2001-12-23 14:08:17

by Andreas Gietl

[permalink] [raw]
Subject: serial console on > 2.4.14

hi,

i am exploring problems on linux-smp kernels (did not test single-processor)
with version greater 2.4.14. In Fact i just tested 2.4.16 and 2.4.17.

Here's my problem:

I compiled all kernels with the same configuration with serial console-option
enabled. With 2.4.14 everything is just fine: I see the kernel-output and can
type in things during startup esp. do a fsck. With 2.4.16 and 2.4.17 i SEE
everything but no input is accepted.

Did you break anything in 2.4.15 or 2.4.16. I of course did not test 2.4.15
because of the fs-bug.

I will of course provide any information you request.

thanx

andreas
--
e-admin internet gmbh
Andreas Gietl
Roter-Brach-Weg 124a
tel +49 941 3810884
fax +49 941 3810891
mobil +49 171 6070008


2001-12-23 14:24:07

by Russell King

[permalink] [raw]
Subject: Re: serial console on > 2.4.14

On Sun, Dec 23, 2001 at 03:06:58PM +0100, Andreas Gietl wrote:
> I compiled all kernels with the same configuration with serial console-option
> enabled. With 2.4.14 everything is just fine: I see the kernel-output and can
> type in things during startup esp. do a fsck. With 2.4.16 and 2.4.17 i SEE
> everything but no input is accepted.

>From what I've deduced, there are a number of SysVinit implementations out
there that incorrectly clear the terminals CREAD flag.

[-]cread
allow input to be received

Obviously if it is cleared, then you'll see your behaviour. If you can log
into the box some other way, please check the settings using:

stty -aF /dev/ttyS0

--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html

2001-12-23 14:35:07

by Andreas Gietl

[permalink] [raw]
Subject: Re: serial console on > 2.4.14

On Sunday 23 December 2001 15:23, Russell King wrote:

This is what stty gives

speed 38400 baud; rows 0; columns 0; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^X; eof = ^D; eol = <undefiniert>;
eol2 = <undefiniert>; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R; werase =
^W; lnext = ^V;
flush = ^O; min = 1; time = 0;
-parenb -parodd cs8 hupcl -cstopb cread -clocal -crtscts
-ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl -ixon
-ixoff -iuclc -ixany -imaxbel
-opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0
ff0
-isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase -tostop
-echoprt -echoctl -echoke



But if forgot to say that after the startup completed and agetty starts the
console the input works again. It just does not work during the startup. It
works before the startup to go into the bios and it works after the startup.


> On Sun, Dec 23, 2001 at 03:06:58PM +0100, Andreas Gietl wrote:
> > I compiled all kernels with the same configuration with serial
> > console-option enabled. With 2.4.14 everything is just fine: I see the
> > kernel-output and can type in things during startup esp. do a fsck. With
> > 2.4.16 and 2.4.17 i SEE everything but no input is accepted.
>
> From what I've deduced, there are a number of SysVinit implementations out
> there that incorrectly clear the terminals CREAD flag.
>
> [-]cread
> allow input to be received
>
> Obviously if it is cleared, then you'll see your behaviour. If you can log
> into the box some other way, please check the settings using:
>
> stty -aF /dev/ttyS0

--
e-admin internet gmbh
Andreas Gietl
Roter-Brach-Weg 124a
tel +49 941 3810884
fax +49 941 3810891
mobil +49 171 6070008

2001-12-23 19:39:49

by Eric W. Biederman

[permalink] [raw]
Subject: Re: serial console on > 2.4.14

Andreas Gietl <[email protected]> writes:

> On Sunday 23 December 2001 15:23, Russell King wrote:
>
> This is what stty gives
>
> speed 38400 baud; rows 0; columns 0; line = 0;
> intr = ^C; quit = ^\; erase = ^?; kill = ^X; eof = ^D; eol = <undefiniert>;
> eol2 = <undefiniert>; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R; werase =
> ^W; lnext = ^V;
> flush = ^O; min = 1; time = 0;
> -parenb -parodd cs8 hupcl -cstopb cread -clocal -crtscts
> -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl -ixon
> -ixoff -iuclc -ixany -imaxbel
> -opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0
> ff0
> -isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase -tostop
> -echoprt -echoctl -echoke
>
>
>
> But if forgot to say that after the startup completed and agetty starts the
> console the input works again. It just does not work during the startup. It
> works before the startup to go into the bios and it works after the startup.

You almost certainly have an init that is clearing cread. There are patched
versions that have this problem fixed. Apparently cread actually being implemented
in the kernel serial drivers is new, and only started appearing since 2.4.2.

Eric