Leftover mentions of gfx-hal

[memoryruins]

Currently, there are a few leftover mentions of gfx-hal after the switch to wgpu-hal in #1471. Originally I was going to mention the "About" section in matrix, but then I noticed some other places so I'm opening this in case it's easier to track.

---

The "About" section in github's sidebar mentions gfx-hal. The use of "native" might be worth removing to avoid confusing with wgpu-native which is in another repo.

Native WebGPU implementation based on gfx-hal

---

The "Development" section in https://github.com/gfx-rs/wgpu/blob/master/wgpu/README.md has instructions for patching gfx-hal. The entire section seems safe to remove.

If you need to test local fixes to gfx or other dependencies, the simplest way is to add a Cargo patch. For example, when working on DX12 backend on Windows, you can check out the latest release branch in the gfx-hal repository (e.g. currently hal-0.8) and add this patch to the end of Cargo.toml

The description in wgpu-core refers to gfx-hal.

wgpu/wgpu-core/Cargo.toml

Line 6 in 0c76819

description = "WebGPU core logic on gfx-hal"

A doc comment in wgpu-core refers to a gfx-hal limitation.

wgpu/wgpu-types/src/lib.rs

Lines 205 to 209 in 0c76819

	/// time for operations between them to finish.
	///
	/// Due to gfx-hal limitations, this is only supported on vulkan for now.
	///
	/// Supported Platforms: