Bug 360921 - serial-USB console does not work properly
Summary: serial-USB console does not work properly
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: 6
Hardware: All
OS: Linux
low
low
Target Milestone: ---
Assignee: Kernel Maintainer List
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks: 427887
TreeView+ depends on / blocked
 
Reported: 2007-10-31 20:16 UTC by Paul B Schroeder
Modified: 2008-02-08 04:26 UTC (History)
2 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2008-02-08 04:26:52 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)

Description Paul B Schroeder 2007-10-31 20:16:39 UTC
Description of problem:
I'm trying to use a USB-serial adapter for my system console and it does not
seem to be working properly.  Using a standard serial port (ttyS1) as the
console does work properly though.  Consequently, I cannot get initscripts
output when using ttyUSB0 as my console.

i.e. Using ttyS1 as my console works fine, here's my grub.conf boot kernel entry:
kernel /boot/vmlinuz-2.6.19-1.2895.uplogix.14 ro root=/dev/hda1 panic=5
console=tty0 console=ttyS1,115200

I get initscripts output and can for example 'echo test >/dev/console' and get
the output via ttyS1.

However, when I try to use a usb-serial console:
kernel /boot/vmlinuz-2.6.19-1.2895.uplogix.14 ro root=/dev/hda1 panic=5
console=tty0 console=ttyUSB0,115200

I get no initscripts output.  And when I 'echo test >/dev/console', I get
absolutely nothing out of ttyUSB0.

When I add 'CONSOLE=/dev/ttyUSB0' to my kernel command line, init seems to read
CONSOLE from the environment and will use that instead of /dev/console.  But the
output comes out at 9600 instead of the 115200 baud rate I wanted.

This seems to be an issue regardless of the USB-serial driver.  I've used the
io_ti driver with my IO Edgeport adapter and mos7840 driver for my Moschip
USB-serial device.  I've tried it with the default kernel build and I've done my
on compile with my kernel config that built in:
#
# USB Serial Converter support
#
CONFIG_USB_SERIAL=y
CONFIG_USB_SERIAL_CONSOLE=y
CONFIG_USB_SERIAL_EDGEPORT_TI=y
CONFIG_USB_SERIAL_MOS7840=y

FYI, partial kernel boot messages will come out of ttyUSB0.  But /dev/console
never seems to work properly.

Version-Release number of selected component (if applicable):
kernel-2.6.19-1.2895
SysVinit-2.86-14
initscripts-8.45.3-1

How reproducible:
Plug in a USB-serial adapter and add console=ttyUSB0 to the kernel command line.
 Use minicom or whatnot to try to get the output and see that no initscripts
output comes.

Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:

Comment 1 Bill Nottingham 2007-10-31 20:58:26 UTC
AFAIK, the console= argument cannot be provided by something that is modular (as
usb is.)

Comment 2 Paul B Schroeder 2007-11-01 19:39:21 UTC
As noted above, I did compile in the modules proper..  And it still didn't work.
 At any rate, to get around the problem, I've patched my init:

--- sysvinit-2.86.orig/src/init.c       2007-11-01 13:16:02.000000000 -0500
+++ sysvinit-2.86/src/init.c    2007-11-01 14:06:49.000000000 -0500
@@ -694,7 +694,7 @@
        (void) tcgetattr(fd, &tty);
 
        tty.c_cflag &= CBAUD|CBAUDEX|CSIZE|CSTOPB|PARENB|PARODD;
-       tty.c_cflag |= HUPCL|CLOCAL|CREAD;
+       tty.c_cflag |= HUPCL|CLOCAL|CREAD|B115200;
 
        tty.c_cc[VINTR]  = 3;   /* ctrl('c') */
        tty.c_cc[VQUIT]  = 28;  /* ctrl('\\') */


So now init is using /dev/ttyUSB0 as I've set 'CONSOLE=/dev/ttyUSB0' on my
kernel command line and it's getting passed to the environment.  The patch
simply has init setting the baud rate to 115200 every time.

It would be nice if init could read another environment variable, say
CONSOLE_BAUD or something similar and set that if it exists.  Or maybe make init
smart enough to parse CONSOLE like 'console' is parsed from the kernel command
line.  i.e. CONSOLE=/dev/ttyUSB0,115200 would tell init to use ttyUSB0 and set
it's baud rate to 115200..


Comment 3 Bill Nottingham 2007-11-01 19:57:55 UTC
Did you compile in the usb host adapter, etc. as well?

If the usb driver isn't honoring the baud rate, that's probably a kernel bug. I
really don't think it belongs in init to set it.


Comment 4 Paul B Schroeder 2007-11-01 20:20:39 UTC
Yup.  Compiled in everything USB.

I'm not sure I'd say the driver isn't honoring the baud rate.  I get the kernel
boot messages at 115200 which has nothing to do with my little hack.  Somewhere
along the way, something is changing the baud rate to 9600.


Comment 5 Bill Nottingham 2007-11-01 20:34:22 UTC
Right, but if it's not needed for 'normal' serial, it shouldn't be needed for
USB. What's your inittab line look like for it, if there is one?

Comment 6 Paul B Schroeder 2007-11-01 20:39:48 UTC
Agreed, that's why I opened the bug..  :)

I have an entry to run agetty, but that's it:
a0:2345:once:/sbin/agetty /dev/ttyUSB0 -t 60 115200

That part has always worked okay though..


Comment 7 Pete Zaitcev 2007-11-01 20:58:33 UTC
I'm too swamped to look at this promptly. The USB console is just not a good
idea. Consider that it must have interrupts open to function, which is not
true for any normal console. It's entirely useless in case of any crash.
But if you do not have a crash, why have it at all? Just use a netconsole
instead.

That said, please attach /proc/bus/usb/devices. The usb-serial has a legion
of subdrivers, each of which may fail to set termios right.

BTW, please try F8. It has massive cleanups to termios handling by Alan Cox.
It's VERY likely that Alan cleared any improper and/or erroneous termios
handling scattered across usb-serial subdrivers. Even if you cannot deploy
F8 on that box now, we need to know if this bug is a dead-end.


Comment 8 Jon Stanley 2008-01-08 01:53:59 UTC
(This is a mass-update to all current FC6 kernel bugs in NEW state)

Hello,

I'm reviewing this bug list as part of the kernel bug triage project, an attempt
to isolate current bugs in the Fedora kernel.

http://fedoraproject.org/wiki/KernelBugTriage

I am CC'ing myself to this bug, however this version of Fedora is no longer
maintained.

Please attempt to reproduce this bug with a current version of Fedora (presently
Fedora 8). If the bug no longer exists, please close the bug or I'll do so in a
few days if there is no further information lodged.

Thanks for using Fedora!

Comment 9 Jon Stanley 2008-02-08 04:26:52 UTC
Per the previous comment in this bug, I am closing it as INSUFFICIENT_DATA,
since no information has been lodged for over 30 days.

Please re-open this bug or file a new one if you can provide the requested data,
and thanks for filing the original report!


Note You need to log in before you can comment on or make changes to this bug.