OpenSolaris

Printable Version Enter a New Search
Bug ID 4028137
Synopsis serial line hangs and gets "link down" message
State 10-Fix Delivered (Fix available in build)
Category:Subcategory kernel:tty
Keywords
Responsible Engineer James D Carlson
Reported Against 2.5 , 2.6 , 3.2 , 5.6 , 5.8 , 5.5.1 , 3.0fcs , s10_30 , s10_31 , s81_50 , s297_33 , s297_fcs , 2.6_patch , s297_beta , 3.1beta1.1 , 4.1beta1.0 , s28_27_64bit
Duplicate Of
Introduced In solaris_2.0
Commit to Fix s10_50
Fixed In s10_50
Release Fixed solaris_10(s10_50) , solaris_9(s9patch) (Bug ID:2127815) solaris_8(s8patch) (Bug ID:2127816,)
Related Bugs 6334314 , 1205352 , 1227655 , 4053843 , 4118072 , 4121909 , 4128595 , 4162664 , 4167761 , 4176234 , 4186727 , 4243600 , 4269418 , 4306957 , 4327173 , 4348839 , 4510118 , 4526160 , 4527582 , 4545882 , 4639739 , 4652318 , 4825252 , 4841355 , 4954740 , 4954747 , 6453651 , 4954716 , 4954751 , 4954733 , 4954735 , 4954712 , 4954725 , 4954744 , 6222697
Submit Date 27-January-1997
Last Update Date 3-April-2007
Description
I use /dev/term/b of my electron as the console for another ultra machine.
From time to time the serial line seems to hang.  I can kill the
tip process but the serial line seems to still be hung.  Any attempts
to tip to the line gets the "link down" message.

I am tired of having to reboot my machine because the serial line is hung.

I am assigning a priority 4 to this bug because I don't have a reproduceable
case.  Normally it happens when I am doing more of a file.  I am using
an xterm.   It has happenned twice in the last week.

This is running Solaris 2.6 beta with openwindows (under cde).  I don't
remember seeing this on Solaris 2.5.

----

This happens to me all the time, especially if CDE dies while I have a tip running.  Only a reboot can clear it if the tip process is in the <defunct>
state.

dean.long@Eng 1998-04-15

---------------------------------------------
(1/20/99 - Albert S.)
I am updating this bug report since my original bug report 4186727 was closed as
duplicate of this bug. I also will jack up the priority since we can reproduce
the behaviour.

sptest would appear as a hanged process which would not allow sunvts kernel
to switch system status from "stopping" to "idle". sptest process is actually
turned into a <defunct> process and the only way to make vtsk become idle,
would be to deselect the sptest from the option screen. Configurations were:

ctech130	Ultra-1 (4u)		SunVTS,3.1 REV=09.99.01.07	5.7 Generic_106541-02 (s998s_u1SunDesktop_06)	sparcv8

ctech234	SPARCstation-20 (4m)	SunVTS,3.1 REV=09.99.01.07	5.7 Generic_106541-02 (s998s_u1SunDesktop_08)	sparcv8

 xxxxx@xxxxx.com 2000-07-13
If we have tip line connected to port A and a loopback connector at port B and while configuring SUNVTS sptest we say 
test type  		a_b ( correctly it should be b )
loopback_ type 		Plug_a_to_a__b_to_b

and start the test; system hard hangs in first 10 minutes if we try to do anything at tip line. 
It hangs every system irrespective of VTS version. THe only difference is some Vts version are ok with lower baud rate; but once I increase the baud rate. System hangs. 
Work Around
ali [1/27/96]:
reboot the machine.  I think this is unacceptable but don't know of any other
way to get around it for now.
Comments
N/A