[Gluster-devel] One more regression

[Emmanuel Dreyfus]

Hello

I cannot keep up with the flow of regression that hit NetBSD since two 
weeks. There have been some new test cases added withuout looking at
NetBSD result. It may be easy to fix but it is still frustrating.

There have also been real new bugs introduced. The fact that they did
not appear in Linux is not a guarantee that you will never see them 
bite you later.

Consider tests/basic/quota.t, il now fails on NetBSD because the
brick's glusterfsd crashes on a unintialized  client structure (xl=0x3
below)

Core was generated by `glusterfsd'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0xbb792021 in create_frame (xl=0x3, pool=0xbb188078) at stack.c:45
45              stack->ctx = xl->ctx;
(gdb) bt
#0  0xbb792021 in create_frame (xl=0x3, pool=0xbb188078) at stack.c:45
#1  0xb9ada0d4 in server_alloc_frame (req=0xb750e028) at server-helpers.c:384
#2  0xb9ada257 in get_frame_from_request (req=0xb750e028)
    at server-helpers.c:424
#3  0xb9af3f44 in server3_3_statfs (req=0xb750e028) at server-rpc-fops.c:6177
#4  0xbb71d6a5 in rpcsvc_handle_rpc_call (svc=0xbb1a4218, trans=0xb750d018, 
    msg=0xb75150d8) at rpcsvc.c:690
#5  0xbb71d932 in rpcsvc_notify (trans=0xb750d018, mydata=0xbb1a4218, 
    event=RPC_TRANSPORT_MSG_RECEIVED, data=0xb75150d8) at rpcsvc.c:784
#6  0xbb722b7f in rpc_transport_notify (this=0xb750d018, 
    event=RPC_TRANSPORT_MSG_RECEIVED, data=0xb75150d8) at rpc-transport.c:543
#7  0xbb28e90c in socket_event_poll_in (this=0xb750d018) at socket.c:2290
#8  0xbb28edc1 in socket_event_handler (fd=24, idx=10, data=0xb750d018, 
    poll_in=1, poll_out=0, poll_err=0) at socket.c:2403
#9  0xbb7b8730 in event_dispatch_poll_handler (event_pool=0xbb144018, 
    ufds=0xb7510098, i=10) at event-poll.c:391
#10 0xbb7b8a54 in event_dispatch_poll (event_pool=0xbb144018)
    at event-poll.c:487
#11 0xbb789752 in event_dispatch (event_pool=0xbb144018) at event.c:127
#12 0xb9bd6e99 in changelog_rpc_poller (arg=0xbb1d2018)
    at changelog-rpc-common.c:29
(gdb) frame 1
#1  0xb9ada0d4 in server_alloc_frame (req=0xb750e028) at server-helpers.c:384
384             frame = create_frame (client->this, req->svc->ctx->pool);
(gdb) print *client
$1 = {scratch_ctx = {lock = {pts_magic = 3, pts_spin = 3 '\003', 
      pts_flags = 8}, count = 3, ctx = 0xc}, ref = {lock = {pts_magic = 3, 
      pts_spin = 14 '\016', pts_flags = 3}, bind = 9, count = 3}, 
  bound_xl = 0x11, this = 0x3, tbl_index = 13, 
  client_uid = 0x3 <error: Cannot access memory at address 0x3>, auth = {
    flavour = 15, len = 3, 
    data = 0x16 <error: Cannot access memory at address 0x16>, 
    username = 0x3 <error: Cannot access memory at address 0x3>, 
    passwd = 0x17 <error: Cannot access memory at address 0x17>}}



-- 
Emmanuel Dreyfus
manu at netbsd.org