Tate, Hongliang Tian

# The problem Since NGO has shifted to an architecture that schedules threads by itself, the implementation inherently needs to determine the number of CPU cores on which these threads...

Benchmarks

2

Tracking status of the micro-benchmarks and application benchmarks that we plan to use to measure the performance of NGO. Reply to this issue to propose new benchmarks. # Micro-benchmarks ##...

Tracking the areas of improvement for NGO. 1. I/O syscalls (e.g., `write` and `read`) a. io_uring-based switchless technique @shuocheng b. Page cache @jessehui c. File system optimization (See ELFS) @liqinggd...

The latter name is more readable.

After the project is made open source, all the code, issues, and docs should be written in English.

This plan captures our work in July. This is a 4-week iteration. For each iteration, we want to make balanced progress in two directions, one is to add new functionalities,...

Improve the ergonomics when using spinlocks with IRQ disabled

2

This issue proposes to refactor the API of `SpinLock` to improve the ergonomics when using spinlocks with IRQ disabled。 ### Current Feature In many situations, `SpinLock` must be acquired with...

Replace `Pod` with `zerocopy::FromBytes`

1

Recently, I found that we invented the `Pod` trait independently from [Google's `zerocopy` crate](https://github.com/google/zerocopy), in which the same concept is called `FromBytes`. ### Current Feature The `Pod` trait represents a...

Inject the page allocator into ostd

3

To achieve the design goal of keeping the size of `ostd` to a bare minimum, this proposal suggests extracting all page allocation code from `ostd`. This can be done by...

### Describe the bug This is how futex wait is implemented. ```rust pub fn futex_wait_bitset( futex_addr: Vaddr, futex_val: i32, timeout: &Option, bitset: FutexBitSet, ) -> Result { debug!( "futex_wait_bitset addr:...