Re: kern/49398 (i386-current GENERIC (i915drmkms) boot hangs if monitor powered off)

["John D. Baker" <jdbaker%mylinuxisp.com@localhost>]

The following reply was made to PR kern/49398; it has been noted by GNATS.

From: "John D. Baker" <jdbaker%mylinuxisp.com@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: riastradh%NetBSD.org@localhost
Subject: Re: kern/49398 (i386-current GENERIC (i915drmkms) boot hangs if
 monitor powered off)
Date: Tue, 9 Jun 2015 20:20:52 -0500 (CDT)

 On Thu, 14 May 2015, John D. Baker wrote:
 
 > On Thu, 14 May 2015, riastradh%NetBSD.org@localhost wrote:
 > 
 > > State-Changed-Why:
 > > Via the serial console, can you get a stack trace from ddb?
 > 
 > When last I tried, I could not.  The machine was completely hung and
 > required power-cycle to regain control.
 
 I've finally gotten things updated to the latest HEAD.  There has been
 no change in behavior since the last update of this PR.  As I noted
 above, the machine is completely wedged at this point, so dropping to
 DDB is not possible.  Power-cycle is the only way out.
 
 > > If not, we can try printf debugging.  I'd start by adding printfs
 > > around all the calls to drm_add_edid_modes in i915, and throughout
 > > drmfb_attach.  That should be enough to get started, but I can supply
 > > a patch if you'd like.
 
 If you could supply said patch, I think that would be the simplest way
 forward.
 
 The only other complication I see is that when the console is on the
 serial port, the VGA output is disabled when inteldrm attaches, so
 there's no video output anyway.  I read a note about "genfb" stealing
 registers, but I'm using a custom kernel which omits "genfb" entirely
 and the video output is still disabled.
 
 
 > > Also: does this happen on any machine other than your i845?
 > 
 > I may have access to other intel-graphics-equipped machines I can poke
 > at, but it will be early June before I can try.
 
 I have access to at least one, possibly two, other machines with intel
 graphics that I can test.  I'm in the process of setting those up now.
 
 
 One other possibility has occurred to me.  Perhaps the monitor I've
 been using is flaky (IIUC, EDID data should be readable even if the
 monitor is off).  Unfortunately, testing the machine in question with
 another monitor and the monitor with another machine is a rather
 difficult undertaking.
 
 -- 
 |/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
 |\ / jdbaker[snail]mylinuxisp[flyspeck]com    OpenBSD            FreeBSD
 | X  No HTML/proprietary data in email.   BSD just sits there and works!
 |/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645