Re: detecting all possible deadlock conditions

Posted May 31, 2006 15:08 UTC (Wed) by NAR (subscriber, #1313)
In reply to: Re: detecting all possible deadlock conditions by mingo
Parent article: The kernel lock validator

if LTP doesnt trigger it, users will eventually - and will likely not trigger the deadlock itself, but only the codepath that makes it possible. (hence there is much better debug info and a still working system.)

If this feature is turned on in the user's kernel - isn't the performance hit too much for users to keep this configuration option turned on?

If some code is never triggered in practice, why is the code there and how was it tested to begin with?

Error handling (especially hardware-related error handling) might be triggered extremely rarely, but I don't think it should be thrown out. The code could be tested by manually setting some error flag, but obviously the setting of this flag is not included in the production kernel.

Bye,NAR