Re: acpica5 and acpi sub-modules (Re: cvs commit: src/sys/dev Makefile)

[YONETANI Tomokazu <qhwt+dragonfly-submit@xxxxxxxxxx>]

On Mon, Jun 14, 2004 at 11:10:36AM -0700, Matthew Dillon wrote:
> 
> :It surely does, the only problem is that I have to press reset button
> :(or power-cycle) when it supposed to resume :) For suspend/resume to
> :work properly, other drivers must support it. And I think the status of
> :suspend/resume support is almost the same as that of FreeBSD 4.x, and
> :I doubt such changes will be MFC'ed to 4.x soon.
> :
> :My main interest has been on CPU throttling and poweroff, so I didn't
> :care much about suspend/resume. Does suspend/resume work reliably on
> :recent -CURRENT? I haven't used it after having experienced panics
> :(I believe that was something related to ppp code) back in 4-CURRENT days,
> :but if can resume reliably, that's definately a nice feature to have.
> 
>    We are actually fairly close to 5.x here.  5.x has been doing a lot of
>    recent PCI power-state work which is not yet in DragonFly, but we aren't
>    as far behind as 4.x is.
> 
>    So eventually we will have a working resume (well, at least as well as 
>    fbsd-5), but it isn't a priority right now.

You're right. I didn't notice it was working because the screen turned
into white and nothing showed up when my laptop resumed from the sleep
state. Turned out that

hw.acpi.reset_video=1 (default)

doesn't work properly on my laptop, and I should have turned it off.
Now my laptop can resume from S3 state without panic or lock-up. And I'm
so glad that I don't have to restart moused after the resume.
By the way, this was when I put my laptop into S3 using acpiconf command,
which is not in our tree; if I set

hw.acpi.lid_switch_state=S3

to use lid to put into sleep state, the machine panicked when it tried
to resume. I'll try to track this down when I have time.

I also noticed that, if I boot the machine without AC adapter plugged
in and interrupt dhclient by pressing ctrl-C, it locks up suddenly and
even ctrl-alt-esc doesn't work. It doesn't lock up when AC adapter is
plugged in, so I doubt this is caused by the networking code. This is
not new and I've been seeing this lock up since last year(not sure when
I saw it the first time).

I updated my patch to compile pmtimer in acpi module so that the system
clock is adjusted when it resumed.