lkml.org 
[lkml]   [1998]   [May]   [29]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: IDE spin-up improvement?
Date
Hi!

> I don't know if there is any real value behind it but 2.1.99 seems a lot
> more patient to wait for the drives to spin up. Or then my EIDE drives
> have gotten over their menopause :)

<g> Nice, I had an old Maxtor drive here doing this. Anyway.

The kernel is patient enough to detect my drives, that's not
the problem. I just wonder if it is necessary to *wait* for
the drive to spin up (yes, it was sleeping) when the actual
detection is done. Of course, the kernel cannot know whether
the disk is sleeping. But could it be possible to call a
little function in the IDE driver much earlier in the kernel
startup process which just sends a NOP instruction to all
[E]IDE channels without waiting for them to "complete"? Or
a bogus "read" command. This would cause all drives to spin
up while the kernel meanwhile could continue its initialization.

Then, when the IDE bus is scanned for what drives are connected,
the kernel would not have to wait. This could save several
seconds.

Another thought: There are filesystems (affs, for example) that
don't need to write any additional data to the drive when
they are unmounted, assuming that a recent sync has been done.
ext2 *does* need this, however (to write a "cleanly unmounted"
stamp, I think). Okay, some fs don't need to, but on some of
my drives there are only such fs, so why does the kernel send
something to them when it unmounts these partitions?

Because again the kernel has to wait for the drives to spin up
when the partitions are unmounted (on shutdown/unmount). Why?
It's even more crazier since these partitions are mounted read-only.

ciao,
johnny
--
Trust nobody.
## CrossPoint v3.11 ##

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

\
 
 \ /
  Last update: 2005-03-22 13:42    [W:0.070 / U:0.096 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site