> On Jan. 21, 2016, 7:36 p.m., Steve Reinhardt wrote: > > Sorry it took a while for me to get to this... overall this is nice but I'm > > a little confused on the high-level design, and what the FabricLinks are > > adding, and how flow-control is working via the FabricLinks. For example, > > let's say everything is idle, and then two sources (A and B) simultaneously > > transmit to the same destination (C). If I'm understanding the code > > correctly, since the A->C and B->C FabricLinks are both idle, both the A > > and B sender interfaces will transmit their packets along these respective > > links, but then when the links go to deliver to C's receive interface, one > > will win and the other will find the interface busy and drop the packet, > > right? If so, then that means we drop a packet despite having plenty of > > buffering available. If not, then what am I missing? Thanks!
In the current implementation we just drop packets when there is no buffer space at the Fifo-s. In the scenario that you mentioned, both packet will be pushed into the outputFifo and scheduled for delivery later on. > On Jan. 21, 2016, 7:36 p.m., Steve Reinhardt wrote: > > src/dev/net/etherswitch.hh, line 5 > > <http://reviews.gem5.org/r/3230/diff/4/?file=53085#file53085line5> > > > > Is this license necessary? I believe this extra paragraph only applies > > to ARM contributions, so there's no reason to include it here. (Same goes > > for .cc file of course.) I'm Sorry, I am not familiar with gem5 licensing stuff. I removed that part, is that what you mean? - Mohammad ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/3230/#review7940 ----------------------------------------------------------- On Jan. 24, 2016, 8:35 p.m., Mohammad Alian wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/3230/ > ----------------------------------------------------------- > > (Updated Jan. 24, 2016, 8:35 p.m.) > > > Review request for Default. > > > Repository: gem5 > > > Description > ------- > > Changeset 11292:9ed37b34da3a > --------------------------- > dist,dev: add an ethernet switch model > > > Diffs > ----- > > src/dev/net/Ethernet.py 9d2364203316 > src/dev/net/SConscript 9d2364203316 > src/dev/net/etherswitch.hh PRE-CREATION > src/dev/net/etherswitch.cc PRE-CREATION > > Diff: http://reviews.gem5.org/r/3230/diff/ > > > Testing > ------- > > several testing done with different benchmarks and different switch sizes > > > Thanks, > > Mohammad Alian > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
