Ledger unresponsive for api calls

[r4mmer]

Description

On the new unreleased version the ledger device is not responding to the getVersion call, blocking the queue and not timing out.

Initial thoughts

There could be some issues with the transport version, since we use an outdated and possibly broken version.
To attest this I created a simple script to send the same command to the device and it responded correctly.
This means the device is reachable (some users may have issues connecting and ledger already has a troubleshoot) and that the transport version is working as intended.

Investigation

As described above the ledger device did respond when calling manually but not when called from the electron main process.
This was fixed when we updated the ledger transport dependency (and its subdependencies which include a major usb update).

Another issue popped after re-establishing communication with ledger, the device would not start the wallet on the first try, the issue was that locked wallets would be redirected to the locked screen and hardware wallets when entering the locked screen would call the LOCAL_STORE.cleanWallet method which cleans the storage (and access data).
Since the locked screen makes no sense for hardware wallets, we can just ignore this redirect for hardware wallets.
The better solution would be to untangle the flow of screens during wallet initialization but this requires a larger re-work of the routing system which is outside the scope of this issue.