Re: Fwd: [PATCH] [-mm] ACPI: export ACPI events via netlink

[Patrick McHardy <kaber@trash.net>]

Johannes Berg wrote:
> On Mon, 2007-07-02 at 16:48 +0200, Johannes Berg wrote:
> 
> 
>>If I find time I might
>>actually fix the unregistration bug too, but I have a feeling digging in
>>the socket code might take more time than I have right now.
> 
> 
> Hmm. I started digging into the af_netlink.c code and realised that the
> whole thing I've been doing cannot possibly work completely since the
> genl socket is created with GENL_MAX_ID as the "groups" parameter to
> netlink_kernel_create() and that limits the groups, and the af_netlink
> code really wants to know the number of groups up-front.


Good point, I missed that.

> So some deeper surgery is required to lift the limit of 1023 multicast
> group now. Not that I like the current genetlink code, we allocate 256
> bytes for the in-kernel socket just for the listeners bitmap, and just
> as many for each socket's groups bitmaps while it's unlikely a regular
> system right now will ever reach that limit.


The kernel socket doesn't have a bitmap allocated, only user sockets
have, and only if they actually bind to a multicast group. In case
of genetlink that still almost as bad of course.