|
Description
|
OS configure info: project binaries, which e1000g is implemented as a
Nemo driver
Although it is already associated with an aggregation, an e1000g0 device
can be plumbed successfully. Then try to plumb & unplumb the aggregation
for several times, system panicks.
...
panic[cpu0]/thread=ffffffff80e2eda0:
BAD TRAP: type=e (#pf Page fault) rp=fffffe8000703030 addr=0 occurred in module "<unknown>" due to a NULL pointer dereference
...
fffffe8000702f40 unix:die+da ()
fffffe8000703020 unix:trap+5ea ()
fffffe8000703030 unix:cmntrap+13f ()
fffffe8000703150 0 ()
fffffe80007031a0 e1000g:e1000g_m_resources+3a ()
fffffe80007031b0 mac:mac_resources+11 ()
fffffe80007031d0 aggr:aggr_m_resources+1c ()
fffffe80007031e0 mac:mac_resources+11 ()
fffffe8000703210 dld:proto_poll_enable+4b ()
fffffe80007032b0 dld:proto_capability_enable+143 ()
fffffe80007032c0 dld:proto_capability_req+ab ()
fffffe8000703300 dld:dld_proto+ed ()
fffffe8000703310 dld:str_mproto_put+9 ()
fffffe8000703320 dld:dld_str_put+25 ()
fffffe8000703330 dld:drv_uw_put+d ()
fffffe8000703390 unix:putnext+1f1 ()
<snipped>
xxxxx@xxxxx.com 2005-05-18 10:45:50 GMT
This bug can also be reproduced in snv_14 with a BGE card. In V60z or a
computer with CK804 motherboard, you can successfully plumb a BGE card
even it belongs to an aggregation. Below are the steps to reproduce
the bug:
#dladm create-aggr -d bge0 1
#ifconfig bge0 plumb //The bge was plumbed successfully
#ifconfig bge0 191.0.0.1/24 up
#ifconfig aggr1 plumb
#ifconfig aggr1 192.0.0.1/24 up
#ifconfig aggr1 unplumb
... (Executed the following commands several times)
#ifconfig aggr1 unplumb
#ifconfig aggr1 plumb
#ifconfig aggr1 192.0.0.1/24 up
...
Then, the system crashed.
I attached the core files in Attachment item of this bug.
xxxxx@xxxxx.com 2005-05-25 10:12:56 GMT
xxxxx@xxxxx.com 2005-05-31 06:55:23 GMT
|