From Bugzilla Helper: User-Agent: Mozilla/4.76 [en] (X11; U; Linux 2.2.16-22smp i686) When ifup'ing a pcnet32 based interface with the pcnet32 module not loaded the following oops results: Unable to handle kernel paging request at virtual address c888ee40 printing eip: c01aa65d pgd entry c4c70c88: 00000000007c92063 pmd entry c4c70c88: 00000000007c92063 pte entry c7c92238: 00000000000000000 ... pte not present! Oops: 0002 CPU: 0 EIP: 0010:[<c01aa65d>] EFLAGS: 00010246 eax: c888ee40 ebx: c889a000 ecx: 00000000 edx: c113ed20 esi: c889ce40 edi: 00000000 ebp: 080a0188 esp: c4ab5f08 ds: 0018 es: 0018 ss: 0018 Process modprobe (pid: 1741, stackpage=c4ab5000) Stack: c889a000 00000000 00000000 c889c0d2 c889ce40 c889a000 00000000 00000000 080a0188 c011a3b5 00000000 c4d0c000 00002e80 c4d0d000 00000060 ffffffea 00000007 c65c0de0 00000060 c8890000 c889a060 00002ff0 00000000 00000000 Call Trace: [<c889a000>] [<c889c0d2>] [<c889ce40>] [<c889a000>] [<c011a3b5>] [<c8890000>] [<c889a0\60>] [<c01091cb>] Code: 89 30 8b 1d a8 a0 28 c0 81 fb a8 a0 28 c0 74 23 8d 76 00 53 Reproducible: Always Steps to Reproduce: 1. Boot system 2. ifdown interface 3. rmmod pcnet32 4. ifup interface Actual Results: lsmod shows the pcnet32 module in initializing state system is not totally hung Additional information: This nic is built in on several motherboards including the IBM eServer xSeries 340
This sounds like the driver leaves the interface "alive" after rmmod. I'll look into this.
Looks like an identical message with IBM PC-Server 325 with onboard AM79C790 ethernet. I was trying to upgrade from 6.2 to 7.1 - I ended up falling back to 7.0 (2.2 kernel, not 2.4 kernel) which works fine except for the X display. My error is just slightly different, attached here just in case it helps (I was using insmod, not ifup): Unable to handle kernel paging request at virtual address c887ee40 printing eip: c01aa65d pgd entry c62d1c88: 0000000007c9d063 pmd entry c62d1c88: 0000000007c9d063 pte entry c7c9d1f8: 0000000000000000 ... pte not present! Oops: 0002 CPU: 1 EIP: 0010:[pci_register_driver+29/96] EIP: 0010:[<c01aa65d>] EFLAGS: 00010246 eax: c887ee40 ebx: c8884000 ecx: 00000000 edx: c027bd18 esi: c8886e40 edi: 00000000 ebp: 080763d8 esp: c62d3f08 ds: 0018 es: 0018 ss: 0018 Process insmod (pid: 943, stackpage=c62d3000) Stack: c8884000 00000000 00000000 c88860d2 c8886e40 c8884000 00000000 00000000 080763d8 c011a3b5 00000000 c6294000 00002e80 c6295000 00000060 ffffffea 00000007 c6eefca0 00000060 c8880000 c8884060 00002ff0 00000000 00000000 Call Trace: [ipchains:__insmod_ipchains_S.bss_L1072+53696/30003232] [ipchains:__insmod_ipchains_S.bss_L1072+62098/29994830] [ipchains:__insmod_ipchains_S.bss_L1072+65536/29991392] [ipchains:__insmod_ipchains_S.bss_L1072+53696/30003232] [sys_init_module+1349/1584] [ipchains:__insmod_ipchains_S.bss_L1072+37312/30019616] [ipchains:__insmod_ipchains_S.bss_L1072+53792/30003136] [<c8884000>] [<c88860d2>] [<c8886e40>] [<c8884000>] [<c011a3b5>] [<c8880000>] [<c8884060>] [system_call+51/56] [<c01091cb>] Code: 89 30 8b 1d a8 a0 28 c0 81 fb a8 a0 28 c0 74 23 8d 76 00 53
We hope to have fixed this in the latest nightly build. I hope it will be available in rawhide soon.