NetBSD Looking To Improve QEMU Support

[phoronix]

Phoronix: NetBSD Looking To Improve QEMU Support

Upstream NetBSD developers are looking at dropping support for sub-optimally supported hosts/platforms if there are not maintainers willing to take over the responsibilities. As such, there's now a NetBSD volunteer looking to improve their OS support on QEMU with this being an important piece of the open-source virtualization stack...

http://www.phoronix.com/scan.php?page=news_item&px=QEMU-Better-NetBSD

[illwieckz]

Michael

Upstream NetBSD developers are looking at dropping support for sub-optimally supported hosts/platforms if there are not maintainers willing to take over the responsibilities.

You mean “Upstream QEMU developers”?

The warning comes from QEMU: http://wiki.qemu.org/ChangeLog/2.9#W...d_host_systems

It's about running QEMU on NetBSD, not about running NetBSD on QEMU.

[mulenmar]

Given that most of the systems in question are free software in active development, the QEMU developers should stop whining about "not having those systems" and:

1. &@#$ing download the OS
2. &@#$ing incorporate automated testing
3. &@#$ing communicate with the developers of the target platform
4. &@#$ing hire people willing to do the work, if they're too lazy.

A working emulation layer is just a *little bit vital* for developers trying to write and test cross-platform native code, given that the whole "Wintel" anti-competition debacle crushed nearly all providers of desktop-performent hardware not x86/x86_64-based.* Can't test your code in other CPU architectures -- and crush otherwise-unseen bugs -- without having the hardware or an accurate emulator to run it!

Additionally, saying "We don't want to support OpenBSD" is pretty much advertising that you either know if or don't care that your code would never pass security inspection.



* Show me an ARM-based desktop-level board, and I'll show you a toy with soldered-on RAM and CPU you can't upgrade, and in most cases, doesn't have a SATA port. Show me the Sam440/460 and AmgiaOne X1000/X5000 boards, and I'll show you where said board *used* to be, then run away to install Debian on it.

[illwieckz]

hire people willing to do the work, if they're too lazy.

That's exactly what they just did, not because they're lazy, but because it's not their job.

Additionally, saying "We don't want to support OpenBSD" is pretty much advertising

1. It's not about OpenBSD but NetBSD.
2. It's not about QEMU supporting NetBSD, but NetBSD supporting QEMU.

The issue is about NetBSD not contributing enough to QEMU to be able to run QEMU on NetBSD (that's their problem after all).

[danieru]

Awesome! Good to see NetBSD and Minix with improvements lately.

[mulenmar]

That's exactly what they just did, not because they're lazy, but because it's not their job.

Maybe you should read the article and links again? The QEMU developers aren't hiring outsiders to make sure their software works on other UNIX platforms, they're telling the developers of the other platforms, "We can't be arsed to make sure our software works on anything but our own systems". Not even trying to get them to contribute to a common, cross-UNIX-kernel shim layer, either.

1. It's not about OpenBSD but NetBSD.
2. It's not about QEMU supporting NetBSD, but NetBSD supporting QEMU.

The issue is about NetBSD not contributing enough to QEMU to be able to run QEMU on NetBSD (that's their problem after all).

Reread http://wiki.qemu.org/ChangeLog/2.9#W...d_host_systems and you'll clearly see that QEMU told OpenBSD the same damn thing.

[gilboa]

Maybe you should read the article and links again? The QEMU developers aren't hiring outsiders to make sure their software works on other UNIX platforms, they're telling the developers of the other platforms, "We can't be arsed to make sure our software works on anything but our own systems". Not even trying to get them to contribute to a common, cross-UNIX-kernel shim layer, either.

You seem to be incapable of grasping the basic concept of open source development.
Those who do the work, get to make the decisions.
The rest, can bitch and moan phoronix forum...

If OpenBSD and/or NetBSD (Or RedHat for that matter) want a seat at the table, *they* should be *arsed* to make sure QEMU works (as a host) on their OS (or pay someone if they are too lazy). Not the other way around.

- Gilboa