Xorg/xfce4 failing on Dual Processor G4 PowerMac's BUT Single Processor G4 PowerMac works (same boot SSD)...

Mark Millard markmi at dsl-only.net
Fri Sep 5 01:00:15 UTC 2014


The below includes it reporting getting a "failed to set mtrr: File exists" notice. (A Intel P6+ concept?)


I have since rebuilt (buildworld, kernel, installworld; portmaster -af) from the same /usr/src/... and /usr/ports/... on that same SSD based on /etc/make.conf having:

WITH_DEBUG_FILES=
WITHOUT_CLANG= # because otherwise it leads to buildworld failures with the above enabled
WITH_DEBUG=

Unfortunately uname -a loses the r268571 reference in the process, leaving less of a clue what the build is based on.

In this context the script logfile for running on a PowerMac G4 ends up recording "failed to set mtrr: File exists":

> root at FBSDG4S0:~ # startxfce4 -- -logverbose 9^M
> /usr/local/bin/startxfce4: Starting X server
> 
> 
> X.Org X Server 1.12.4
> Release Date: 2012-08-27
> X Protocol Version 11, Revision 0
> Build Operating System: FreeBSD 10.0-STABLE powerpc 
> Current Operating System: FreeBSD FBSDG4S0 10.0-STABLE FreeBSD 10.0-STABLE #0: Thu Sep  4 00:50:31 PDT 2014     root at FBSDG4S0:/usr/obj/usr/src/sys/GENERIC powerpc
> Build Date: 04 September 2014  06:13:34AM
>  
> Current version of pixman: 0.32.4
>         Before reporting problems, check http://wiki.x.org
>         to make sure that you have the latest version.
> Markers: (--) probed, (**) from config file, (==) default setting,
>         (++) from command line, (!!) notice, (II) informational,
>         (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
> (==) Log file: "/var/log/Xorg.0.log", Time: Thu Sep  4 15:14:07 2014
> (==) Using config file: "/etc/X11/xorg.conf"
> failed to set mtrr: File exists
> xfce4-session: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.
> xinit: connection to X server lost
> root at FBSDG4S0:~ # xfwm4: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.0.
> root at FBSDG4S0:~ # 

The G5 PowerMac does not display the "failed to set mtrr: File exists" message when booted from the same boot SSD and startxfce4 -- -logverbose 9 is used (but with xorg.conf's BusID changed to match the PowerMac7,11's BusID). The G5 context works fine. (I've not tried the single processor G4 context yet.)

Note: I'm using the G4 NVIDIA PowerMac for most G4 expirements because it allows me to Option-Fn to get to VTn and see thing there and so I can avoid reboots.


The last thing that dmesg -a shows after the G4 startxfce4 -- -logverbose 9 is my logging in as root:

> $ dmesg -a | tail
> Configuring syscons: blanktime.
> Performing sanity check on sshd configuration.
> Starting sshd.
> Starting sendmail_submit.
> Starting sendmail_msp_queue.
> Starting cron.
> Starting background file system checks in 60 seconds.
> 
> Thu Sep  4 14:21:42 PDT 2014
> Sep  4 14:21:48 FBSDG4S0 login: ROOT LOGIN (root) ON ttyv0

I do not see anything earlier that looks to be of interest. About the only thing remotely error-like are some examples of "hid_get_item: Number of items truncated to 255".


/var/log/Xorg.0.log's ends in:

> [   182.354] (II) Module mouse: vendor="X.Org Foundation"
> [   182.354]    compiled for 1.12.4, module version = 1.9.0
> [   182.354]    Module class: X.Org XInput Driver
> [   182.354]    ABI class: X.Org XInput driver, version 16.0
> [   182.354] (II) Using input driver 'mouse' for 'Apple Optical USB Mouse'
> [   182.354] (**) Apple Optical USB Mouse: always reports core events
> [   182.354] (**) Option "Device" "/dev/sysmouse"
> [   182.354] (==) Apple Optical USB Mouse: Protocol: "Auto"
> [   182.354] (**) Apple Optical USB Mouse: always reports core events
> [   182.355] (==) Apple Optical USB Mouse: Emulate3Buttons, Emulate3Timeout: 50
> [   182.355] (**) Apple Optical USB Mouse: ZAxisMapping: buttons 4 and 5
> [   182.355] (**) Apple Optical USB Mouse: Buttons: 5
> [   182.355] (**) Option "config_info" "hal:/org/freedesktop/Hal/devices/usb_device_5ac_304_noserial_if0"
> [   182.355] (II) XINPUT: Adding extended input device "Apple Optical USB Mouse" (type: MOUSE, id 7)
> [   182.356] (**) Apple Optical USB Mouse: (accel) keeping acceleration scheme 1
> [   182.356] (**) Apple Optical USB Mouse: (accel) acceleration profile 0
> [   182.356] (**) Apple Optical USB Mouse: (accel) acceleration factor: 2.000
> [   182.356] (**) Apple Optical USB Mouse: (accel) acceleration threshold: 4
> [   182.356] (II) Apple Optical USB Mouse: SetupAuto: hw.iftype is 4, hw.model is 0
> [   182.356] (II) Apple Optical USB Mouse: SetupAuto: protocol is SysMouse

I.e., the Xorg quit is silent. Again I do not see anything earlier that looks to be of interest.



Xorg is suddenly, silently quitting. And it is leaving all associated log files with older modification date/times/content from what I can tell. No core file either. "failed to set mtrr: File exists" may explain this. I've not started investigating it yet.



Side notes:

I've not investigated yet if there is a way around gdb Xorg's run :0 under powerpc/GENERIC getting:

> cannot get thread event message: generic error


and gdb's cont then getting:

> suspend error: generic error


I've also seen for other things I've tried with Xorg involved:

> cannot find new threads: generic error


and

> no thread to satisfy query


On the G5 PowerMac's using powerpc64/GENERIC64 boot SSD's I have no such problems with threading for gdb Xorg: only the powerpc/GENERIC boot SSD's have this problem and the gdb thread issues happen on G4's and G5's when booted from powerpc/GENERIC boot SSD's.



===
Mark Millard
markmi at dsl-only.net

On Sep 4, 2014, at 9:46 AM, Nathan Whitehorn <nwhitehorn at freebsd.org> wrote:

Anything in dmesg or in /var/log/Xorg.0.log? What happens if you set logverbose 9?
-Nathan

On 08/31/14 17:27, Mark Millard wrote:
> As an example of how sudden and arbitrary the silent-quit of the X server is:
> 
> The partial display update currently on my ADC display has approximating the top 1/2 updated and the bottom half not updated (still black).
> 
> But the boundary is somewhat interesting: The last updated raster/pixel line has its left side updated (a uniform grey desktop color) and its right side not updated (still black).
> 
> So the silent abort stopped the update mid-raster/pixel line.
> 
> ===
> Mark Millard
> markmi at dsl-only.net
> 
> On Aug 31, 2014, at 4:49 PM, Mark Millard <markmi at dsl-only.net> wrote:
> 
> [E-mail history trimmed this time.]
> 
> I used script to log my startxfce4 command and what it produces. With only the ADC connected this results in various "Unhandled monitor type 0" notices as part of normal operation.
> 
> The script output shows that the X server quit. There is no core file left in the directory.
> 
> Thunar: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.
> xfwm4: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.0.
> xinit: connection to X server lost
> xfdesktop: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.0.
> 
> (xfsettingsd:984): GVFS-RemoteVolumeMonitor-WARNING **: invoking IsSupported() failed for remote volume monitor with dbus name org.gtk.Private.GPhoto2VolumeMonitor: org.freedesktop.DBus.Error.Disconnected: Connection was disconnected before a reply was received
> root at FBSDG4S0:~ # xfsettingsd: Fatal IO error 2 (No such file or directory) on X server :0.0.
> 
> The full script file contents are:
> 
> Script started on Sun Aug 31 16:34:12 2014
> You have mail.
> root at FBSDG4S0:~ # startxfce4^M
> /usr/local/bin/startxfce4: Starting X server
> 
> 
> X.Org X Server 1.12.4
> Release Date: 2012-08-27
> X Protocol Version 11, Revision 0
> Build Operating System: FreeBSD 10.0-STABLE powerpc 
> Current Operating System: FreeBSD FBSDG4S0 10.0-STABLE FreeBSD 10.0-STABLE #0 r268571: Sun Jul 13 05:15:31 UTC 2014     root at grind.freebsd.org:/usr/obj/powerpc.powerpc/usr/src/sys/GENERIC powerpc
> Build Date: 21 July 2014  07:18:49PM
>  
> Current version of pixman: 0.32.4
>         Before reporting problems, check http://wiki.x.org
>         to make sure that you have the latest version.
> Markers: (--) probed, (**) from config file, (==) default setting,
>         (++) from command line, (!!) notice, (II) informational,
>         (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
> (==) Log file: "/var/log/Xorg.0.log", Time: Sun Aug 31 16:34:23 2014
> (==) Using config file: "/etc/X11/xorg.conf"
>   XRANDR name: DVI-1
>   Connector: DVI-I
>   CRT2: INTERNAL_DAC2
>   DFP1: INTERNAL_TMDS1
>   DDC reg: 0x64
>   XRANDR name: DVI-0
>   Connector: DVI-I
>   CRT1: INTERNAL_DAC1
>   DFP2: INTERNAL_DVO1
>   DDC reg: 0x60
>   XRANDR name: S-video
>   Connector: S-video
>   TV1: INTERNAL_DAC2
>   DDC reg: 0x0
> finished output detect: 0
> Unhandled monitor type 0
> finished output detect: 1
> finished output detect: 2
> finished all detect
> Unhandled monitor type 0
> Entering TV Save
> Save TV timing tables
> saveTimingTables: reading timing tables
> TV Save done
> disable FP1
> disable FP1
> disable TV
> disable FP1
> init memmap
> init common
> init crtc1
> init pll1
> freq: 117130000
> best_freq: 117132353
> best_feedback_div: 295
> best_frac_feedback_div: 0
> best_ref_div: 34
> best_post_div: 2
> restore memmap
> restore common
> restore crtc1
> restore pll1
> finished PLL1
> set RMX
> set FP1
> enable FP1
> disable TV
> Unhandled monitor type 0
> Thunar: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.
> xfwm4: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.0.
> xinit: connection to X server lost
> xfdesktop: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.0.
> 
> (xfsettingsd:984): GVFS-RemoteVolumeMonitor-WARNING **: invoking IsSupported() failed for remote volume monitor with dbus name org.gtk.Private.GPhoto2VolumeMonitor: org.freedesktop.DBus.Error.Disconnected: Connection was disconnected before a reply was received
> root at FBSDG4S0:~ # xfsettingsd: Fatal IO error 2 (No such file or directory) on X server :0.0.
> 
> ===
> Mark Millard
> markmi at dsl-only.net
> 
> 
> 




More information about the freebsd-ppc mailing list