On Wed, May 18, 2011 at 09:27:23PM -0400, Genes MailLists wrote:
On 05/18/2011 09:06 PM, Matthew Garrett wrote:
> On Wed, May 18, 2011 at 07:42:17PM -0400, Simo Sorce wrote:
>>
>> You know, it's not like there is a choice for many models ...
>
> That's really not a given. For anything short of us having to send http
> requests, there's no fundamental reason why this can't be done
> in-kernel. For cases where we do have to send http requests, well,
> that's kind of a separate issue (you want to run java after you've
> unmounted all the filesystems?)
>
A bunch of individual UPS kernel drivers feels inappropriate to me.
Drivers have to live somewhere. We've traditionally put them in the
kernel rather than in userland.
But if there is some common kernel functionality that could be used
by
and help all UPS devices that feels more reasonable. What functionality
do think needs be added to the kernel so that the userspace UPS code
could be better ?
If the kernel is able to send requests itself then it can do so when it
would otherwise be requesting that the firmware halt the system. That
means you can guarantee that the system is in a consistent state.
--
Matthew Garrett | mjg59(a)srcf.ucam.org