[openib-general] error messages for debugging

[Sayantan Sur]

Hi,

Upon more debugging, I seem to be hitting this error condition in file
libmthca/src:qp.c [SVN revsision: 2455]


    592         if (wr->opcode >= sizeof mthca_opcode / sizeof
mthca_opcode[0]) {
    593             ret = -1;
    594             *bad_wr = wr;
    595             goto out;
    596         }

After the post, a dump of the descriptor looks like this:

(gdb) print *bad_wr
$2 = {next = 0x0, wr_id = 69288064, sg_list = 0x4217050, num_sge = 1, 
  opcode = IBV_WR_RDMA_WRITE, send_flags = 10, imm_data = 0, wr = {rdma
= {
      remote_addr = 69681408, rkey = 2046832896}, atomic = {
      remote_addr = 69681408, compare_add = 2046832896, swap = 0, rkey =
0}, 
    ud = {ah = 0x4274100, remote_qpn = 2046832896, remote_qkey = 0}}}


This doesn't seem to be an error with remote address or keys, since this
error is generated locally just after a post operation.

Any clues to why I am hitting this condition will be appreciated!!

Thanks,
Sayantan.


* On Jun,1 Sayantan Sur<surs at cse.ohio-state.edu> wrote :
> Hi,
> 
> Is there any way of knowing the reason ibv_blah_blah() failed? All I get
> on return is a -1/NULL. VAPI provides quite helpful error codes on
> return, with even string equivalents with VAPI_strerror(). I find that
> method a very easy way to debug.
> 
> Thanks,
> Sayantan.
> 
> 
> -- 
> http://www.cse.ohio-state.edu/~surs
> _______________________________________________
> openib-general mailing list
> openib-general at openib.org
> http://openib.org/mailman/listinfo/openib-general
> 
> To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

-- 
http://www.cse.ohio-state.edu/~surs