| 00:33.21 | *** join/#edev rickytaylor26 (~quassel@80.83.125.94) |
| 01:17.08 | *** join/#edev prpplague (~prpplague@ppp-70-242-116-151.dsl.rcsntx.swbell.net) |
| 01:57.33 | *** join/#edev s1axter (~s1axter@c-98-245-7-56.hsd1.co.comcast.net) |
| 02:16.01 | *** mode/#edev [+o prpplague] by ChanServ |
| 04:49.43 | *** join/#edev mbuf (user@nat/redhat/x-pbyniqmxrthtbpin) |
| 07:50.53 | *** join/#edev calculu5 (~calculus@gentoo/user/calculus) |
| 09:01.56 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 09:04.41 | *** join/#edev risca (~risca@f-static-78-70-87-29.business.telia.com) |
| 09:33.42 | *** join/#edev yakov_ (~ysz@95-27-65-11.broadband.corbina.ru) |
| 10:34.46 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 10:52.50 | *** join/#edev thannoy (~aloiseau@231.39.0.109.rev.sfr.net) |
| 11:37.25 | *** join/#edev g1powermac (~g1powerma@rrcs-173-197-150-24.west.biz.rr.com) |
| 11:37.25 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 12:03.32 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 12:14.19 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 12:51.52 | *** join/#edev MonMotha (~monmotha@2001:470:c5f2:1:224:7eff:fedc:893e) |
| 13:14.56 | *** join/#edev yakov (~ysz@77.72.120.66) |
| 13:26.37 | *** join/#edev yakov (~ysz@77.72.120.66) |
| 14:08.54 | *** join/#edev yakov (~ysz@77.72.120.66) |
| 14:28.07 | *** join/#edev htns (~htns@175.139.208.57) |
| 14:46.05 | *** join/#edev yakov (~ysz@77.72.120.66) |
| 14:52.36 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 14:52.36 | *** mode/#edev [+o GPSFan] by ChanServ |
| 15:38.37 | *** join/#edev JViz (~JViz@cpe-071-076-188-130.triad.res.rr.com) |
| 15:44.19 | *** join/#edev itdnhr (~itdnhr@cpe-72-177-5-204.austin.res.rr.com) |
| 16:34.09 | *** join/#edev prpplague (~danders@nat/ti/x-xsxbcbzgmsmvhvha) |
| 16:35.53 | *** join/#edev CosmicPenguin (~nobody@soa.codeaurora.org) |
| 16:35.53 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 16:50.41 | *** mode/#edev [+o prpplague] by ChanServ |
| 16:53.19 | CosmicPenguin | Here's a question. What order are file descriptors freed in when a process crashes? |
| 17:13.57 | *** join/#edev JViz (~JViz@rrcs-70-63-118-85.midsouth.biz.rr.com) |
| 17:20.45 | *** join/#edev risca (~risca@m90-129-171-120.cust.tele2.se) |
| 17:59.52 | *** join/#edev risca (~risca@m90-129-42-177.cust.tele2.se) |
| 18:14.19 | *** join/#edev landley (~landley@140.242.26.2) |
| 18:17.43 | landley | Why does this not work? |
| 18:17.49 | landley | #include <sys/types.h> |
| 18:18.02 | landley | int main(int argc, char *argv[]) { return gettid(); } |
| 18:18.14 | landley | It says undefined reference to gettid, but man 2 gettid says it's in sys/types.h. |
| 18:18.27 | landley | I've dug down into the ubuntu headers to find out that SYSCALL() is apparently never defined anywhere. |
| 18:18.43 | landley | And that if it's undefined at the top of asm/unistd.h it'll be #defined to an empty stub. |
| 18:19.13 | landley | But... If there was no gettid() ubuntu 10.04 (despite it being a system call and having a man page), sure somebody would have NOTICED by now? |
| 18:19.32 | landley | If NO SYSCALL WRAPPERS WORK AT ALL, that would be career limiting. Surely there's something I'm misunderstanding here... |
| 18:19.47 | landley | Nevertheless, programming according to the man page ain't working for me here... |
| 18:20.26 | landley | Oh, found it: |
| 18:20.36 | landley | Glibc does not provide a wrapper for this system call; call it using |
| 18:20.37 | landley | <PROTECTED> |
| 18:20.48 | landley | Thanks, glibc! You horrible bastard abonimation you. |
| 18:21.37 | prpplague | hehe |
| 18:51.30 | Mirell | Daft Punk cures all ailes. |
| 18:52.03 | Mirell | Like figuring out how to nicely convert an Amazon AMI to a VirtualBox instance |
| 18:56.01 | kergoth | is listening to the trent razor & karen o cover of zeppelin's immigrant song |
| 18:56.02 | kergoth | <3 |
| 18:56.06 | kergoth | so good |
| 19:20.48 | Mirell | Hmm. I wonder if it's possible to specify a linux kernel to boot from on VirtualBox |
| 19:33.14 | landley | Mirclue. |
| 19:33.20 | landley | Mirell: not a clue. |
| 19:33.55 | landley | Sigh. Now the xubuntu vm on the windows box is doing the same "buffering keystrokes when I lag? No, go ahead and discard them" thing all the other ubuntu things do. |
| 19:34.14 | landley | It's not like my commodore 64 had a keyboard input buffer that WORKED back in 1982... Grrr. |
| 19:34.28 | landley | Lateral progress. 30 years thereof. |
| 19:39.11 | g1powermac | landley, doesn't surprise me |
| 19:39.55 | g1powermac | landley, after multiple install attempts, I've yet to get xubuntu to dual boot on my macbook pro |
| 19:40.24 | g1powermac | I've had less issues with getting debian installed on a powermac 6100 back when I barely knew how to operate a command line |
| 19:40.47 | g1powermac | (and that's a tough install even for someone fairly experienced in it) |
| 19:43.06 | *** join/#edev florian (~fuchs@eduroam.uni-siegen.de) |
| 19:43.06 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 19:44.20 | landley | g1powermac: yeah, but the PC was pretty standardized, especially in the late 90's. |
| 19:44.27 | wmat | g1powermac: didn't you used to hang out in the powerpc channels? I think I remember your nick. |
| 19:44.49 | landley | Interchangeable parts available from multiple sources, mail-order assembling your own white box was the standard way to do it back when Computer Shopper was big. |
| 19:45.10 | landley | The mac, in contrast, is a nonstandard device made by a company that's raised "not invented here" syndrome to an art form. |
| 19:45.33 | landley | Just because it switched to x86 processors doesn't mean it's making a PC. |
| 19:46.41 | g1powermac | landley, this is true, however, supposedly installing on an intel mac should be the easiest of them all, however, xubuntu keeps screwing with the partition tables and refuses to boot |
| 19:47.31 | g1powermac | wmat, not really, back when I worked on the old ppc machines I was on the mailing lists, I found irc later |
| 19:48.14 | wmat | i swear i remember your nick |
| 19:48.33 | wmat | maybe it was related to yellowdog linux |
| 19:48.41 | wmat | anyhoo, gotta run |
| 19:49.43 | g1powermac | landley, I've done install on the powermac 6100 (which is basically an m68k mac but with a ppc chip), a new world ppc powerbook, and this intel macbook |
| 19:50.01 | g1powermac | the powerbook was the absolutely simplest, worked like a charm |
| 19:50.44 | g1powermac | well, except the wifi card, but that was always a problem across all of the linux world at the time |
| 19:51.20 | g1powermac | still is to a degree |
| 19:51.26 | landley | This was a power mac from before Jobs came back and redid everything, I take it? |
| 19:51.36 | landley | Keep in mind Apple was essentially stagnant for a decade before Jobs came back. |
| 19:51.55 | landley | (There was the m68k->powerpc switch, but motorola and IBM drove rather a lot of that.) |
| 19:52.15 | g1powermac | yes, I think so, basically, it was the first ppc mac |
| 19:52.15 | landley | And on some designs they swapped the processor out and otherwise reused the board. |
| 19:52.38 | g1powermac | so it had primarily everything a m68k mac had minus the processor |
| 19:52.46 | landley | I'm actually poking at QEMU trying to get an m68k board to work. (1.0 shipped this morning, I should see what that's got.) |
| 19:53.02 | g1powermac | it had fun PDS and NuBus slots |
| 19:53.10 | landley | And the closest thing to usable seems to be the last m68k board they shipped, which is identical to the PPC equivalent just with the older processor. Quadra somethingorother. |
| 19:53.26 | landley | has it written down at home somewhere. |
| 19:53.29 | g1powermac | ya, the old quadras |
| 19:53.57 | landley | Alas, it may only have 128 megs of ram, and I need 256 megs to make current gcc happy. :P |
| 19:54.04 | g1powermac | hehe |
| 19:54.06 | g1powermac | probably |
| 19:54.08 | landley | (Native compiling: the breakfast of champions.) |
| 19:54.17 | g1powermac | the old powermac I have only has 64mb and its maxed out |
| 19:54.24 | g1powermac | and that is newer than a quadra |
| 19:54.27 | landley | Yes, I prefer to natively compile under an emulator rather than extensively cross compiling lots of stuff. |
| 19:54.36 | g1powermac | me too |
| 19:54.36 | landley | I'll take "slow" vs "crazy" any day. |
| 19:54.46 | g1powermac | its a pain otherwise to setup a toolchain |
| 19:54.50 | *** join/#edev fean0r (~Feanor_@p54A43DF6.dip.t-dialin.net) |
| 19:55.05 | landley | I may have to whip up a fake board both in the emulator and the kernel config in order to splice enough memory into the sucker. |
| 19:55.08 | g1powermac | has spent days setting up toolchains |
| 19:55.30 | landley | I've already been looking to do that to versatilepb on arm in order to get more than 256 megs there. (512 would be nice. 1-2 gigs would be nicer.) |
| 19:55.37 | g1powermac | hands landley a handful of 68030 chips he has laying on his desk. . . |
| 19:55.54 | g1powermac | I got 140 of them. . . |
| 19:55.55 | landley | I prefer emulators. |
| 19:56.08 | landley | Although architecture manuals are nice. |
| 19:56.35 | landley | Wikipedia often has an opinion on things like the m68k instruction set, but wikipedia's opinion is more often "firmly held" than "accurate". |
| 19:56.40 | g1powermac | yea, but I got enough to make an ancient super computer out of them ;-) |
| 19:56.47 | landley | Woot. |
| 19:56.55 | landley | Send 'em to Jeri Ellsworth. |
| 19:57.18 | landley | You need a fire breathing hardware geek to really take advantage of that sort of thing. |
| 19:57.22 | landley | Me, I do software. |
| 19:57.27 | landley | Emulators let me do software for the hardware. |
| 19:57.40 | landley | #turtlesalltheway |
| 19:58.37 | g1powermac | likes messing with hardware till the physics of electricity starts popping up. . . |
| 19:59.20 | g1powermac | I made a cooperative 'cluster' of 8051s one time to generate a cellular automata |
| 19:59.41 | g1powermac | also developed a custom bus for the comms between them too |
| 20:03.53 | Mirell | landley: The simplest way may be to make a small disk image with grub and boot from that to the second hard disk image. Oy. |
| 20:22.23 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 20:31.18 | landley | Mirell: you can always do a tarball transplant. |
| 20:31.35 | *** join/#edev MonMotha (~monmotha@2001:470:c5f2:1:224:7eff:fedc:893e) |
| 20:32.10 | landley | Make a tarball of the old root filesystem (trimming for /proc and such), extract it into a newly formatted one, arrange for bootloader, kernel, and appropriate init glue to mount stuff on the way up. |
| 20:32.21 | landley | That's how I used to install knoppix on my hard drive, and so on. |
| 20:33.07 | landley | The old "tar cpz | ssh me@elsewhere 'tar xpvz'" trick is often helpfulthere... |
| 20:36.51 | Mirell | I don't really want to touch the pristine filesystem though |
| 20:37.01 | MonMotha | landley: I find rsync useful in that situation nowadays, but I used the "tar over ssh" trick a lot in the past |
| 20:37.08 | Mirell | That requires far more time/effort |
| 20:37.20 | MonMotha | well, I guess maybe if you don't already have rsync on both ends :) |
| 20:37.38 | MonMotha | otherwise, rsync -rP * host:path/foo |
| 20:38.18 | Mirell | MonMotha: Was meant at landley's suggestion |
| 20:38.26 | MonMotha | Mirell: ah |
| 20:38.36 | MonMotha | prpplague: you around? |
| 20:38.38 | Mirell | I'm taking Ubuntu's filesystems used for the Amazon EC2 AMIs, and wanting to boot them under VirtualBox |
| 20:38.43 | prpplague | MonMotha: yep |
| 20:38.46 | prpplague | MonMotha: what up? |
| 20:39.07 | MonMotha | prpplague: you said you had some resources for using e.g. flyswatter with non-ARM targets for stuff like flash programming? |
| 20:39.21 | prpplague | MonMotha: oh yea sorry, let me locate that |
| 20:39.32 | MonMotha | prpplague: I'm looking for a "general purpose JTAG pod" for bench and small production |
| 20:39.38 | MonMotha | and flyswatter is priced right, and essentially open |
| 20:44.22 | MonMotha | oh, and nice. Flyswatter2 apparently supports 5V? |
| 20:44.26 | prpplague | yea |
| 20:44.50 | MonMotha | prpplague: nice. That was one of my reservations about the flyswatter. I do occasionally need 5V JTAG, though it's usually on AVR, and I have one of the Atmel pods. |
| 20:45.07 | prpplague | MonMotha: yea we spent alot of time working that out |
| 20:45.19 | MonMotha | about the only reason I still ever use 8-bit AVR, actually, is the ability to actually run the thing at 5V |
| 20:45.30 | MonMotha | prpplague: cool |
| 20:47.04 | prpplague | MonMotha: iirc this is what people are using with the flyswatter for xilinix - http://sourceforge.net/projects/xilprg/ |
| 20:47.18 | prpplague | MonMotha: you'd need to patch it with flyswatter2 support, but nothing major |
| 20:47.23 | MonMotha | unfortunately, Atmel never released the "official" docs on doing debug on 8-bit AVR, but it's been reversed |
| 20:47.42 | MonMotha | prpplague: I assume most of the FTDI based devices are roughly the same. Just pin ordering on the FIFO? |
| 20:47.51 | MonMotha | (in terms of software side) |
| 20:48.08 | landley | Mirell: the filesystem's easy. heck, you can cat the block device of the partition and use that as your raw image (I know how to do that in qemu, there's probably a way in virtual box). The tarball trick just allows you to change format and size and partition layout and such. |
| 20:48.20 | prpplague | MonMotha: most of them are all the same for the data and clock, it is the power/reset controls via the gpios that are slightly different |
| 20:48.36 | MonMotha | prpplague: gotcha. Sounds like it shouldn't be a difficult patch. |
| 20:48.43 | prpplague | MonMotha: i sent an email to the person i know who is using it |
| 20:48.43 | landley | Hooking a bootloader up to a kernel and then getting it to mount the filesystems together properly in the init scripts are generally fiddly. Not hard, just gotta adjust that plumbing manually if you perturb anything. |
| 20:48.50 | prpplague | MonMotha: let me see if i hear back from them |
| 20:49.00 | MonMotha | prpplague: sounds good |
| 20:49.12 | landley | prpplague: if i wanted to get a flyswatter and try to hook it up to an arbitraryish board... what would I do? |
| 20:49.22 | MonMotha | prpplague: btw, do you know of any good ARMs with USB and a well developed free/oss USB device software stack? |
| 20:49.40 | prpplague | landley: first determine the type of jtag interface available on the board |
| 20:49.46 | landley | For example, at work we've got a TI nightmare based on a netra chip (armv7 with 8 gazillion peripherals) in a custom board layout. Presumably I confugre openocd. In practice, I have no clue how. |
| 20:49.47 | MonMotha | prpplague: I'm currently stuck on Atmel AVR/UC3 because I'm using LUFA, but that's because it's darned good |
| 20:49.47 | prpplague | landley: most boards have a standard pinout |
| 20:50.10 | roxfan | there's lpcusb for lpc2xxx |
| 20:50.10 | landley | plus I dunno the physical connector setup. They have a windows jtag they can hook up, but it requires windows sotware to run. :( |
| 20:50.31 | prpplague | landley: yea, you would need to identify the connector type |
| 20:50.33 | roxfan | and lufa got ported to 1xxx |
| 20:50.34 | landley | prpplague: is there a connector list for jtags? |
| 20:50.38 | landley | Some web page somewhere, maybe? |
| 20:50.39 | prpplague | landley: yea |
| 20:50.42 | prpplague | one sec |
| 20:50.47 | landley | Lemme back up. |
| 20:50.48 | roxfan | http://hackaday.com/2011/11/11/lufa-open-source-usb-stack-now-for-nxp-arm-processors/ |
| 20:50.51 | landley | Where's the JTAG HOWTO? |
| 20:50.57 | landley | I need to start at the beginning, I expect. |
| 20:51.14 | roxfan | http://www.freelabs.com/~whitis/electronics/jtag/ |
| 20:51.17 | prpplague | http://elinux.org/JTAG |
| 20:51.31 | prpplague | landley: http://elinux.org/images/e/ea/JTAGpinouts.pdf |
| 20:51.50 | MonMotha | roxfan: oh interesting...that would be nice (not having to dig into a new API) |
| 20:52.16 | prpplague | landley: in addition there is a TI variant of the arm 14 |
| 20:52.28 | prpplague | landley: tincantools sells adapters for those |
| 20:52.57 | MonMotha | the NXP LPC series does seem to have won the war for the low end ARM market at this point |
| 20:53.08 | roxfan | mm i don't know |
| 20:53.28 | MonMotha | well at least amongst the oss/hacker community, maybe? |
| 20:53.29 | landley | I'll have to read those links over the weekend. My brain is kinda fried right now. |
| 20:53.33 | landley | Friday, long week. |
| 20:53.43 | MonMotha | dunno...I've been pushed out of it for a couple years at this point |
| 20:54.04 | landley | I have to have POST working by the 7th and the microcontroller I'm supposed to be bit-banging a serial protocol with in order to SAVE the data is not, as of yet, working right. And the guy who does it was out yesteday and today. |
| 20:54.25 | roxfan | nxp is late with their M4 chips... |
| 20:55.23 | MonMotha | roxfan: who are the major players left at this point? There were tons when I was looking ~5 years ago |
| 20:55.31 | landley | And I need to poke at QEMU 1.0, now that's out. |
| 20:56.00 | roxfan | atmel with STM32, TI with stellaris and Freecale with Kinetis |
| 20:56.15 | MonMotha | STM32 is ST, I thought? Atmel has SAM. |
| 20:56.24 | roxfan | er right sorry |
| 20:56.34 | MonMotha | is Sharp still playing? |
| 20:56.46 | MonMotha | I remember they had some ARM7TDMIs back when those were cool stuff |
| 20:56.47 | roxfan | yeah they have SAM3 but it's not very visible imo |
| 20:57.03 | roxfan | there's also fujitsu, nuvoton etc but they don't seem to be targeting hobbyists at all |
| 20:57.04 | MonMotha | I used a SAM7. It was...quirky. |
| 20:57.41 | MonMotha | I had a good Atmel rep, so I had lots of experience with Atmel stuff :) |
| 20:58.44 | roxfan | oh right the new arm arduino will use an atmel chip i think |
| 20:58.53 | roxfan | sam3u |
| 21:06.17 | MonMotha | I wonder if they've fixed the SPI controller by now :) |
| 21:06.27 | MonMotha | The SAM7 SPI controller was rather buggy, but it's fixed on UC3. |
| 21:06.45 | *** join/#edev bsdfox (~h36sa@unaffiliated/bsdfox) |
| 21:09.27 | calculus | I remember SAMs from command and conquer, they took down enemy aircraft |
| 21:09.48 | roxfan | surface-to-air missile, yep |
| 21:18.00 | roxfan | finally announced http://www.nxp.com/news/press-releases/2011/11/nxp-delivers-industry-first-arm-cortex-m0-microcontrollers-with-integrated-usb-drivers.html |
| 21:19.00 | *** join/#edev bgamari (~ben@gamari.physics.umass.edu) |
| 21:23.18 | MonMotha | roxfan: sounds like the old Anchor (prior to being bought by Cypress) EZ-USB but with a modern core |
| 21:23.49 | roxfan | not familiar with those |
| 21:23.52 | MonMotha | also, is it even remotely feasible to get an XIP ARM9 Linux kernel down to <512k ? |
| 21:24.42 | roxfan | oh interesting, integrated EEPROM |
| 21:25.38 | roxfan | vid/pid program too |
| 21:26.05 | MonMotha | there are some interesting little MCUs out there with ARM926 cores and integrated flash/sram. The CPU could obviously run Linux, but the memories are usually pretty limited. |
| 21:27.21 | roxfan | http://www.cygwin.com/ml/ecos-discuss/2002-12/msg00183.html from 2002 |
| 21:28.24 | roxfan | also http://elinux.org/Kernel_Size_Tuning_Guide |
| 21:28.26 | MonMotha | yeah, that's unfortunately so old as to be effectively obsolete |
| 21:28.37 | MonMotha | (that mail list thread) |
| 21:32.04 | MonMotha | my major interest is being able to use some drivers and (mostly) the networking code from Linux, as it's all rather complete and well tested, on "quite small" systems |
| 21:32.05 | roxfan | heh http://www.mentby.com/rob-landley/what-happened-to-linux-tiny.html |
| 21:34.27 | MonMotha | I'm inclined to agree with all that |
| 21:34.37 | MonMotha | or at least the first chunk (I haven't read the rest) |
| 21:35.36 | MonMotha | I'd argue that allnoconfig shouldn't do anything. You should be able to turn on/off every little bit. And yes, I know that's a pain in the butt to modularize out and maintain, which is probably why it hasn't been done. |
| 21:36.12 | *** join/#edev Matth_FA (~Matth_FA@mtl93-14-78-241-192-235.fbx.proxad.net) |
| 21:38.01 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 21:38.01 | *** mode/#edev [+o GPSFan] by ChanServ |
| 21:39.29 | roxfan | http://groups.google.com/group/linux.kernel/browse_frm/thread/d688fe97611dcc7e/7e685339b152b7f3 |
| 21:39.34 | roxfan | mentions "371K uncompressed" |
| 21:45.43 | MonMotha | unfortunately, all this is hinting that none of this probably applies to remotely modern kernels |
| 22:30.31 | *** join/#edev itdnhr (~itdnhr@cpe-72-177-5-204.austin.res.rr.com) |
| 22:56.21 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 23:44.23 | *** join/#edev flo_lap (~fuchs@Maemo/community/contributor/florian) |
| 23:55.33 | *** join/#edev JViz (~JViz@cpe-071-076-188-130.triad.res.rr.com) |