Re: Security issues with local filesystem caching

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

Ar Mer, 2006-10-25 am 11:14 +0100, ysgrifennodd David Howells:
> Currently, CacheFiles temporarily changes fsuid and fsgid to 0 whilst doing its
> own pathwalk through the cache and whilst creating files and directories in the
> cache.  This allows it to deal with DAC security directly.  All the directories
> it creates are given permissions mask 0700 and all files 0000.

That seems sensible and fine. It is precisely why we added a separate
fsuid in the first place so that the user space nfsd could take on an fs
identity without breaking signal and other security based forms.

>  (1) Do all the cache operations in their own thread (sort of like knfsd).

Slow it down to keep Christoph happy seems iffy

>  (2) Add further security ops for the caching code to call.  These might be of
>      use elsewhere in the kernel.  These would set cache-specific security
>      labels and check for them.

I can see good arguments for this in some cases where you want strict
divisons in extremely secure computing cases but not usually.

> Thoughts anyone?

I'd like to know more about why Christoph is objecting, whether he has
actual real world examples of races/problems it introduces by altering
fsuid or what his concern is.