kern/38316: tcp nfs mount of solaris filesystem stucks

[yamt%mwd.biglobe.ne.jp@localhost]

>Number:         38316
>Category:       kern
>Synopsis:       tcp nfs mount of solaris filesystem stucks
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Mar 27 11:00:00 +0000 2008
>Originator:     YAMAMOTO Takashi <yamt%mwd.biglobe.ne.jp@localhost>
>Release:        NetBSD 4.99.58
>Organization:
        
>Environment:
Architecture: i386
Machine: i386
>Description:
        an nfs mount of a solaris filesystem sometimes stucks.
        the client and the server keep sending tcp window probes each other.

        netstat on server:
TCP: IPv4
   Local Address        Remote Address    Swind Send-Q Rwind Recv-Q    State
-------------------- -------------------- ----- ------ ----- ------ -----------
172.17.14.11.2049    172.17.80.56.1012        1      0     0      0 ESTABLISHED

        netstat on client:
Proto Recv-Q Send-Q  Local Address          Foreign Address        State
tcp    99120  59016  172.17.80.56.1012      172.17.14.11.2049      ESTABLISHED

>How-To-Repeat:
        it's reproducable on my environment with the following commands.

        # mount -o tcp solaris10_box:/zfs_volume /mnt
        # cd /mnt/some_directory
        # iozone -i0 -B -t128
>Fix:
        

>Unformatted: