SSD on the nNetwinder

D. Hugh Redelmeier hugh at mimosa.com
Mon Apr 1 06:48:34 PDT 2019


| From: Jon Bradbury <dr_zee at ntlworld.com>

| I’m beginning to think this has been a monumental waste of time.. unless 
| anyone has any other suggestions?

I've been reading your experiments with interest.  Thanks for being a
pioneer and thanks for keeping us informed.

ATA gradually evolved, all the time keeping backward compatibility.
Mostly.  Some niggling worries:

- disks grew bigger and broke several limits.  This would break
  compatibility when disks were big.  But there were often hacks.
  Some limits are documented in
	<https://en.wikipedia.org/wiki/Parallel_ATA>

  What geometry does the netwinder see for the SSD?

- Some accidental limits in systems are due to code not being tested
  for such sizes.  (Think: overflow when too short a representation
  is used.)  I imagine that 120G has been tried before in a
  Netwinder.  Consider a test: allocate partitions only within
  the first 10G of you SSD.

- I've had flaky PATA to SATA adapters.  Mostly, I thought, due to
  buying cheap no-name adapters.  I never did find a pattern to the
  sporadic failures.

- ATA defines some pages of information that the drive provides and
  the computer can query.  As time went on, I got the feeling that
  some drives (especially SSDs) didn't bother to provide some of the
  older, superseded stuff.

- [This does not seem to be your problem, but I'll mention it]
  "Advanced Formatting" was introduced and sectors were no longer
  512 bytes.  This would likely throw the Netwinder for a loop.
	<https://en.wikipedia.org/wiki/Advanced_Format>


More information about the Netwinder mailing list