00:30
_whitelogger has joined #m-labs
00:48
mauz555 has joined #m-labs
00:53
mauz555 has quit [Ping timeout: 252 seconds]
00:57
futarisIRCcloud has joined #m-labs
01:21
_whitelogger has joined #m-labs
01:35
mauz555 has joined #m-labs
01:35
mauz555 has quit [Remote host closed the connection]
01:35
mauz555_ has joined #m-labs
02:06
mauz555_ has quit [Remote host closed the connection]
02:11
balrog has quit [Quit: Bye]
02:16
balrog has joined #m-labs
04:07
mauz555 has joined #m-labs
04:11
mauz555 has quit [Ping timeout: 258 seconds]
04:40
sb0 has joined #m-labs
04:51
_whitelogger has joined #m-labs
06:37
<
_whitenotifier-9 >
[m-labs/nmigen] whitequark ce1eff5 - hdl.rec: implement Record.connect.
06:40
rohitksingh has joined #m-labs
06:43
<
_whitenotifier-9 >
[m-labs/nmigen] whitequark 083016d - back.rtlil: only expand legalized values in Array/Part context on RHS.
06:48
<
whitequark >
migen apparently creates clock domains that are used anywhere in the design if they are not explicitly instantiated
06:48
<
whitequark >
is this actually desirable? to me it feels like it could read to non-obvious bugs
06:50
<
whitequark >
e.g. forgetting to rename AsyncFIFO's domains would leave them hanging
07:10
rohitksingh has quit [Ping timeout: 246 seconds]
07:20
<
_whitenotifier-9 >
[m-labs/nmigen] whitequark 33f9bd2 - hdl.ast: accept Signals with identical min/max bounds.
07:20
<
_whitenotifier-9 >
[m-labs/nmigen] whitequark 360bc9b - hdl.ast: improve tests for exceptional conditions.
07:55
<
_whitenotifier-9 >
[m-labs/nmigen] whitequark 85ae99c - back.rtlil: emit `nmigen.hierarchy` attribute.
07:55
<
_whitenotifier-9 >
[nmigen] whitequark closed issue #54: Hirarchy of submodules is not obvious from verilog -
https://git.io/fjkXd
07:56
<
_whitenotifier-9 >
[nmigen] whitequark commented on issue #54: Hirarchy of submodules is not obvious from verilog -
https://git.io/fjOdQ
08:19
<
_whitenotifier-9 >
[nmigen] whitequark commented on issue #3: Ensure that all submodules are added to the design -
https://git.io/fjOdN
08:32
mauz555 has joined #m-labs
08:37
mauz555 has quit [Ping timeout: 240 seconds]
08:45
mauz555 has joined #m-labs
08:53
<
_whitenotifier-9 >
[m-labs/nmigen] whitequark 44711b7 - hdl.ir: detect elaboratables that are created but not used.
08:53
<
_whitenotifier-9 >
[m-labs/nmigen] whitequark aed2062 - Remove examples/tbuf.py.
08:53
<
_whitenotifier-9 >
[nmigen] whitequark closed issue #3: Ensure that all submodules are added to the design -
https://git.io/fpbth
08:57
mauz555 has quit [Ping timeout: 258 seconds]
09:01
mauz555 has joined #m-labs
09:09
mauz555 has quit []
10:15
rohitksingh has joined #m-labs
10:46
sb0 has quit [Quit: Leaving]
11:00
_whitelogger has joined #m-labs
12:46
cr1901_modern has quit [Ping timeout: 246 seconds]
12:49
cr1901_modern has joined #m-labs
12:54
cr1901_modern1 has joined #m-labs
12:55
cr1901_modern has quit [Ping timeout: 246 seconds]
13:39
mauz555 has joined #m-labs
13:53
mauz555 has quit []
13:56
<
mtrbot-ml >
[mattermost] <sb10q> whitequark: that's done to make migen produce a useful output by default - export clock and reset signals as inputs
13:56
<
mtrbot-ml >
[mattermost] <sb10q> otherwise they would be undriven by default
13:56
<
whitequark >
sb: why not a hard error?
13:57
<
mtrbot-ml >
[mattermost] <sb10q> so you can demonstrate migen with fewer lines of code. not terribly important though, having a hard error is also acceptable
14:01
<
mtrbot-ml >
[mattermost] <sb10q> iirc that behavior is disabled if the user creates any clock domains
14:01
<
whitequark >
oh I see, I misunderstood how that worked
14:01
<
mtrbot-ml >
[mattermost] <sb10q> so, in most real-world designs, that won't cause issues with things like forgetting to add ClockDomainsRenamer on a FIFO
14:02
<
mtrbot-ml >
[mattermost] <sb10q> also, Vivado errors out if you have I/O pins that are not constrained
14:02
<
mtrbot-ml >
[mattermost] <sb10q> so even with the default behavior, that would be automatically caught eventually
14:02
<
whitequark >
that varies by toolchain
14:02
<
whitequark >
so you can't rely on it
14:02
<
whitequark >
i think it is OK if the platform always creates some clock domain
14:04
rohitksingh has quit [Ping timeout: 246 seconds]
14:12
X-Scale has quit [Ping timeout: 246 seconds]
14:16
cr1901_modern1 has quit [Quit: Leaving.]
14:16
cr1901_modern has joined #m-labs
14:19
<
cr1901_modern >
err, double link
14:25
X-Scale has joined #m-labs
14:26
rohitksingh has joined #m-labs
15:19
<
mtrbot-ml >
[mattermost] <sb10q> is sci-hub down?
15:21
<
cr1901_modern >
Website exists, but I can't download it either
15:22
<
whitequark >
cr1901_modern: thanks, this helps
15:29
rohitksingh has quit [Ping timeout: 244 seconds]
17:42
rohitksingh has joined #m-labs
18:24
_florent_ has left #m-labs [#m-labs]
18:26
rohitksingh has quit [Ping timeout: 246 seconds]
19:26
ZirconiumX has quit [Ping timeout: 258 seconds]
20:39
<
lkcl >
hmmm, it
*might* be the case that the removal of unnecessary elaboratables is being overaggressive
20:40
<
lkcl >
m.submodules += Array() where the Array is a suite of modules results in important signals being destroyed
20:40
<
lkcl >
whereas m.submodules.child1 = Array()[1]
20:40
<
lkcl >
m.submodules.child2 = Array()[2]
20:41
<
lkcl >
this since a git pull from about... 36 hours ago
20:46
<
lkcl >
self.rs = Array(rs)
20:46
<
lkcl >
m.submodules += self.rs
20:46
<
lkcl >
used to work fine
23:35
<
mtrbot-ml >
[mattermost] <astro> artiq windows-no-hardware-tests work on hydra now
23:36
<
mtrbot-ml >
[mattermost] <astro> but extended-tests (kc705) hang on test_rtio_log