Roy Bixler | 27 Nov 16:15 2015

can't get Nouveau console

I have an old Dell Precision M70 laptop with Nvidia NV40 series
graphics.  I tried to enable the Nouveau framebuffer on recent HEAD

nouveau*        at pci? dev ? function ?
nouveaufb*      at nouveaufbbus?

but the framebuffer doesn't start on bootup.  It does work with a
Knoppix Linux CD.  I'll attach dmesg outputs for Linux and NetBSD HEAD
booting without Nouveau.  I think the following lines may be

pchb0 at pci0 dev 0 function 0: vendor 8086 product 2590 (rev. 0x03)
agp0 at pchb0: can't find internal VGA config space

I also attach a screenshot I took showing what happens with Nouveau,
where I end up in the debugger and don't think I can otherwise save
the "dmesg" output.  I did try once booting a Nouveau-enabled kernel
under VESA mode.  I saw Nouveau driver lines similar to what I saw
under Linux and I could try taking a screenshot of those, except that
there are many of them and I'm not sure which ones would be most
useful.  How should I proceed?


Roy Bixler <rcbixler <at>>
"The fundamental principle of science, the definition almost, is this: the
sole test of the validity of any idea is experiment."
-- Richard P. Feynman
(Continue reading)

Richard PALO | 24 Nov 16:40 2015

bad karma between mouse and ati6 driver

Looking for advice on the following.
> ...
> [  4175.385] (II) RADEON(1): RADEONScreenInit d0000000 0 0
> [  4175.522] (II) RADEON(1): Dynamic Power Management Disabled
> [  4175.522] mc fb loc is 00df00d0
> [  4175.522] (II) RADEON(1): RADEONInitMemoryMap() : 
> [  4175.523] (II) RADEON(1):   mem_size         : 0x10000000
> [  4175.523] (II) RADEON(1):   MC_FB_LOCATION   : 0x00df00d0
> [  4175.523] (II) RADEON(1):   MC_AGP_LOCATION  : 0x003f0000
> [  4175.523] (II) RADEON(1): Depth moves disabled by default
> [  4175.531] (II) RADEON(1): RADEONRestoreMemMapRegisters() : 
> [  4175.531] (II) RADEON(1):   MC_FB_LOCATION   : 0x00df00d0 0x0f3f0f00
> [  4175.531] (II) RADEON(1):   MC_AGP_LOCATION  : 0x003f0000
> [  4175.542] (==) RADEON(1): Backing store enabled
> [  4175.542] (WW) RADEON(1): Direct rendering disabled
> [  4175.542] (II) RADEON(1): Acceleration enabled
> [  4175.542] (==) RADEON(1): DPMS enabled
> [  4175.542] (==) RADEON(1): Silken mouse disabled
> [  4175.542] (EE) RADEON(1): Hardware cursor initialization failed
> [  4175.542] (II) RADEON(1): Using software cursor
> [  4175.542] (II) RADEON(1): Textured video requires CP on R5xx/R6xx/R7xx/IGP
> [  4175.550] (II) RADEON(1): RADEONRestoreMemMapRegisters() : 
> [  4175.551] (II) RADEON(1):   MC_FB_LOCATION   : 0x00df00d0 0x00df00d0
> [  4175.551] (II) RADEON(1):   MC_AGP_LOCATION  : 0x003f0000
> [  4175.561] (II) RADEON(1): crtc(0) Clock: mode 94500, PLL 945000
> [  4175.561] (II) RADEON(1): crtc(0) PLL  : refdiv 2, fbdiv 0x54(84), fracfbdiv 0, pdiv 12
> [  4175.642] (II) RADEON(1): RandR 1.2 enabled, ignore the following RandR disabled message.
> [  4175.643] (--) RandR disabled
> [  4175.666] (II) AIGLX: Screen 0 is not DRI2 capable
> [  4175.666] (EE) AIGLX: reverting to software rendering
(Continue reading)

Ryosuke Moro | 6 Nov 22:41 2015

typo in src/external/mit/xorg/lib/pixman/Makefile


Index: Makefile
RCS file: /cvsroot/src/external/mit/xorg/lib/pixman/Makefile,v
retrieving revision 1.33
diff -u -p -r1.33 Makefile
--- Makefile	6 Nov 2015 21:32:22 -0000	1.33
+++ Makefile	6 Nov 2015 21:36:24 -0000
 <at>  <at>  -61,7 +61,7  <at>  <at>  SRCS+=	pixman-sse2.c pixman-ssse3.c
 COPTS.pixman-sse2.c=	-msse2 -fvisibility=hidden
 COPTS.pixman-ssse3.c=	-msse3 -mssse3 -fvisibility=hidden
-MKDEPFLAGS+=		-msse2 -mssse3 -mssse3 -fvisibility=hidden
+MKDEPFLAGS+=		-msse2 -msse3 -mssse3 -fvisibility=hidden

 .if ${MACHINE_ARCH} == "powerpc"



Ottavio Caruso | 4 Nov 09:55 2015

xdm prevents me from shutting the system down clean


I've been using 7.0-rc3 on amd64 for a few months. I haven't bothered
upgrading 7.0 proper so far. I am using native x11.

The only way I can shutdown the system clean is to open a virtual
console, log in as root, kill xdm manually and then issue the poweroff

Not your usual user friendly set up, is it?

In all other cases, whether by pressing the power button or issuing
"sudo poweroff" or "sudo shutdown -h now",  the system hangs forever.

When I enable debugging messages, I see that it is repeatedly try to
kill a pid that is associated with xdm.

Where would I start debugging this issue?



Robert Swindells | 31 Oct 18:44 2015

gallium build problem

Is anyone looking at the build problem in gallium that is happening
for most ports ?

It seems to me that there is no need to build gallium at all on most
of the ports that are failing.

I have disabled the build of it locally for a couple of ports but
it would be cleaner if there was something defined in share/mk to
enable building the 3d stuff on selected ports.

Robert Swindells

Ryosuke Moro | 30 Oct 16:37 2015



i don't have nVidia cards, though.

Index: Makefile.defines
RCS file: /cvsroot/src/external/mit/xorg/lib/libdrm/Makefile.defines,v
retrieving revision 1.1
diff -u -p -r1.1 Makefile.defines
--- Makefile.defines	17 Mar 2014 08:01:18 -0000	1.1
+++ Makefile.defines	30 Oct 2015 15:24:15 -0000
 <at>  <at>  -10,7 +10,5  <at>  <at>  CPPFLAGS+=	-DHAVE_INTTYPES_H \

right ?


matthew green | 29 Oct 09:36 2015

nouveaudrm -- actually worth trying now!

hi folks.

i seem to have figured out the final bits for getting nouveau limping
along in a working state.  i can't stress enough that Taylor and Chuq
really did the vast majority of this work, i feel like the final 5%
at best...

it's probably worth testing in -current now if you want.  you'll need
to enable these lines in your config:

   nouveau*        at pci? dev ? function ?
   nouveaufb*      at nouveaufbbus?

and you might want to disable nouveau debug -- my console spews a really
large amount, but only if it works ;)  you'll have to do this by editing
files.nouveau to have this instead:

   makeoptions     nouveau CPPFLAGS+="-DCONFIG_NOUVEAU_DEBUG=0"
   makeoptions     nouveau CPPFLAGS+="-DCONFIG_NOUVEAU_DEBUG_DEFAULT=0"

i've so far only testing with i386, but i'll be building an amd64
world to test next.  you'll need rebuilt kernel and xsrc from
- -current to test (i don't know about pkgsrc x11 offhand.)



Mariusz Wawrzyniak | 26 Oct 15:35 2015

Nouveau driver on nvidia 8500GT (amd64)


Today I built the kernel from the latest sources with nouveau, LOCKDEBUG,
DEBUG and KMEM_GUARD_DEPTH=30000 enabled. It gave me no console
and the relevant fragment of /var/log/messages looks like this:

Oct 26 15:00:12 netbsd syslogd[580]: Exiting on signal 15
Oct 26 15:04:19 netbsd syslogd[606]: restart
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[  
VBIOS][nouveau0] 0xd016[ ]: CONDITION_TIME	0x01 0xff
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[  
VBIOS][nouveau0] 0xd019[ ]: RESUME
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[  
VBIOS][nouveau0] 0xd01a[ ]: COPY_NV_REG	R[0x61000c] &= 0xffff0000 |=
((R[0x610000] >> 0x00) & 0x0000ffff ^ 0x00000000)
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[  
VBIOS][nouveau0] 0xd030[ ]: NV_REG	R[0x61000c] &= 0xbfffffff |= 0x40000000
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[  
VBIOS][nouveau0] 0xd03d[ ]: CONDITION_TIME	0x02 0xff
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[  
VBIOS][nouveau0] 0xd040[ ]: RESUME
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[  
VBIOS][nouveau0] 0xd041[ ]: DONE
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau D[
DEVINIT][nouveau0] initialised
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[   
GPIO][nouveau0] use(+1) == 1
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[   
GPIO][nouveau0] initialising...
Oct 26 15:04:19 netbsd /netbsd: drm kern debug: nouveau T[ 
(Continue reading)

Adrien Fernandes | 25 Oct 17:05 2015

xfree86 build error

I am building meta-pkgs/modular-xorg from current (2015/10/24) and, when it builds x11/modular-xorg-server, it stops on this error :

Making all in .
gmake[4]: Entering directory '/root/HDD/pkgsrc/x11/modular-xorg-server/work/xorg-server-1.17.2/hw/xfree86'
  CCLD     Xorg
sdksyms.o:(.data+0x1a88): undefined reference to `xf86EnableIO'
sdksyms.o:(.data+0x1a8c): undefined reference to `xf86DisableIO'
sdksyms.o:(.data+0x1a90): undefined reference to `xf86SetTVOut'
sdksyms.o:(.data+0x1a94): undefined reference to `xf86SetRGBOut'
common/.libs/libcommon.a(xf86Configure.o): In function `DoConfigure':
xf86Configure.c:(.text+0x564): undefined reference to `xf86EnableIO'
common/.libs/libcommon.a(xf86Events.o): In function `xf86VTLeave':
xf86Events.c:(.text+0xb0c): undefined reference to `xf86DisableIO'
common/.libs/libcommon.a(xf86Events.o): In function `xf86VTEnter':
xf86Events.c:(.text+0xec8): undefined reference to `xf86EnableIO'
common/.libs/libcommon.a(xf86Init.o): In function `InitOutput':
xf86Init.c:(.text+0x23a0): undefined reference to `xf86EnableIO'
xf86Init.c:(.text+0x2540): undefined reference to `xf86EnableIO'
Makefile:794: recipe for target 'Xorg' failed
gmake[4]: *** [Xorg] Error 1
gmake[4]: Leaving directory '/root/HDD/pkgsrc/x11/modular-xorg-server/work/xorg-server-1.17.2/hw/xfree86'
Makefile:844: recipe for target 'all-recursive' failed
gmake[3]: *** [all-recursive] Error 1
gmake[3]: Leaving directory '/root/HDD/pkgsrc/x11/modular-xorg-server/work/xorg-server-1.17.2/hw/xfree86'
Makefile:659: recipe for target 'all' failed
gmake[2]: *** [all] Error 2
gmake[2]: Leaving directory '/root/HDD/pkgsrc/x11/modular-xorg-server/work/xorg-server-1.17.2/hw/xfree86'
Makefile:603: recipe for target 'all-recursive' failed
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory '/root/HDD/pkgsrc/x11/modular-xorg-server/work/xorg-server-1.17.2/hw'
Makefile:771: recipe for target 'all-recursive' failed
gmake: *** [all-recursive] Error 1
*** Error code 2

make[2]: stopped in /root/HDD/pkgsrc/x11/modular-xorg-server
*** Error code 1

make[1]: stopped in /root/HDD/pkgsrc/x11/modular-xorg-server
*** Error code 1

make: stopped in /root/HDD/pkgsrc/meta-pkgs/modular-xorg

NetBSD 7.99.21 evbarm-earmv7hf
built from sources without X
matthew green | 13 Oct 06:44 2015

nouveau driver runs and crashes!

hi folks.

just a quick status update.

those of you watching source-changes may notice that we recently got the
nouveaudrm driver able to attach a text console.

starting X crashes, but it's at a point you can either use the text console
or help debug the crash :-)  you'll need the very latest -current, as all
the components were only just recently commited.

thanks to fair <at> netbsd for sending me 3 nvidia graphics cards!  also to
Taylor and chuq who ported most of the code.  so far, i've only tested on
a GTX 750 card, but i plan to try the GTX 450(?) and 610 soon enough.

it's vaguely possible i broke the build or other parts of X.  if so,
please file a PR or contact me via email.  thanks!


Thomas Klausner | 29 Sep 11:31 2015

kqueue: SIGIO?


During the XDC last week, Keith Packard talked about a select(2) issue
in xserver he would like to fix with epoll and its support for SIGIO.
Is there a similar feature in kqueue in NetBSD?