Re: [PATCH 1/2] timer: Avoid waking up an idle-core by migrate running timer

On Tue, Mar 31, 2015 at 12:25:16PM +0530, Viresh Kumar wrote:

Cc: Steven Miao realmz6@gmail.com

+#define TIMER_FLAG_MASK 0x7LU

So Steven, this will break compilation on blackfin because that makes ____cacheline_aligned a NOP while we assume it will generate __attribute__((__aligned__(SMP_CACHE_BYTES))) with the further assumption that SMP_CACHE_BYTES >= 8.

Can you explain why blackfin chooses to break ____cacheline_align for UP? We have the *_in_smp variants to distinguish these cases.