On Fri, Nov 30, 2012 at 4:39 PM, Paulo Zanoni <przanoni at gmail.com> wrote: > The "FDI retrain" and "FDI disable" sequences don't match, just like > the "DP retrain" and "DP disable" sequences also don't match. One is a > full disable, the other is just a link retrain in the middle of the > enable sequence, disabling only what's needed. We're following the > spec for the disable sequences and we're also following the spec for > the "retrain" sequence (except for the lack of the "disable > DDI_BUF_CTL" instruction introduced by this patch). What do you > suggest to change? It's just that I've sacrified too many kittens yesterday trying to get fdi link training to work on ivb (to no avail), so when I see too much magic code I start to freak out. And on a quick hunch I don't see why disabling fdi when the link train fails or when we disable the thing for other reasons should be different. But I've you're convinced that this is the right thing, I'm ok. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch
