Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Microsoft's Linux Kernel

github.com

Post reply on HN
Print view View original

111–120 of 532 posts

Re: Microsoft's Linux Kernel

#111

Post by sjruckle »

AnIdiotOnTheNet wrote: post #92
cameronbrown wrote: post #75

Earlier quoted context omitted.

I wouldn't be so cynical. As with all things, tech comes and goes in cycles. I think with the limit to Moore's law rapidly approaching for current silicon, there's a good chance that efficiency & simplicity will become front and centre again at some point. Not sure if this will mean new architectures, operating systems & better tools.. but one can dream!

It is difficult to continue to hope for that while watching people cheer on additional layers of complexity like the garbage fire that is the modern web.

Yes, but if garbage never caught fire nobody would bother inventing better fire suppression techniques.

Also, nobody ever saw a fireman by hiding inside with their window blinds shut. Take a peek outside, what you see might bring you hope.

Re: Microsoft's Linux Kernel

#112

Post by DoreenMichele »

AnIdiotOnTheNet wrote: post #92
cameronbrown wrote: post #75

Earlier quoted context omitted.

I wouldn't be so cynical. As with all things, tech comes and goes in cycles. I think with the limit to Moore's law rapidly approaching for current silicon, there's a good chance that efficiency & simplicity will become front and centre again at some point. Not sure if this will mean new architectures, operating systems & better tools.. but one can dream!

It is difficult to continue to hope for that while watching people cheer on additional layers of complexity like the garbage fire that is the modern web.

Try roasting marshmallows over it once in a while. See if that helps you feel better about it.

Re: Microsoft's Linux Kernel

#113

Post by polyomino »

briffle wrote: post #102
rolph wrote: post #27

Earlier quoted context omitted.

problem is the kernel is sitting on top of a binary blob, and that manipulates the access to the real machine. we shouldnt be so worried about the kernel, we should be concerned about what the binary between the kernel and the hardware is doing/notdoing , and what degree of control do we have over that layer.

The same problem exists in VMware, or really any cloud provider...

Or on any proprietary firmware...

Re: Microsoft's Linux Kernel

#114

Post by scardine »

ChuckMcM wrote: post #51

Well that is pretty nice. WSL gets that much more exactly like Linux. What I really appreciate about WSL is that you get the accessibility of a bunch of OSS projects and a machine which has legit drivers for all of its component bits. What this means to me is that searching for a "linux laptop" won't be a chore, if it runs the latest Windows it will run Linux. And I can do development on the Linux side while communic…

There are a few annoyances with WSL2. For me the most important are:

1. You can't connect to a port listening on WSL localhost like in WSL1, you have to figure out the WSL IP address and use that.

2. From WSL you can't connect to a Windows TCP port on localhost, you must figure out the Windows IP (cat /etc/resolv.conf) and use that.

3. The WSL remote interpreter on PyCharm is not working anymore. The suggested workaround is "use SSH remote interpreter" but given #1 you can't connect to localhost (and the WSL IP changes every time you restart it)

In order to use the SSH remote interpreter in my favorite IDE I'm using this script (on Windows):

    import subprocess
    import re

    output = subprocess.check_output('wsl.exe ifconfig')
    match = re.search(r"\sinet\s+(\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3})\s", output.decode('utf-8'))
    if match:
        ip = match.group(1)
        with open(r"C:\Windows\System32\drivers\etc\hosts") as i:
            content = i.read()
            new_content = re.sub(r"(172.\d+\.\d+\.\d+)\s+wsl", f"{ip} wsl", content, re.M)       
        if new_content != content:
            try:
                with open(r"C:\Windows\System32\drivers\etc\hosts", "w") as o:
                    o.write(new_content)
            except PermissionError:
                print(new_content)
First I use the Windows Scheduler to start SSH (wsl service start ssh) on logon. Then this script is executed on logon with elevated privileges to update the hosts file. Instead of "localhost", I use "wsl" and it works well enough for me while I wait for the fixes.

Despite the above, the combo Windows 10 + WSL2 is the best "Linux Desktop" for me.

Re: Microsoft's Linux Kernel

#115

Post by filmgirlcw »

mlevental wrote: post #98
filmgirlcw wrote: post #42

Earlier quoted context omitted.

Disclosure: I work at Microsoft but not on Windows (and my work machine is a MacBook Pro FWIW) WSL 2 is a lot like that -- it's a VM in the sense that it uses Hyper-V, but it's super lightweight. And in fact, when WSL 2 is enabled, both Windows AND Linux are running in the same type of VM (separate, obviously), so it's near-native performance in a lot of ways. There are still some I/O issues that the team is actively…

do you know if there's hardware pass through for gpu? ie can you do cuda stuff in wsl2?

On Twitter, the team has said [1] it won’t be supported at launch but that this will make it easier to add that support in the future.

[1]: https://twitter.com/tara_msft/status/1125888319974400000?s=2...

Re: Microsoft's Linux Kernel

#116

Post by kevin_thibedeau »

Osiris wrote: post #88
gorio wrote: post #85

Earlier quoted context omitted.

Despite the wording I don't think that is the original definition? The idea is to make Linux viable as a desktop operating system, and whatever comes with that. Not running Linux as an application on another desktop. Which is almost the opposite as in leveraging capabilities outside of Linux. Maybe more importantly redefining the goal isn't helping Linux much. I remember everyone raving about Android, but now look at…

Given that drivers are Linux's biggest problem, isn't there benefit to leveraging Windows drivers and then running Linux on top of that virtualized hardware? I don't see why you couldn't run x11 in the window manager with the new Linux kernel in Windows.

You can run X11 today with Cygwin.

Re: Microsoft's Linux Kernel

#117

Post by oil25 »

We've come a long way from Bill Gates' original characterization of open source (long before it was called that) as thievery - https://worldhistoryproject.org/1976/2/bill-gates-writes-ope...

I'm still sure as ever Microsoft is no friend of free (as in Freedom, not price) software and can't help but shake my head at this latest publicity stunt. I wouldn't trust their software any further than I throw it.

Re: Microsoft's Linux Kernel

#118

Post by mlevental »

filmgirlcw wrote: post #115
mlevental wrote: post #98

Earlier quoted context omitted.

do you know if there's hardware pass through for gpu? ie can you do cuda stuff in wsl2?

On Twitter, the team has said [1] it won’t be supported at launch but that this will make it easier to add that support in the future. [1]: https://twitter.com/tara_msft/status/1125888319974400000?s=2...

thanks

Re: Microsoft's Linux Kernel

#120

Post by meruru »

ISL wrote: post #57
AdmiralAsshat wrote: post #36

All the Linux with none of the freedom. Yaaaay?

The real advantage is that, if all of the applications are written for linux, one day, society may notice that the binary blob is vestigial/removable.

That's a big "if". Microsoft will keep pushing their proprietary platforms.
Post reply on HN