[driverloader] Kernel Oops [eid-20050704-ed6]

Wed Jul 6 23:22:41 EDT 2005

Hi Jonathan,

i should have been more elaborate. You are right, it is possible to set the 
compiler using the CC variable. But this is of not importance, because of the 
the modules/kernelcompiler.sh script. This script is called twice from 
modules/makefile. If you replace it in the CC initialization it remains one 
time. This script checks - to make a long story short - the version of the  
kernel compiler. It does this by finding gcc using the tool 'which', which 
finds the first occurence of gcc in the PATH. The PATH is not the PATH 
variable i set on my workstation but hardcoded in dldrconfig, Line 2070
PATH=@DLDRSBINDIR@:/usr/sbin:/sbin:/usr/bin:/bin:/usr/local/sbin:/usr/local/bin

So, kernelcompiler.sh compares the first gcc found in 
PATH=@DLDRSBINDIR@:/usr/sbin:/sbin:/usr/bin:/bin:/usr/local/sbin:/usr/local/bin 
and compares it against the kernel. Because my gcc 2.95 is not in this PATH, 
kernelcompiler.sh fails and the installation stops. The most pragramatic 
solution for me was to add the path to my compiler to the PATH in dldrconfig, 
but this is only a dirty workaround. A reasonble solution would be to enhance 
kernelcompiler.sh to check first the CC variable, before it searches the 
path. Additionally i'm in doubt that it is wise to override the PATH 
variable, i think it would be better to add @DLDRSBINDIR@ to the PATH:
PATH=@DLDRSBINDIR@:${PATH}

Yours
Olaf

> Hi,
>
> if you need a special compiler to compile kernel modules, you can change
> this setting in the '/usr/lib/driverloader/modules/Makefile' file. You
> should look for a variable named CC.
>
> Regards,
>
>
> Jonathan
> Technical specialist / Linuxant
> www.linuxant.com
> support at linuxant.com
>
> Olaf Groeger wrote:
> > ok,
> >
> > thank you. I had a look onto TI's homepage. After that i suspect that it
> > is the 1450 chip. I heard another rumour, that avm will provide a linux
> > driver in the 3rd quarter although i couldn't find an official
> > announcement. Anyway,i'll wait and see. Thanks again for your interest.
> >
> > Yours
> > Olaf
> >
> > P.S.: one point for the FAQ: the installation fails if the kernel
> > compiler is not in the (hard coded) PATH in dldrconfig. This happens with
> > many LFS systems, because they use (as the kernel team recommends) gcc
> > 2.95 for compiling the kernel, which resides somewhere in /opt, but an
> > up-to-date compiler as system compiler. Therefor the gcc 2.95 compiler
> > isn't in the path and it need not to be, because make uses the CC
> > variable which can be set accurate when compiling the kernel (make
> > CC=/opt/... bzImage). So when it happens, that the installation fails
> > with "gcc version ... or similar is required" and you know you have this
> > version, you should adapt dldrconfig.
> >
> >>Hi,
> >>
> >>it is quite possible that you are right that the chipset is provided by
> >>Texas Instruments (TI). The only hint is the following strings in the
> >>.sys file:
> >>
> >>---
> >>\Device\TIWLNUSB
> >>\DosDevices\TIWLNUSB
> >>ACXUseInterruptThreshold
> >>ACXMemoryBlockSize
> >>---
> >>
> >>The ACX100 and ACX111 are wireless chips from TI.
> >>
> >>DriverLoader doesn't officially support this new USB chipset and we will
> >>contact you back when official support for this chipset will be added to
> >>DriverLoader.
> >>
> >>Regards,
> >>
> >>
> >>Jonathan
> >>Technical specialist / Linuxant
> >>www.linuxant.com
> >>support at linuxant.com
> >>
> >>Olaf Groeger wrote:
> >>>Hello,
> >>>
> >>>as you wished:
> >>>the attached file driverloaderdiag.txt has been recorded while the avm
> >>>driver was uploaded, but the hotplug was disabled. This avoids loading
> >>>the ehci driver for the D-Link USB Cardbus card and the driver for the
> >>>usb stick wan device. The file driverloaderdiag_without_driver.txt has
> >>>been recorded after removing the avm drivers from driverloader and
> >>>starting the hotplug. Thus ehci was loaded and the usb stick was
> >>> attached to the usb bus.
> >>>
> >>>Yours
> >>>Olaf
> >>>
> >>>Am Montag, 4. Juli 2005 16:56 schrieben Sie:
> >>>>Hi,
> >>>>
> >>>>could you please send at support at linuxant.com the output of 'dumpdiag'?
> >>>>Type the following in a root shell:
> >>>>
> >>>>---
> >>>>dldrconfig --dumpdiag
> >>>>---
> >>>>
> >>>>If the previous command crashes your machine, please try this command
> >>>>instead:
> >>>>
> >>>>---
> >>>>dldrconfig --dumpdiag --noprobe
> >>>>---
> >>>>
> >>>>Just send us the generated file located in /tmp
> >>>>('driverloaderdiag.txt'). It'll help us solve the problem.
> >>>>
> >>>>Regards,
> >>>>
> >>>>
> >>>>Jonathan
> >>>>Technical specialist / Linuxant
> >>>>www.linuxant.com
> >>>>support at linuxant.com
> >>>>
> >>>>Olaf Groeger wrote:
> >>>>>Hello,
> >>>>>
> >>>>>i installed Driverloader, current version, and these drivers from avm:
> >>>>>ftp://ftp.avm.de/cardware/fritzwlanusb.stick/windows.xp/deutsch/avm_fr
> >>>>>it z !wlan_usb_stick_build_050608.exe
> >>>>>
> >>>>>The device is a Fritz!WLAN USB Stick, which contains as a rumor said,
> >>>>> a recent version of TI. At least it supports  802.11g++ and WPA2.
> >>>>>
> >>>>>As soon as driverloader loads the fwlanusb driver the kernel crashes.
> >>>>> I tried both a LinuxFromScratch with kernel 2.4.31 and a Fedore Core
> >>>>> 3 with the Linuxant patched kernel 2.6.11-1.35_FC3.stk16.
> >>>>>
> >>>>>The kernel 2.4.31 stack trace is:
> >>>>>driverloader: fwlanusb.sys: 0xe0db6000..0xe0e24a80 entry = e0e21700
> >>>>>(pbase = 0x10000)
> >>>>>Unable to handle kernel NULL pointer dereference at virtual address
> >>>>>0000000c printing eip:
> >>>>>e0e11229
> >>>>>*pde = 00000000
> >>>>>Oops: 0000
> >>>>>CPU:    0
> >>>>>EIP:    0010:[<e0e11229>]    Tainted: P
> >>>>>EFLAGS: 00010286
> >>>>>eax: 00000000   ebx: 00000100   ecx: 00000000   edx: 00000000
> >>>>>esi: d5bfb808   edi: d5bfb808   ebp: dade3e6c   esp: dade3e50
> >>>>>ds: 0018   es: 0018 ss: 0018
> >>>>>Process modprobe (pid: 1300, stackpage=dade3000)
> >>>>>Stack: db1cf008 00000000 00000000 db1cf448 00000000 db1cf008 00000000
> >>>>>dade3e84 e0e13dcd db1cf008 db1cf008 c0000002 db1cf448 dade3ecc
> >>>>> e0e21312 db1cf008 00000100 d5bfb808 d5bfb808 dade3ecc 00000000
> >>>>> 23222120 e0e20dc0 00de3ecc Call Trace:    [<e0d82c16>] [<e0d84661>]
> >>>>> [<c01345c8>] [<c0115c88>] [<c0119e48>]
> >>>>> [<c010849a>] [<e0d84580>] [<c0106ed8>]
> >>>>>
> >>>>>Code: 3b 51 0c 73 35 8b 45 f4 89 45 f8 8b 4d f8 0f b6 11 85 d2 74
> >>>>><0>Kernel panic: Aiee, killing interrupt handler!
> >>>>>In interrupt handler - not syncing
> >>>>>
> >>>>>Is there any hope to get this device working? The driver itself is
> >>>>>working with Windows XP.
> >>>>>
> >>>>>Yours Olaf
> >>>>>_______________________________________________
> >>>>>driverloader mailing list
> >>>>>driverloader at lists.linuxant.com
> >>>>>https://www.linuxant.com/mailman/listinfo/driverloader
> >
> > _______________________________________________
> > driverloader mailing list
> > driverloader at lists.linuxant.com
> > https://www.linuxant.com/mailman/listinfo/driverloader