Date: Sat, 18 Apr 2015 14:37:56 -0700 From: Rui Paulo <rpaulo@me.com> To: freebsd-net@freebsd.org Cc: Mark Johnston <markj@freebsd.org> Subject: Re: moving struct bpf_if to bpf.c Message-ID: <3983189.kvm2TuS0U7@akita> In-Reply-To: <20150418210855.GA50409@charmander.west.isilon.com> References: <20150418210855.GA50409@charmander.west.isilon.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Saturday 18 April 2015 14:08:56 Mark Johnston wrote: > Hello, > > At the moment, bpf.h defines struct bpf_if differently depending on > whether BPF_INTERNAL is #defined. This causes problems with CTF, as it > results in a sort of bifurcation within the type graph: CTF sees two > different struct bpf_ifs, and so every struct/union containing a struct > bpf_if is duplicated, and so on. CTF currently imposes a limit of 2^15 > distinct types within a container, and several people have been running > into this limit. The type duplication exacerbates this problem. > > The change here fixes the issue by moving the definition of > struct bpf_if to bpf.c, and making its externally-used fields > available via struct bpf_if_ext: https://reviews.freebsd.org/D2319 > > In particular, the ext fields are at the same offsets within struct > bpf_if as before, so this change should have no functional impact. > Moreover, it reduces the number of types from 20879 to 15725 with my > (stripped-down) kernel config on amd64. Would anyone be willing to > review the proposed change? I've also placed the raw diff here: > https://people.freebsd.org/~markj/patches/bpf_entrails.diff Looks good to me. Thanks for fixing this! -- Rui Paulo
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3983189.kvm2TuS0U7>