AtaPassThru: Add read_pio method for AtaRequestBuilder

[seijikun]

Every warning sign has a story behind it.
And boy ... do I have stories of actual server hardware doing "things".

Let's call this time's offending vendor Super M.
Or better yet! Let's call it: S. Micro

---

On some mainboards, we have the problem that using UDMA to send the IDENTIFY command to the HDD just hangs indefinitely until the watchdog kills the app. Sending the same command using PIO works.

Hangs:

	let request = AtaRequestBuilder::read_udma(scsi_pt.io_align(), 0xEC)?
		.with_timeout(Duration::from_millis(500))
		.use_read_buffer(bfr)?
		.build();
      let response = device.execute_command(request)?;

Works:

	let request = AtaRequestBuilder::read_pio(scsi_pt.io_align(), 0xEC)?
		.with_timeout(Duration::from_millis(500))
		.with_device_head(0xA0)
		.with_sector_count(0)
		.use_read_buffer(bfr)?
		.build();
      let response = device.execute_command(request)?;

At first I thought it hangs because there were no drives connected to the port. But no ... turns out there actually are drives connected.

Checklist

• Sensible git history (for example, squash "typo" or "fix" commits). See the Rewriting History guide for help.
• Update the changelog (if necessary)

[phip1611]

oh no.. - thanks for your ongoing valuable contributions!

[seijikun]

I added more documentation. This ATA part of the UEFI spec is really confusing.

[phip1611]

thanks!