Re: [RFC] [PATCH] cpuset: fix wrong calculation of relax domainlevel

From: Paul Jackson
Date: Thu Jul 17 2008 - 16:09:32 EST

In looking at this, I notice something I should have questioned before.

The include/linux/sched.h code:

struct sched_domain_attr {
int relax_domain_level;

#define SD_ATTR_INIT (struct sched_domain_attr) { \
.relax_domain_level = -1, \

and the associated passing of relax_domain_level's embedded inside
a kmalloc'c struct sched_domain_attr 'dattr' seems like excessive
obfuscating apparatus to me. Unless someone has short term plans
to be adding some other attributes to this sched_domain_attr, I
suspect it would make more sense just to pass relax_domain_level's
as explicit lvalues, dropping all this attr stuff.

Adding unnecessary abstractions 'for future growth' is usually a
bad idea. It impedes current code understanding more than it aids
future code growth.

I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <pj@xxxxxxx> 1.940.382.4214
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at