<br><br><div><span class="gmail_quote">On 9/27/05, <b class="gmail_sendername">James Lentini</b> <<a href="mailto:jlentini@netapp.com">jlentini@netapp.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br><br>On Tue, 27 Sep 2005, Todd Bowman wrote:<br><br>> On 9/27/05, James Lentini <<a href="mailto:jlentini@netapp.com">jlentini@netapp.com</a>> wrote:<br>><br>> What is the output of cat /sys/class/net/*/ifindex?
<br>> ><br>> > cat /sys/class/net/*/ifindex<br>> 1 #eth0<br>> 10 #ib0<br>> 11 #ib1<br>> 2 #lo<br>> 3 #tunl0<br><br>This looks like the problem I reported last week:<br><br><a href="http://openib.org/pipermail/openib-general/2005-September/011668.html">
http://openib.org/pipermail/openib-general/2005-September/011668.html</a><br><br>If so, this is fixed in the current subversion sources. Could you<br>update your sources, specifically infiniband/core/at.c, and try again?<br>
<br>james</blockquote><div><br>
That worked. Thanks.<br>
<br>
It is printing the addresses backwards. But I assume this is also a bug in printing.<br>
<br>
ips_by_gid: RET 0 at_rec 0x1ffff9e5f40 -> id 1<br>
dapli_at_event_cb()<br>
ip_comp_handler: rec 0x1ffff9e5f40 ->id 1 id 1 num 1 b0f0add<br>
open_hca: mthca0, port 1, AF_INET <a href="http://221.10.15.11">221.10.15.11</a> INLINE_MAX=128<br>
query_hca: mthca0 AF_INET <a href="http://221.10.15.11">221.10.15.11</a><br>
query_hca: (0.30002) ep 64512 ep_q 65535 evd 65408 evd_q 65535<br>
query_hca: msg 2147483648 rdma 2147483648 iov 28 lmr 131056 rmr 0<br>
Bus error<br>
<br>
<br>
The bus error is probably due to changes in ibv_context. I didn't recompile libdapl. <br>
<br>
The changes in ibv_context that broke udapl are:<br>
num_comp is now num_comp_vectors<br>
I don't know what is used in place of cq_fd[1]. Any pointers?<br>
<br>
Todd <br>
<br>
<br>
</div><br></div><br>