> -----Original Message----- > From: Patrick Georgi [mailto:[email protected]] > Sent: Saturday, April 25, 2009 9:20 AM > To: Myles Watson > Cc: [email protected] > Subject: Re: [coreboot] [PATCH] more intelligent cbfs walker > > Am 25.04.2009 17:10, schrieb Myles Watson: > > Have you seen this happen? If so delete is broken. Delete zeroes freed > > space. > > > Can we be sure to have only one implementation that writes cbfs?
I hope so, but you're right it's not guaranteed right now. > I've seen quite some weird stuff happen (for example, the first version > of the patch should _always_ end up in an infinite loop, but it didn't > at first), and given that this part of the code can make the difference > between "works" and "hangs with black screen", I'll rather be paranoid > about these issues. Well when you put it that way :) My only concern is that if we get too paranoid it gets harder to maintain and test all of the code paths. Thanks for your patch. It was definitely an improvement. Myles -- coreboot mailing list: [email protected] http://www.coreboot.org/mailman/listinfo/coreboot
