Fixed a deadlock caused by locking the device's snatchable lock after locking the queue's pending writes

[RedMindZ]

Connections
May be connected to some of the open deadlock issues, but I haven't investigated them thoroughly.

Description
There is a deadlock in Queue::write_buffer and Queue::write_staging_buffer that is caused by taking the device's snatchable lock after locking the queue's pending writes. This violates the lock ranking, which states that the device's snatchable lock must be taken first. This is fixed by simply reversing the order in which the locking occurs.

Testing
I initially caught this deadlock by updating wgpu from v0.20 to to v25.0, and having my application deadlock afterwards.
After implementing the change, the application no longer deadlocks.

Squash or Rebase?
This is a single commit.

Checklist

• Run cargo fmt.
• Run cargo test.
• Run cargo xtask test to run tests.
• If this contains user-facing changes, add a CHANGELOG.md entry.

[ErichDonGubler]

A great first contribution to WGPU. 🙂 Glad you could figure out the problem!