| 00:25.47 | *** join/#edev ce_geek (n=ce_geek@roger-blanchette-4.umelst.maine.edu) |
| 00:26.15 | ce_geek | Is anyone around for some linux kernel device driver discussion? |
| 00:35.42 | ce_geek | I'll take silence as a "no." :) |
| 00:37.42 | cajaka | Ask away you never know who reads there back logs |
| 00:37.58 | ce_geek | True |
| 00:39.20 | ce_geek | I'm just curious about writing time-sensitive software on the Hammer board. I wonder what sort of issues one might run into writing code to toggle GPIO pins to do things like control stepper motors. Let's say you have your program making the motor go round and round, and all of a sudden someone access the web server running on the Hammer board. Does the motor movement get interrupted? |
| 00:39.49 | ce_geek | How difficult is it to guarantee that something will happen with a given frequency? |
| 00:39.51 | ds2 | in userland w/o the preempt patches, yep |
| 00:40.00 | ce_geek | preempt patches? |
| 00:40.30 | ds2 | yeah, it might be integrated into the current mainline now but for 2.6.21 I think there are still some patches - it gives you the ability to give processes hard realtime |
| 00:41.07 | ds2 | in the kernel, you are a little better with a timer but there could be other code that would slow things down (latency issues) |
| 00:41.19 | ce_geek | Oh cool. So, with some patches, in userland, it is possible to get hard realtime for a process? |
| 00:41.38 | ce_geek | Hrm okay |
| 00:41.43 | ce_geek | Can you read the timer in userspace? |
| 00:41.45 | ds2 | yes, BUT there is a limit to how real the real time is... i.e. don't expect 1 clock response |
| 00:41.57 | ce_geek | Right. |
| 00:42.21 | ds2 | how fast do you need this? there are timers readable from userland....i.e. the RTC's (don't think hammer comes with a RTC but they should be easy enough to emulate/add) |
| 00:43.26 | ce_geek | I'm more just asking in general. I got talking to a faculty member about the differences between controlling hardware on an embedded device running linux and code that isn't running in an operating system evironment, for example on a PIC chip. |
| 00:43.52 | ce_geek | Not much to worry about on the PIC chip. You tell it to delay a certain number of clock cycles and then step the motor, and it's gonna be dead nuts every single time. |
| 00:44.04 | ds2 | it all comes down to how fast you need things |
| 00:44.06 | ce_geek | But if you've got an operating system swapping processes in and out, it's not quite that simple. |
| 00:44.19 | ds2 | actually, no... even on a PIC, you can have jitter due to interrupts |
| 00:44.20 | ce_geek | Yeah. I mean, to drive a motor for a robot to roam around, that shouldn't be an issue. |
| 00:44.35 | ce_geek | Ah, good call. That's absolutely right. |
| 00:45.01 | ce_geek | Actually I've used timer overflow interrupts to control stuff like that now that I think back to some previous projects. |
| 00:45.10 | ds2 | the approach the EMC folks did it is with yet another different RT patch to the kernel and do most of the critical stuff in the kernel side |
| 00:45.22 | ce_geek | EMC? |
| 00:45.30 | ce_geek | Embedded MicroController? |
| 00:45.33 | ds2 | but for slower stuff, the other RT stuff should be fine |
| 00:45.51 | ds2 | forgot what it stands for but it is software to control machines |
| 00:45.56 | ds2 | i.e. a mill or a lathe |
| 00:46.01 | ce_geek | Oh okay. |
| 00:46.14 | ds2 | it was originally a NIST project |
| 00:46.21 | ce_geek | Not familiar with NIST. |
| 00:46.28 | ds2 | but they are more interested in high speed servos instead of just steppers |
| 00:46.39 | ds2 | NIST = National Institute of Standards and Time |
| 00:46.55 | ce_geek | I'd actually like to write some code for the Hammer board that will run without Linux, and then do the same thing with Linux running, just as some examples. |
| 00:46.56 | ds2 | but EMC is currently x86 only |
| 00:46.57 | ce_geek | Oh, very cool. |
| 00:47.01 | MonMotha | ce_geek: using timer interrupts is a standard practice |
| 00:47.20 | ce_geek | And the devices that controlled these high speed servos and such were running a real time operating system? |
| 00:47.21 | MonMotha | many of my timer based uC projects consist of a "while(true)" at the end of main and everything is handled from timer interrupts |
| 00:48.04 | ce_geek | MonMotha: Yeah, that makes sense. |
| 00:48.40 | MonMotha | assuming you don't have any timers that can fire on top of each other, it eliminates the timing jitter |
| 00:48.54 | MonMotha | you just have to make sure they don't want to fire on top of each other and/or the ISRs complete fast enough |
| 00:49.14 | ds2 | ce_geek: yep. it counts pulses from a rotary encoder as the feedback |
| 00:49.45 | ds2 | the PREEMPT_RT patches lets you preempt most ISRs so that can be worked around |
| 00:50.01 | ce_geek | Interesting. |
| 00:50.16 | ds2 | just becareful if you have PM enabled ;) |
| 00:50.25 | ce_geek | I'd like to get a better understanding of when a given task requires the use of an embedded device that can run a real time operating system, and when you can just "throw a PIC chip at it." |
| 00:50.33 | ce_geek | PM? |
| 00:51.02 | MonMotha | the answer is that you should never throw a PIC at it |
| 00:51.07 | ds2 | Power Management |
| 00:51.11 | ce_geek | Oh ;) |
| 00:51.12 | MonMotha | instead, consider any of a multitude of better 8-bit archs :) |
| 00:51.19 | ce_geek | HAHA. |
| 00:51.20 | ds2 | throw a FPGA at it ;) |
| 00:51.24 | MonMotha | that works, too |
| 00:51.30 | ds2 | softcore }:-) |
| 00:51.34 | MonMotha | I personally am a big fan of AVR, though |
| 00:51.40 | ce_geek | Okay, well, let's come up with a more generic term than PIC then. |
| 00:51.50 | MonMotha | 8-bit uC |
| 00:51.51 | ce_geek | We just use a lot of PICs here because they're free :) |
| 00:51.54 | ce_geek | Perfect. |
| 00:52.04 | MonMotha | you don't usually run an RTOS on those |
| 00:52.08 | MonMotha | (though you can...) |
| 00:52.11 | ce_geek | So where's the line where an 8-bit uC isn't acceptable? |
| 00:52.26 | MonMotha | um, pretty much when it isn't fast enough or the code doesn't fit :) |
| 00:52.46 | MonMotha | or you REALLY need to work on lots of >8-bit quantities |
| 00:53.07 | MonMotha | at that point, grab yourself a small ARM and go to town |
| 00:53.15 | MonMotha | you don't have to run an RTOS on those, but many do |
| 00:53.19 | ds2 | or if you also got a UI to deal with (network stack/webserver/etc) |
| 00:54.39 | ce_geek | I suppose those are sufficient criteria. |
| 00:56.26 | ds2 | make life fun...celebrate legacy and use a 8051 ;) |
| 00:56.37 | ce_geek | I used a PIC 16F877A to balance an inverted pendulum. There was a lot going on there. I had to monitor two encoders, one for pendulum angle and one for car position, and then set the duty cycle and direction for a DC motor with PWM control to keep the pendulum upright. I just wonder how I'd do that differently if I used the Arm9 with Linux. |
| 00:56.40 | MonMotha | hey, I like 8051 |
| 00:56.48 | MonMotha | for how old it is, it's remarkably useful (and surprisingly clean) |
| 00:56.57 | ce_geek | I'm studying the 68000 right now |
| 00:56.58 | ds2 | the SiLabs toolstick is really cool for that |
| 00:57.06 | ce_geek | SiLabs toolstick? |
| 00:57.16 | ds2 | look at the SiLabs site |
| 00:57.23 | ce_geek | Yeah lookin' now |
| 00:57.36 | ds2 | little USB programmer stick with a detachable daughterboard |
| 00:57.42 | ce_geek | Oh yeah, slick. |
| 00:57.47 | ds2 | no SMT soldering |
| 00:58.14 | ds2 | my only complaint is the toolstick takes over the serial port :/ |
| 01:04.48 | ce_geek | Cool. Well, thanks the information. I appreciate it. |
| 01:05.19 | ce_geek | In the next week or two I hope to get to doing some low-level programming with the Hammer board. I've got to finish this project with the 68000 first so I don't get super confused. |
| 01:20.43 | ds2 | who still sells a 68000 core based chip? |
| 01:25.17 | ce_geek | Pfft, beats me. |
| 01:25.32 | ce_geek | The whole reason I'm doing this hammer project is to try to get my professor to teach something current :) |
| 01:25.58 | ce_geek | our text book is actually pretty good. It does a good job of talking about microprocessor systems and such. |
| 01:27.02 | ds2 | what dept is this? |
| 01:27.11 | ce_geek | Electrical and Computer Engineering |
| 01:28.10 | ds2 | we had a cross listed class with the Aeronautics dept to do some PIC board to control model airplane servos... that was fun |
| 01:28.29 | ce_geek | Cool |
| 01:28.57 | ce_geek | I've been trying to think of a fun project to do with RC helicopters. |
| 01:29.11 | ds2 | is the hammer for a structured class (one where everything is spelled out on what you need to do and what is expected)? |
| 01:29.12 | ce_geek | But I'm not sure I want to do much in the way of flight control with a helicopter. |
| 01:29.55 | ce_geek | No, I've got a lot of flexibility. We just had to do "a project," so I decided to basically create some example lab assignments for an undergraduate class in microprocessors. |
| 01:30.08 | ce_geek | It's a graduate class I'm taking in Microprocessor Systems Design. |
| 01:30.38 | ds2 | I see... we had that sort of thing for undergrad -propose a project and have it implemented by the end of the term |
| 01:30.44 | ds2 | they were fun |
| 01:30.46 | ce_geek | Haha |
| 01:30.54 | ce_geek | Sounds like you're in academia? |
| 01:31.05 | ds2 | not anymore |
| 01:31.25 | ds2 | this was ages ago before the ARMs ruled |
| 01:31.40 | ce_geek | Well, the pendulum thing I described was my capstone project. I had a signed contract saying what the project would do, had to meet the contract by the end of the semester, etc. People did all sorts of stuff. A 3 semester sequence. |
| 01:32.07 | ce_geek | Graduated in '06 with my BS in Computer Engineering. |
| 01:32.11 | ce_geek | Graduating in May with my MS. |
| 01:33.06 | ce_geek | And I probably sound like a useless toolbag in this forum, but we don't have any embedded systems classes. I'm trying to help change that. I know the faculty here really well, and some fellows students and I are trying to get some changes made around here. |
| 01:33.54 | Mirell | That would probably be different if they just started teaching C in Intro to CS classes. |
| 01:34.16 | ds2 | what school doesn't teach C in intro to CS classes?! |
| 01:34.21 | ce_geek | Our computer science department is completely useless. |
| 01:34.27 | ce_geek | ds2: Our CS department. |
| 01:34.31 | ds2 | !!! |
| 01:34.31 | ce_geek | They learn Java now. |
| 01:34.35 | ds2 | ewwwwwwwwwwwwww |
| 01:34.37 | ce_geek | Yuh. |
| 01:34.55 | ce_geek | My sophomore programming classes that I was required to take over there were C++. I learned C in our ECE department. |
| 01:34.56 | Mirell | ce_geek: It was the same at my university, University of Texas. |
| 01:35.53 | ce_geek | The ECE and CS departments here kinda butt heads a bit. They get all territorial because we teach C. My opinion is that we wouldn't have to if they didn't suck at it. |
| 01:35.59 | ce_geek | Anyway, that's a local politics issue. |
| 01:37.30 | ce_geek | Is anyone familiar with any embedded systems courses taught at other universities offering Computer Engineering degrees? |
| 01:37.46 | ce_geek | I'd be interested in collecting some details for the CE faculty here. |
| 01:38.47 | ce_geek | ds2: Actually, I think they might learn scheme. |
| 01:39.07 | ce_geek | Huge AI geeks over there. They <3 lisp. |
| 01:39.27 | ce_geek | I'm not particularly interested myself. |
| 01:39.34 | ds2 | my course work had embedded stuff but it is a pure EE program |
| 01:39.44 | ce_geek | Really? Interesting. |
| 01:40.18 | ds2 | I think I had 3 classes of where you propose at the beginning, implement, and present results at the end |
| 01:40.37 | ds2 | the intro the the series was a semicanned project to get some practice |
| 01:40.43 | ds2 | intro to the series |
| 01:41.33 | ce_geek | Cool |
| 01:42.38 | ce_geek | In this case, it was just super open ended. There are only 3 of us in the class, and I've known the professor for 6 years now. We don't have any good relevant 68000 hardware to actually do anything with, so he suggested we can pick another processor to do somethign with if we wanted. |
| 01:43.02 | ce_geek | The other students are going to study something about memory, or maybe do some sort of paper-based design. I was not interested. |
| 01:43.09 | ds2 | then why not a Microblaze softcore ;) |
| 01:43.46 | ce_geek | A co-worker of mine and I both thought it would be neat to learn about the Arm9, and we just happened to trip over the Hammer kit. |
| 01:44.37 | ce_geek | Mind sending a link to an example development platform? I'm curious ab out this Microblaze product. |
| 01:44.52 | ds2 | goto Xilinx's site and search for it |
| 01:45.09 | ds2 | you build a CPU by point and clicking |
| 01:45.18 | ce_geek | Is the software available for Linux? |
| 01:45.33 | ds2 | the processor you create has a Linux port for it |
| 01:45.53 | ce_geek | If the whole project can't be done in Windows, it's just about a no-go. |
| 01:46.13 | ce_geek | er |
| 01:46.20 | ce_geek | If it can't be done entirely in Linux, that is. |
| 01:46.53 | ce_geek | The department has almost managed to eliminate Windows completely. A few minor apps hanging around that require Win2k running in vmware that we still have to use. |
| 01:47.00 | ds2 | the only way it would work is if there is a university support program...the package starts at like $10K+ |
| 01:47.07 | ce_geek | Mostly VHDL stuff, and microcap. |
| 01:47.15 | ce_geek | Yep, forget it. Never going to happen. |
| 01:47.16 | ds2 | but the demo at the trade show for it was kind of neat |
| 01:48.26 | ce_geek | Looks cool. It's really hard for a university department to spend money like that, unless it's a product they really need. I think they looked into getting a commercial version of htsoft's PIC compiler, because they do a ton of stuff with PICs in their classes. |
| 01:48.52 | ce_geek | Hmmmn. Not sure what's wrong with C18 though. I'll have to ask my prof about that. |
| 01:49.20 | ds2 | some places give a deep discount for universities or donate it entirely |
| 01:49.34 | ce_geek | Yup! And we appreciate it too :) |
| 01:49.59 | ce_geek | Some students get crazy things for free for their capstone projects. |
| 01:50.18 | Mirell | I still use my Student ID for discounts in various places |
| 01:50.24 | Mirell | Despite the fact I graduated 6 months ago. |
| 01:50.26 | ce_geek | A co-worker built a coil gun a few years ago. He got a huge $1k switch to turn the coil off for free. |
| 01:51.06 | ce_geek | I forget what kind of device it was or where he got it, but it was impressive. I never actually saw it, he graduated long before I was around, he just uses it as an example of getting free stuff. |
| 01:51.24 | ce_geek | Might have been an IGBT |
| 01:52.36 | ce_geek | Oh cool, the Prius uses a 50kW IFBT inverter for controlling two AC motor/generators. |
| 01:52.48 | ce_geek | s/IFBT/IGBT/ |
| 01:53.19 | ce_geek | HAHA. I never noticed that happened. I thought people just typed the regular expression to be cute. |
| 01:53.49 | ce_geek | Alright, back to my take-home assignment. TTYL. |
| 02:04.20 | MonMotha | I really want to build an electric car |
| 02:04.29 | MonMotha | the sheer amount of power involved is kinda scary though |
| 02:04.35 | MonMotha | even 25HP is a LOT |
| 02:05.50 | ds2 | MonMotha: did you follow the prius dissection last year at ESC? |
| 02:06.10 | MonMotha | ds2: I did not |
| 02:06.35 | ds2 | they published everything in one of rags a few months later |
| 02:06.55 | ds2 | can't recall the name but it was interesting seeing the different bits and how much customization/legacy parts is in there |
| 02:07.16 | MonMotha | I want a pure plug-in |
| 02:07.44 | MonMotha | even a 20 mile range would be enough for me to entirely eliminate my use of gasoline for commuting and shopping |
| 02:07.49 | ds2 | I toyed with the idea of putting together a electric ATV |
| 02:08.04 | MonMotha | and 50 miles is entirely possible even just using a bunch of lead acids |
| 02:08.15 | MonMotha | especially if you use a small vehicle |
| 02:08.25 | MonMotha | I was thinking of finding an old, dead Geo |
| 02:08.33 | MonMotha | or something of similar size |
| 02:08.43 | ds2 | heh and have more HP in there then the original engine? ;) |
| 02:08.49 | MonMotha | quite possibly |
| 02:08.58 | MonMotha | big thing is that you don't NEED all that peak HP |
| 02:09.06 | MonMotha | you've got torque out the wazoo |
| 02:09.27 | MonMotha | so unless you really need that high-speed performance, you don't need to come anywhere close to the peak HP rating of the gas engine |
| 02:09.35 | MonMotha | 10-20HP is amazingly sufficient, from what I'm told |
| 02:10.41 | MonMotha | and that's not terrible. Figure 15kW, and that's 125A off a 120V battery stack |
| 02:10.55 | MonMotha | now, there's no conversion loss accounted in there, but still |
| 02:11.23 | MonMotha | that's a manageable level of power |
| 02:12.05 | ds2 | but like you said, that's a scary amount of power to switch around |
| 02:14.06 | SpeedEvil | 15-20Hp will get you a top speed of 60MPH with a very small car. |
| 02:14.48 | SpeedEvil | It's not too bad - you can do 20Kw with very small semiconductors |
| 02:15.07 | ce_geek | UMaine used to have a very active solar vehicle team. They would go to the Tour de Sol every year and clean house in their division, with the same budget the big guys used for coffee and donuts. Unfortunately, the project went unfunded due to budget and politics. |
| 02:15.36 | ce_geek | The Phantom was damn impressive: |
| 02:15.40 | ce_geek | http://www.evalbum.com/837 |
| 02:15.40 | SpeedEvil | That's only 50A at 400V. - maybe 4 devices costinf $5 each. The problem is that if you get the design ever so slightly wrong, you will lose all 4 FETs. |
| 02:15.41 | MonMotha | SpeedEvil: which is perfectly sufficient for tooling around town |
| 02:15.45 | SpeedEvil | <PROTECTED> |
| 02:16.02 | SpeedEvil | MonMotha: google g-whizz |
| 02:16.07 | MonMotha | yeah, I've recently gotten into power conversion :) |
| 02:16.14 | MonMotha | had to do a small flyback design at work |
| 02:16.22 | ce_geek | 400HP and 500lb-ft of torque, at just about 0 RPM is kinda nuts :) |
| 02:16.23 | SpeedEvil | And then you've got the problem of batteries. |
| 02:16.30 | ce_geek | Damn heavy batteries :( |
| 02:16.33 | MonMotha | and I have to say it's even more intriguing now than when I took my power electronics clas |
| 02:17.17 | MonMotha | my college had a solar car program, but it got canceled just before I went there |
| 02:17.21 | SpeedEvil | The right batteries to use are the ones from A123 systems - but those are expensive and relatively unavailable. |
| 02:17.21 | MonMotha | they had the "Solar Phantom" |
| 02:17.32 | *** join/#edev dijenerate (n=dijenera@72.22.143.79) |
| 02:18.28 | SpeedEvil | LiFeP cells out of china may change this in the next few years. |
| 02:19.36 | ds2 | I'd think something smaller for offroad recreation (small ATV) would be easier to do |
| 02:20.20 | ce_geek | I heard a rough figure a while back that said it would cost about $10k for all the goodies to make an electric truck. |
| 02:20.28 | ce_geek | Batteries, motor, controller, etc. |
| 02:20.29 | ce_geek | :-/ |
| 02:21.57 | SpeedEvil | I did some numbers for my car. (Nissan Micra - a very small car) has a range of 400 miles on 42l of fuel in the tank. |
| 02:22.09 | SpeedEvil | It costs around $8K(us) new. |
| 02:22.15 | ds2 | even if you did build it, can you even drive it on the road? |
| 02:22.33 | ce_geek | The pickup truck UMaine had is still driven by the advisor every day. |
| 02:22.44 | ce_geek | He uses it to get to get back and forth to work. |
| 02:22.45 | SpeedEvil | If you rip out the fuel tank and engine, and replace with Li-Ion batteries, you get a range of around 100 miles, and about double the price. |
| 02:22.55 | ds2 | ce_geek: does main have a required inspection before they issue license tags? |
| 02:22.55 | ce_geek | yeah. |
| 02:23.20 | ce_geek | ds2: Yeeaaaaaaaah.... I'm not sure how that works though, because I think that truck has a University plate. |
| 02:23.42 | ce_geek | ds2: Do you think the conversion would fail inspection? |
| 02:23.42 | SpeedEvil | With lead-acid batteries it gets cheaper, the problem is that the lifetime of those suck |
| 02:23.58 | ce_geek | Surely emissions aren't a problem :) |
| 02:23.58 | SpeedEvil | ce_geek: it's tottally going to depend on locality. |
| 02:24.05 | ds2 | ce_geek: in CA, it definitely would |
| 02:24.23 | ce_geek | ds2: What criteria would it fail? |
| 02:24.28 | ds2 | might even get hit with a federal thing for "tampering" with emissions |
| 02:24.28 | SpeedEvil | ce_geek: generally, in pretty much all locations you can do anything to a vehicle, but the paperwork required may be extensive in some case. |
| 02:24.37 | SpeedEvil | cases |
| 02:24.54 | ds2 | ce_geek: it would fail cuz it does not visually match the OEM setup and it lacks CARB approval numbers for the deviation |
| 02:25.07 | SpeedEvil | You may need to - for example - demonstrate or prove that you have not worsened the crash performance. |
| 02:25.20 | SpeedEvil | In the UK, we have a _very_ lax setup. |
| 02:25.45 | SpeedEvil | There are a very limited number of things on which the vehicle can fail. |
| 02:25.52 | SpeedEvil | (at mandated tests) |
| 02:26.01 | ds2 | SpeedEvil: that's safety, I am talking about the braindead emissions here in CA |
| 02:26.23 | SpeedEvil | Emissions are simply tailpipe sensors at idle and fast idle |
| 02:27.31 | ce_geek | Speaking of efficient vehciles.... did I read somewhere that some states tax owners of hybrids more because they don't contribute as much to the taxes included with fuel purchases that help maintain roadways? |
| 02:27.32 | SpeedEvil | A big problem with electric cars of any form is that it's a lot cheaper to make 100 than 1 |
| 02:27.45 | ce_geek | Yeah. That is true. :-/ |
| 02:28.00 | SpeedEvil | (cheaper per car obviously) |
| 02:28.00 | MonMotha | A big problem with * of any form is that it's a lot cheaper to make 100 than 1 |
| 02:28.09 | SpeedEvil | Not babies! |
| 02:28.38 | MonMotha | I wouldn't be surprised if per unit cost still goes down, though maybe not by the time you get to 100 |
| 02:28.48 | SpeedEvil | But yes, most products. |
| 02:29.08 | SpeedEvil | The problem with EVs is the stuff you want isn't really sold. |
| 02:29.26 | SpeedEvil | Unless you have contracts to buy a few million worth. |
| 02:29.42 | SpeedEvil | As (forex) toyota has bought up world production. |
| 02:30.24 | MonMotha | yes |
| 02:30.35 | MonMotha | you have to make due with inferior parts (esp. batteries) |
| 02:31.06 | ce_geek | Hmn. I don't hear that spin about Toyota much. |
| 02:31.55 | SpeedEvil | http://zeva.com.au/tech/LiFePO4.php may be of use |
| 02:32.06 | SpeedEvil | I don't mean toyota in specific |
| 02:32.31 | SpeedEvil | I mean that if you're a battery maker, you really want to if you can simply make a nice fat contract with an EV maker. |
| 02:32.40 | SpeedEvil | Rather than selling in ones and thousands to every hobbyist. |
| 02:32.50 | ce_geek | Oh, yes, that makes perfect sense. |
| 02:33.15 | SpeedEvil | That above url goes into affordability and performance of new cells |
| 02:36.26 | *** join/#edev dijenerate (n=dijenera@72.22.143.79) |
| 02:43.12 | SpeedEvil | http://zeva.com.au/tech/headway/ gives a link to $.50/Wh cells, which works out to $8K or so for the above vehicle. |
| 02:45.35 | SpeedEvil | http://headway-cn.en.alibaba.com/product/0/200607688/12V20Ah_lithium_ion_battery.html I love the high-tech spotless picture in http://headway-cn.en.alibaba.com/product/0/200607688/12V20Ah_lithium_ion_battery.html |
| 04:44.03 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 04:57.35 | *** join/#edev harpal (n=harpal@124.125.76.195) |
| 05:06.09 | flyback- | wtf |
| 05:06.45 | flyback- | must have been the split or something reset the ban |
| 07:31.38 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 07:46.29 | *** join/#edev steliosk (n=Stelios@79.131.65.61) |
| 08:59.49 | *** join/#edev lyakh (n=lyakh@p57BD27BB.dip0.t-ipconnect.de) |
| 10:07.13 | *** join/#edev steliosk (n=Stelios@athedsl-281704.home.otenet.gr) |
| 10:53.21 | *** join/#edev linac (n=lin@121.41.70.78) |
| 10:57.39 | *** join/#edev KazaLite (n=kazaLite@203.215.180.205) |
| 12:45.13 | *** join/#edev uski (n=uski@uski.ST.HMC.Edu) |
| 12:45.46 | uski | (the following question has been asked in #uclibc with no reply, so i'm trying here :)) |
| 12:45.47 | uski | hi; i am thinking of building a cross-compilation toolchain for a target arch arm-cortexm3-elf (cortexm3 support is new in gcc 4.3.0); i have read stuff about a so-called "ARM EABI" and I am confused if it is something i should download or if the current snapshot of uclibc already supports it. I have read that the ARM EABI improves floating-point performance and it matters to me |
| 12:45.54 | uski | should i do anything to support the ARM EABI with uclibc ? |
| 13:12.42 | *** join/#edev sjhill (n=sjhill@real.realitydiluted.com) |
| 13:12.42 | *** mode/#edev [+o sjhill] by ChanServ |
| 13:40.45 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 13:42.22 | *** join/#edev chouimat|work (n=dieu@209.217.106.98) |
| 14:12.29 | chouimat|work | morning |
| 14:13.48 | *** join/#edev ce_geek (n=ce_geek@roger-blanchette-4.umelst.maine.edu) |
| 14:21.11 | *** join/#edev prpplague^2 (n=dave@mail.americanmicrosystems.com) |
| 14:21.58 | ce_geek | prpplague: one isn't enough? |
| 14:22.00 | ce_geek | :-P |
| 14:42.12 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 14:42.12 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 14:48.30 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 14:48.52 | prpplague | ce_geek: one of me is never enough |
| 15:20.56 | RobotGuy | Oops, clone didn't make it.. I guess prpplague is one of a kind.. :D:D |
| 15:32.39 | *** join/#edev MothaWork_ (n=monmotha@mail.forwardengineering.com) |
| 16:18.32 | ce_geek | Heh, cool: http://lwn.net/Kernel/LDD3/ |
| 16:19.26 | *** join/#edev dijenerate (n=dijenera@66.54.118.16) |
| 16:19.28 | prpplague | ce_geek: yep, pretty much the bible of kernel drivers |
| 16:20.06 | ce_geek | prpplague: I love that it's free. |
| 16:20.13 | ce_geek | prpplague: I will probably buy a print version of it though. |
| 16:20.26 | prpplague | ce_geek: yea well worth the money |
| 16:23.00 | ce_geek | prpplague: At some point this will probably become just a hobby. I'm not sure that I'll get that deep into it for my class project, I might do other lower level stuff with jtag and assembley and such, not sure yet. After the class project, (graduation), I have no idea what I'll be doing at my new job, and chances are I'll focus some energy on stuff to support whatever tasks I'm doing there. |
| 16:23.11 | ce_geek | prpplague: But this stuff is just too cool not to play with :) |
| 16:23.55 | prpplague | ce_geek: hehe, well, embedded linux dev pays pretty good |
| 16:24.58 | ce_geek | prpplague: I'm hoping I get to do that at my new job. The job is such that they can't really tell me a whole lot about what I'll be doing. :-/ |
| 16:25.11 | prpplague | ce_geek: oh? |
| 16:25.21 | ce_geek | prpplague: Defense stuff. |
| 16:25.27 | prpplague | ce_geek: ahh |
| 16:26.04 | ce_geek | prpplague: Kind of a funky job to accept.... just have to hope it doesn't suck :) |
| 16:27.44 | *** join/#edev emil|k (i=emil@f054014155.adsl.alicedsl.de) |
| 16:28.47 | ce_geek | prpplague: Forgive me, I haven't done the research / reading that I need to do to help support my project yet, been working on some more urgent stuff. Anyway, will OpenOCD let me look at things like stack pointers and memory locations? Let me run raw assembly code, or even compiled C code? Step the code? |
| 16:29.56 | prpplague | ce_geek: yea, openocd allows you to step break points and view all the register values |
| 16:30.02 | ce_geek | Excellent. |
| 16:30.28 | ce_geek | I'll be doing that at some point just to demonstrate that it can be done, and to help learn some low level things about the ARM |
| 16:30.58 | ce_geek | prpplague; For this midterm exam I'm working on i have to examine a listing file for an example program and draw a memory map of the stack :-/ |
| 16:31.26 | prpplague | fun fun |
| 16:31.28 | ce_geek | I was thinking it'd be interesting to do a more interesting problem with the ARM and actually look at stuff. |
| 16:32.07 | ce_geek | Interesting? <--- seemed like I had to use it one more time. jeez. |
| 16:32.32 | ce_geek | Book problems are kinda dry. It's fun to actually DO things.... something this class lacks. |
| 16:35.55 | ce_geek | Hmmn. The ARM instruction set doesn't look TOO scary.... |
| 16:37.45 | ce_geek | prpplague: "Thumb" instruction set? wtf? Cute.... |
| 16:39.13 | prpplague | ce_geek: hehe thumb is mainly use on highly embedded devices with very low resoruces |
| 16:47.48 | prpplague | ce_geek: no the arm ISA is actually very similiar to old 8050 isa |
| 16:49.08 | *** join/#edev ico2 (n=ico2@modem-2105.cougar.dialup.pol.co.uk) |
| 16:52.44 | prpplague | ce_geek: btw the documentation you started looks really good |
| 16:57.49 | *** join/#edev kazaLite (n=danishJa@116.71.161.194) |
| 17:11.56 | *** join/#edev kajaca (n=rakhus@32.137.243.193) |
| 17:20.30 | *** join/#edev dijenerate (n=dijenera@66.54.118.16) |
| 17:20.51 | *** join/#edev emil|k_ (i=emil@e179234213.adsl.alicedsl.de) |
| 17:34.22 | ce_geek | prpplague: Cool |
| 17:54.15 | *** join/#edev bbradley (n=bbradley@82-32-35-42.cable.ubr05.azte.blueyonder.co.uk) |
| 18:40.51 | *** join/#edev ico2_ (n=ico2@modem-1426.chimpanzee.dialup.pol.co.uk) |
| 19:07.11 | prpplague | SpeedEvil / ds2 : btw, i got that 320x240 lcd up and running on the hammer, the only thing that you have to do is remember the exact framebuffer format to code you applications |
| 19:08.14 | *** join/#edev dijenerate (n=dijenera@66.54.118.16) |
| 19:10.07 | *** join/#edev tom_ (n=tom@c-71-224-74-24.hsd1.pa.comcast.net) |
| 19:10.22 | tom_ | ~seen sjhill |
| 19:10.29 | purl | sjhill <n=sjhill@real.realitydiluted.com> was last seen on IRC in channel #uclibc, 6d 4h 26m 9s ago, saying: 'i'll talk to the admins'. |
| 19:10.47 | tom_ | prpplague: hey dave, you around? |
| 19:10.53 | prpplague | tom_: yep |
| 19:10.57 | prpplague | tom_: what up? |
| 19:11.22 | prpplague | T0mW: recovering from a really bad flu |
| 19:11.37 | T0mW | I've got a "debian" arm system here that looks as if it were wind-river or timesys, or RHEL |
| 19:11.56 | prpplague | T0mW: yea? |
| 19:12.02 | T0mW | prpplague: it is really complete, /usr/local/bin has pango, etc. |
| 19:12.21 | T0mW | prpplague: any ideas how I can ID what this system is? Where it came from? |
| 19:12.37 | prpplague | T0mW: yea WR and TS usually spend alot of time making sure its complete |
| 19:12.49 | prpplague | T0mW: ID it in what way? |
| 19:13.00 | T0mW | prpplague: I know you roll your own, but, maybe you have seen some others. |
| 19:13.21 | T0mW | prpplague: it does have a 'Debian 3.1' version in there. |
| 19:13.34 | prpplague | T0mW: well for starters look at the kernel version |
| 19:14.15 | prpplague | T0mW: cat /proc/version |
| 19:14.45 | T0mW | prpplague: well, client has had this dumped into his lap from consultant. That consultant was to give him the o/s sources so he could build his product up. But, the consultant has backed off on giving the sources to him. |
| 19:15.22 | prpplague | T0mW: ugh |
| 19:15.24 | T0mW | prpplague: I looked this thing over and it is just too complete to have come from a roll-your-own. It looks like a commercial offering of an ARM distro |
| 19:15.41 | prpplague | T0mW: check the /proc/version |
| 19:15.48 | prpplague | T0mW: that will give you a good start |
| 19:15.53 | T0mW | k |
| 19:16.28 | T0mW | Linux version 2.6.22-81348 (root@nfsdeb) (gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)) #2 Thu Nov 15 16:37:34 MST 2007 |
| 19:16.53 | prpplague | T0mW: my guess is its probably the standard debian for arm build |
| 19:16.53 | T0mW | is this a debian ARM distro? |
| 19:16.58 | T0mW | ah |
| 19:17.42 | T0mW | WTF do you get ARM Debian? Whenever I go to the debian site it looks so darned primitive that you'd think it was something from the 1990's |
| 19:17.53 | prpplague | hehe |
| 19:17.58 | prpplague | there is a debian arm channel |
| 19:18.07 | T0mW | prpplague: you know what I'm saying, right? heh |
| 19:18.16 | T0mW | yeahl, debianj-arm |
| 19:18.22 | prpplague | just be prepared for the normal rtfm stuff |
| 19:18.59 | ohsix | debootstrap --arch=arm :D |
| 19:19.04 | T0mW | prpplague: heh, just point me to where they hid the darned manuals, I'll go |
| 19:19.54 | prpplague | T0mW: yea, sorry, thats an area i'm not familiar with |
| 19:20.03 | prpplague | T0mW: as i never got along with debian |
| 19:24.46 | CosmicPenguin | well then you shouldn't have insulted its mother |
| 19:24.50 | CosmicPenguin | that would make me angry too |
| 19:27.42 | prpplague | CosmicPenguin: hehe |
| 19:39.36 | *** join/#edev Sascha (n=sisawess@cssun.rrze.uni-erlangen.de) |
| 20:18.15 | *** join/#edev ico2_ (n=ico2@modem-1426.chimpanzee.dialup.pol.co.uk) |
| 20:23.47 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 20:23.47 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 20:27.57 | *** join/#edev linac (n=lin@121.41.15.183) |
| 20:55.18 | *** join/#edev davidc__ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 21:06.44 | prpplague | ho ho hum |
| 21:06.57 | prpplague | CosmicPenguin: hows the new child process? |
| 21:08.16 | CosmicPenguin | maturing nicely |
| 21:09.52 | prpplague | good deal |
| 21:10.05 | prpplague | how mom? |
| 21:10.16 | CosmicPenguin | bored |
| 21:10.21 | prpplague | hehe |
| 21:30.49 | *** join/#edev landley (n=landley@adsl-68-93-84-18.dsl.hstntx.swbell.net) |
| 21:30.49 | *** mode/#edev [+o landley] by ChanServ |
| 21:31.42 | landley | Anything new? |
| 21:34.27 | prpplague | landley: other than i've been deathly ill for the last 10 days, no |
| 21:34.47 | prpplague | landley: i assume you got wifi working at hotel slez |
| 21:36.47 | landley | Actually I'm borrowing work's net connection on the way out the door. |
| 21:36.53 | landley | I clocked out at 4:30. :) |
| 21:37.04 | landley | (Meeting at 8:30 am, and I didn't go to lunch today.) |
| 21:37.29 | landley | Off to collect the keys at my new apartment in Houston. I may have stability to thump on hammer a bit soon. |
| 21:37.45 | landley | (Got a tinycc release-ish thing out on tuesday.) |
| 21:38.05 | landley | (The -c part of arm works, the linking part doesn't, but I haven't tracked down the problem yet.) |
| 21:38.30 | landley | It's more a snapshot than a really polished release, but it's a known set of bugs to fix and regression test against... |
| 21:39.28 | prpplague | landley: okie dokie, drop me an address to send a case to when you are ready |
| 21:41.17 | landley | Will do. Thanks. |
| 22:02.54 | *** join/#edev GPSFan (n=kenm@12.10.255.246) |
| 22:02.54 | *** mode/#edev [+o GPSFan] by ChanServ |
| 22:18.47 | *** join/#edev ico2_ (n=ico2@modem-1426.chimpanzee.dialup.pol.co.uk) |
| 22:19.44 | flyback- | after a month of running off the same boot of ubuntu livecd and ramfile system my main box just got wacked when I was cleaning the keyboard, going to try to dumo as many files to flash I can before pulling the plug, bbl |
| 23:18.07 | *** join/#edev linac (n=lin@121.41.37.235) |
| 23:27.01 | *** join/#edev redblue (i=star@ppp059.109-253-207.mtl.mt.videotron.ca) |
| 23:50.16 | *** join/#edev redb1ue (i=star@ppp030.108-253-207.mtl.mt.videotron.ca) |