Re: [PATCH] drivers/net: Use RCU_INIT_POINTER(x, NULL) in tun.c

[Eric Dumazet]

On Mon, 2014-03-24 at 08:22 +0200, Michael S. Tsirkin wrote:

> What happens if someone then changes that NULL to something else?
> Things will start to break in subtle way, won't they?

Sure. As anything else can break when/if using wrong API or in any
change. Particularly in RCU protected code. Even a kzalloc() can be
buggy.

Note that smp_wmb() is hardly expensive in all these paths,
so thats why I do not care anymore asking guys using RCU_INIT_POINTER(x,
NULL) instead of rcu_assign_pointer() as I used to.

When doing tree wide change (like in commit 2cfa5a0471fef43), I usually
not do a dumb one, I take the time to read the code and check it for
defects.

I do not remember what kind of compiler warnings we had 3 years ago,
But I do not want to spend more cycles on this subject...


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/