in initrd "exec sh /dev/console 2>&1" does not respond to keyboard

David Mathog mathog at caltech.edu
Sat Mar 2 00:32:38 UTC 2019


On 28-Feb-2019 17:46, David Mathog wrote:
> On 27-Feb-2019 18:13, Michael Conrad wrote:
>> Well, just a guess, but I'd say your problem is with kernel device
>> drivers for input.  Usually people have the basic text showing up on
>> their console and the complaints are about how to get ^C and other
>> tty-specific features to work while using /dev/console.  Your problem
>> sounds more like the Linux virtual console subsystem isn't connecting
>> to your input devices. I'd suggest poking around in the input drivers
>> section of the config and make sure things are compiled as "=y" and
>> not as modules.
> 
> Well, in 3 different kernels there were:
> 
> CONFIG_INPUT=y
> CONFIG_INPUT_MOUSEDEV=y
> CONFIG_INPUT_KEYBOARD=y
> 
> but only 2 of the 3 work, the third one ignores input. Oddly 3.10.108
> was the one originally giving me problems, so rebuilt it using "menu
> oldconfig" starting from the config-3.10.0-957.5.1.el7.x86_64 from
> Centos 7 and then 3.10.108 worked.  But the 3.10.108 kernel had
> problems with xfs.  So built kernel 3.14.78, starting from the last
> good 3.10.108 config, and that kernel does not respond (but xfs
> works!)

Correction, xfs did not work, that test was accidentally done with a 
Centos 7 kernel.  In case this bites somebody else, 3.14.* kernels have 
a parameter
CONFIG_MODVERSIONS which is automatically set when upgrading a config 
(here the 3.10.108 config) like so:

yes "" | make oldconfig

The problem is that on the subsequent build hundreds of warnings like 
this appear
in the log file:

WARNING: "schedule_timeout" [drivers/usb/storage/usb-storage.ko] has no 
CRC!

Looks innocuous right, just a warning?  These really should not be 
warnings but errors, because if one of these afflicted modules tries to 
load it will fail.  That happened with gcc  4.8.5 and 5.3.1, for x86_64. 
  Presumably it works for some other combination of compiler/build tools 
for that platform.  The solution for that was to change the config for 
3.14.78 to:

# CONFIG_MODVERSIONS is not set

and do a clean build.  With 3.14.78 built like that the keyboard works 
when it should and so does xfs.

Thanks,

David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech


More information about the busybox mailing list