xqemu / xqemu-manager

Simple graphical user interface to manage XQEMU
http://xqemu.com
GNU General Public License v2.0
34 stars 10 forks source link

Fix accelerator arguments to use fallback chaining #41

Closed BigBrainAFK closed 5 years ago

BigBrainAFK commented 5 years ago

This PR removes the accelerator generation function and replaces it with an accelerator fallback chain that's always present.

The kernel_irqchip argument is added if the platform is Linux otherwise its omitted.

Also replaced the import of platform by just using the already imported sys and it's platform property to determine if the current operating system is Linux or not. (this still detects Linux in WSL)

JayFoxRox commented 5 years ago

After merge, we'll have to add an issue about adding HVF once the QEMU issues with it have been resolved. We didn't need this before because we were selective, but with a catch-all approach, we should also list all options (if they don't outright crash during early boot).

Speaking of early-crashes: I'm also still a bit torn about addition of WHPX at this point - considering the ongoing discussion in https://github.com/xqemu/xqemu/issues/112#issuecomment-462393393; WHPX might still simply cause a crash, so falling back on it might lead to preventing emulation - a problem this PR tries to address (and why we don't have HVF in the fallback list either).

JayFoxRox commented 5 years ago

Apparently this was closed because author accidentally added info from master. Generally this is not a good reason to close; however, according to author, GitHub auto-closed as they temporarily removed too many commits.

In general: please contact maintainers if you need assistence with git - the history on this PR still looked fine in the end. However, an accidental auto-close is unfortunately not avoidable - mistakes do happen.

However, once closed GitHub can't re-open after force pushes (and re-opening should be avoided in general due to other GitHub oddities).

Continued in #42

In your continuation-PRs, always mention the previous PR, and in the previous PR, state a reason for the close (as I've done now).