Martin Kröning
Martin Kröning
> edit: not sure what the failing test is nor why it fails... Firecracker is a bit flaky on CI nowadays, unfortunately.
@dependabot recreate
Also see https://github.com/hermit-os/kernel/pull/1885.
What do you mean by breaking out the commits? You can rebase if you like, but I still need to write down my opinion about this before this is ready...
This is failing starting at [`fcd60b5`](https://git.trustedfirmware.org/plugins/gitiles/arm-firmware-crates/arm-gic.git/+/fcd60b5f7d2a6613ba066f54e6c373866a6009ca), so we need to adapt to using `GicRedistributor`.
https://github.com/hermit-os/kernel/pull/1982 will resolve the single-thread regressions (https://github.com/hermit-os/kernel/issues/1968). After that is done, we should measure this again. Two things might slow us down here, though: 1. A lot of task switching...
Thank you so much for putting the work into creating this issue! :) The plan looks good to me. You are welcome to start working on this. The first step...
Ah, that explanation makes sense. 👍 Our scheduler may be suboptimal in some scenarios. We currently assign threads to cores when spawning those and don't have work-stealing. Furthermore, we don't...
Why would we want this to be read-only and not implement it using the kernel's existing Ramfs/Romfs VFS nodes?
Are the small string allocations due to `ThinTree::Directory(BTreeMap)`? Would it be possible to avoid intermediate allocations here and go from parsing to the device tree without the intermediate step? Or...