I just tried to mount a NFS file system on a system with no local portmap running.
mount ended up in an endless uninterruptible sleep in xprt_transmit.
It only recovered when I eventually started portmap
Shouldn't there be a shorter timeout for this?
Happened on 2.6.2rc2 on AMD64.
That was the hanging mount:
mount D 0000000045aaaf57 0 2012 2043 (NOTLB)
0000010017b278b8 0000000000000006 0000010016d50448 ffffffff803601a2
0000010015670b30 000000000000001d ffffffff803ec4c0 0000010015670e68
0000000000000000 00000100191b2138
Call Trace:<ffffffff803601a2>{xprt_transmit+1202} <ffffffff803630a0>{__rpc_execute+432}
<ffffffff80130900>{default_wake_function+0} <ffffffff8035e9ee>{rpc_call_sync+142}
<ffffffff803601d0>{xprt_timer+0} <ffffffff8035d8a0>{call_status+0}
<ffffffff80362020>{rpc_run_timer+0} <ffffffff803694be>{rpc_register+286}
<ffffffff80365094>{svc_register+148} <ffffffff801d4276>{nfs3_rpc_wrapper+118}
<ffffffff80365200>{svc_create+224} <ffffffff801ea086>{lockd_up+118}
<ffffffff801cd92c>{nfs_get_sb+2828} <ffffffff80328b13>{inet_recvmsg+51}
<ffffffff8017554b>{do_kern_mount+107} <ffffffff8018b527>{do_mount+1463}
<ffffffff801562b4>{__alloc_pages+164} <ffffffff80156634>{__get_free_pages+68}
<ffffffff8018b71b>{sys_mount+203} <ffffffff8010f704>{system_call+124}
-Andi
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo _at_ vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/