On Tue, Nov 08, 2022 at 04:24:48PM -0600, rama nichanamatlu wrote: > in-line. <...> > > The thing is that "vendor" failed to explain internally if this debug > > code is useful. Like I said, extremely rare debug code shouldn't be part > > of main data path. > > > > Thanks > > thank you very much for you insights into the relevance of the patch. before > we close this topic, do want to ask on this. what is the expectation of the > nic hardware when it signatures / checksum's an eqe or cqe ? > > if it not to be verified by the receiver host for what ever reasons, then > why even do the checksum computation on the hardware ? mlx5 data sheet has more than 3000 pages in it. Not everything there is needed now. Thanks
