[Freeipa-devel] [PATCH] refactor nss_ctx and other bugfixes

[Simo Sorce]

On Fri, 2009-03-27 at 14:09 -0400, Stephen Gallagher wrote:

> Not your day, my friend:
> 
> Core was generated by `/usr/libexec/sssd/sssd_nss'.
> Program terminated with signal 11, Segmentation fault.
> [New process 2801]
> #0  0x000000000040b2cd in nss_dp_send_acct_req (rctx=0x1877a40,
> memctx=0x187f680, callback=0x4069f0 <nss_cmd_setpw_dp_callback>,
> callback_ctx=0x187fad0,
>     timeout=100000, domain=0x1878e90 "TEST", type=1, opt_name=0x0,
> opt_id=0) at responder/nss/nsssrv_dp.c:145
> 145	    conn = sbus_get_connection(rctx->dp_ctx->scon_ctx);
> Missing separate debuginfos, use: debuginfo-install
> libselinux-2.0.78-1.fc10.x86_64
> (gdb) bt
> #0  0x000000000040b2cd in nss_dp_send_acct_req (rctx=0x1877a40,
> memctx=0x187f680, callback=0x4069f0 <nss_cmd_setpw_dp_callback>,
> callback_ctx=0x187fad0,
>     timeout=100000, domain=0x1878e90 "TEST", type=1, opt_name=0x0,
> opt_id=0) at responder/nss/nsssrv_dp.c:145
> #1  0x0000000000406e55 in nss_cmd_setpwent_ext (cctx=0x187d210,
> immediate=false) at responder/nss/nsssrv_cmd.c:967
> #2  0x0000000000406fc7 in nss_cmd_setpwent (cctx=0x187d210) at
> responder/nss/nsssrv_cmd.c:1006
> #3  0x000000000041b248 in sss_cmd_execute (cctx=0x187d210,
> sss_cmds=0x622860) at responder/common/responder_cmd.c:66
> #4  0x0000000000419d5a in client_recv (ev=0x1876880, cctx=0x187d210) at
> responder/common/responder_common.c:120
> #5  0x0000000000419efb in client_fd_handler (ev=0x1876880,
> fde=0x187cfb0, flags=1, ptr=0x187d210) at
> responder/common/responder_common.c:158
> #6  0x0000003daba042bd in std_event_loop_select (std_ev=0x1876920,
> tvalp=0x7fffadd9c270) at tevent_standard.c:523
> #7  0x0000003daba047fb in std_event_loop_once (ev=0x1876880) at
> tevent_standard.c:554
> #8  0x0000003daba048ae in std_event_loop_wait (ev=0x1876880) at
> tevent_standard.c:567
> #9  0x000000000040cd4a in server_loop (main_ctx=0x18769e0) at
> util/server.c:325
> #10 0x000000000040465d in main (argc=1, argv=0x7fffadd9c4c8) at
> responder/nss/nsssrv.c:227
> 

Ok I have a theory that the initial reconnection code may not be kicking
in in time in some cases that I do not experience.

attached a bandaid patch, the code responsible for this should be
replaced by your new reconnection code anyway.

Simo.

-- 
Simo Sorce * Red Hat, Inc * New York
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Fix-potential-segfault-if-dp_ctx-is-still-NULL.patch
Type: text/x-patch
Size: 1166 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/freeipa-devel/attachments/20090327/8cc8b015/attachment.bin>