On Mon, 2 Aug 2010, Paul Goyette wrote:
Yes, the cv mechanism, coupled with changing the semantics of module_autoload() (to not require the caller to obtain the lock) makes good sense here. I'm working on it and I should be able to post new diffs tomorrow.
OK, it's time for Round #3! Attached are diffs that implement Adam's suggested cv-based solution,along with moving the locking for module_autolock() into the routine itself (for consistency with other routines). I've also updated the module(4) man page this time around.
------------------------------------------------------------------------- | Paul Goyette | PGP Key fingerprint: | E-mail addresses: | | Customer Service | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com | | Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette at juniper.net | | Kernel Developer | | pgoyette at netbsd.org | -------------------------------------------------------------------------
Attachment:
MODULE_LOCKING.diff.gz
Description: Binary data