linux-kernel - Re: [PATCH v2 net-next 1/2] bpf: allow extended BPF programs access skb fields

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]

Date:	Sat, 14 Mar 2015 19:02:57 -0700
From:	Alexei Starovoitov <ast@...mgrid.com>
To:	Daniel Borkmann <daniel@...earbox.net>,
	"David S. Miller" <davem@...emloft.net>
CC:	Thomas Graf <tgraf@...g.ch>, linux-api@...r.kernel.org,
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 net-next 1/2] bpf: allow extended BPF programs access
 skb fields

On 3/14/15 4:51 PM, Daniel Borkmann wrote:
> On 03/14/2015 04:55 PM, Alexei Starovoitov wrote:
> ...
>> so from there I saw two options: either copy paste all
>> build_bug_on and have the same *insn=... and build_bug_on in
>> two places or consolidate them in single helper function.
>> Obviously single helper function is a preferred method.
>> I'm not sure what are you still arguing about.
>
> I'm repeating myself here, but fair enough. To me the v1
> code in sk_filter_convert_ctx_access() was more sound. So
> taking out the ifindex issue, that's 4 BUILD_BUG_ON()s in
> addition.

correct. it's 4 build_bug_on to several lines of copy pasted code.
That copy-paste between two functions was already bugging me
when I posted v1, but back then I didn't see a way to create
a common helper.
Adding this 4 extra lines pushed it over my internal bar of
'acceptable' copied code :)
so in v2 I figured a way to create a helper.

> I actually think the current filter code is in a reasonable
> shape. convert_bpf_extensions() is full of BPF to eBPF
> conversions, so going over convert_bpf_extensions() some of
> them would now use convert_skb_access(), some other ``skb
> accesses''use macros directly in place, the reading-flow of
> this code now is inconsistent to me and it would have been
> more sound if that's just left as is in convert_bpf_extensions().

I still don't see this 'inconsistency'.
convert_bpf_extensions() has code to convert classic only accesses.
convert_skb_access() has code to convert both classic and extended.
sk_filter_convert_ctx_access() has code to convert extended only.

In this patch convert_skb_access() has to deal with 3 fields.
Later we may allow more field to be accessed by extended, so
more lines will simply move from convert_bpf_extensions() into
convert_skb_access(). So it will save us from copy-pasting in the
future.

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