Re: wg(4): encapsulated transport checksums

2020-06-30 Thread Matt Dunwoodie
On Tue, 30 Jun 2020 20:40:10 -0600 "Theo de Raadt" wrote: > Matt Dunwoodie wrote: > > > Depends on your definition of significant, I've seen 1-3% throughput > > improvement without the patch. > > > Real networks require statistics, which you want to throw away. > > > Overall, it is still

Re: wg(4): encapsulated transport checksums

2020-06-30 Thread Theo de Raadt
Matt Dunwoodie wrote: > Depends on your definition of significant, I've seen 1-3% throughput > improvement without the patch. > Real networks require statistics, which you want to throw away. > Overall, it is still debatable whether to skip the IPv4 checksum as > modern crypto certainly offers

Re: wg(4): encapsulated transport checksums

2020-06-30 Thread Matt Dunwoodie
On Tue, 30 Jun 2020 09:22:18 +1200 (NZST) richard.n.proc...@gmail.com wrote: > Hi Jason, > > On 27/06/2020, at 10:09 PM, Jason A. Donenfeld > wrote: > > [...] I thought I'd lay out my understanding of the matter, > > and you can let me know whether or not this corresponds with what > > you had

Re: wg(4): encapsulated transport checksums

2020-06-30 Thread Claudio Jeker
On Tue, Jun 30, 2020 at 09:22:18AM +1200, richard.n.proc...@gmail.com wrote: > Hi Jason, > > On 27/06/2020, at 10:09 PM, Jason A. Donenfeld wrote: > > [...] I thought I'd lay out my understanding of the matter, > > and you can let me know whether or not this corresponds with what you > > had

Re: wg(4): encapsulated transport checksums

2020-06-29 Thread richard . n . procter
Hi Jason, On 27/06/2020, at 10:09 PM, Jason A. Donenfeld wrote: > [...] I thought I'd lay out my understanding of the matter, > and you can let me know whether or not this corresponds with what you > had in mind: > [...] My main concern is the end-to-end TCP or UDP transport checksum of the

Re: wg(4): encapsulated transport checksums

2020-06-27 Thread Theo de Raadt
> - Therefore, it's not necessary to check the IP checksum on ingress because: There is actually a really good reason. There are various counters (of all packets) which people observe to debug network problems. Now, if lower-level packets carrying wg with corruption don't increment those

Re: wg(4): encapsulated transport checksums

2020-06-27 Thread Jason A. Donenfeld
Hi Richard, Thanks for the patch. I had problems parsing some terminology in your description, so I thought I'd lay out my understanding of the matter, and you can let me know whether or not this corresponds with what you had in mind: - On egress, we must compute the packet checksum, because it