[BUG] S3 Module / State Configuration Not Aligned

Description

The S3 module expects the s3 configuration to be:

s3.key: key
s3.location: eu-west-1

Code: https://github.com/saltstack/salt/blob/master/salt/modules/s3.py

Docs:

Warning

This is literally the pillar key s3.keyid or the config option s3.keyid, not:

s3:
  keyid: blah

However, when using s3:// as a file source - s3_opt uses ['s3']['key']

https://github.com/saltstack/salt/blob/4f3cb36955d84c6793efcbeabf53963067922d90/salt/fileclient.py#L554

Setup

This means that a backup and restore state like this will never work unless the s3 configuration is duplicated.

letsencrypt-create-backup:
  cmd.run:
    - name: tar -cvf /tmp/letsencrypt.tar ./letsencrypt
    - cwd: /etc

letsencrypt-upload-backup:
  module.run:
    - name: s3.put
    - bucket: {{ pillar['s3']['bucketName'] }}
    - local_file: /tmp/letsencrypt.tar
    - path: {{ grains.nodename }}/letsencrypt.tar

teleport-s3-sync-down:
  archive.extracted:
    - name: /etc
    - source: s3://{{ pillar['s3']['bucketName'] }}/{{ grains.nodename }}/letsencrypt.tar
    - skip_verify: True
    - user: root
    - group: root
    - if_missing: /etc/letsencrypt
    

Start by comparing configuration handling in salt/modules/s3.py with the s3_opt lookup in salt/fileclient.py around line 554. Reproduce the documented S3 settings with the backup and restore state shown in the issue, then verify that the same configuration works for both s3.put and an s3:// file source without duplication.