#milkymist IRC log for Tuesday, 2012-06-12

wolfspraulso embarassing. I have two serious customers who are thinking about or have already bought an m1, and I know there is a > 80% chance they will regret their purchase later and/or sooner or later run into such a long list of frustrating bugs and missing features that they will just give up.06:22
wolfspraulmy job is to give them some fake support before this 'give up' and get them to the point that they at least don't want their money back and think "ok, I was stupid in buying this"06:23
wolfspraulgreat work :-)06:23
wpwrakhow not to run an enterprise, rule #2: put a pessimist in charge of marketing and sales :)06:40
wolfspraulI will try to sell m1 to them, and i will cordially invite them to this channel06:49
wolfspraulthen I have the right problem in front of me making them happy, rather than discussing random new tech features06:50
wolfspraul:-)06:50
wolfspraulthere's the optimist! :-)06:50
wpwrakthat sounds better :)06:59
wolfspraullarsc said migen is an "ast generator" - what is ast? abstract syntax tree?07:01
wolfsprauland what's the point, or rather difference between an "ast generator" and "language"?07:01
JiaAST is abstract syntax tree, yes. But AST is not language.07:03
JiaFE have a AST generator07:03
wolfspraulah Jia, hi :-)07:03
wolfspraulI don't understand the difference between "ast generator" and "language" and the point lars was trying to make07:03
Jiamake what? a new compiler?07:04
wolfspraulit was in reference to migen07:05
Jialars must be a man, I think :-)07:05
wolfspraulany update on the gcc-lm32 bugs you looked into?07:05
JiaI'm just back when I finish qemu-or32.07:06
JiaI nerver giev up fix gcc-lm32. and I'm still working on it.07:07
Jiabut, it is really a BUUUUUUUUUUUUUUUG07:07
Jiasorry, migen is?07:09
wolfspraulJia: thanks for asking, I think it's best if you read for yourself https://github.com/milkymist/migen07:17
wolfspraulthe README is right there07:17
JiaWow! cool!07:18
Jiareplace verilog?07:18
lekernelwolfspraul: and what do you expect from these two people?08:04
wolfspraulexpect? I would hope that m1 satisfies them.08:06
wolfspraulobviously08:06
wolfspraulor maybe m1 is depression training?08:06
wpwrakmaybe one of them is a magic frog, waiting to be awakened by a kiss from M1. then he'll be eternally grateful and grant you everlasting fame, mountains of gold, and beautiful princesses :)08:07
lekernelwpwrak: :))08:09
Fallenoulekernel: what's your plan about milkymist-ng and mmu exactly ?09:39
lekernelwell, the only big plans that I have are about -ng. it will replace the current soc at some point.10:01
wolfspraulI don't really see that on m110:11
wolfspraulit sounds like many m1 features will not work, and some are hard to represent via -ng? or am I wrong?10:11
wolfspraulwe just wait and see10:12
lekerneljust reuse the legacy verilog code10:44
lekernelthis is much simpler than reverse engineering fpga architectures and bitstream formats by the way ...10:55
lekernelso, unless you're not convinced that -ng will work or brings any benefit at all, "wait and see" doesn't exactly sound like the right thing to do10:56
Fallenoulekernel: I mean what do you plan using the MMU for on milkymist-ng ? Running Linux ?11:36
lekernelnot me - but nevertheless, the current soc will become obsolete.11:41
lekernelunless you insist that MM should have slow memory, difficult to connect buses, low video resolution, etc.11:42
Fallenousure I would rather use milkymist-ng on M112:11
Fallenouif it's not too difficult to integrate all the other cores (for instance those doing DMA on FML)12:11
lekernelthere's only video-in12:34
Fallenouok so upgrading M1 to milkymist-ng would only break video-in feature ?12:38
lekernelif someone ports the other cores (midi, dmx, etc.), yes13:07
lekernelwhich are quite easy to port13:07
lekernelwhat I'll do for -ng on m1 is read-only flash, uart, ddr, vga, ethernet, tmu, and usb. probably nothing else.13:09
lekerneloh and dvi on r413:09
kristianpaulobsolete if migen aceptance is high.. have anyone check google trends latelly? ;-)14:18
kristianpaulalso rtems still just suported on the legacy m1 soc :), i'm not agaisnt lua, but for wider porpuses it may get hang14:32
lekernelthen port it or use bare metal or linux14:33
lekerneland btw http://www.google.com/trends/?q=lua,rtems =]14:34
kristianpaullol14:34
lekernelthe lua results are probably biased though14:35
kristianpaulas rtems return zero, yeah14:35
kristianpaulhttp://www.google.com/trends/?q=milkymist,+nanonote :p14:36
GitHub74[migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/a5915101890ecf7052c058d860aad546170f792c14:59
GitHub74[migen/master] ASMI simulation models - Sebastien Bourdeauducq14:59
wpwrakof course, ... http://www.googlefight.com/index.php?lang=en_GB&word1=rtems&word2=linux15:09
wpwrakthe rtems bar definitely isn't to scale :)15:10
kristianpaullol http://www.googlefight.com/index.php?lang=en_GB&word1=rtems&word2=baremetal15:21
kristianpaulnice website wpwrak15:22
wpwrakbare metal barely wins :)15:48
GitHub36[migen] sbourdeauducq pushed 3 new commits to master: https://github.com/milkymist/migen/compare/a5915101890e...ce9e35b8efe915:54
GitHub36[migen/master] PureSimulable - Sebastien Bourdeauducq15:54
GitHub36[migen/master] Reorganize examples folder - Sebastien Bourdeauducq15:54
GitHub36[migen/master] fix SimActor get_fragment - Sebastien Bourdeauducq15:54
Fallenoulekernel: I suggest you wait that I push my refactoring before trying to integrate mmu in milkymist-ng/mmubranch15:56
FallenouI think I will push it today or tomorrow15:56
FallenouI move a lot of code around15:56
GitHub150[migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/3a58916a4ffc6917a634530e16337699963d986417:58
GitHub150[migen/master] examples/dataflow/dma: refactor - Sebastien Bourdeauducq17:58
GitHub109[migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/b14be4c8a339eb1a39d9abb8576d990d71518fd019:06
GitHub109[migen/master] actorlib: ASMI sequential reader - Sebastien Bourdeauducq19:06
GitHub64[milkymist-mmu-simulation] fallen pushed 2 new commits to master: http://git.io/CmrJgg19:14
GitHub64[milkymist-mmu-simulation/master] Cleanup and refactor DTLB code - Yann Sionneau19:14
GitHub64[milkymist-mmu-simulation/master] Cleanup and refactor ITLB code - Yann Sionneau19:14
Fallenousynthetizing, let's try ITLB on real fpga19:29
Fallenoudoes not work :)19:53
Fallenougn8!21:54
GitHub68[milkymist-mmu-simulation] fallen pushed 1 new commit to master: http://git.io/zA4ejQ21:55
GitHub68[milkymist-mmu-simulation/master] Add support for ITLB miss generating page fault - Yann Sionneau21:55
GitHub81[milkymist-mmu] fallen pushed 1 new commit to mmu-bios: http://git.io/RW2Y3Q21:59
GitHub81[milkymist-mmu/mmu-bios] Add BIOS support for ITLB page fault handling - Yann Sionneau21:59
--- Wed Jun 13 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!