Blocking access to removable drives that haven't spun up yet

Posted Oct 8, 2011 23:49 UTC (Sat) by Los__D (guest, #15263)
In reply to: Blocking access to removable drives that haven't spun up yet by Richard_J_Neill
Parent article: A Plumber's Wish List for Linux

But does the drive tell the host controller that something is happening, and does the controller have the ability pass that information on? Without that, the OS can't do much.

to post comments

Blocking access to removable drives that haven't spun up yet

Posted Oct 9, 2011 11:44 UTC (Sun) by tialaramex (subscriber, #21167) [Link] (3 responses)

I won't swear to it, but I have a feeling my older optical drives (when I still used one more than once in a blue moon) would actually block an I/O operation while spinning up with a new disc.

So I suspect this is a drive firmware decision and not amenable to influence by the operating system.

Blocking access to removable drives that haven't spun up yet

Posted Oct 10, 2011 22:55 UTC (Mon) by smoogen (subscriber, #97) [Link] (2 responses)

I believe this can happen with the new Green drives also. Your drive is spun down and you are talking to "cache" which if what you are wanting was something it had gotten recently is good enough. Otherwise you are blocked until the drive spins up.

Blocking access to removable drives that haven't spun up yet

Posted Oct 12, 2011 7:09 UTC (Wed) by Los__D (guest, #15263) [Link] (1 responses)

Yes, that is pretty standard, and I guess that the drive just lets the controller wait for the data. Problem is that CD/DVD/Blu-Ray drives doesn't do that, they send a no-disc (or something to that effect) instead, even if it is spinning up the disc.

I have no idea why it's done that way, it seems rather silly when the drive knows that data is (probably) on the way.

Blocking access to removable drives that haven't spun up yet

Posted Oct 15, 2011 0:11 UTC (Sat) by giraffedata (guest, #1954) [Link]

I'm not sure which drives we're talking about, but I've had CD drives for about 10 years that spin down after a period of inactivity, and when the client sends a read command, the drive spins up while the command executes.

However, when you load a disk, before it spins up and the drive otherwise loads the disk, commands fail immediately.

I don't know any reason a "green" drive would be different.

But it does seem like an obvious idea to have the drive hold off on completing any commands while loading is taking place.