On Thu, 2010-11-11 at 15:27 -0800, Andrew Morton wrote: > On Fri, 12 Nov 2010 00:19:54 +0100 > "Ricardo M. Correia" <ricardo.correia@xxxxxxxxxx> wrote: > > > On Thu, 2010-11-11 at 23:45 +0100, Ricardo M. Correia wrote: > > > On Thu, 2010-11-11 at 14:25 -0800, Andrew Morton wrote: > > > > And then we can set current->gfp_mask to GFP_ATOMIC when we take an > > > > interrupt, or take a spinlock. > > > > Also, doesn't this mean that spin_lock() would now have to save > > current->gfp_flags in the stack? > > > > So that we can restore the allocation mode when we do spin_unlock()? > > If we wanted to go that far, yes. Who's up for editing every spin_lock() > callsite in the kernel? Hmm... Coccinelle? ;) [1] [1] http://coccinelle.lip6.fr/ -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxxx For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>