On Tue, 2012-03-13 at 11:04 +0100, Indan Zupancic wrote:
> Hello,
>
> I made a quick pseudo-patch for BPF JIT support. As far as I can tell,
> the actual code itself is very simple, just:
>
> - bpf_jit_compile(fp);
> + jit = bpf_jit_compile(fp->insns, fp->len, 1);
> + if (jit) {
> + fp->bpf_func = jit;
> + /* Free unused insns memory */
> + newfp = krealloc(fp, sizeof(*fp), GFP_KERNEL);
> + if (newfp)
> + fp = newfp;
> + }
>
> old_fp = rcu_dereference_protected(sk->sk_filter,
> sock_owned_by_user(sk));
Dont mix unrelated changes in a single patch.
Current krealloc() doesnt alloc a new area if the current one is bigger
than 'new_size', but this might change in next kernel versions.
So this part of your patch does nothing.
If it did, this kind of 'optimization' can actually be not good, because
sizeof(*fp) is small enough (less than half cache line size) to trigger
a possible false sharing issue. (other part of the cache line could be
used to hold a often dirtied object)
--
To unsubscribe from this list: send the line "unsubscribe linux-arch" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
