edited by archimboldo (2019-10-19 13:00:18)
Rules for problems.
Everyone has problems. Animals have problems. And buildings. And cats, and trees.
Problems are your friends. Treat them well.
Offline
@zebulon - I see this behavior with nvidia-390xx-dkms, build for all installed kernels that also have headers package.
@archimboldo - You're welcome!
Last edited by graysky (2019-10-19 21:20:24)
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
@zebulon - I see this behavior with nvidia-390xx-dkms, build for all installed kernels that also have headers package.
Well in that case you can probably spare yourself the maintaining of -cm-dkms packages. That said many thanks for all your work.
Offline
graysky wrote:@zebulon - I see this behavior with nvidia-390xx-dkms, build for all installed kernels that also have headers package.
Well in that case you can probably spare yourself the maintaining of -cm-dkms packages. That said many thanks for all your work.
Last time I checked it didn't work for nvidia latest
Offline
Last time I checked it didn't work for nvidia latest
You mean nvidia-dkms? On my system it compiles correctly with Linux-ck, and always has. The purpose of dkms is to facilitate driver compilation against kernel headers for every custom kernels. If however you meant nvidia package (non dkms) then yes it only works with stock linux package. It is already compiled and does not even need kernel headers.
What is your error message?
Last edited by zebulon (2019-10-19 20:59:09)
Offline
Tharbad wrote:Last time I checked it didn't work for nvidia latest
You mean nvidia-dkms? On my system it compiles correctly with Linux-ck, and always has. The purpose of dkms is to allow driver compilation against kernel headers for every custom kernels.
What is your error message?
Not the dkms. This one: https://www.archlinux.org/packages/extra/x86_64/nvidia/
No error. sddm stuck.
Offline
Not the dkms. This one: https://www.archlinux.org/packages/extra/x86_64/nvidia/
No error. sddm stuck.
Well as I wrote this is normal. You need either nvidia-ck or nvidia-dkms (but dkms requires linux-ck-headers. I recommend using dkms when you want to use custom kernels from AUR.
I was just advising graysky that he does not need to maintain nvidia-ck-dkms since it is identical to nvidia-dkms in extra.
Offline
Anything I have to apart from installing nvidia-dkms?
(I'll also edit the wiki)
Last edited by Tharbad (2019-10-20 12:48:22)
Offline
I believe you need the linux-ck-xxx-headers and dkms for it to work. I requested to have all the nvidia-ck* and broadcom-wl-ck* AUR packages deleted. No need to confuse people.
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
I believe you need the linux-ck-xxx-headers and dkms for it to work. I requested to have all the nvidia-ck* and broadcom-wl-ck* AUR packages deleted. No need to confuse people.
nvidia-ck* was deleted
Offline
@graysky: You deleted the link to this thread from the wiki. It was intentional?
I've added a paragraph about nvidia.
Offline
@tharbad - I tried to clean up both the linux-ck and repo-ck articles. I believe there was some text on the linux-ck article that applied to repo-ck vaguely so I deleted it.
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
Hi graysky, am I correct that you will not maintain the non-dkms version of nvidia-ck driver then? Is it because there are too many variants to maintain (not a criticism here )
Offline
Anything I have to apart from installing nvidia-dkms?
(I'll also edit the wiki)
Not only the headers, but also the compilation tools (package base-devel will install everything you need).
Offline
@zebul- right: https://bbs.archlinux.org/viewtopic.php … 1#p1868051
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
@zebul- right: https://bbs.archlinux.org/viewtopic.php … 1#p1868051
Thanks, I missed this piece of news.
Offline
@tharbad - I tried to clean up both the linux-ck and repo-ck articles. I believe there was some text on the linux-ck article that applied to repo-ck vaguely so I deleted it.
OK. I'll add the link the see also.
Maybe I'll add nvidia-dkms later if people will complain here.
Last edited by Tharbad (2019-10-21 10:58:44)
Offline
Just a heads up, because I don't know if it's specific to the architecture, the ck patchset, or even the build, and it's too late here to go into it further right now, but there might be a power management issue with linux-ck-piledriver 5.2.19-1. I experienced a couple of sudden hard shutdowns this evening, which I initially assumed to be a hardware issue (gulp!), but while trying to diagnose it I noticed that my CPU was constantly drawing power uncomfortably close to the “critical” level. Switching to the stock 5.3.6-arch1 kernel fixes this; I don't yet know if it's also solved the shutdowns, but it seems likely that the two issues are connected. It possibly also affects earlier 5.2 kernels, because I saw a similar unexplained shutdown a week or two ago.
Obviously there's not much graysky can do about it, but if anyone else is experiencing the same thing, it might be best to avoid 5.2.
Edit, next day: Okay, it looks like the sudden shutdowns might actually be a hardware issue on my machine (I think - I hope - it's as simple as a dodgy power cord), but 5.2.19 is definitely over-using the CPU, compared to the stock kernel.
Last edited by dunc (2019-10-22 12:45:27)
0 Ok, 0:1
Offline