On Wed, Jun 24, 2009 at 11:48, Øyvind Harboe<[email protected]> wrote: > On Wed, Jun 24, 2009 at 11:46 AM, Michael Bruck<[email protected]> wrote: >> On Wed, Jun 24, 2009 at 11:30, Øyvind Harboe<[email protected]> wrote: >>> But why should we go for such an inferior and specif solution when a more >>> general one is proposed and worked on? >> >> What are the speed/roundtrip time implications of passing all data >> through the Windows socket interfaces for the TCP/IP solution? > > Possibily neglible as OpenOCD is already highly geared towards > long latency interfaces. It would have to be measured.
This may be valid for high-speed data transfers (but on arm11 for example only in the fast memwrite mode that replaces status polling with fixed delays). I doubt that for single-stepping, where the whole processor state is fetched register-by-register, the long-latency-optimizations make much of a difference. Michael _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
