Re: [PATCH] 9p: fix crash when transaction killed

[asmadeus]

Christian Schoenebeck wrote on Wed, Nov 30, 2022 at 02:25:59PM +0100:
> > I'm also not convinced it'd fix anything here, we're not talking about a
> > real server but about a potential attacker -- if a reply comes in with
> > the next tag while we're allocating it, we'll get the exact same problem
> > as we have right now.
> > Frankly, 9p has no security at all so I'm not sure this is something we
> > really need to worry about, but bugs are bugs so we might as well fix
> > them if someone has the time for that...
> > 
> > Anyway, I can appreciate that logs will definitely be easier to read, so
> > an option to voluntarily switch to cyclic allocation would be more than
> > welcome as a first step and shouldn't be too hard to do...
> 
> I would actually do it the other way around: generating continuous sequential
> tags by default and only reverting back to dense tags if requested by mount
> option.
> 
> Is there any server implementation known to rely on current dense tag
> generation?

No, I thought ganesha did when we discussed it last time, but checked
just now and it appears to be correct.

I had a quick look at other servers I have around (diod uses a plain
list, libixp uses a bucket list like ganesha...), but there are so many
9p servers out here that I'm far from keeping track...

Happy to give it a try and see who complains...

> If there is really some exotic server somewhere that uses e.g. a simple
> constant size array to lookup tags and nobody is able to replace that array by
> a hash table or something for whatever reason, then I am pretty sure that
> server is limited at other ends as well (e.g. small 'msize'). So what we could
> do is adjusting the default behaviour according to the other side and allow to
> explicitly set both sequential and dense tags by mount option (i.e. not just
> a boolean mount option).

Well, TVERSION doesn't have much negotiation capability aside of msize,
not sure what to suggest here...