Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 19 Feb 2016 17:55:26 +0600
From:      Alexey Dokuchaev <danfe@nsu.ru>
To:        John Baldwin <jhb@freebsd.org>
Cc:        Kevin Oberman <rkoberman@gmail.com>, Adrian Chadd <adrian@freebsd.org>, "current@freebsd.org" <current@freebsd.org>, Stefan Ehmann <shoesoft@gmx.net>, "freebsd-mobile@freebsd.org" <freebsd-mobile@freebsd.org>
Subject:   Re: Thinkpad T410: resume broken
Message-ID:  <20160219115525.GA12543@regency.nsu.ru>
In-Reply-To: <9693306.i8x5an8c0H@ralph.baldwin.cx>
References:  <53762216.8020205@gmx.net> <2846385.IUIOUlEDYf@ralph.baldwin.cx> <20160218154712.GA56997@regency.nsu.ru> <9693306.i8x5an8c0H@ralph.baldwin.cx>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Feb 18, 2016 at 09:48:40AM -0800, John Baldwin wrote:
> On Thursday, February 18, 2016 09:47:12 PM Alexey Dokuchaev wrote:
> > On Thu, Feb 18, 2016 at 06:50:34AM -0800, John Baldwin wrote:
> > > On Thursday, February 18, 2016 08:13:38 PM Alexey Dokuchaev wrote:
> > > vesa.ko shouldn't be working with KMS.  KMS turns off the legacy VGA
> > > emulation in the hardware when it starts which prevents VESA from
> > > working (VESA requires the legacy VGA interface).
> > 
> > Hmm, could this explain why vesa.ko can be loaded from loader.conf, but
> > not via kldload(8)?  OTOH, kldload'ing it later does not seem to work
> > (ir)regardless of whether i915kms.ko is loaded or not...
> 
> You might load it from loader.conf but it might then fail with an error
> about failing to register in dmesg.  In that case it is effectively
> ignored.

By "can be loaded from loader.conf" I mean that it's reported in kldstat(8)
output.  I'm still unsure if I understand all implications of having or not
having VESA kernel option or vesa.ko loaded in vt(4)+KMS world.  E.g., do I
understand correctly that vesa.ko is not needed (or might even cause certain
problems) with vt(4)+KMS, and "options VESA" is left in GENERIC as part of
syscons(4) support, and will likely go south together with syscons(4) some
day?

> > I'm about to try fresh -CURRENT on some HP AMD APU-based laptops, hence
> > I'll ask: do [Intel or Radeon graphics] laptops suspend/resume without
> > issues in X11 or on the naked console as well?  (In a shop, I could only
> > quickly test our X11-less memstick image.)
> 
> Once KMS is loaded they resume fine.  They require the KMS driver to turn
> the LCD panel back on.  Without KMS they also "resume" but the screen is off
> so you can't see anything.  However, you can type blind and run commands.
> If the network is up you can ssh into the laptop, etc. after resume.
> [...]
> On HEAD you need to set kern.vty=sc to use sc(4) instead of vt(4).  For
> suspend/resume without X that should be sufficient.  You would only need
> the old drm drivers for X.
> [...]
> Note that for the HP netbook, resume in console with vt(4) does not work
> without KMS (due to vt(4) not supporting VESA).

OK I see, makes sense.  So my best bet would be to try suspend/resume with
vt(4)+KMS, then if it fails try with kern.vty=sc (in the context of GENERIC
kernel and pure console).

./danfe



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160219115525.GA12543>