SteveTech @ SteveTech @programming.dev

New account: @SteveTech@aussie.zone

---

Read more

Posts

4

Comments

461

Joined

3 yr. ago

You can host a page with an iframe, but you can't directly change the DNS record to point to something that isn't GitHub.

Guys you're not gonna believe this:

The common sketchy performance advice is to disable mitigations in the kernel, this post is about disabling mitigations in Intel's userspace graphics stack because it's already checked in the kernel.

Assuming you meant disabling kernel mitigations, since AFAIK audio stuff doesn't usually use OpenCL:

Has anyone else here disabled it?

Nah, my understanding is it's not worth it on newer CPUs, and in some cases, the microcode expects things to be mitigated for best performance. Older CPUs (pre-2019ish) it does make a difference though.

But you're welcome to benchmark it, and see if it makes a worthwhile difference on your CPU. Kernel mitigations are easy enough to turn on and off.

I think they were trying to say that the cage in front with the AP behind, acts as a directional antenna. Similar to how Yagi antennas have metal elements that aren't connected in front of the actual antenna.

But I don't know enough antenna theory to know if that's correct.

I've previously found OpenRGB's udev rules to be a really good example since there's a bit of everything in there: https://openrgb.org/releases/release_0.9/60-openrgb.rules

But I think you'd want something like: SUBSYSTEMS=="usb|hidraw", ATTRS{idVendor}=="REPLACE WITH USB VENDOR", ATTRS{idProduct}=="REPLACE WITH USB PRODUCT", TAG+="uaccess"

It's not strictly Linux anymore, but I wrote a library (or userspace driver?) in Python that interacts with a ChromeOS Embedded Controller found in Framework Laptops and Chromebooks. The driver part of it interacts with the EC directly over the IO ports, which was originally written for Linux but later ported to FreeBSD and Windows since IO ports aren't at all OS specific. It can also talk to the cros_ec_dev driver on Linux if it's loaded.

https://github.com/Steve-Tech/CrOS_EC_Python

I wrote a GUI utility for Framework Laptops too, which also serves as the example for CrOS_EC_Python: https://github.com/Steve-Tech/YAFI

help now actually opens the help utility on Python 3.13!

Thanks TIL! Although I prefer this diagram that has all the wifi channels on it, instead of just the 3 common ones.

Source

Perhaps there was an easier lighter-weight way of doing this?

Yeah, SSH tunneling. What I would do (and have done in the past) is something like:

    
ssh -L 8080:192.168.0.1:80 myserver


  

That will forward port 8080 on your host to port 80 on 192.168.0.1, so you can access your router's web UI with http://localhost:8080/ in your own web browser.

You can also setup full tunneling with SSH, but that requires messing around with SOCKS and I usually can't be bothered.

If we're suggesting a GUI for basic trimming and splicing, I prefer Avidemux, it supports cutting without transcoding the whole video (as long as you cut on an I-frame), saving time and reducing artefacts.

My understanding is previously the kernel would crash on systems with more RAM than the address space, so there's now a patch to ignore the anything above the max address supported (e.g. 32bit without PAE, 36bit with PAE). More RAM was never supported, so I think the author of the article has misunderstood or oversimplified what's been done.

You can always install activate-linux, and it even works on Windows.

Epic!

You should be able to add options it87 force_id=0x8688 ignore_resource_conflict=1 to /etc/modprobe.d/it87.conf (or whatever filename) and it87 to /etc/modules. To get it to run at startup.

You can try ignore_resource_conflict which is it87 specific, rather than a system wide acpi_enforce_resources.

modprobe it87 force_id=0x8628 ignore_resource_conflict=1

The reason why this is needed is ACPI claims the I/O ports required to talk to the it87, and Linux doesn't want to override that.

The master branch works well with Intel ARC, I contributed a lot of the ARC changes. I don't think they've made it into a release yet though.

Edit: 3.2.0 has them: https://github.com/Syllo/nvtop/releases/tag/3.2.0

I got through University running Debian testing. It was mostly fine, some Linux based subjects were way easier without dealing with a VM (they recommended against WSL for some reason).

However there were a couple units that absolutely required you to use Visual Studio (non-code), I occasionally used a VM, the Uni IT also provided me with a remote VM (there's a form to fill and and it's all automated). But I mostly used Rider, which for one unit it confused their CI and I got marked down for (otherwise got top marks so it's fine).

For office, it didn't matter. Group projects mostly used Google Docs, occasionally Microsoft Office where the online version worked fine. All my units wanted PDFs at the end anyway, so it does not matter that you used LibreOffice or whatever. Some units provided you with DOCX templates, I had no issues opening them with LibreOffice.

Edit: People are mentioning online exams, my Uni did 'online quizzes' which worked fine, and some had to be done in class on their PCs anyway. Final exams where always done on paper.

Ahh sorry, I thought you meant you plugged it into the input side. If that's the case then are you running anything that measures CPU usage? I run the TIG stack, it might be able to give you some hits. Also back to my original point which is already unlikely, if it's a modified sinewave UPS, it can confuse some measuring devices while it's on battery.

It's weird to do this daily, but it's possible that the UPS is doing a self test, which would drain the battery a little and the load is from charging it back up.

It's not immediately a DNS issue. Usually if there's no response within less then a second, then a browser will skip IPv6 and use IPv4 (Happy Eyeballs). But in this case the server responds with an SSL error over IPv6.

    
curl -v -6 "https://rebeltechalliance.org/"
* Host rebeltechalliance.org:443 was resolved.
* IPv6: 2a10:e000:1::10
* IPv4: (none)
*   Trying [2a10:e000:1::10]:443...
* ALPN: curl offers h2,http/1.1
* TLSv1.3 (OUT), TLS handshake, Client hello (1):
*  CAfile: /data/data/com.termux/files/usr/etc/tls/cert.pem
*  CApath: /data/data/com.termux/files/usr/etc/tls/certs
* TLSv1.3 (IN), TLS handshake, Server hello (2):
* TLSv1.3 (OUT), TLS change cipher, Change cipher spec (1):
* TLSv1.3 (OUT), TLS handshake, Client hello (1):
* TLSv1.3 (IN), TLS handshake, Server hello (2):
* TLS connect error: error:00000000:lib(0)::reason(0)
* OpenSSL SSL_connect: SSL_ERROR_SYSCALL in connection to rebeltechalliance.org:443
* closing connection #0
curl: (35) TLS connect error: error:00000000:lib(0)::reason(0)

  

The symbol they defined out is not the equals symbol but rather U+2550, so the for loop is fine.