| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
More packages have been renamed recently.
|
|
|
|
|
| |
Second attempt at getting this right. Step 1: Read the thrown
assertions. Step 2: Don't be tired. Step 3: Test changes.
|
|
|
|
| |
This reverts commit 20ee65d79602b6412994dc52907715b274daeebe.
|
|
|
|
| |
This reverts commit 2332e49b29528325ad1ff2bc0a279275c08ea266.
|
|
|
|
| |
Dunno why it was at 8, but 16 is what it should be.
|
|
|
|
|
| |
This must have been a recent change where an assert is now thrown
because some nix.* settings have been regrouped as nix.settings.*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The boot.kernelPackages option is already defined in
devhell/profiles/base.nix to be linuxPackages_latest with the same
priority (100) as in the machine configuration of "gunnr".
While generally I leave it for respective owners of the machine
configuration to fix, this however breaks evaluation of all of our
tests.
Given that "latestCompatibleLinuxPackages" is used on a ZFS system, I
assume that this is what the author wanted with higher priority back in
678754b4a5ac5a468e2223e1b701b8e9e2375c30.
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @devhell
|
|
|
|
|
| |
This commit also includes a bit of drying the `auto-optimise-store` Nix
option.
|
|
|
|
|
| |
I don't know if this will work, especially considering I'm adding this
without having access to `gunnr` at the moment.
|
|
|
|
| |
This has been removed a while ago, so might as well remove this too.
|
|
|
|
|
|
| |
I didn't notice this at first, but `thermald` only really works for
Intel CPUs. This is an AMD machine, and the daemon can't find anything
to work with, so there's no point in keeping it enabled here.
|
|
|
|
|
|
| |
Duh.
This reverts commit 2d8ad8d8d158f5c4d1e3085062ec4619a18416c9.
|
|
|
|
| |
There's really no need to keep this exposed on the internal network.
|
|
|
|
|
| |
Without this, the kernel tries to use `mq-deadline` for gunnr. The
invocation for eir was wrong anyway, so this should fix it.
|
|
|
|
|
|
|
|
|
| |
Both, `herja` and `gunnr` have no spinning rust in them, and so they
won't need bfq anymore. According to this old benchmark, the `none`
scheduler works overall better for most workloads than other
schedulers[1].
[1]: https://www.phoronix.com/scan.php?page=article&item=linux-56-nvme&num=1
|
|
|
|
|
|
| |
It's been a while since I last used this machine. At some point I need
to go through it with a fine-toothed comb, but for now these changes
will be enough.
|
|
|
|
|
| |
As it turns out, rocm isn't supported on integrated Ryzen GPUs. No point
of having it around on herja then.
|
|
|
|
| |
I've done this before with herja, and then forgot to update the others.
|
|
|
|
|
| |
A few things have changed since the recent NixOS 21.05 release. Mostly
naming stuff, so nothing dramatic.
|
|
|
|
|
|
|
| |
I'm going to start using `Zettlr` to try and keep on top of the
gazillion notes I have. Zettlr doesn't have built-in sync to anything,
so I'll probably have to use `syncthing`. I'd prefer `git`, but if the
file versioning works halfway decently then that'll be fine.
|
|
|
|
| |
It's been renamed recently and throws a warning saying as much.
|
|
|
|
| |
Just like with herja.
|
|
|
|
|
|
|
| |
There's no need for this as I've reverted back to `pulseaudio` for the
time being.
This reverts commit a520414825f672eadd25017f06bf13fb9de2db8d.
|
|
|
|
|
|
|
|
| |
While I'm generally happy with `pipewire`, some things broke, such as
volume control. I've tried to fix it using a range of front-ends for PA,
but so far only `pactl` seems to be behaving correctly (but broke mic
mute again). I might have to switch back to PA in the end as it's also
broken my setup on gunnr somewhat.
|
|
|
|
|
|
|
| |
Every boot the kernel will silently dump some a trace referencing an
issue with AMDGPU. Some sleuthing suggests that the module needs to be
loaded much earlier. Funnily NixOS actually has a wiki specifically on
AMDGPU, so these settings have been taken from there.
|
|
|
|
|
|
|
|
| |
Finally, after a lot of hussle with the IT department they've come to
the conclusion that, in fact, their eduroam cert was not generated
properly thereby preventing me from using iwd. It looks like this only
affected a minority of devices (presumably those with proper
implementation).
|
| |
|
|
|
|
| |
Let's keep it simple, everything else seems to just end in tears.
|
|
|
|
| |
Sigh.
|
|
|
|
|
| |
So the previous idea didn't work as expected, instead I'm going to use
the directory name Syncthing would use anyway.
|
|
|
|
| |
This reverts commit d46aa4ec882dcb9ac0e09da6e1aecb361d90d287.
|
|
|
|
|
| |
This doesn't seem to work properly, Syncthing will still add its default
`Sync` directory.
|
|
|
|
|
|
|
|
|
| |
I've been using `illum` to make the screen brightness keys work on this
machine. It's better than invoking light manually or via a xbindrc call.
Although, `illum` is a bit outdated in nixpkgs, and I'm guessing that's
the reason why it crashes when the machine boots. After restarting the
service it's fine. If `illum` isn't updated in nixpkgs soon, I'll have a
look at it.
|
|
|
|
|
|
| |
`xbindkeys` seems to be the solution, at least via user-space.
This reverts commit 59dbc84cb214ed3df6506e95d3bf59b89f9d3548.
|
| |
|
|
|
|
|
|
| |
For some reason this doesn't work yet. It works fine when run manually
as root or my own user, but when acpid runs it nothing happens. Dunno
yet why.
|
|
|
|
| |
No more Brother crap.
|
|
|
|
| |
I don't know what I'm doing, but it's fun.
|
|
|
|
|
| |
So, turns out that enp6s0f3u1 sometimes also is initialised as
enp6s0f4u1. Why? WHyyyyyyyyy?????
|
|
|
|
| |
Dunno why I thought it was the former.
|
|
|
|
|
| |
The T14s is an awesome machine, but unfortunately it needs an Ethernet
dongle.
|
|
|
|
|
|
|
|
| |
After some experimentation and looking at the NixOS' hardware repo, this
configuration seems to work quite well. There are no major issues to
worry about for now.
This is also the first machine where `fwupd` is available and working.
|
|
|
|
| |
Shiny new machine needs NixOS love. <3
|
|
|
|
| |
Recommended by aszlig to switch from module to package.
|
|
|
|
|
| |
Who'd have thought. Work has finally embraced a way of printing no
matter what OS one runs. What a time to be alive.
|
| |
|
|
|
|
|
| |
I'm finding myself use containers more and more these days, primarily
for testing stuff, so it's useful to have the tools installed.
|
|
|
|
|
|
| |
Trying to figure out a better, more consistent approach to theming my
terminals and terminal programs, so let's start from scratch and remove
as much as possible.
|
|
|
|
| |
Let's see if this does anything good.
|