[PATCH] rxrpc: Fix smp_processor_id() usage
Sven Schnelle
svens at stackframe.org
Wed Apr 2 15:40:48 EDT 2008
Pete Zaitcev <zaitcev at redhat.com> writes:
>> +#define dbgprintk(FMT,...) do {\
>> + preempt_disable(); \
>> + printk("[%x%-6.6s] "FMT"\n", smp_processor_id(), current->comm ,##__VA_ARGS__); \
>> + preempt_enable(); \
>> +} while(0)
>
> Hmm, this seems to be a bit exessive. Why do we even need
> the CPU ID in printouts?
Maybe it is helpful in tracking down race conditions (even though my
experience is that race conditions are gone with debugging enabled ;-)).
I share your oppinion that this is excessive - that's why i tend to
simply remove the processor id as written in my first mail.
More information about the linux-afs
mailing list