Re: [linux-pm] what the patches do Re: [RFC 10/15] PM / Hibernate: user, implement user_ops reader

[Nigel Cunningham <ncunningham@crca.org.au>]

Hi.

On 26/03/10 08:26, Pavel Machek wrote:
> Hi!
>
>>>>> I'm asking because if we're going to make a go of getting the in-kernel
>>>>> code in much better shape, and we have Rafael, Jiri and I - and you? -
>>>>> all pulling in the same direction to improve it, there's going to come a
>>>>> point (hopefully not too far away) where uswsusp is just making life too
>>>>> difficult, and getting rid of it will be a big help.
>>>>
>>>> We're not dropping user space interfaces used by every distro I know of.
>
> Good.
>
>>> So what's your long term plan then?
>>
>> First, improve the in-kernel thing, second, switch people to it, _then_ remove
>> the s2disk interface (after we're reasonably sure it's not used by any major
>> distro) and _finally_ simplify things after it's been removed.
>
> I'd really prefer to keep s2disk interface. It allows advanced stuff
> like internet suspend resume (yep someone is doing it), crazy stuff
> like multiple resumes from same image for fast booting (yep, some
> embedded people are doing that) and might-be-useful stuff like
> s2both...

Neither of those are impossible with in-kernel code, so I'd argue that 
there's no need to keep the s2disk interface long-term. Userspace 
helpers might be necessary for the first one (to manage the network 
interface), but I already have multiple-resumes-from-the-same-image 
support in TuxOnice (and more 'crazy stuff' like support for resuming a 
different image after writing one - that can be used to switch to an 
initramfs containing the binaries needed to power down a UPS).

Regards,

Nigel