Hi, I can be live with using little endian. This is not the first time I see using little endian with IEEE.
However with the current code the endian depends on machine. So a linuxptp daemon that runs on a big endian (there are yet some) will send the message in big endian. May be we should limit this TLV message to UDS only? Since UDS always run on the same machine. Erez -----Original Message----- From: Richard Cochran <richardcoch...@gmail.com> Sent: Tuesday, 9 March 2021 17:57 To: Geva, Erez (ext) (DI PA DCP R&D 3) <erez.geva....@siemens.com> Cc: linuxptp-devel@lists.sourceforge.net Subject: Re: PORT_STATS_NP management TLV message On Tue, Mar 09, 2021 at 01:59:33PM +0000, Geva, Erez wrote: > As I add the PORT_STATS_NP message, > I notice that the statistics is stored in unsigned integer 64 bits values, > but using host order instead of network order. Oops, that was an oversight. > As I know you follow the IEEE standard, It would be nice if you could > elaborate on this message. This message is a custom message, and it is not covered by IEEE 1588 or any other standard. > Can we fix it and rebase it to network order? Unfortunately, simply "fixing" it unconditionally will break existing deployments. > May be add a flag for backward compatible. If we go with that, then the default value of "compatible" will have to be "true". > What do you think? Maybe it is better to simply leave it as is. I myself don't use this TLV. Since the pmc tool prints the correct values, probably nobody would insist on having big endian in the on-the-wire format. Thanks, Richard _______________________________________________ Linuxptp-devel mailing list Linuxptp-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
