[brussels-dev] Gldv3 network Driver crashes with a BAD TRAP error=e ( error in mac_ether)

[sowmini.varadhan at sun.com]

Pradeep, 

I think you have filed a bug about this:

  http://bugs.opensolaris.org/view_bug.do?bug_id=6654638

Looks like the data being passed to mac_ether_header_info
is bad:
 "Bad kernel fault at addr=0x4300000c"

Is it possible that your driver is either passing a bad mblk
or is otherwise trashing the input packet? 

--Sowmini

On (01/28/08 21:38), gopanapalli pradeep wrote:
> 
>   Hi,
>  I am porting my 10 gig network driver to solaris( gldv3 base ) .
> Everything seems to be fine until this crash happens , this is after a
> one hour of TCP traffic between the network interface and a client .
>  
>  panic[cpu1]/thread=fffffe8000425c80:
>  BAD TRAP: type=e (#pf Page fault) rp=fffffe8000425820 addr=4300000c occurred in
>  module "mac_ether" due to an illegal access to a user address
>   sched:
>  #pf Page fault
>  Bad kernel fault at addr=0x4300000c
>  pid=0, pc=0xfffffffff3dbace5, sp=0xfffffe8000425910, eflags=0x10216
>  cr0: 8005003b<pg,wp,ne,et,ts,mp,pe> cr4: 6f0<xmme,fxsr,pge,mce,pae,pse>
>  cr2: 4300000c cr3: 68fe000 cr8: c
>  rdi: ffffffff85bf63c0 rsi: 0 rdx: 43000006
>  rcx: ffffffff85bf63c0 r8: fffffe8000425a6c r9: 1
>  rax: 5ea rbx: fffffe8000425a70 rbp: fffffe8000425920
>  r10: ffffffff8bdb1a00 r11: 0 r12: 43000000
>  r13: 1 r14: fffffe8000425a6a r15: ffffffff85bf63c0
>  fsb: ffffffff80000000 gsb: ffffffff81432800 ds: 43
>  es: 43 fs: 0 gs: 1c3
>  trp: e err: 0 rip: fffffffff3dbace5
>  cs: 28 rfl: 10216 rsp: fffffe8000425910
>  ss: 30
>  
>  fffffe8000425730 unix:die+da ()
>  fffffe8000425810 unix:trap+d86 ()
>  fffffe8000425820 unix:cmntrap+13f ()
>  fffffe8000425920 mac_ether:mac_ether_header_info+25 ()
>  fffffe8000425930 mac:mac_header_info+15 ()
>  fffffe8000425970 dls:dls_link_header_info+45 ()
>  fffffe8000425a10 dls:i_dls_link_subchain+38 ()
>  fffffe8000425ad0 dls:i_dls_link_rx+59 ()
>  fffffe8000425b00 mac:mac_rx+45 ()
>  fffffe8000425b70 benet:process_nic_rx_completion+3f7 ()
>  fffffe8000425bc0 benet:process_ucast_rx_completion+cf ()
>  fffffe8000425c20 benet:osm_process_sts+141 ()
>  fffffe8000425c60 unix:av_dispatch_softvect+62 ()
>  fffffe8000425c70 unix:intr_thread+a5 ()
>   Where osm_process_sts is my interrupt function . I tried to debug using 
>  mdb but iam clueless. 
>  Iam using 
>  uname -a
>  SunOS unknown 5.10 Generic_120012-14 i86pc i386 i86pc
>  Solaris 10 8/07 s10x_u4wos_12b X86
>  
>  Solaris 08/07..
>  
>  I remember running this in solaris 10 previous release ( 02/07)
>  
>  Any help will be greatly appreciated 
>  
>  PradeeP
>