Re: kern/49744: blank screen DL-360 radeon

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

The following reply was made to PR kern/49744; it has been noted by GNATS.

From: "John D. Baker" <jdbaker%mylinuxisp.com@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: matthew green <mrg%eterna.com.au@localhost>
Subject: Re: kern/49744: blank screen DL-360 radeon
Date: Sun, 26 Apr 2015 09:09:19 -0500 (CDT)

 I'm not the originator of this PR, so I'm not on its default distribution
 list when updates are posted.  The following was copy/pasted from the
 netbsd-bugs@ mailing list archive with some fixups/formatting.
 
 On Sun, 26 Apr 2015 08:02:07 +1000, matthew green <mrg%eterna.com.au@localhost>
 wrote:
 
 > On Wed, 8 Apr 2015, John D. Baker wrote:
 >
 > > The previously-noted hack/workaround adversely affects my ThinkPad
 > > A31p.  It does not suffer from an almost-black-on-black framebuffer
 > > console display, but the hack still produces unpleasant color effects
 > > on the X display.
 
 > hmmm, this is an RV200 chipset.  the RV250 in my old laptop doesn't
 > work at all still, with the hack in place (to not set the LUT), i get
 > correct green/white on black display, but X fails (AFAIK, the GPU
 > never acknowledges a job completing, so X keeps asking in a loop.)
 
 As I noted in an aside to kern/49710, the black-console hack causes the
 white console text appear to have red/blue pixels sprinkled through it
 and the green kernel text is unreadable as if pixels are missing.  It
 reminds me of trying to display 80-column output from an Apple ][+ or
 //e on an NTSC composite color monitor.
 
 > if you remove the hack, what happens on your system?  this is in
 > legacy_crtc_load_lut(), make sure you take the original code path (the
 > 2nd loop, with 20/10/0 shifts.)  if that works OK, then we probably
 > need to work on this hack better, since it isn't related to the actual
 > chipset, but something else.   note that the macro "ASIC_IS_RV100"
 > matches almost all the r100 and r200 chipsets.
 
 I was pondering just reverting to r1.2 of radeon_display.c (at least
 in netbsd-7 now that the hack has been pulled up).  I think you're
 suggesting to carefully "#if 0" out a portion of the code, yes?
 
 > > While tolerable for virtual terminasls (mine at least), it makes
 > > colored text on some simple web sites hard to read and graphics-rich
 > > websites practically unusable.
 
 > the hack broke X11 for my RV280 card.  all my colours were wrong.
 > having it run the original code makes it all happy.
 
 I suspect the same will be true for the graphics device in my A31p
 (modulo having to keep acceleration disabled).
 
 In contrast, my ThinkPad T42 does suffer from the black console
 problem, so needs this or a better hack, as it screws up the colors
 on X (and has to run with acceleration disabled).
 
 -- 
 |/"\ 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