Failure to boot and dead processes

Craig Dooley cd5697 at albany.edu
Sat Dec 6 16:24:17 PST 2003


Actually, I have fixed almost every issue on the list.  I built a kernel
without ACPI, and no hung processes, every boot completes, and rebooting
works properly.  The X Server crashing was due to the 855 chipset not
supported by AGP, but using the PCI IDs from here and merging some more
changes from FreeBSD5, it works, including DRI with the
XFree86-4-Server-snap.  I want to try to add AGP 8x from FreeBSD 5 and
then send off the patch.  Also, there are a lot of PCI ID's that can be
added to the ac97 code without any other modification I think.  I have
most of the 20031203 ACPI code merged also, so I'm going to see if I can
get that working also.

-Craig

On Sat, Dec 06, 2003 at 12:45:36PM -0800, Matthew Dillon wrote:
>     Well, this isn't going to be easy.  First thing is to see if
>     you can get it to work with FreeBSD-stable.  If you can then there
>     is a good chance we can get it to work with DragonFly.
> 
>     ACPI is governed by the acpica device.  If you remove the device
>     it should disable acpi.
> 
> 					-Matt
> 					Matthew Dillon 
> 					<dillon at xxxxxxxxxxxxx>
> 
> 
> :Since I've gotten this new laptop, I've had a couple problems.  One is
> :no apm showing up, but I can deal with that.  ACPI though, has some
> :major issues.  Over 50% of the time, boot will not finish.  Things will
> :hang after ad0 and acd0 are found, but when it does finish the next
> :message is about an ACPI embedded controller failure to connect.  Is
> :there a way I can disable this subsystem to figure out if it is the
> :problem?  Next is after heavy disk io, especially cvs and compilation,
> :processes will start being stuck in state inode, but taking 0% cpu.  At
> :this point they are unkillable and the kernel will lose buffers when
> :trying to shutdown.  Is this related to the ata controller?  Should
> :atang help this, and if so is there a testable patch?  Third, rebooting
> :does not work.  I get the rebooting prompt, but it won't go anywhere.  I
> :tried taking out the DISABLE_BAD_KEYBOARD define or something like that
> :in cpu_restart_real (I think thats the function) but it does not print
> :the debug printfs or help.  I have to hold the power button for 4
> :seconds to get a real reboot.  The final problem is with power
> :management.  Since apm does not work and acpi has no device, there is no
> :way to suspend or hibernate or whatever, and any sleep event (power
> :button, lid close) will just hang the laptop.  If there is a way to
> :disable acpi subsystems, should I try lid, or event?  I want to try
> :importing acpica 20031203 and adding back /dev/acpi to see if that
> :helps, but does anybody have any hints on the other problems?  I flashed
> :my radeon bios since thats the only warning X was giving before
> :segfaulting, but I dont think that will help, so I'm rebuilding
> :everything after pkg_delete -rf "XFree86*" and rm -rf /usr/X11R6. Thanks
> :for the help.  Any info will be given thats needed.  Heres the laptop
> :specs.
> :
> :HP Pavilion zt3020us
> :Intel Centrino
> :	1.4Ghz Pentium M
> :	Intel 855 Chipset (ICH4)
> :	Intel 2100 802.11b Wireless
> :	Radeon Mobility 9200
> :	ICH Sound (Soundmax AC97)
> :	Realtek Lan
> :	Intel USB 2.0
> :	VIA Firewire
> :
> :-Craig
> :-- 
> :------------------------------------------------------------------------
> :Craig Dooley                                            craig at xxxxxxxxxx

-- 
------------------------------------------------------------------------
Craig Dooley                                            craig at xxxxxxxxxx
------------------------------------------------------------------------





More information about the Kernel mailing list