On Tue, Aug 19, 2025 at 12:27:08PM +0200, Alexandr Nedvedicky wrote:
> Hello,
> 
> On Tue, Aug 19, 2025 at 11:02:37AM +0200, Claudio Jeker wrote:
> > I want to remove most use of RT_TABLEID_MAX (or actually make it a pure
> > userland thing and allowing it to be much larger).
> > 
> > sppp(4) still uses it in a for loop where most other callers use
> > rtmap_limit. So this diff does the same as code in route.c.
> 
>     there is yet another one use of RT_TABLEID_MAX in
>     for loop like we see here in sppp_update_gw() here
>     in  mrt_sysctl_mfc() in ip_mroute.c, shall we take care of it too?

I have a much larger diff for ip_mroute.c that completely reworks the
rdomain / rtable support there. So all those bits will be changed.
 
>     I'm not sure about the other occurrences of RT_TABLEID_MAX we tree,
>     but those are perhaps out of scope at least for now.

Most other occurrences are to validate values sent from userland. So right
now I want to keep those.

My plan i to redo the crazy lookup structures in rtable.c and also the
even worse bits in the multicast routing code. I have a diff for that but
I want to get at least the trivial bits out of the way first.

-- 
:wq Claudio