Re: amd64 radeondrmkms near-invisble text

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

On Mon, 23 Feb 2015 14:04:22 -0600, eric%cirr.com@localhost (Eric Schnoebelen)
wrote:

> I've seen this on a DELL 1950.

Mostly the 1U version of a 2950, yes?  So mostly the same hardware.

> My solution was to add "userconf disable radeondrm" to
> /boot.conf for persistence (after doing it manually the first
> time to be able to see sysinst and anything else.

I did that as well for a short while.  Then I configured a custom
kernel that disabled all the DRMKMS stuff and reinstated the UMS
radeondrm driver.

The Dell PowerEdge 2950 is very happy in this state.  There was some
initial trouble starting X because of insufficient "/dev/pci??" nodes
(default 12, required 16).  After making the additional nodes, X ran
very well.

Thinking that might have had some bearing on the dim-text problem, I
booted a GENERIC kernel again, but the radeondrmkmsfb text was still
near-invisible.

The HP ProLiant DL380 G5 similarly required more "/dev/pci??" nodes
(17 of them) and X worked, but was strangely sluggish.  It would take
several seconds to redraw objects.  Not what I'd expect from dual-quad
2.5GHz Xeons.  I'm thinking it might have to do with the ciss/sd0
SAS RAID-1 claiming that one of the units is in danger of imminent
failure (just using the disks the machine happened to come with--probably
why it was discarded by its previous owner).

The Dell PowerEdge 2850 did not like running X at all.  I thought it
was completely hung.  After regaining control, I put a serial console
on it and discovered that Xorg claimed 104% of CPU (and rising) while
the console spewed the following message:

[...]
info: [drm] wait idle failed status : 0x80010140 0x00000000
info: [drm] wait idle failed status : 0x80010140 0x00000000
info: [drm] wait idle failed status : 0x80010140 0x00000000
[...]

> I'd not realized there was anything vaguely readable on the
> screen.  To me, it just  looked like the cursor was in the lower
> left quadrant, about half size (or less) and blinking..

:   It's the wild colour scheme that freaks me.  When you try and
:   operate one of these weird black controls which are labelled in
:   black on a black background a small black light lights up black to
:   let you know you've done it.  What is this?  Some kind of
:   intergalactic hyperhearse?
:
:     -- Zaphod Beeblebrox  "The Restaurant at the End of the Universe"

I suppose it depends on your display settings and ambient lighting (and
I suppose user's vision).  I have to just about put my nose on the screen
to distinguish characters rather than just "that's not quite black".  The
"white" text looks like a very dim light blue.   The green kernel messages
are completely invisble.

This partly reminds me of a problem from netbsd-[56] and Rage128 cards
that after switching to graphics mode, returning to text mode left
the text almost-black-on-black.  There was supposedly some utility
that would restore the text, but I was never able to find reference
to it again.

-- 
|/"\ 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