linux-kernel - Re: [PATCH 6/9] bpf: arraymap: introduce BPF_MAP_TYPE_ARRAY

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20160111191433.GC26495@ast-mbp.thefacebook.com>
Date:	Mon, 11 Jan 2016 11:14:34 -0800
From:	Alexei Starovoitov <alexei.starovoitov@...il.com>
To:	Ming Lei <tom.leiming@...il.com>
Cc:	linux-kernel@...r.kernel.org, Alexei Starovoitov <ast@...nel.org>,
	"David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
	Daniel Borkmann <daniel@...earbox.net>,
	Martin KaFai Lau <kafai@...com>
Subject: Re: [PATCH 6/9] bpf: arraymap: introduce BPF_MAP_TYPE_ARRAY_PERCPU

On Mon, Jan 11, 2016 at 11:56:58PM +0800, Ming Lei wrote:
> This patch introduces percpu array map so that expensive
> atomic operations can be avoided in eBPF prog in case of
> ARRAY map.
> 
> PERCPU MAP uses the percpu version of update/lookup element
> helpers and callbacks to access element in the map, and
> the previous update/lookup element helpers and callbacks
> don't work at the same time.
> 
> Signed-off-by: Ming Lei <tom.leiming@...il.com>

useful stuff!

> +	if (percpu) {
> +		if (alloc_percpu_array(array, attr->max_entries,
> +				       attr->value_size)) {
> +			kvfree(array);
> +			return ERR_PTR(-ENOMEM);
> +		}
> +		array->map.pages = round_up(attr->max_entries *
> +				attr->value_size * num_possible_cpus(),
> +				PAGE_SIZE) >> PAGE_SHIFT;

I think it would be more accurate to add it to array_size instead of
doing page rounding here.

> -	array->map.pages = round_up(array_size, PAGE_SIZE) >> PAGE_SHIFT;
> +	array->map.pages += round_up(array_size, PAGE_SIZE) >> PAGE_SHIFT;

and this line wouldn't need to change...