From: Jeff Dike <jdike@ka...> - 2003-01-20 16:32:18
> If you do the oldconfig and the oldconfig has no reference to SKAS or
> TT, then it defaults both of them to false. I am claiming that at
> least TT and probably both should default to true.
I don't see a way of making them both default to Y in the 2.4 Kconfig.
What I did was force CONFIG_MODE_TT to Y if CONFIG_MODE_SKAS is disabled.
So, if you're paying attention and say Y to CONFIG_MODE_SKAS, you are probably
still paying attention to CONFIG_MODE_TT, which is the next question.
If you're not paying attention, you will get CONFIG_MODE_SKAS=N and
CONFIG_MODE_TT=Y, which will give you a working UML.
From: Michael Richardson <mcr@sa...> - 2003-01-24 18:37:12
-----BEGIN PGP SIGNED MESSAGE-----
>>>>> "David" == David Coulson <david@...> writes:
>> If you copy a config.in from an older uml (eg from the
>> 2.4.19-5um rpm) then do make oldconfig in a new one,
>> both tt and skas default to no. I believe they should
>> both default to yes since there is no harm in having
>> them, and you have to have at least one.
David> config.in? Don't you mean '.config'? Why would you copy the .config in
David> from 2.4.19-5um to a more recent kernel, when you can get the default
David> configuration with 'make mrproper'?
Because I (and my team) do automated kernel builds, and I do not want to
have them die each time that UML adds a new option without a reasonable
As it is, I have to update my stock UML config files before my team can try
a newer patch case, and that means that we rarely run the latest patches.
UML has a very nice feature - the hardware is always the same (Unix
process). It is therefore a very well defined test environment. skas changes
this slightly, but as long as there is TT still available, things are okay.
] ON HUMILITY: to err is human. To moo, bovine. | firewalls [
] Michael Richardson, Sandelman Software Works, Ottawa, ON |net architect[
] mcr@... http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another Debian GNU/Linux using, kernel hacking, security guy"); [
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)
Comment: Finger me for keys
-----END PGP SIGNATURE-----
From: Jeff Dike <jdike@ka...> - 2003-01-24 18:53:24
> Because I (and my team) do automated kernel builds, and I do not want
> to have them die each time that UML adds a new option without a
> reasonable default...
This was a bug, and I've fixed it in the current UML.