| 01:25.04 | *** join/#edev GPSFan (n=kenm@12.10.255.248) |
| 01:25.05 | *** mode/#edev [+o GPSFan] by ChanServ |
| 01:27.35 | *** join/#edev dijenerate_ (n=dijenera@69.73.234.205) |
| 02:30.37 | *** join/#edev dijenerate (n=dijenera@69.73.216.217) |
| 03:29.47 | *** join/#edev nighty (n=nighty@sushi.rural-networks.com) |
| 03:40.30 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 04:55.14 | *** join/#edev LyosNorezel (n=LyosNore@75-95-6-177.day.clearwire-dns.net) |
| 05:27.53 | *** join/#edev lyakh (n=lyakh@dynamic-unidsl-85-197-26-24.westend.de) |
| 05:49.28 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 06:26.52 | *** join/#edev steliosk (n=Stelios@ipa226.211.tellas.gr) |
| 10:24.15 | *** join/#edev redblue (i=star@ppp230.108-253-207.mtl.mt.videotron.ca) |
| 11:21.22 | *** join/#edev dijenerate (n=dijenera@69.73.216.217) |
| 11:44.11 | *** join/#edev dijenerate_ (n=dijenera@69.73.252.220) |
| 12:41.09 | *** join/#edev steliosk (n=Stelios@ipa226.211.tellas.gr) |
| 12:55.22 | *** join/#edev GPSFan (n=kenm@12.10.255.248) |
| 12:55.22 | *** mode/#edev [+o GPSFan] by ChanServ |
| 13:20.27 | *** join/#edev mallum (n=mallum@host86-138-105-250.range86-138.btcentralplus.com) |
| 13:21.16 | *** join/#edev silverpower (n=shrike@atheme/bofh/silverpower) |
| 13:33.09 | *** join/#edev prpplague (n=billybob@mail.americanmicrosystems.com) |
| 13:36.28 | prpplague | i miss anything good yesterday? |
| 13:47.32 | *** join/#edev ekiller200 (n=Eric@mail.newport-ent.com) |
| 13:48.25 | *** join/#edev dijenerate (n=dijenera@69.73.252.220) |
| 14:01.35 | cbrake | prpplague: apparently not :-) |
| 14:04.29 | wmat | heh |
| 14:08.37 | prpplague | cbrake: hehe |
| 14:22.54 | chouimat | morning |
| 14:29.00 | *** join/#edev keesj_ (n=keesj@ip49-193-210-87.adsl2.versatel.nl) |
| 14:40.58 | *** join/#edev CosmicPenguin (n=nobody@crank.laptop.org) |
| 14:42.08 | prpplague | CosmicPenguin: boo |
| 14:42.39 | CosmicPenguin | yeah |
| 14:45.00 | *** mode/#edev [+o prpplague] by ChanServ |
| 14:49.56 | prpplague | CosmicPenguin: hmm, i seem to be getting 3 interrupts when what i should expect is 1 |
| 14:50.53 | CosmicPenguin | Sucks to be you |
| 14:50.59 | prpplague | CosmicPenguin: after the first interrupt i disable the imask and clear the status bits, but i then receive two more interrupts |
| 14:51.03 | prpplague | CosmicPenguin: every day |
| 14:51.06 | CosmicPenguin | heh |
| 14:51.19 | CosmicPenguin | you get the interrupts with the mask cleared? |
| 14:51.30 | prpplague | yea |
| 14:51.31 | CosmicPenguin | are you failing to clare something within your processor? |
| 14:51.59 | prpplague | CosmicPenguin: that was my first thought, but i've double check several times that the status are cleared |
| 14:52.46 | CosmicPenguin | that is odd |
| 14:53.02 | chouimat | hi CosmicPenguin prpplague |
| 14:53.06 | CosmicPenguin | And you are masking it in the processor, and it still comes? |
| 14:55.28 | prpplague | CosmicPenguin: yea |
| 14:55.43 | CosmicPenguin | thats not right |
| 14:55.54 | CosmicPenguin | that sounds like kernel issues |
| 14:56.37 | prpplague | CosmicPenguin: the only thing i can think of is there might be a timer that is kicking off the interrupt |
| 14:57.10 | CosmicPenguin | interrupt handler you mean? |
| 14:57.31 | prpplague | yea |
| 14:58.14 | CosmicPenguin | it wouldn't do that unless you told it to, though |
| 14:58.40 | prpplague | CosmicPenguin: thats the thing, this is marvell code, not all mine |
| 14:58.58 | prpplague | CosmicPenguin: so it is possible they could have that called somewhere |
| 14:59.01 | CosmicPenguin | ahh - therein laies the rub |
| 14:59.04 | CosmicPenguin | lies even |
| 15:04.02 | *** join/#edev lyakh (n=lyakh@dynamic-unidsl-85-197-26-24.westend.de) |
| 15:07.59 | prpplague | CosmicPenguin: hmm, i checked, doesn't look like a timer is set anywhere |
| 15:08.06 | prpplague | CosmicPenguin: mystery |
| 15:11.32 | CosmicPenguin | kprobe it |
| 15:14.21 | prpplague | interrupt IMASK=0x10000 CSTA=0xa34 DSTA=0x0 |
| 15:14.21 | prpplague | interrupt IMASK=0x0 CSTA=0x34 DSTA=0x0 |
| 15:14.59 | CosmicPenguin | whats your ret? |
| 15:15.09 | prpplague | after the first irq, the mask is set to zero |
| 15:15.14 | CosmicPenguin | er - return address |
| 15:15.19 | CosmicPenguin | sorry - you're not an x86er |
| 15:19.28 | prpplague | wonder if i should just disable the irq after |
| 15:27.53 | CosmicPenguin | ugh, I think the http pull for linus is broken |
| 15:40.44 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 16:14.29 | MonMotha | ok, so who the hell is responsible for the pinout of the SAM7X256 from Atmel? |
| 16:14.42 | MonMotha | they need a lesson in making things nice for the layout guys |
| 16:15.29 | prpplague | MonMotha: hehe |
| 16:17.20 | MonMotha | prpplague: yeah, doing an unusually dense board and it's causing problems |
| 16:17.44 | MonMotha | to get a feel for the density, if it weren't for the fact that there are some through-hole components on the top, you probably wouldn't be able to tell which is the top and which is the bottom of the board |
| 16:18.04 | prpplague | MonMotha: lovely |
| 16:18.07 | MonMotha | yeah |
| 16:18.10 | MonMotha | and I'm cramming it into 4 layers |
| 16:18.22 | prpplague | lovely indeed |
| 16:18.36 | MonMotha | my autorouter gives up with about 75% of the connections routed. I've got about 85% routed now manually |
| 16:19.03 | prpplague | MonMotha: yea, i usually set the autorouter for 50% |
| 16:19.09 | prpplague | MonMotha: do the rest by hand |
| 16:19.44 | MonMotha | this is a reasonably good autorouter, but the component density on both sides means it can't get good layer directions going |
| 16:21.14 | MonMotha | also, do you know a good source for LCDs ~10"? |
| 16:22.37 | *** join/#edev ccjoe (n=ccjoe@c-67-166-124-229.hsd1.co.comcast.net) |
| 16:34.08 | prpplague | MonMotha: what size and capabilities? |
| 16:34.48 | MonMotha | prpplague: case is 4x4, board is 3.75x3.75" |
| 16:35.20 | MonMotha | ethernet, zigbee, usb slave, RS-485, CAN, 24V AC/DC power input, and several inputs/outputs designed for use in HVAC systems |
| 16:35.27 | CosmicPenguin | ohh - sezy |
| 16:35.32 | MonMotha | oh, you mean the lCD :) |
| 16:35.42 | MonMotha | LCD would be ~10", color, STN is fine |
| 16:36.09 | MonMotha | 4:3 aspect, 10" diagonal |
| 16:37.12 | prpplague | what kind of $$ you need? |
| 16:37.19 | prpplague | and qty? |
| 16:38.04 | MonMotha | $$ is not a primary concern, and a few hundred (<1000) |
| 16:38.51 | prpplague | 10" is right on the big $$ line |
| 16:39.03 | prpplague | 7.8" do? |
| 16:39.16 | MonMotha | 7.8 would probably suffice |
| 16:39.36 | prpplague | http://store.earthlcd.com/EDMGRB8KJF |
| 16:39.44 | prpplague | same lcd as in the tuxscreen |
| 16:39.54 | prpplague | has pretty good clarity and well documented |
| 16:40.16 | MonMotha | gotcha |
| 16:40.51 | MonMotha | thanks |
| 16:40.53 | MonMotha | I need to get back to work |
| 16:40.59 | MonMotha | I might hop on from there |
| 16:41.10 | prpplague | MonMotha: okie dokie, later |
| 16:42.25 | *** join/#edev redblue (i=star@ppp053.108-253-207.mtl.mt.videotron.ca) |
| 16:53.22 | ekiller200 | I need an embedded system with 3 NICs, does anyone make an ARM that can support 3 PHYs, any suggestions would be great. |
| 17:05.09 | Exiles | I've seen mips systems with multi nics |
| 17:05.16 | *** join/#edev lyakh (n=lyakh@dynamic-unidsl-85-197-26-24.westend.de) |
| 17:06.13 | Exiles | I'm using one atm with two phy's and I can split one of the phy's into multiple interfaces |
| 17:09.03 | steliosk | ekiller200 : The is a mips one with 5 phy's RouterBOARD 150 (miniROUTER) |
| 17:09.20 | steliosk | ekiller200 : http://www.routerboard.com/pdf/mrugA.pdf |
| 17:21.19 | prpplague | ekiller200: generally speaking, most MIPS SoC's are network targeted |
| 17:21.27 | prpplague | ekiller200: arm is more portable/wireless |
| 17:21.47 | prpplague | ekiller200: so you don't see too many arm SoC's with wired ethernet support |
| 17:22.22 | prpplague | ekiller200: so i'd say you'd have better luck getting a MIPS or x86 based system to do what you want |
| 17:23.08 | Exiles | I have a samsung ltn121xj-l02 LCD screen sitting here |
| 17:23.39 | Exiles | 12.1 in |
| 17:26.01 | prpplague | ekiller200: http://www.acrosser.com/product/Networking%20applicance/Index1.htm |
| 17:26.21 | prpplague | ekiller200: acrosser has nice selection of devices |
| 17:26.54 | prpplague | ekiller200: i've used this one before - http://www.acrosser.com/Product/Networking%20applicance/Wiressless-W-Series/Firewall_eden_m9922.html |
| 17:33.10 | prpplague | CosmicPenguin: still haven't found the reason for the multiple interrupts |
| 17:38.56 | *** join/#edev lyakh (n=lyakh@dynamic-unidsl-85-197-26-24.westend.de) |
| 17:47.00 | *** join/#edev SpaceMonkey (n=andrew@client-82-2-143-77.manc.adsl.virgin.net) |
| 17:53.06 | steliosk | prpplague : Is acrosser making the boards/appliances or just selling them ? |
| 17:53.24 | prpplague | steliosk: iirc they make them in china |
| 17:55.11 | steliosk | prpplague : They look interesting. I am also in the market for a 3/5 ethernet port appliance (1u) |
| 17:55.47 | prpplague | steliosk: they generally are nice x86 embedded boards |
| 17:56.03 | prpplague | steliosk: my experience with them was good |
| 17:56.48 | steliosk | prpplague : Do they have a minimum for an order ? probably will try one or two devices before we order |
| 17:57.05 | prpplague | steliosk: no i don't believe they do |
| 17:57.14 | prpplague | steliosk: or atleast i don't recall that being an issue |
| 17:57.27 | prpplague | steliosk: it has been about 3.5 years since i did business with them |
| 18:03.09 | steliosk | prpplague : I am talking to these guys -> http://win-ent.com/ |
| 18:03.38 | steliosk | prpplague : I am waiting for a price quote from their European agent |
| 18:03.53 | prpplague | steliosk: the thing i liked about acrosser is they have some really nice boxes |
| 18:08.05 | steliosk | prpplague : Yeah, this is what *a lot * of the customers uses to "judge" if a product is good |
| 18:09.19 | prpplague | steliosk: well, i don't use that as the only qualification, it just makes it easier to use |
| 18:10.24 | Exiles | the router I use atm can do 5 ports sepratly, or as a single switch |
| 18:11.13 | steliosk | Exiles : Does it use a switch chip or seperate phys ? |
| 18:11.19 | Exiles | switch chip |
| 18:11.58 | Exiles | so all depends if you need highspeed or not |
| 18:12.11 | Exiles | it already has linux on it |
| 18:13.11 | steliosk | Exiles : switches are good if you need vlans setup and they route traffic within them, without having the cpu bother much |
| 18:13.37 | Exiles | what are you needing to do with the device your wanting to build? |
| 18:14.31 | Exiles | I can seperate the idividual ports of this switch out, by default one ports pulled out to create the wan port |
| 18:16.36 | Exiles | but you can get the at low cost |
| 18:20.02 | Exiles | so if it dosn't need alot of bandwidth, might be worth looking into |
| 18:28.12 | *** join/#edev prpplague (n=billybob@mail.americanmicrosystems.com) |
| 18:29.41 | *** join/#edev mallum (n=mallum@host86-141-46-90.range86-141.btcentralplus.com) |
| 18:34.08 | steliosk | Exiles : We are building a voip pbx |
| 18:34.41 | Exiles | oh, you want a x86 machine then |
| 18:34.56 | steliosk | Exiles : no :) |
| 18:35.32 | Exiles | gonna use hardware to take the phone calls? |
| 18:36.11 | steliosk | Exiles : yes. can't say much but its connected to ethernet so don't need pci etc |
| 18:36.33 | steliosk | Exiles : just additional ethernet ports |
| 18:36.38 | Exiles | I tried a pbx, it needed alot of cpu power to take phone calls |
| 18:36.49 | Exiles | I used digium hardware |
| 18:37.32 | steliosk | you need cpu for the echo cancellation and/or codec translation |
| 18:38.58 | steliosk | on a powerpc @266Mhz i can run 10 calls with alaw/ulaw/gsm codecs or 3 g729/speex |
| 18:39.11 | Exiles | intresting |
| 18:39.31 | Exiles | thats from a analog phoneline, or from a voip line? |
| 18:39.31 | steliosk | g729/speex are fixed point optimized for ppc |
| 18:40.32 | Exiles | nice |
| 18:40.33 | steliosk | voip. For the analogue we handle echo cancelation outside the main pbx |
| 18:40.45 | Exiles | ah ok |
| 18:40.52 | Exiles | voip makes it simpler |
| 18:41.20 | Exiles | I had to deal with analog |
| 18:41.36 | steliosk | well its not that difficult with the analogue part either, but you have to make choices |
| 18:42.04 | steliosk | in general the Digium code is not optimized for embedded stuff |
| 18:43.04 | steliosk | so if you take the code and just recompiled it don't expect much in term of performance |
| 18:45.35 | prpplague | i just can't figure out whats trippin this irq |
| 18:47.40 | Exiles | what irq's do you have? |
| 18:49.31 | prpplague | Exiles: its a sdio controller interrupt |
| 18:50.13 | Exiles | ah, did you look over the drivers? |
| 18:53.01 | Exiles | sdio I know has both a driver for the interface, then the device drivers |
| 18:54.16 | prpplague | Exiles: yea |
| 18:55.21 | prpplague | Exiles: this is on the low level interface |
| 18:55.36 | prpplague | Exiles: getting some interrupts even when the interrupt mask is set to 0 |
| 18:55.45 | Exiles | is it possible the card is spitting out something? |
| 18:55.48 | prpplague | <PROTECTED> |
| 18:56.05 | prpplague | Exiles: the imask is suppose to ignore all interrupts |
| 18:56.11 | Exiles | yeah |
| 18:56.50 | Exiles | maybe stick a basic interupt handler in your program to clear them? |
| 18:57.05 | ds2 | add a NAND gate + GPIO pin =) |
| 18:57.10 | prpplague | ~lart Exiles |
| 18:57.10 | purl | acting on orders from an unspecified client drags Exiles into court suing for $200 million |
| 18:57.31 | prpplague | Exiles: this is the kernel module i'm working on |
| 18:57.37 | Exiles | I'm just trying to help |
| 18:57.42 | prpplague | Exiles: hehe |
| 18:58.09 | Exiles | I admit I've never worked with building in linux before |
| 18:58.19 | Exiles | morning ds2 |
| 18:58.26 | ds2 | morning Exiles |
| 19:04.49 | *** join/#edev purl_ (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 19:04.50 | *** topic/#edev is a channel for the discussion of hacking small consumer electronic devices, mostly ARM based |
| 19:05.05 | ds2 | hmmm so it is better to pay the CA-US bridge tolls with USD thesedays |
| 19:09.05 | chouimat|busy | ds2: I guess we will have 1 CAD = 1USD before the end of 2007 |
| 19:13.55 | ds2 | level or edge? |
| 19:14.28 | chouimat|busy | prpplague: probably a few gremlins ;) |
| 19:15.12 | prpplague | chouimat|busy: or probably bugger sdio controller |
| 19:15.15 | prpplague | ds2: neither |
| 19:15.17 | prpplague | ds2: SoC |
| 19:15.26 | ds2 | ack |
| 19:16.08 | ds2 | checked for an errata saying interrupt masking is broken on all revs? ;) |
| 19:17.20 | prpplague | ds2: yea, problem is, samsung doesn't post alot of info on their SDI controller |
| 19:17.39 | ds2 | that sucks |
| 19:19.57 | *** join/#edev FireEgl (i=FireEgl@Sebastian.Atlantica.CJB.Net) |
| 19:20.01 | CosmicPenguin | How can it be a buggy sdio controller if you are masking in the processor and they're still coming? |
| 19:20.14 | CosmicPenguin | Thats a buggy processor if anything |
| 19:22.31 | prpplague | CosmicPenguin: sdi is part of the SoC |
| 19:22.35 | prpplague | CosmicPenguin: intergrated |
| 19:23.52 | CosmicPenguin | ahhh - evil |
| 19:25.20 | prpplague | CosmicPenguin: when i mask for one interrupt, i get three |
| 19:25.26 | chouimat|busy | WOW |
| 19:25.27 | prpplague | CosmicPenguin: with no aparent source |
| 19:25.33 | CosmicPenguin | ahh - the status isn't set? |
| 19:26.05 | prpplague | CosmicPenguin: right, first interrupt shows to be correct, but i get 2 additional interrupts with no apparent source |
| 19:26.31 | ds2 | if they are finite,why not just ignore them as spurious? |
| 19:27.47 | CosmicPenguin | Strange though |
| 19:28.00 | CosmicPenguin | becuase spurious interrupts will be handled at the interrupt controller level |
| 19:28.26 | CosmicPenguin | If there is no status bit for you, then how did the IRQ handler know to send it to you? |
| 19:29.02 | ds2 | ah good point... thinking of x86 where you get an IRQ # that can be exclusive to you |
| 19:29.39 | prpplague | CosmicPenguin: i'm working directly with the low level irq handler for controller |
| 19:30.07 | CosmicPenguin | so you're not using the standard mechanism> |
| 19:30.42 | prpplague | CosmicPenguin: as far as i know, i am using the standard irq mec |
| 19:30.47 | prpplague | hanism |
| 19:31.26 | CosmicPenguin | Okay - we're talking past each other |
| 19:31.53 | CosmicPenguin | I'm not for sure on ARM, but if its anything like x86, then there will be a master interrupt controller routine that will get called when the IRQ pin (or whatever) is asserted |
| 19:32.12 | CosmicPenguin | It will check a status register, and dispatch interrupts to various controllers based on the value of that register |
| 19:32.27 | ds2 | don't think ARM has a interrupt # that gets sent over the bus like x86 |
| 19:33.18 | CosmicPenguin | But there is a status register somewhere |
| 19:33.20 | prpplague | CosmicPenguin: yea, there is a register that tells which irq was trigger, but its up to the int_handler to determine what aspect of the irq was trigger |
| 19:33.29 | CosmicPenguin | stop concentrating on the how, and start thinking about the what |
| 19:33.46 | CosmicPenguin | prpplague: so if your status bit isn't being set - then how is it that your handler is being called? |
| 19:34.17 | prpplague | CosmicPenguin: the "master" irq handler doesn't look at the status, only that the irq was trigger |
| 19:34.21 | prpplague | triggered |
| 19:34.37 | CosmicPenguin | Then how does it know which interrupt handler to call? |
| 19:34.53 | prpplague | based on the irq number that was triggered |
| 19:35.31 | CosmicPenguin | So you are saying that the interrupt is indeed being trigged? |
| 19:35.54 | prpplague | correct, that atleast the SoC interrupt controller thinks that it is being triggered |
| 19:36.17 | CosmicPenguin | edge or level? |
| 19:36.48 | prpplague | CosmicPenguin: doesn't really work that way, its all internal, its either on or off |
| 19:37.17 | CosmicPenguin | fair enough |
| 19:41.37 | prpplague | CosmicPenguin: basically its suppose to go like this: main irq controller gets an interrupt, it checks its internal registers to determine which irq was trigger, passed that irq number to the "master handler", which passes it to the appropriate driver irq handler, the "local" irq handler then checks the status registers to determine what trigger the interrupt in the first place |
| 19:42.20 | prpplague | CosmicPenguin: i guess in x86 terms it would more like a "shared irq" |
| 19:42.49 | CosmicPenguin | sure |
| 19:42.55 | CosmicPenguin | everybody acts like that |
| 19:43.02 | CosmicPenguin | okay - well, then ds2's suggestion is very accurate |
| 19:43.09 | CosmicPenguin | if you know the last few are spurious, ignore em |
| 19:44.14 | prpplague | CosmicPenguin: yea, thats what i've been doing, it's just that it makes me think i have coded something wrong |
| 19:46.57 | prpplague | CosmicPenguin: i'm just wondering that if the interrupt isn't serviced fast enough, that it might get additional interrupts before its actually serviced |
| 19:48.07 | *** join/#edev pierrelux (n=pierre-l@144-125.sh.cgocable.ca) |
| 19:53.49 | CosmicPenguin | ew... |
| 19:54.50 | Exiles | I've seen that crash systems before |
| 19:57.28 | CosmicPenguin | I guess clear the status as the first thing you do in the handler |
| 19:58.21 | prpplague | CosmicPenguin: yea, already tried that |
| 19:58.31 | prpplague | CosmicPenguin: didn't seem to make a different |
| 19:58.46 | prpplague | CosmicPenguin: i'm thinking i need to do a spin_lock_irqsave() |
| 19:58.47 | CosmicPenguin | This is why SOCs are a bitch |
| 19:59.05 | CosmicPenguin | if something breaks, there's no where you can get a scope on it |
| 19:59.10 | prpplague | CosmicPenguin: hehe yea |
| 20:06.02 | prpplague | CosmicPenguin: hmm, the mmc driver for the s3c2410 does use spin_lock_irqsave() |
| 20:14.51 | prpplague | CosmicPenguin: hmm, that didn't seem to work either |
| 21:58.44 | *** join/#edev lyakh (n=lyakh@dynamic-unidsl-85-197-26-24.westend.de) |
| 22:12.15 | *** join/#edev LyosNorezel (n=LyosNore@unaffiliated/lyosnorezel) |
| 22:27.37 | *** join/#edev GPSFan (n=kenm@12.10.255.248) |
| 22:27.37 | *** mode/#edev [+o GPSFan] by ChanServ |
| 22:44.29 | *** join/#edev redblue (i=star@ppp054.109-253-207.mtl.mt.videotron.ca) |
| 23:15.37 | *** join/#edev FireEgl (i=FireEgl@Sebastian.Atlantica.IRCNut.Com) |
| 23:18.15 | *** join/#edev joe_bleau (n=joe_blea@adsl-67-66-232-230.dsl.fyvlar.swbell.net) |
| 23:30.51 | joe_bleau | Anyone playing with the AT91SAM9263? |