On Friday 06 July 2007, Mark Lord wrote: > Robert Hancock wrote: > >.. > > Currently the DMA, multi-sector mode, etc. are not controllable with > > hdparm with libata. libata is designed to use the fastest settings > > possible by default. In a lot of cases this messing with hdparm was only > > needed because of stupidity with the old IDE code (like DMA not being > > automatically enabled if the low-level driver was built modular). > > Actually, most of the hdparm flags were put there to help test the IDE > subsystem and to help debug the much stranger hardware it had to deal with. > > "DMA off by default" was a Linus Torvalds request, to help ensure data safety > with all of the weird and wonderful crap pre-standardization.
Exactly as Mark says. I just want to add that "DMA off by default" hasn't been a case for a long time now and since 2.6.21 the config option for "DMA on by default" is also finally gone (to make a long story short: this was the best way to fix some bugs while not causing regressions, not to mention that it was defaulting to Y for years). Thanks, Bart - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
