panic when removing pccard

Niki Denev nike_d at
Wed Jun 20 14:08:45 UTC 2007

Hash: SHA1

Hans Petter Selasky wrote:
> On Tuesday 19 June 2007 14:28, Niki Denev wrote:
>> Hans Petter Selasky wrote:
>>> On Tuesday 19 June 2007 13:25, Niki Denev wrote:
>>>> Hans Petter Selasky wrote:
>>>>> On Tuesday 19 June 2007 11:06, Niki Denev wrote:
>>>>>> Hans Petter Selasky wrote:
>>>>>>> On Friday 08 June 2007 10:31, Niki Denev wrote:
>>>>>>>> -----BEGIN PGP SIGNED MESSAGE-----
>>>>>>>> Hash: SHA1
>>>>>>>> I experience the following panic on a few days old -current :
>>>>>>>> If i insert and then remove a pcmcia card using the ubsa module
>>>>>>>> (Vodafone Mobile Connect, which actually is Huawei E630 UMTS/HSDPA
>>>>>>>> modem) the machine panics, because i think the order of removal of
>>>>>>>> the devices by the kernel is not correct. I'm not sure because i
>>>>>>>> don't have -STABLE machine with pcmcia now, but i think this problem
>>>>>>>> does not exist there.
>>>>>>>> Here is what happens when card is inserted :
>>>>>>> Hi!
>>>>>>> If you want to be able to remove the card you need to install the new
>>>>>>> USB stack. The old USB stack is completely broken when it comes to
>>>>>>> this point!
>>>>>>> --HPS
>>>>>>> Install the SVN version.
>>>>>>> _______________________________________________
>> I did make S=../src package the first time, but then for some reason i
>> was trying without defining the S variable.
>> Making it properly this time i went further, but still the build failed
>> in if_rum with error if_rum.c:1118 too few arguments to function
>> 'ieee80211_input' and several warnings about nested extern declarations
>> and implicit declarations of the ieee80211 functions.
>> I'll try disconnecting if_rum from the build to see if will help.
> Just create an empty "/sys/dev/usb/if_rum.c" rum file.
> rm -f "/sys/dev/usb/if_rum.c"
> touch "/sys/dev/usb/if_rum.c"
> Then try to compile again.
> --HPS

After i disabled if_ural and if_rum and fixed a minor thing in umass.c
(added the missing argument to one cam/xpt function which api apparently
changed in -current)
the new stack compiled without problems, and now the card works
perfectly and don't have more usb related panics :) Thanks!
Version: GnuPG v1.4.7 (FreeBSD)
Comment: Using GnuPG with Mozilla -


More information about the freebsd-current mailing list