On 12 January 2015 at 19:21, Mike Holmes <[email protected]> wrote: > This raises a general president question for ODP and a more specific one. > > My feeling is that ODP wants to abstract network acceleration capabilities > rather than OS features and so leave things like atomics etc to the > application if possible. Is that also your view on the change Ola ? IMO ODP should abstract what is often HW capabilities and their interfaces. These can be both SoC accelerations and features on CPU architecture level. Atomics support fits in here (it is not related to the OS).
> Have we hidden needed functionality, it was in a real external application > afterall. There's nothing stopping an application from using whatever atomics support is available from the compiler. But I don't mind exposing more of the ODP atomics support, as was done in my original proposal. The original atomics API with cmpset operations wasn't great (not abstract enough, working on ints directly) and could not be implemented on all architectures. > > Mike > > > On 12 January 2015 at 04:17, Ola Liljedahl <[email protected]> wrote: >> >> On 12 January 2015 at 08:54, Prashant Upadhyaya >> <[email protected]> wrote: >> > Hi, >> > >> > >> > >> > I am migrating my application code from ODP v0.2 to v0.7 >> > >> > In v0.2, there used to be a function like the following – >> > >> > odp_atomic_cmpset_u32 >> > >> > >> > >> > How do I realize the above in v0.7 ? >> It was decided that the atomic operations not related to counters and >> counting should be kept internal. So there is a header file >> odp_atomic_internal.h (at least in the linux-generic implementation of >> ODP). Here is a function _odp_atomic_u32_cmp_xchg_strong_mm() defined. >> You will have to specify the memory model (or memory order), e.g. >> _ODP_MEMMODEL_RLX (relaxed) or _ODP_MEMMODEL_SC (sequential >> consistency). >> >> > >> > >> > >> > Regards >> > >> > -Prashant >> > >> > >> > >> > "DISCLAIMER: This message is proprietary to Aricent and is intended >> > solely >> > for the use of the individual to whom it is addressed. It may contain >> > privileged or confidential information and should not be circulated or >> > used >> > for any purpose other than for what it is intended. If you have received >> > this message in error, please notify the originator immediately. If you >> > are >> > not the intended recipient, you are notified that you are strictly >> > prohibited from using, copying, altering, or disclosing the contents of >> > this >> > message. Aricent accepts no responsibility for loss or damage arising >> > from >> > the use of the information transmitted by this email including damage >> > from >> > virus." >> > >> > _______________________________________________ >> > lng-odp mailing list >> > [email protected] >> > http://lists.linaro.org/mailman/listinfo/lng-odp >> > >> >> _______________________________________________ >> lng-odp mailing list >> [email protected] >> http://lists.linaro.org/mailman/listinfo/lng-odp > > > > > -- > Mike Holmes > Linaro Sr Technical Manager > LNG - ODP _______________________________________________ lng-odp mailing list [email protected] http://lists.linaro.org/mailman/listinfo/lng-odp
