#milkymist IRC log for Saturday, 2011-03-12

kristianpaulk00:03
kristianpaullekernel: you made this crossbar swich and conbus motifications because navre? if, yes why?00:04
Action: kristianpaul install gitg in order to do a better browsing of changes00:06
Fallenouhehe gitg is great00:07
Fallenougitk too00:07
Fallenousomething funny is gource00:12
kristianpaulwow00:15
kristianpaulgource is awesome00:15
kristianpaul(not saying how usefull)00:15
kristianpaulthe little human moving and jumping around is sebastien? :_)00:16
Fallenouits not written "lekernel" on it ? :p00:16
kristianpauloh yes00:17
kristianpaulfinally clicked on it00:17
kristianpaulis moving fast00:17
Fallenouoh my gource is broken00:17
Fallenougsoc organization applications are now closed00:18
FallenouI hope we submitted00:18
kristianpaullets wait rejon00:22
kristianpaulok xbar is the new conbus00:45
lekernelFallenou: rejon submitted it, yeah04:45
lekernelkristianpaul: what jumpy person?04:45
kristianpauljumpy?04:53
kristianpauli used word jump for when some one will join something04:54
kristianpaulmay be is not well miss undernstoof for the enlgish concept04:54
Action: kristianpaul zzz04:56
kristianpaulFallenou: remenber MLEN it was defined in ./cpukit/libnetworking/sys/mbuf.h but it you already knew it..11:37
kristianpaulFallenou:About ethernet driver, where are defined DMA "parameters" wich point the right address from system memory to read?11:45
kristianpaulIs that the procedure isnt?11:48
kristianpaulor wait, question is how to you read the the rx slots12:00
kristianpaul?12:00
Action: kristianpaul thinks that have a misscopnception about DMA12:01
kristianpaulah i fount somethin..12:03
kristianpaulfound*12:03
kristianpaul0x0C 0x18 0x24 0x30 are slots DMA adress12:03
Action: kristianpaul rtfm12:04
kristianpaulhmm but thosre are adress at the CSR bridge..12:08
kristianpaullet see12:08
kristianpaulacording to system_conf.h  MM_MINIMAC_ADDR0 0xe000800C12:09
kristianpaulPeripheral                      WB base (L1 cached/uncached)12:15
kristianpaulFML bridge (SDRAM)              0x40000000 / 0xc000000012:15
kristianpaulah !12:15
kristianpaulergg no12:15
kristianpaulCSR bridge                      0x60000000 / 0xe000000012:15
kristianpaulbut from xbar12:20
kristianpaul // SDRAM        0x40000000 (shadow @0xc0000000)12:20
kristianpaulargg again12:20
kristianpaul // CSR bridge   0x60000000 (shadow @0xe0000000)12:20
kristianpaulFallenou: when you read MM_MINIMAC_ADDR0, this reg store an adress not actually data?13:32
kristianpaullekernel: resuming, can you explain me a bit how is implemented the DMA Controller in the mm1 soc?14:05
kristianpaulif there is one..14:09
lekernelwell just send wishbone or fml master cycles?14:10
lekernelcores supporting dma are simply additional bus masters, whose requests get routed to the peripherals through the arbiters14:11
lekernelthere's no special DMA infrastructure (yet)14:11
kristianpaulah14:12
kristianpaulhmm, so potentially devices like the ethernet core can create bottle neck in for busy env on wich the ethernet handle big amount of data..14:14
kristianpaulof course not the case for a VJ ;-)14:14
lekernelyou have this problem with any centralized shared memory system14:16
kristianpaulis the case also of fml? (note i dint read too much about this yet)14:17
kristianpauls(¿/of/for14:17
kristianpauloops14:17
lekernelyes, of course14:17
lekernelbut fml has higher throughput than the wishbone bus14:18
kristianpaulso, the swich code you added with the cross bar will help on this shared bus problem, isnt? but is working right now?14:22
lekernelall the crossbar allows is simultaneous processor<->flash (which uses the bus inefficiently due to the flash's slowness) and ethernet<->ram communications14:26
lekernelthis way the processor<->flash communication, e.g. when running the bios, doesn't suck all the memory bandwidth from the ethernet core14:27
kristianpaulif the pfpu is not in the fml,  it means the caculations it have are no so troughtput demanding? or..14:28
kristianpaul(ethernet<->ram communications) ah. nice hint ! less fifo overflow issues ;-)14:28
kristianpauls/fifo/buffer14:28
lekernelwell, I'll leave it as an exercise for you to compute how much memory bandwidth the pfpu is using14:29
kristianpaul;-)14:29
kristianpaulIf you're are teacher i wish the best for your students14:30
lekerneloh come on, it's not hard14:31
kristianpaulBut if remenber well you compile the pach before run it..14:32
lekerneltwo 32-bit integers on each vertex of a 128x128 grid, 30 times per second, how much bandwidth is that?14:32
kristianpaul1 megabyte per second i think14:33
kristianpaulbut i may wrong i dont know what the vertex part is about in the rendering process14:34
kristianpaulbut i guess (128*128*30)*214:36
lekernel*3214:36
kristianpauloh, yes i asumed data word byte but is a integer, true14:42
Fallenoukristianpaul: yes the reg stores address15:57
kristianpaulFallenou: yeah ;-) oh well i still learning :-)16:00
kristianpaullekernel: Fallenou mwalle  wpwrak  What you think about http://sphinx.pocoo.org/ in order to create a central base for milkymist soc documentation16:00
kristianpaulThat could be esilly update rather than individuals pdf.. latex...16:01
kristianpaulof course that, project is not a wiki, and should be friendly for programmers16:01
lekernelpersonally, even though I recognize the value of documentation (and I think milkymist has second-best documentation for open source SoCs after GRLIB), I have little time to take care of it...16:02
lekernelnote that this doesn't mean i'm against switching to sphinx, just that it's not high on my priority list16:08
kristianpauloh, sure, i just asking if you like this platform or have suguestions for other?16:10
kristianpaulmigration could be done even slowly, but wiki could be a good point for decent manuals, also pdf is not easilly updated within the time16:11
kristianpaulwiki could not*16:11
lekernelwell, at first sight it looks better than doxygen16:12
Fallenoukristianpaul: it seems great16:17
Fallenoufor the moment we have a wiki, which is good too16:18
FallenouI think we can first focus on writing some valuable documentation on the wiki16:18
Fallenoubefore thinking about migrating stuff to another plateform16:18
Fallenoubut yes it looks nice :)16:18
Fallenouwiki pages looks nice too imo16:18
kristianpaullekernel: Are you aware of a way/program to convert verilog to a DOT graphiz output?16:28
lekernelno16:29
wpwrakkristianpaul: (doc) dunno. i find post-processing-friendly comments in source code hard to read.16:32
lekernelkristianpaul: also, DOT would be a mess with large and complex verilog files16:42
lekerneland maybe not even relevant16:42
lekernelit works great for LLHDL to show the tree like structures (which e.g. can help explaining/understanding how libmapkit plug-ins work) but for verilog...16:44
kristianpaullekernel: yeah in llhdl looks awesome16:57
kristianpaul(relevant) indeed, actually as the DOT ouput should be done after the sintesis process i think16:58
lekernelkristianpaul: sure, we can do that. that's what it does: http://milkymist.org/fpgatools/blinker.pdf17:44
lekernelas you can see, it doesn't work so well except for designs of trivial complexity17:45
kristianpaulpuff, what a mess17:46
Action: kristianpaul remenber when trying to undesrstand some self generated networks maps by ntop + graphiz17:47
kristianpaulnow i read with care sebastien tesis i'm please to read how a FFT process done by software on a PC is replaced by an accumulator :-). So the new zelands were not so crazy after all when they tought that for the gold code tracking18:12
kristianpaulFallenou: how big a ethernet frame is?20:50
kristianpaul#define ETHERNET_FRAME_LENGTH 153220:50
Fallenousomething like that20:57
Fallenou1500+20:57
Fallenouit's the max size btw20:58
Fallenouit can be less20:58
Fallenoumaybe 151820:59
Fallenouhttp://fr.wikipedia.org/wiki/Fichier:Ethernet_Type_II_Frame_format.svg20:59
Fallenouanyone here is richard sharpe ?22:10
kristianpaulFallenou: is he, from the oshg?22:24
kristianpaulthe question about ddr2 should be in the FAW22:24
kristianpaulFAQ*22:24
Fallenoudon't think so22:25
Fallenouit's just to ask him if he is porting milkymist on genesys22:25
Fallenouor what he' gonna do22:25
kristianpauli see22:25
kristianpaulFallenou: you made the driver for the softcore usb, or whatever that is, or was sebastien?22:33
Fallenousebastien did22:33
kristianpauli dont remenber well, in that time your logs were unknow for me22:33
kristianpaulk22:33
kristianpaulAnd is all in flicernoise repo? that was you're talking about last time?22:34
Fallenoui'm sorry what ?22:34
kristianpaulergg yes sorry22:35
kristianpaulwhere is that driver?22:35
Fallenouthe usb driver ?22:51
kristianpaulsoftub core..22:52
kristianpauli can dig np..22:52
kristianpauljust i dont used yet to rtem directory structure22:52
Fallenouc/src/lib/libbsp/lm32/shared/milkymist_usbinput22:52
Fallenouall milkymist drivers are in the same directory22:52
Fallenouin "shared"22:52
kristianpaulah ok22:52
Action: kristianpaul blind22:52
kristianpaul // TODO: move softusb firmware to the filesystem22:53
kristianpaulFallenou: what that means?22:53
Fallenoudunno22:53
kristianpaulah, i guess he mean the microcode22:54
Fallenouyes22:54
Fallenouthe code that runs on navre22:54
kristianpaulwich actually es load by cpu22:54
kristianpauland using some xilinx libraries i think..22:54
kristianpaulyeah TOOD22:54
--- Sun Mar 13 201100:00

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