14:12 <sb0> _florent_, "set property DRIVE " + + str(c.strength) ?

14:14 <GitHub172> [migen] sbourdeauducq pushed 2 new commits to master: http://git.io/G8swMA

14:14 <GitHub172> migen/master f6dfabf Florent Kermarrec: mibuild/xilinx_vivado.py: add set property to misc constraint

14:14 <GitHub172> migen/master 506ac0f Sebastien Bourdeauducq: Merge branch 'master' of github.com:m-labs/migen

14:16 <_florent_> yes sorry, I fixed the underscore but missed the double +...

14:16 <_florent_> thanks

14:17 <sb0> _florent_, I recommend using flen(rmap.bus.dat_w)

14:17 <sb0> it's done like that for other buses except LASMI that needs more info

14:18 <sb0> ...in general, things should be done consistently in migen, and data should not be duplicated in different attributes

14:19 <_florent_> ok I can change that

14:20 <sb0> in LASMI, we already have "aw" which is difficult to obtain otherwise (as the signal is inside the bank groups)

14:20 <sb0> so dw is acceptable

14:21 <_florent_> so if I change "busword" to "dw", it's ok for you?

14:21 <sb0> no, since this is not LASMI; additionally this has absolutely nothing to do in Bank

14:23 <sb0> "bank.bus.dw" is acceptable (for consistency with LASMI), but in that case the "dw" attribute should be set on *all* bus interfaces (and it needs to be "dw" everywhere), and all code that uses flen(bus.dat_w) changed

14:24 <sb0> the current solution is to use flen(bus.dat_w) and make an exception for LASMI as there's a lot of data to store and the address signal width is difficult to optain

14:24 <_florent_> hmm ok, so for now I will use flen(bus.dat_w) for now

15:39 <GitHub120> [misoc] sbourdeauducq pushed 1 new commit to master: http://git.io/QJpeHA

15:39 <GitHub120> misoc/master d4833cb Florent Kermarrec: cpuif: remove limitations on csr data_width