#milkymist IRC log for Sunday, 2011-02-13

lekernellarsc: how is your linux port going btw? any progress recently?16:27
lekernelrejon: do you feel like submitting a gsoc application this year?16:30
lekernel(there have been many disappointing things last year and i'd not do it)16:35
wpwraklekernel: (gsoc) similar experience on my side from some years ago. i think it would be more useful if google just handed out some money to pay people who are already active in the project :)17:51
larsclekernel: nothing new. it still works more or less. I'm waiting for mwalle uclibc port, so we can overhaul syscalls and signal handling18:45
rejonlekernel yes, lets do it20:43
rejonthey have some new people running it20:43
Fallenourejon: on what are you porting linux ?20:47
rejonhuh?20:47
Fallenouoh it's larsc sorry20:47
kristianpaulFallenou: hi20:49
Fallenouhi kristianpaul  !20:49
kristianpaulFallenou: Do you have some idea about the Minimac RX FIFO overflow ramdon bug?20:50
kristianpaulAs you developed the driver i want to read your comments about it :-)20:50
Fallenouno I really don't20:50
Fallenoubut i am looking for a scenario to reproduce the bug20:51
FallenouIs the bug present using qemu ?20:51
Fallenoubecause I don't have a board20:51
lekernelno, it probably won't be present with qemu20:51
Fallenouarg20:51
Fallenouhow would you process to track the bug ? can kristianpaul send me informations that could help finding out what's going on ?20:53
kristianpaulyday i experienced the bug when doing a _ping_ nothing hangup but every time i pinged the board i got the Minimac RX FIFO overflow20:53
Fallenouwo20:53
Fallenouthe driver has serious issue then =(20:53
kristianpaulthis was just before configure minimac0 networking using ifconfig ....20:53
kristianpaulI must do more test and send it back to you, yes20:53
Fallenousure sure send me your test scenario and test results20:54
Fallenouand if you find a way to reproduce the bug at 100% chance, please gimme the exact scenario :)20:54
Fallenouit would really help a lot20:54
Fallenoubecause searching randomly in the source code (even if the code of the driver is not that big) is not an easy task20:54
kristianpaul100% chance i doubt it... it something happes just after powerup..20:54
Fallenouhumm ok :x20:55
Fallenouanyway, send me all the information you have about it please20:55
Fallenou(send it to the ML actually)20:55
kristianpaulyeah sure20:55
Fallenouso that everyone can read it20:55
kristianpaulah i forgot said, i never got ping reply21:01
Fallenouhummm i think i did, but not sore21:02
Fallenousure*21:02
Fallenoudon't remember21:02
Fallenoui tried ARP, TCP21:03
Fallenoudon't remember if i tried icmp21:03
lekerneltry with large frames too21:04
lekernelthose seem rather prone to triggering bugs in the ethernet drive21:04
Fallenouactually i tried with TFTP21:04
Fallenouand I transfered like 20 MB21:04
lekernelbut to be honest, I believe large parts of the ethernet driver needs to be cleaned up and rewritten21:04
kristianpaulFallenou: tftp is not very smart, i just burst data from thr board to the other end i think21:05
kristianpaulFallenou: i wonder if you never tried the oposite?21:05
Fallenoudo you think we can keep the RTEMS_EVENT principle lekernel ?21:05
Fallenouor rewrite everything using another principle (i don't know what)21:06
Fallenoukristianpaul: actually it was from the host computer to the board21:06
kristianpaulFallenou: oh21:06
Fallenouand the board was printing the data on serial output21:06
Fallenouyou can find the source code i was using and do the test yourself21:06
kristianpaulwait no, i was talking about the 20 MB transfer21:07
Fallenous/was using/made/21:07
Fallenouyes so do I21:07
Fallenoucheckout this https://github.com/fallen/network-demos21:07
lekernelfirst, for the tenth time or so, use the stock crc32 for fuck's sake21:07
Fallenouand use the tftpclient21:07
lekernelthen don't mix softc and global variables, this just doens't make sense (I told you that several times, too)21:08
Fallenoulekernel: yes it would be better, but the last time I tried it didn't work21:09
lekernelonce those problems are fixed, to avoid overflows, i'd keep a pool of buffers and refill the ethernet core in the interrupt handler21:10
Fallenouhumm ok instead of delaying it21:10
Fallenoudoing the change of dma address on the top half21:11
lekernelyes21:11
Fallenouthat should improve network performance yes21:11
lekernelthe crc32 is an easy problem to track down, you can run both functions (working and supposedly non-working) side by side in a x86 C program on your computer21:11
lekerneli'd be very surprised if there would be no stock crc32 suitable for ethernet in rtems21:12
Fallenouthere are two actually, one for big endians and one for little endians21:14
Fallenoumaybe I did it wrong while testing21:14
Fallenoubut both of them failed :/21:14
lekerneleverything is big endian there21:14
larscFallenou: for the milkymist21:39
Fallenouoh nice21:39
Fallenouare you using the original port of Lattice ? or the modified one by the japanese guy ?21:40
larscits based on takeshis tree21:41
Fallenouyes ok21:41
larscit basically rewrote most of the code from lattice/theobroma because it was either broken or outdated21:42
larscso now it is quite stable21:43
Fallenoudo you plan on making the shared library and relocation stuff working ?21:43
Fallenouoh good :)21:43
larscmwalle has been working on uclibc and stuff21:44
Action: lekernel just discovered that C support complex numbers22:07
Fallenouhehe22:11
FallenouI was surprised too when I discovered that22:11
wpwrak(weird C features) one of the more galling features of C is that math.h provides y0, y1, and yn ...22:18
kristianpaulok tftpclient worked well it seems..22:44
kristianpaullets see tftpd :-)22:44
Fallenoutftpd ?22:49
Fallenouon milkymist ?22:49
kristianpaulyes22:50
Fallenouit's not in netwok-demos is it ?22:50
kristianpaultftpTest22:50
kristianpaulit is22:50
Fallenouoh ok22:51
Fallenoutftpclient is something i wrote because i never got tftpTest to work :)22:51
kristianpaulBut i guess i requires a working fs22:51
Fallenouso if you do, tell me :p22:51
Fallenouyes22:51
Fallenouit is using a tftp file system22:51
kristianpauls/i/it22:51
Fallenoui never managed to get it working22:51
Fallenouit mounts the tftp directory in the filesystem22:52
Fallenouit's kind of weird22:52
Fallenouto test the network driver i had to write the tftpclient22:52
kristianpaulwhy you dont used  ttcp to test? it inmeadiatly fire the overflow bug ! ;)22:54
Fallenouoh !22:55
Fallenounever tried22:55
Fallenoucan you repeat several times ?22:55
Fallenouand confirm it triggers the problem ?22:55
Fallenouthanks !22:55
Action: kristianpaul dint said yes yet22:55
kristianpaulok22:56
kristianpaulFallenou: you already know that ifconfig trigger the problem, why no start from there?22:56
kristianpaulThere are more that one trgger for sure22:56
Fallenouif i have several things to test it's better22:57
Fallenouand i think ifconfig may be another bug22:57
kristianpaulhmm, ok22:57
Fallenoubecause it's obviously not a full fifo problem22:57
kristianpaulyeah, that call  my atention with ifconfig (fifo)22:57
--- Mon Feb 14 201100:00

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