IRC logs for #openrisc Tuesday, 2015-04-14

--- Log opened Tue Apr 14 00:00:06 2015
-!- Netsplit *.net <-> *.split quits: knz, martinboehnert1, Amadiro03:06
-!- Netsplit over, joins: martinboehnert1, knz, Amadiro03:13
HeshamAre these 4 asm instructions enough to light up the LEDS on Atlys board (regardless of setting the direction regs)?
bandvigHesham: I think, you should configure PIO as outputs by writing FF into 0x91000001 before playing with ON/OFF LEDs.08:32
Heshambandvig: Yeah I did, that's why I said (regardless of setting the direction regs).08:33
HeshamEven though, they produce no output when dumping them to rom.v instead of the bootloader.08:34
bandvigHesham: My ROM-code with round switching LEDs could be found in 1st comment here It worked after correction ROM address size in wb-interconnect.08:36
HeshamThanks, I made this correction according to your help before ;)08:39
_franck__Hesham: I also have a led blinker asm code here:
HeshamPerfect, let me try again.08:40
HeshamSeems like it will be the only way to debug on my Atlys board08:42
-!- Netsplit *.net <-> *.split quits: bentley`10:02
HeshamGreat. RTEMS is finally alive and working on real hardware now (Atlys) :)12:38
poke53281I don't have a clue, what I should do with rtems.12:47
poke53281But it's running :)12:48
Heshampoke53281: Yeah, the current RTEMS can run on or1ksim, QEMU and jor1k, but not on real hardware.12:49
HeshamMainly because simulators don't care about UART init code.12:50
HeshamThere was a bug at the RTEMS UART driver, which I am submitting a patch for now.12:50
olofkHesham: Great news!12:54
olofkAre you still using the complicated way to download software by embedding it in a u-boot image?12:54
poke53281@Hesham: Do you have a good demo, which can be run in rtems?12:54
Heshamolofk: Yes for now, now I can hack around and upload some data via UART from RTEMS, which can be useful for downloading software later.12:58
HeshamBut now, I am not using u-boot12:59
HeshamI replaced u-boot with RTEMS image (i.e. ticker.exe)12:59
Heshampoke53281: on hardware?12:59
HeshamI wrote a tutorial how to build/run RTEMS on or1ksim and QEMU recently13:00
HeshamThat's it
poke53281@Hesham. I mean something like a game, something like bash, a video player, ....13:02
Heshampoke53281: The jor1k RTEMS demo you provided a link for is bash-like app13:04
poke53281@Hesham: Nothing more fancy?13:55
HeshamI don't think so, it's a real-time OS. It's great you can see some output :)13:56
poke53281No comparison between a car run by rtems vs. a car run by windows?14:08
poke53281A car, which would stall for 5 seconds in windows, when doing a full brake,14:08
poke53281because the screen update routine cannot show the braking symbol, because it14:08
poke53281has to wait for an update of the network filesystem in the explorer?14:08
HeshamThat's why they are using RTEMS instead of Windows on space missions ;)14:12
HeshamAs I expected, the problem was with the USB-UART Linux driver from Exar!15:45
HeshamIt wasn't working properly, that's why I couldn't send commands to u-boot.15:45
HeshamThey released a new update early this year and it's working now.15:46
HeshamNext step is to send some binaries via USB-UART. olofk, any hints how to send data to a ttyUSB device instead of minicom, putty, etc?15:52
sb0we also found shittons of bugs in RTEMS16:31
Heshamsb0: Which hardware platform are you using? or1k?16:32
sb0for rtems? lm32. but the bugs I'm talking about were in platform independent code.16:33
HeshamI see, of course you can file tickets with any bugs you find there.16:34
sb0we filed them, sent patches even.16:37
sb0at least those bugs are gone now16:38
sb0but there were pretty serious ones, e.g. races due to the ISRs when using message queues16:38
Heshamolofk: Have you come across the riscv HTIF interface? I think it can serve as general purpose host/target interface for debugging, loading programs, console output, etc19:14
HeshamIt requires addition to the hardware19:14
HeshamBut it can work with either Ethernet and/or rs23219:16
olofkWishbone async ack got me confused again. Do I need to do anything about them on the master side?20:58
olofkI found a bug in my BFM that was caused by earlier confusion when I thought that (Classic cycles) === (cti == 000) === (async ack)21:01
stekernno, you don't need to do anything special on the master side21:30
--- Log closed Wed Apr 15 00:00:07 2015

Generated by 2.15.2 by Marius Gedminas - find it at!