Forward proxy in 2021?

I have a requirement to provision a forward web proxy for my datacentre. It's been a while since I rolled out such a beast - and previously squid would have been my automatic choice. But when I google for "Open Source Forward Proxy" I find links to lots of projects, none of which are squid.

Is squid now frowned upon? (why?)

What should I be using?

(just to clarify my requirements - my immediate need is to handle logging and address translation, but in future I hope to provide caching - and therefore would need SSL MITM capability. Please don't tell me that a product works for you without telling me why).

https://redd.it/lg3881
@r_linux

I have been using Ubuntu for 5 years and I finally figured out how to install arch! It works great!
https://redd.it/lg45d4
@r_linux

Not a support question: Peacock streaming service doesn't work on linux



Third attempt:

This is not a support question

Linux is not the issue here, this is an issue with the Peacock streaming service and its DRM.

(This is not a support question)

I am not asking how I can get it working, I already know the solution... change my OS to windows or use wine or use my mobile device as currently those are the things I have that peacock works on.

(This is not a support question)

No, what I am asking for is my fellow Linux users to give their feedback to https://www.peacocktv.com/help/web-to-case

And see if we can convince them to remove the anti linux DRM restriction

(This is not a support question)

Other streaming services work fine, but peacock is a hold out, and I would love to have it working natively in linux without me (or anyone else who wishes to use peacock with linux) having to fiddle with wine (which doesn't always co-operate) or setup a VM or turn my dedicated HTPC linux machine into a windows one.

(This is not a support question)

Also yes I can pirate, but I would like to do this legit.

Again the issue is not me, I know how to use Linux, have used it for 25 years and even though it's not on my main gaming machine anymore I still use it on my other machines.

The issue is peacocks DRM and there is nothing I as the user can do about it other than use windows or wine or a VM thus this is not a support question.

The only "help" I ask for here is to get other linux users to visit https://www.peacocktv.com/help/web-to-case and tell them to support linux, linux is not the problem.... they are

(This is not a support question)

https://redd.it/lg5jyi
@r_linux

PSA: ACS Override patch now included in Linux Zen kernel.

As of now the ACS Override patch is now included in the Linux Zen kernel. So you just have to add the respective kernel parameter pcie\_acs\_override=.

I'm running 5.10.13-zen1-2-zen instead of linux-zen-vfio on Arch Linux and my usual VM with GPU-passthrough works just fine.

Source: https://queuecumber.gitlab.io/linux-acs-override/

https://redd.it/lg6wtw
@r_linux

Turning an old Amazon Kindle into a eink development platform
https://blog.lidskialf.net/2021/02/08/turning-an-old-kindle-into-a-eink-development-platform/

https://redd.it/lfxvsr
@r_linux

Seriously, Fuck Nvidia

So unless you want to reminisce about the mid 90's, when you had to go to a website to search for drivers -'avoid Nvidia like the plague. Unless you know exactly what the marginal improvement you are using Nvidia for, and which/how to install the drivers - it is nowhere near the marginal benefits. I am not speaking to gamers here (who maybe know why they want to make the extra lift that turns out to be harder than installing the base operating system).

But for the rest of us that expect things to just work - know that Nvidia is the mark of either a lot of wasted effort - or a lot of wasted hardware that won't be fully utilized.

https://redd.it/lg9j0m
@r_linux

Just out of curiosity, did anyone try installing every package available?

What i mean is did anyone try to install every package available to the distribution they are using (except obvious stuff like all versions of the kernel or multiple network drivers etc.) Just wondering if it would even if it’s even possible or break the system or slow it down to the point of being unusable.

https://redd.it/lg98c8
@r_linux

Impressions after switching to pipewire for audio

So I did switched to pipewire with pipewire-pulse (Archlinux) on my main workstation.

So far, I finally have no more frequency bugs. With pulse, sometimes the frequency would shift randomly when starting streams (especially when joining a discord call while music was playing) since the switch I never had any frequency bugs. I'm using a Focusrite 2i2 as my sound card.

Some bugs I noticed: when switching my main device using the KDE sound indicator, sometimes some applications don't switch. Switching device usually is very smooth, but sometimes don't happen.

It may not be related, but rarely I have a considerable delay when talking on discord, but only happened twice since in two months and was resolved each time by restarting discord. It may be my old 4790k sweating because I'm compiling stuff often and on all cores.

I have a huge thank you for everyone in this community and also everyone that made pipewire happen. It's already in a pretty good shape and simply removed most of the problems I had with pulse.

https://redd.it/lgc1wp
@r_linux

Error on manually mounting LUKS encrypted device

Hello, i have a LUKS encrypted device. I want to mount it manually, but im running in one error: 'Name: "/media/nooby/USB" invalid. It contains "/".'. I couldnt find anything on the web about this. The command i used was "sudo cryptsetup luksOpen /dev/sdc /media/nooby/usb". What did i do wrong? Thanks for evey help

https://redd.it/lgdsyi
@r_linux

Pulseaudio -k, or a pro audio user's perspective on Linux's sound stack

# A bit about me.

I\'m a professional Physicist, Composer, Musician and I code in some off
time. You have probably not used my programs, but it is nonetheless
possible, maybe you do some Bayesian inference, or like to do some
org-ing in emacs. I\'ve done my share of audio mixing on Windows,
I\'ve written a bunch of Midi stuff in Cakewalk Pro Audio (yeah, that
relic), and I currently own a Macbook Pro that I do some serious heavy
lifting in. I don\'t just write music, though I happen to like tinkering
with sound in wave editors, and I often create mods for \'90s shooters.

I also happen to be running Arch Linux since forever, but 2011 is the
earliest date since when I ran it on bare metal on any of the machines
that I own. I\'ve since branched out to Gentoo, and recently NixOS,
because reproducible builds make your developer life a lot easier. I\'ve
ran Arch for a very long time, it would have been five consecutive
years, were it not for hardware failure. In other words, I know what
I\'m doing, and I can follow a man page.

# Life before pulseaudio

I don\'t hate it out of thin air. I was running Ubuntu before Arch and I
remember the days before pulseaudio was a thing. I didn\'t hate it
because change=bad, I\'ve grown to hate over the span of several
sleepless nights, trying to figure out if the sound artifacts that I can
hear are due to hardware problems or another in a series of bugs.

Things used to be simple: you had to do most things manually, no
hotplugging, no autodetect and sadly, no way of adjusting program volume
levels individually. ALSA isn\'t perfect, but it\'s workable if you know
your way around alsamixer. Certainly, setting up a DAW (digital audio
workstation) is a bit more of an adventure than one might expect it to
be on Windows or Mac OS X.

You see, on Windows, in the olden days, there used to be this thing
called a VST, that could come with an effect like reverb, or it could
come with a full instrument like surge. It\'s mostly plug and play,
however, in some particular instances, there are mild issues, that can
be resolved with ccleaner plus some reinstalling. Mac OS X, is
notoriously good in this regard. Not only do things arrange themselves
neatly, they rarely if ever misbehave. Logic Pro is a DAW that almost
never malfunctions by itself, and there\'s not a single piece of manual
intervention between downloading a plugin, and listening to it in
action.

On linux, you need to have Jack set up properly. It used to be easy:
Jack sits on top of Alsa, and lets you do things like route different
Jack Outputs to different inputs, including the midi tracks in say
Ardour, to say the VST instrument. Not easy, but doable. You can
actually just autostart Jack at login, and be fine. Then pulseaudio came
along, and... things broke. It\'s fine, it\'s new, it\'s got shiny new
ideas and a talented young developer behind it. It\'ll be alright two
years down the line.

Almost ten years after, and things are not fine. Not in the slightest.

# Pulseaudio at its best.

I mean no disrespect to Lennart Poettering. I really don\'t. It\'s not
that he has bad ideas (they\'re not bad), or that two out of three of
his major projects are so bad, that not having this software is a
selling point. The trouble is that under pressure from Red Hat, and
Arch, and other cool new shiny distros, pulseaudio was adopted by
default. Gnome cut out the option of running on pure alsa, because of
course it did. Some major applications outright require the presence of
pulseaudio, namely Steam, making the proposition of not running this
Alpha quality, Early-access CoreAudio asset flip of an abomination a bit
difficult to swallow for most users. However, much of the problems with
pulseaudio ten years ago are still there, so I can equally well describe
the situation back then as now.

First let\'s start with the bugs. Pulse is a meme at this point. The
biggest indicator of the abject failure of pulseaudio is the fact that
if a typical user has no sound, they never

think \"maybe it\'s JACK\",
or even \"Maybe it\'s ALSA\". This mission-critical-middleware of a
system, is responsible for the vast majority of incompatibilities, and
headaches around sound.

You can\'t reliably make pulseaudio work with JACK. This is a deal
breaker for me, because I need to use VSTis for my projects. I have
tried Ubuntu Studio, Arch and it\'s pro-audio oriented spin-off. None of
them work out of the box, and the only semi-workable solution is to run
pulseaudio -k before starting JACK. The best solutions to the
incompatibility is to get rid of pulse. Things like pulseeffects
work... sometimes... and not that well.

But those problems can be fixed. You can theoretically assign a team of
good programmers to fix the mess that currently is pulseaudio, and maybe
fix some of the recurring issues. The architecture of how it works,
however, cannot. How many command-line tools does pulseaudio come with.
What is the syntax of pamixer, vs pactl, vs pacmd. Why couldn\'t
those utilities be merged? Minimalism? Independence? OK. Fine. Why do we
have sources and sinks? Why do I never get shown a decent name like
onboard audio, vs Family 17h
HD AUDIO CONTROLLER. I\'m a tech savvy person, and even I have doubts
that this is the right audio output. But then you only identify it with
that name in some obscure cases, every real application requires you to
remember the number of the card. I have one onboard audio controller,
and one inside my video-card that I never use. Guess which number is the
one that is active? 1? 0? 4! I guess the internal enumeration
makes sense, but you might have as well used a UUID hash, because a
human would never have guessed that without looking it up.

Why do we make it so obtuse? Why do we have to assume that the end user
is looking at the source code of pulseaudio and already figured out,
that the sink is a recording device and a source is a playback device.
Why add terminology that makes no sense. Moreover, the only thing that
the bloatware of pulseaudio could have done, is at least fix the audible
pop when volume is adjusted. OSS, had this feature twenty years ago!

There is still hope. Pipewire is a bit of a necessity in the post X11
world. It handles many of the tasks that are impossible otherwise, and
handles them well. While I had very moderate success in replacing the
aforementioned pulseaudio and JACK. I have yet to attempt any
serious work on the new stack, and I expect things to not be as simple
as they used to be, but I\'m hopeful. I do, however worry, that pipewire
shall inherit all of the shortcomings of pulseaudio. It very likely
shall fall victim to the same hubris of \"It\'s already adopted on every
Linux machine, why should I fix the glaring issues and omissions?\"

However, I have a list of small things that as a pro audio user I would
like to change.

1. VSTs should support pipewire directly. If you can drop-in-replace
JACK, then you can use pw, as a substitute.

2. Pipewire should probably have a more sane model of the audio
hardware. I understand that it inherits most of that from ALSA,
which is itself an overengineered mess of spaghetti protocols, but
still, a bit more simplicity for the end user, means fewer bug
reports.

3. There should be a single set of command-line tools, that come with
pipewire. I shouldn\'t have to search the repos for the package that
contains the tools I need to debug a problem.

4. On that note, latency is a big problem with pulseaudio. I don\'t
care if it\'s bloated, but I do care if I hit a Midi key and get the
sound noticeably later. This makes improvisation impossible.

5. Popping and cracking. There\'s something called envelopes. There\'s
something called smooth transitions. These can make it so you don\'t
have to use a sound effect to mask the audible pop in every GUI
audio mixer.

6. GUIs for adjusting common settings would also be nice. kmix has a
bare bones system that barely does what it\'s supposed to.
pavucontrol is much better in this

regard, but still too
bare-bones. I want something where I can tweak the sample rate, and
set the avoid-resampling parameter. Most of the people who do Pro
Audio aren\'t me, and won\'t even dare touch
/etc/pulse/default.pa.

7. Less attitude would be welcome. I guess, I\'m one of the many
\"a**holes\" on the internet, that like to target Lennart
Poettering. This sort of hostile attitude makes submitting actual
bug reports very, very, very tense. And Pulse would be in a very
different position, had most of the architectural blunders been
addressed in earnest, and fixed.

Overall, I hope that you will find this rant of a post, as a bit of
a... Ok. This is mostly me venting, but I tried to keep things
objective and relevant.

https://redd.it/lge6fe
@r_linux

security things in Linux v5.8
https://outflux.net/blog/archives/2021/02/08/security-things-in-linux-v5-8/

https://redd.it/lgcu5a
@r_linux

Native Linux software is even better than before

Majority of the tools I have been using to run my small business is either attached to a web service over the browser, via Electron, Snap or AppImage. So far I'm liking AppImage but after spending time browsing the native software installer, I came across Evolution, Dia, GThumb, GnuCash, Liferea, and Umbrello. The crazy improvement has made me reduce the tabs on Firefox, which kills my small laptop and weirdly has reduced my distractions. I can put the browser far away in the background and focus on the task at hand. When working with Firefox, you get that little blue dot when something happens on a tab and that easily leads down a rabbit hole.

Should have done this a long time ago.

https://redd.it/lgg2ie
@r_linux

System76: This Thursday, 2/11: Two-factor authentication, Argon2 password hashing coming to System76.com accounts!

https://blog.system76.com/post/642658662169657345/coming-0211-to-our-website-two-factor

Finally System76 joins the 2FA master race! That's awesome, can't wait for the day 2FA becomes de-facto standard and all online services will be secured this way.

With that said I'm not sure if it's a good idea security-wise to share what hashing algorithm is going to be used to hash passwords. If somebody would break into their servers, the attacker would probably be able to recognize that passwords were hashed using argon2 anyway even without prior knowledge, so it's not much of a damage and probably doesn't make the system less secure, but still it feels like publicly sharing hashing algorithm you use is super wrong.

https://redd.it/lgelb4
@r_linux

I made an ultra-minimal alternative to playerctl (as if that wasn't already minimal enough)

Hey guys, I'm new here but love Linux and have recently started programming for it, which has been a huge step forward in my Linux life :D

First thing I've programmed is mediactl! It does kinda the same thing as playerctl, but with bordering-on-pathological minimalism. It uses the d-bus to broadcast mpris method calls to all compatible players running on your machine; Play, Pause, Stop, Next, Prev... That's it! The x64 binary is 30kb, and spawns no daemons or other... uhmm... "bloat", if you can even call playerctl bloated to begin with hehehe.

# A practical use for this

I wrote this program so that I'd get my media keys functional again after switching to i3wm! After installing my program I added the following lines to my i3 config and I've never been happier lol

bindsym XF86AudioPlay exec mediactl PlayPause
bindsym XF86AudioNext exec mediactl Next
bindsym XF86AudioPrev exec mediactl Previous
bindsym XF86AudioStop exec mediactl Stop

Hope you guys like it, and eh, hi :D

https://redd.it/lgfwst
@r_linux

How are events generated/received/interpreted to turn on laptop keyboard's LED?

I'm not sure if this is the right place to ask this because it is part support help and part how does a subsystem work. The first seems like it shouldn't be here, but the second does. Regardless I'll ask the question:

I have a new Dell XPS 13 9310 running debian 11/testing/bullseye on a custom kernel compiled at 5.11.0-rc7. I've run it for about a week and up until now whenever I touched the mouse or keyboard the keyboard would light up. It has suddenly stopped doing it and I have no idea why. This also made me realize I have no idea how this sort of thing is controlled in the first place. Here is walk-through of my changing the keyboard LED manually (which also shows that it actually does work):

# pwd
/sys/devices/platform/dell-laptop/leds/dell::kbdbacklight
# ls
brightness brightnesshwchanged device maxbrightness power starttriggers stoptimeout subsystem trigger uevent
# echo 2 > brightness # This set's the keyboard LED to high and does work
# echo 1 > brightness # This set's the keyboard LED to low and doeswork
# echo 0 > brightness # This turns off the keyboard LED and doeswork
# cat starttriggers
+keyboard +touchpad
# cat stoptimeout
10s

Given that it works and that the start_triggers and stop_timeout files seem to make sense, I'm not really sure why this suddenly stopped working. Does anyone here have any wisdom? What part of my OS determines this sort of thing? Where is a good place for me to look? Thanks a lot for any help!

https://redd.it/lgit0n
@r_linux

links, shortcuts and maybe screensavers

so I'm just getting back to Linux after a 6-year break and I want to make it my daily driver but I have run it to a slight issue I can't use links and shortcuts in pop os is there any way to make them work I have tried opening .URL files with brave on my Linux system and it just tries to download them again it seems like a very basic thing to be missing from what is otherwise a future rich os also screen savers are missing for some reason which means features have been actively removed

https://redd.it/lgfun9
@r_linux

Why are there no generic VGA mode-setting driver in the kernel

A while back I mistakenly asked why there were no generic mode-setting *VESA* drivers, and I got lots of interesting answers about Real Mode. While very informative, Thinking back what I meant was VGA (or maybe even SVGA) not VESA.

I'm not asking for this as a support request, I have modesetting on all my hardware, and my QEMU VMs.

But there probably are obscure cards that don't have a kernel mode setting driver out there, as well as possibly some VM platforms. Weston and other Wayland based display servers need a /dev/dri/card0 in order to be able to run, as they all require mode setting. (Some do run on framebuffer devices, but not many login managers support falling back to use framebuffers) This could also it make it difficult to kill CONFIG_VT, as you'll likely need kernel mode setting to for a kmscon like terminal to be able to start.


VirtualBox's card used to be one of the the ones where there was no modesetting support, until very recently, no /dev/dri/card0, there was framebuffer support if you passed a vga= boot option. (and can anyone verify if HyperV has modesetting support)

Why is there no generic driver that presents a /dev/dri/card0 for obscurer hardware? Is there some limitation as to why?

https://redd.it/lglxh5
@r_linux

So I'm installing Linux Mint from a Multiboot USB, is this normal?

https://redd.it/lgi94j
@r_linux

Using jq coomand

accountName="demo$RANDOM"
storage=$(az storage account create \
--name ${accountName} \
--resource-group $(echo $group | jq .name -r) \
--location $(echo $group | jq.location -r) \
--sku Standar_LRS \
)

bash: jq.location: command not found error. I'm also using the Azure CLI. This is for a terraform process

https://redd.it/lgo5w3
@r_linux