> On 2025-05-08, Janne Johansson <[email protected]> wrote: > >> Now a lot of my experience here comes from Linux many moons ago. > >> But I found a lot of hardware that wouldn't play nicely unless > >> offloading was disabled, with the most aggravating bugs. There's > >> also minor nuisances, like checksums not matching in tcpdump under > >> some circumstances. And some NICs would seem to "go bad" at some > > > > Well, if you tcpdump packets outwards going which are going to get > > their checksums done by the hw, then tcpdump will correctly state that > > the checksum is bad, because it has not been done yet by the time it > > picks it up. > > One consequence of this (specifically wrt checksum offloading, which has > been done for a *long* time on some NICs): if there's corruption on the > bus, it won't then be picked up and packets dropped. I discovered this > the hard way, sat in a freezing datacentre at 3am.
Yeah, it will cover some bugs, but those should be really rare. Also, like the OP I have "solved" some issues long ago by turning off offloading, but in that case it was cheap crap cards and speeds were so the cpu could handle the extra load. Lots of water under the bridges since then. Then again, if it is possible somehow to turn it off, then I see no serious harm in adding such features, just like all the other development done to any driver it carries a small risk, and hopefully a lots larger gain overall if it works out fine. -- May the most significant bit of your life be positive.
