[sllug-members]: Question For A Linux Kernel Guru WRT Seagate Free Agent

[Michael Heath]

I appologize that this isn't the most 'useful' answer, but I thought
I'd just chime in with a word of advice about my (very limited)
experience with such things.

It has been my experience that subtle improvements to the functioning
of a device are often not direct changes to the device's driver at
all. Even if your device has its own driver in the kernel, that driver
depends on so many other components - usb core, usb mass storage, disk
io, etc - and a change to any of these components can alter the
functionality of your device.

I once spent ages trying to determine a regression in the
functionality of my tablet stylus, only to find that a seemingly
completely unrelated in parallel device support caused the issue, The
stylus isn't even parallel.

Anyway, my point is , its often very hard to grep or browse sources or
changelogs for the change that made yor device work. If you are
really, really interested, though. I would recommend asking on the
LKML.

--
MIchael Heath
[Sent from my phone, so please ignore any horrible formatting]

On 7/26/09, Kevin Benko <benko.kevin at gmail.com> wrote:
> There must have been a change in the Linux kernel that fixed a problem
> withe the Seagate Free Agent external drives, and I am curious to know
> when/if the change occurred.
>
> Way back when the Seagate "Free Agent" drives came out, there was a
> problem with interoperability with the Linux kernel. I used the "hdparm"
> command to turn off time-out/spin-down, as a workaround.
>
> Well.... after "the great shuffling around of computers" event when we
> moved, I took the time to futz around with the dreaded Seagate Free
> Agent drives, again.
>
> They seem to work perfectly, now. I set the spin-down/time-out down to 1
> minute for testing purposes, and I now have the spin-down/time-out set
> at 5 minutes. over the past month, I have not had a single issue with
> these drives as I did back in 2007.
>
> I can only guess that there was some kernel code added to allow Linux to
> work with the Free Agent drives.
>
> The change seems to have happened somewhere between kernel 2.6.26 and
> 2.6.29, but I couldn't find anything that seemed relevant when I grepped
> the kernel source tree (and Google isn't of any help for this particular
> question).
>
> Can any kernel gurus out there tell me when/if there were changes in the
> kernel to correct Seagate's oversight with these drives?
>
> --
> Kevin Benko
>
> "It is not at all simple to understand the simple."
> -- Eric Hoffer
>
>