Re: [PATCH 0/3] hotplug polling, respin

[Tejun Heo <htejun@gmail.com>]

Robin H. Johnson wrote:
> On Wed, Apr 25, 2007 at 04:08:09PM +0900, Tejun Heo wrote:
>>> The last submission of this
>>> (http://thread.gmane.org/gmane.linux.ide/13467) already had a tentative
>>> ACK, pending a link power-saving patchset, that I can't find more
>>> details on.
>> Yeap, the implementation is ACKed but interface to userland is NACKed.
>> We need sysfs nodes for ATA devices to do userland interface properly.
>> I'm working hard to get there.
> I'm not sure I follow you here. The only bit of 'userspace' interface I
> see in this series is hotplug_polling_interval. What other userspace
> bits are in question?

What got NACKed is the lack of per-port interface.  More specifically,
stretching the usage of module parameter too far.

>> Thanks a lot for taking care of these patches.  A lot of people have
>> been asking for patches for newer kernels and I couldn't find time to
>> update them.
> As always, it starts with scratching your own itch. My Fusion 500P works
> like a charm with the SIL 3132 that I have - the previous scanning issue
> where it took forever doing 15 scans (1;1,2;1,2,3;1,2,3,4;1,2,3,4,5)
> instead of just 5 seems to have gone away, but the sluggish performance
> remains (the sil24 seeming to not do SATA2 speed stuff, did you hear
> back from SIL?).

Nope.

> As a gauge of interest in the PMP stuff, I'm presently getting about 3
> emails a week asking for work on it.
> 
>> Jeff, can we make a page for PMP patches in linux-ata.org?  It would be
>> a much better place to direct people to.
> I'd host a git tree, but I don't have sufficient stable bandwidth.

We can host the git tree at kernel.org.  I can get a tree setup there
and pull from you.

-- 
tejun