Re: [PATCH net-next v6 2/7] net: ethtool: add support for symmetric-xor RSS hash

[Jakub Kicinski <kuba@kernel.org>]

On Mon, 27 Nov 2023 07:21:47 -0700 Ahmed Zaki wrote:
> On 2023-11-21 16:33, Jakub Kicinski wrote:
> > On Mon, 20 Nov 2023 13:56:09 -0700 Ahmed Zaki wrote:  
> >> + * @data: Extension for the RSS hash function. Valid values are one of the
> >> + *	%RXH_HFUNC_*.  
> > 
> > @data is way too generic. Can we call this key_xfrm? key_preproc?  
> 
> We manipulate the "input data" (protocol fields) not the key. I will 
> rename to "input_xfrm".

Ugh, right!

> >> +/* RSS hash function data
> >> + * XOR the corresponding source and destination fields of each specified
> >> + * protocol. Both copies of the XOR'ed fields are fed into the RSS and RXHASH
> >> + * calculation.
> >> + */
> >> +#define	RXH_HFUNC_SYM_XOR	(1 << 0)  
> > 
> > We need to mention somewhere that sym-xor is unsafe, per Alex's
> > comments.  
> 
> I already added the following in Documentation/networking/scaling.rst:
> 
> "The Symmetric-XOR" is a type of RSS algorithms that achieves this hash 
> symmetry by XORing the input source and destination fields of the IP
> and/or L4 protocols. This, however, results in reduced input entropy and
> could potentially be exploited."
> 
> Or do you mean add it also to "uapi/linux/ethtool.h" ?

Yes, a short mention in a comment next to the define.
There's a good chance person looking at this will not notice 
the documentation in scaling.