----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/3590/#review8598 -----------------------------------------------------------
This looks good, but I suspect there might be draining issues here. Isn't there a race between a system entering the drained state (DmaPort::handleResp) and calling process? The DmaChunkEvent could potentially get lost in a check point unless the device owning the DmaCallback helper is aware of this. I would suggest that you either implement the Drainable interface in DmaCallback to prevent this from happening or document this behaviour. - Andreas Sandberg On Aug. 4, 2016, 7:44 p.m., Michael LeBeane wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/3590/ > ----------------------------------------------------------- > > (Updated Aug. 4, 2016, 7:44 p.m.) > > > Review request for Default. > > > Repository: gem5 > > > Description > ------- > > Changeset 11556:b91232deb828 > --------------------------- > dev: Add a DmaCallback class to DmaDevice > DmaDevices are able to trigger events on completion of a DMA transfer by > inheriting from this class and implementing the process function. > > > Diffs > ----- > > src/dev/dma_device.hh 91f58918a76abf1a1dedcaa70a9b95789da7b88c > > Diff: http://reviews.gem5.org/r/3590/diff/ > > > Testing > ------- > > > Thanks, > > Michael LeBeane > > _______________________________________________ gem5-dev mailing list gem5-dev@gem5.org http://m5sim.org/mailman/listinfo/gem5-dev