fix(p2p): give each underlay address fresh 10s connection timeout

[gacevicljubisa]

Checklist

• I have read the coding guide.
• My change requires a documentation update, and I have done it.
• I have added tests to cover my changes.
• I have filled out the description and linked the related issues.

Description

• Use context.WithoutCancel to give each underlay address a fresh 10s timeout
• Add explicit parent context cancellation check before each connection attempt
• Prevents later addresses from being starved of connection time
• Still respects upstream cancellations (shutdown, etc.)
• Use 45s total timeout in kademlia

Previously, all connection attempts shared the parent's 15s timeout,
causing later addresses to have insufficient time. Now each address
gets a full 15s attempt while still checking for parent cancellation
between attempts.

Open API Spec Version Changes (if applicable)

Motivation and Context (Optional)

Related Issue (Optional)

Screenshots (if appropriate):

[janos]

Looks ok.