Date: Fri, 18 Apr 2014 12:17:14 +0300 From: Jukka Ukkonen <jau789@gmail.com> To: x11@FreeBSD.org Subject: After the latest update xorg-server (1.12.4_5,1) fails to start Message-ID: <5350ED9A.50008@gmail.com>
next in thread | raw e-mail | index | archive | help
As the subject line already said I have not been able to use X on my
10.0-stable
after the last xorg-server update. The symptoms are shown below.
I guess the whole trouble is rooted on this...
Apr 18 11:13:04 sleipnir kernel: drmn0: <ATI Radeon HD 2600 Pro> on vgapci0
Apr 18 11:13:04 sleipnir kernel: device_attach: drmn0 attach returned 2
For some reason the flippin' radeonkms refuses to talk to the graphics card.
I could imagine it does not find all features it would like to see
supported or
it fails to recognize the chip ID or something like that.
Somehow I am prone to assume, that since the same card has worked just
fine with the ati/radeon drivers before, there is something fishy about
how the
older radeon code and radeonkms try to share their responsibilities. So,
in the
end neither one takes care of HD 2600 Pro.
The xorg.conf file is exactly the same as it has been with this graphics
card
for more than a year now.
Right at the end of Xorg.0.log there is this little additional pearl...
[ 1999.533] (--) Using syscons driver with X support (version 2.0)
[ 1999.533] (--) using VT number 9
[ 1999.549] (II) [KMS] Kernel modesetting enabled.
[ 1999.549] (WW) VGA arbiter: cannot open kernel arbiter, no multi-card
support
[ 1999.549] (**) RADEON(0): Depth 24, (--) framebuffer bpp 32
[ 1999.549] (II) RADEON(0): Pixel depth = 24 bits stored in 4 bytes (32
bpp pix
maps)
[ 1999.549] (==) RADEON(0): Default visual is TrueColor
[ 1999.550] (**) RADEON(0): Option "EnablePageFlip" "1"
[ 1999.550] (**) RADEON(0): Option "ColorTiling" "1"
[ 1999.550] (**) RADEON(0): Option "RenderAccel" "true"
[ 1999.550] (==) RADEON(0): RGB weight 888
[ 1999.550] (II) RADEON(0): Using 8 bits per RGB (8 bit DAC)
[ 1999.550] (--) RADEON(0): Chipset: "ATI Radeon HD 2600 Pro" (ChipID =
0x9589)
[ 1999.550] drmOpenDevice: node name is /dev/dri/card0
[ 1999.554] drmOpenDevice: open result is 9, (OK)
[ 1999.560] drmOpenByBusid: Searching for BusID pci:0000:01:00.0
[ 1999.560] drmOpenDevice: node name is /dev/dri/card0
[ 1999.564] drmOpenDevice: open result is 9, (OK)
[ 1999.564] drmOpenByBusid: drmOpenMinor returns 9
[ 1999.564] drmOpenByBusid: Interface 1.4 failed, trying 1.1
[ 1999.564] drmOpenByBusid: drmGetBusid reports pci:0000:01:00.0
[ 1999.564] (II) RADEON(0): Kernel too old missing accel information,
assuming accel is working
[ 1999.564] (II) Loading sub module "dri2"
[ 1999.564] (II) LoadModule: "dri2"
[ 1999.565] (II) Loading /usr/local/lib/xorg/modules/extensions/libdri2.so
[ 1999.565] (II) Module dri2: vendor="X.Org Foundation"
[ 1999.565] compiled for 1.12.4, module version = 1.2.0
[ 1999.565] ABI class: X.Org Server Extension, version 6.0
[ 1999.565] (II) Loading sub module "exa"
[ 1999.565] (II) LoadModule: "exa"
[ 1999.570] (II) Loading /usr/local/lib/xorg/modules/libexa.so
[ 1999.570] (II) Module exa: vendor="X.Org Foundation"
[ 1999.570] compiled for 1.12.4, module version = 2.5.0
[ 1999.570] ABI class: X.Org Video Driver, version 12.1
[ 1999.570] (II) RADEON(0): KMS Color Tiling: disabled
[ 1999.570] (II) RADEON(0): KMS Color Tiling 2D: enabled
[ 1999.570] (II) RADEON(0): KMS Pageflipping: enabled
[ 1999.570] (II) RADEON(0): SwapBuffers wait for vsync: enabled
[ 1999.570] (EE) RADEON(0): Kernel modesetting setup failed
[ 1999.574] (II) UnloadModule: "radeon"
[ 1999.574] (II) UnloadSubModule: "exa"
[ 1999.574] (II) Unloading exa
[ 1999.575] (II) UnloadSubModule: "dri2"
[ 1999.575] (II) Unloading dri2
[ 1999.575] (EE) Screen(s) found, but none have a usable configuration.
[ 1999.575]
Fatal server error:
[ 1999.575] no screens found
[ 1999.575]
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
[ 1999.575] Please also check the log file at "/var/log/Xorg.0.log" for
additional information.
[ 1999.575]
[ 1999.591] Server terminated with error (1). Closing log file.
The warning about missing VGA arbiter looks suspicious, but why
would a single headed system need an arbiter?
The note "Kernel too old missing accel information, ..." is supposedly
yet another indication of radeonkms having failed.
Then comes an attempt to behave gracefully anyhow until the KMS is
involved again. Since KMS is now unable to provide everything as expected,
xorg gives up and the whole thing essentially falls on its face.
Does anyone have a foggiest idea what should be done to get xorg-server
working again?
Should I somehow disable the KMS feature completely until a later date when
radeonkms might be able to handle the "Radeon HD 2600 Pro" cards properly?
Any pointers to what to do next would be appreciated.
BTW... I assume device_attach alert could report the actual error string
instead of the error code. A string like "No such file or directory"
would be
way more informative than "... returned 2". Someone who has been using
BSD systems for more than 20 years of course guesses that 2 is probably
ENOENT, but for anyone else it is quite as informative as a random number.
I would prefer UNIX systems to keep a healthy distance to reporting every
darn problem as error codes only in the tradition of the old IBM mainframes.
It feels awful to think that UNIX systems would need some specially ordained
mainframe priests to interpret the numerically expressed wishes of a genie
in the magic box.
--jau
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5350ED9A.50008>