| === rsalveti` is now known as rsalveti | ||
| === deffrag__ is now known as deffrag | ||
| TheMuso | c/ | 06:19 |
|---|---|---|
| === phh_ is now known as phh | ||
| === doko_ is now known as doko | ||
| ogra_ | oh, ah ! | 13:31 |
| ogra_ | stgraber, while concentrating on getting the framebuffer to work i didnt actually notice that tablet is already booting fine | 13:32 |
| * ogra_ is just reading /var/log/dmesg from the SD | 13:32 | |
| ogra_ | [ 17.361] ABI class: X.Org Video Driver, version 13.0 | 13:32 |
| ogra_ | [ 17.361] (EE) open /dev/fb0: No such file or directory | 13:32 |
| ogra_ | [ 17.361] (EE) No devices detected. | 13:32 |
| ogra_ | [ 17.361] | 13:32 |
| ogra_ | HA ! | 13:32 |
| stgraber | ogra_: yay! | 13:35 |
| stgraber | ogra_: what did you end up doing to have it boot? | 13:36 |
| ogra_ | damned, but that was an older boot, not the kernel i'm currently trying | 13:36 |
| ogra_ | stgraber, dunno, i'm wildly poking kernel config opts mainly to get some kind of console (HDMI, USB serial or LCD) | 13:37 |
| ogra_ | so if i can get back to that working kernel i should be able to set up wlan in /e/n/i and use ssh | 13:37 |
| stgraber | ogra_: that'd be great! | 13:38 |
| * ogra_ goes back to the defconfig and starts over | 13:38 | |
| ogra_ | luckily cross building only takes a few seconds :) | 13:39 |
| stgraber | ogra_: yeah, once you get through the initial 30-45min build, rebuilding with different options is really quick :) | 13:40 |
| ogra_ | 30-40 min ? | 13:41 |
| ogra_ | what kind of desktop do you have ? | 13:42 |
| ogra_ | i think building it with modules took under 5 min initially here | 13:42 |
| ogra_ | (though i run my development folders in a ramdisk....) | 13:42 |
| zenx | Hi, i am failing to compile latest ti-omap kernel because of a "implicit function declaration" warning. Should I try and disable this or actually instantiate the function? | 13:43 |
| ogra_ | sigh, and now i cant get a single working boot anymore | 13:45 |
| ogra_ | fun | 13:45 |
| ogra_ | i should have checked /var/log eariler | 13:45 |
| zenx | ogra_: When u use a ramdisk do u have a local repo to commit the changes or is this unversioned development? | 13:50 |
| ogra_ | unversioned | 13:51 |
| === cmagina_ is now known as cmagina | ||
| zenx | ok | 13:51 |
| ogra_ | i have the git tree on disk though ... and copy it into the ramdisk as needed | 13:51 |
| === alkisg1 is now known as alkisg | ||
| highvoltage | ogra_: so I hear you have something booting? that's awesome :) | 14:21 |
| ogra_ | i had | 14:21 |
| ogra_ | and i'm not at all sure which kernel that was | 14:21 |
| ogra_ | the timestamps on the logs are 4h old | 14:22 |
| GrueMaster | ogra_: What system are you working on? | 15:18 |
| ogra_ | zatab | 15:18 |
| ogra_ | sigh, i really would like to know how it booted the last time | 15:28 |
| ogra_ | i think i went throuh all combos of u-boot and uImae now | 15:29 |
| ogra_ | *uImage | 15:29 |
| GrueMaster | So, why not just run Ubunto on Android? | 15:32 |
| ogra_ | why would i do such insanity ? :) | 15:32 |
| * ogra_ doesnt really want to carry a monitor with the tablet all the time to make it run ubuntu :) | 15:33 | |
| GrueMaster | I'd better just not comment. May not bode well for future relations. | 15:33 |
| ogra_ | i want to run native ubuntu :) | 15:34 |
| GrueMaster | Is that even possible on arm still? | 15:35 |
| GrueMaster | (yea, I went there). | 15:35 |
| ogra_ | hmm ? | 15:35 |
| ogra_ | sure it is | 15:35 |
| gildean | we're on a channel called ubuntu-arm, hint hint | 15:36 |
| gildean | :D | 15:36 |
| GrueMaster | gildean: If you knew me and my background, you would better understand my reference. | 15:37 |
| gildean | i was just joking, don't take it too seriously | 15:39 |
| * GrueMaster takes nothing seriously on these channels anymore. | 15:42 | |
| geotec | hallo pple! I am trying to compile kernel armel from scratch... cause i need to have CONFIG_AEABI=y and CONFIG_OABI_COMPAT=Y and CONFIG_ARM_THUMB=y (in order to get and older etch ARM(not armel!) chroot working with a needed toolchaing glibc 2.3.2) - Question: Why i fail compiling inside scratchbox with "/scratchbox/compilers/cs2007q3-glibc2.5-arm7/bin/sbox-arm-none-linux-gnueabi-strip: Unable to recognise the format of the input f | 15:43 |
| geotec | here is the output http://pastebin.com/e0uETWfg ... it seems to do its job, but it stops on 'debbing' the headers | 15:44 |
| ogra_ | i doubt anyone in ubuntu land has ever used scratchbox | 15:44 |
| geotec | well, me | 15:44 |
| geotec | ;D | 15:44 |
| LetoThe2nd | was just about to say the same - why not use the standard toolchain? | 15:44 |
| geotec | LetoThe2nd: tell me more | 15:44 |
| LetoThe2nd | geotec: apt-cache search gnueabi ;) | 15:45 |
| ogra_ | right, just a std cross compiler or a plain chroot with qemu-arm-static | 15:45 |
| geotec | it has to be armel | 15:45 |
| LetoThe2nd | geotec: so whats the problem? | 15:46 |
| LetoThe2nd | geotec: just check what arm toolchains are in the repo and pick what ever you like. | 15:46 |
| geotec | interesting.. approach.. | 15:48 |
| geotec | but i need a toolchaing like emdebian? | 15:51 |
| LetoThe2nd | no, why should you? | 15:52 |
| LetoThe2nd | the kernel is one of the things you can compile with mostly just the bare compiler (and some accesory tools) | 15:53 |
| LetoThe2nd | (plus, emdebian is not really what i would call a "toolchain") | 15:53 |
| geotec | yeah right. But how do i make a kernel package of it | 15:54 |
| geotec | ? | 15:54 |
| geotec | i mean, a debian kernel package | 15:55 |
| LetoThe2nd | hmh, easiest is if the source is debianzied. | 15:55 |
| geotec | the source is debianized | 15:56 |
| LetoThe2nd | \o/ for hrw then: | 15:56 |
| LetoThe2nd | http://marcin.juszkiewicz.com.pl/2012/03/26/ubuntu-12-04-precise-and-cross-compilation-of-arm-kernels/ | 15:56 |
| ogra_ | yeah | 15:56 |
| geotec | it gots all its rules and staff goes it way to the kernel package .. ok tnx i'll have a look | 15:56 |
| geotec | we should be more precise on arm architectures.. there are so many differences oabi legacy abi eabi and now hard eabi | 15:57 |
| ogra_ | ubuntu only supports one :) | 15:57 |
| LetoThe2nd | "one architecture to run them all" :P | 15:59 |
| ogra_ | (well, there is technically a rotting armv5 port in the archive nobody uses ... but thats not officially supported) | 16:00 |
| geotec | LetoThe2nd: thaz my problem, i am working with an arm binary which is linked to 2.3.2 glibc | 16:00 |
| geotec | but my kernel is armel and has no old eabi supporto | 16:00 |
| geotec | o_O this is drivin'me crazy | 16:00 |
| LetoThe2nd | geotec: i still do not see the problem. enable OABI and recompile the kernel. | 16:01 |
| geotec | yeah trying to do that, i am working on your tip | 16:01 |
| geotec | dpkg-buildpackage -b -aarmhf (this armhf toolchaing frightens me .. still gonna give it a try) | 16:03 |
| geotec | toolchain* sorry for the tuple errors | 16:03 |
| LetoThe2nd | well what ubuntu are you running? | 16:03 |
| LetoThe2nd | if you're running a hf one, better stick to it. if not, then... not. | 16:04 |
| infinity | geotec: I'd suggest that if you're trying to make OABI binaries work, there's probably something going fundamentally wrong in your life, and you might want to reexamine the choices that led here. :P | 16:16 |
| ogra_ | lol | 16:17 |
| infinity | geotec: (Also, OABI support will likely drop out of the kernel very soon, we've already torn it out of glibc) | 16:17 |
| * ogra_ curses this tablet | 16:19 | |
| ogra_ | it definitely loads the kernel | 16:19 |
| LetoThe2nd | ogra_: shamanic curse? | 16:20 |
| ogra_ | (the LCD brightens up if the kernel initializes it) | 16:20 |
| ogra_ | but all my kernels dont seem to get thruh to booting a rootfs | 16:20 |
| infinity | ogra_: I'm guessing you can't pop the back off and slap a JTAG in? | 16:20 |
| infinity | Pretty hard to debug early kernel failures without some sort of serial love. :/ | 16:21 |
| ogra_ | infinity, well, i just know that stgraber broke his display when opening his device .... | 16:21 |
| ogra_ | i dont really want to do that, the HW is fragile enough | 16:21 |
| infinity | Get stgraber to send you his broken one to attach a JTAG to? ;) | 16:21 |
| ogra_ | infinity, well, i did it on the ac100 and succeeded | 16:21 |
| GrueMaster | Does the device enable any serial ports on the host when usb is plugged in? | 16:21 |
| ogra_ | its not rocket science, just a matter of luck and time | 16:21 |
| stgraber | infinity, ogra_: you can actually get a serial port on the emmc connector, the problem is that it's the same port you're booting from :) | 16:21 |
| infinity | GrueMaster: If it did, that would almost certainly be a kernel driver that does that kind of magic, not firmware, so it would be useless for debugging the kernel itself. | 16:22 |
| ogra_ | stgraber, yeah, i'm experimenting with g_serial compiled in | 16:22 |
| GrueMaster | infinity: Blaze had that feature w/o kernel booting. | 16:22 |
| infinity | stgraber: Oh, that's a special "design". | 16:22 |
| LetoThe2nd | sounds like that thing is totally braindead. apart from the fact that they obviously have been forced to admit right on their webpage that hdmi out is disfunctional. | 16:22 |
| ogra_ | GrueMaster, yeah, TI is lovely, isnt it :) | 16:22 |
| infinity | GrueMaster: Blaze was also wildly overengineered for, like, engineers. I've never seen a consumer device nearly as cool. | 16:23 |
| geotec | infinity: i need to run an old arm binary linked against glibc <=2.3.2 | 16:23 |
| ogra_ | infinity, well, the kernel uses some kind of pre DT devicetree feature, you can re-route any HW you like in a textfile :) | 16:23 |
| infinity | geotec: Yeah, see above about "poor life decisions". :P | 16:23 |
| infinity | geotec: Seriously, why would one need to do such a thing? | 16:23 |
| geotec | cauz the arm binary is worth one year work? | 16:24 |
| infinity | geotec: (But sure, if you do need to, and there's no way around it, recompiling a kernel with OABI flipped on and running an ancient Debian chroot seems "sane") | 16:24 |
| ogra_ | sadly the only documentation for this file format i can find is a chinese PDF :) | 16:24 |
| ogra_ | though i already found out how to switch displays etc | 16:24 |
| geotec | infinity: what i am doing actually | 16:24 |
| geotec | ;) | 16:24 |
| infinity | geotec: It's a binary you have no source to, and no one knows where it came from? :/ | 16:24 |
| ogra_ | (which doesnt help if the kernel doesnt make any use of fbcon) | 16:25 |
| geotec | i need a kernel with old eabi support to run a etch chroot :D inside an arm device | 16:25 |
| infinity | geotec: It's not "old eabi", just "old abi". EABI was what replaced OABI. | 16:26 |
| infinity | (Just nitpicking) | 16:26 |
| geotec | otherwize i am doomed, what other choice do i have? Kill who came before me for choosing closed binaries from the heck, out of the world? | 16:26 |
| ogra_ | old eabi is probably the ganddad of both of them :) | 16:26 |
| LetoThe2nd | "never touch a running system"? what kept netware alive? | 16:26 |
| infinity | geotec: Or find something else that does the job. Or find the people who produced the original binary? | 16:26 |
| geotec | yeah right. Kinda stole the binary ;s | 16:27 |
| geotec | not really stole, but almost | 16:27 |
| * infinity covers his ears. | 16:27 | |
| infinity | LA LA LA CAN'T HEAR YOU. | 16:27 |
| infinity | GOING AWAY NOW. | 16:27 |
| geotec | lol | 16:28 |
| * LetoThe2nd hands infinity some iron maiden cds | 16:28 | |
| LetoThe2nd | anyways, time to leave ;) | 16:29 |
| * geotec wonders what was that thing about 'poor life decisions'.. :S | 16:29 | |
| geotec | you know, steal is something very strange to define on the web. Is getting something sincerly offered, you should not take a steal? Is like leaving donuts to indians saying: keep it there for me, i'll be back in a min | 16:34 |
| geotec | whell i am morally way below indians... i admit | 16:35 |
| geotec | for loggers: dont worry is just an update to an embedded device, which included this binary. They do not let the package without their hardware. I just grabbed the update and took what i needed | 16:37 |
| * geotec wonders where is everyone.. o_O | 16:39 | |
| ogra_ | oha | 16:47 |
| ogra_ | stgraber, ... | 16:48 |
| ogra_ | booting off USB key works flawless | 16:48 |
| ogra_ | i even have a mouse pinter now | 16:48 |
| ogra_ | using my panda USB key | 16:49 |
| stgraber | oh, I never tried usb boot ;) | 16:49 |
| ogra_ | me neither, but it was clear the the kernel comes up fine | 16:50 |
| GrueMaster | For which platform, zatab? | 16:50 |
| ogra_ | so it could only be a panic or issues mounting the rootfs | 16:50 |
| ogra_ | yes | 16:50 |
| stgraber | ok, so the problem is that booting from emmc is that the kernel can't find the rootfs... | 16:50 |
| ogra_ | apparently, i now have a runningg system so i can start inspecing that part :) | 16:50 |
| GrueMaster | sweet. | 16:51 |
| stgraber | nice! | 16:51 |
| * ogra_ goes to find an usb hub so i can attach a kbd | 16:51 | |
| ogra_ | hmm, and i better create an .xsession file so that it doesnt try to boot into unity all the time :) | 16:52 |
| stgraber | ogra_: once you figure out the right kernel config to get emmc boot, can you try and add that to the script (maybe create a zatab branch on LP with the needed kernel/uboot/script.bin/...)? | 16:53 |
| ogra_ | stgraber, sure | 16:54 |
| GrueMaster | But unity works everywhere, right? | 16:55 |
| GrueMaster | :P | 16:55 |
| ogra_ | bah | 16:56 |
| ogra_ | adding a hub doesnt work :/( | 16:56 |
| GrueMaster | You probably need a usb gadget host cable. | 16:57 |
| ogra_ | yes, the device only has micro USB ports | 16:57 |
| ogra_ | (read: i'm already using one) | 16:57 |
| GrueMaster | Kernel probably not configured for it. I know it kept going out of cycle on omap/omap4 images. | 16:59 |
| ogra_ | https://plus.google.com/107109423598372241322/posts | 17:03 |
| ogra_ | dunno if i have shared it the right way | 17:04 |
| * ogra_ isnt such a big g+ user | 17:04 | |
| stgraber | ogra_: "Oliver hasn't shared anything with you." | 17:05 |
| ogra_ | heck, how do i just publish something on g+ ? | 17:05 |
| ogra_ | unrestircted etc | 17:05 |
| stgraber | ogra_: you need to make sure it's shared with "public" | 17:06 |
| ogra_ | if i click share i can only select groups or individuals | 17:06 |
| ogra_ | ah, found it | 17:06 |
| ogra_ | intuitive is something else though | 17:06 |
| ogra_ | stgraber, try again | 17:06 |
| stgraber | yay, login screen on a tablet! | 17:07 |
| stgraber | highvoltage: ^ | 17:07 |
| ogra_ | right | 17:07 |
| ogra_ | no touch yet | 17:07 |
| ogra_ | and i cant attach a uhub, i think that draws to much power | 17:07 |
| ogra_ | so no mouse/kbd yet | 17:07 |
| ogra_ | bah and now DPMS kicked in | 17:08 |
| stgraber | getting it to connect to wifi shouldn't be too difficult | 17:08 |
| ogra_ | oh, indeed, i was just wondering why it didnt | 17:08 |
| ogra_ | i should probably copy the modules onto the panda image :) | 17:08 |
| stgraber | ajmitch managed to get wifi working pretty easily | 17:08 |
| ogra_ | well, i bet he had modules ;) | 17:08 |
| stgraber | yeah :) | 17:09 |
| ogra_ | hmm, intresting, doesnt finish booting with the modules in place | 17:18 |
| * ogra_ removes them again | 17:19 | |
| ogra_ | heh | 17:21 |
| ogra_ | boots fine | 17:21 |
| stgraber | that's odd | 17:22 |
| ogra_ | well, some module or some firmware being evil here | 17:23 |
| ogra_ | ah, awesome, my USB 1.1 hub seems to work | 17:25 |
| ogra_ | hmm, or not | 17:25 |
| ogra_ | seems it goes into constant reboots now | 17:25 |
| GrueMaster | Maybe it needs some QA? | 17:26 |
| ogra_ | no, power | 17:28 |
| ogra_ | i found a powered hub, lets see | 17:28 |
| ogra_ | yeah, mouse and kbd | 17:30 |
| ogra_ | and you can actually use ctrl-alt+t even on non GL systems to get a terminal up ... great | 17:32 |
| ogra_ | aha ! | 17:33 |
| ogra_ | init: failed to open system console | 17:33 |
| ogra_ | all over dmesg | 17:33 |
| ogra_ | stgraber, insmodding the wlan driver gives me a hard hang | 17:38 |
| ogra_ | :/ | 17:38 |
| ogra_ | so that device might have changed between pre-prod and today | 17:38 |
| ogra_ | next i'll try the fbcon module | 17:39 |
| ogra_ | ugh, these modules are also out of date | 17:42 |
| ogra_ | sigh | 17:42 |
| ogra_ | no woder wlan hangs it | 17:42 |
| highvoltage | stgraber: omg | 17:43 |
| ogra_ | yeah, i know, i need to clean my desk | 17:43 |
| ogra_ | :) | 17:43 |
| highvoltage | heh | 17:46 |
| ogra_ | oh cooool ! | 17:46 |
| ogra_ | pressing the power button actually gets me a shutdown dialog, nice | 17:47 |
| highvoltage | stgraber: I suddenly find it weird that we never considered trying USB | 17:47 |
| stgraber | highvoltage: I don't have usb sticks | 17:47 |
| ogra_ | well, i dont think its actually USB vs MMC | 17:47 |
| ogra_ | i guess there is really something weird with the modules | 17:47 |
| highvoltage | stgraber: yeah but they're $4 across the street :) | 17:47 |
| ogra_ | ok,. freshly built modules ... | 17:49 |
| * ogra_ starts over | 17:49 | |
| ogra_ | hmm, hangs again ... | 17:52 |
| ogra_ | even though i should have the matching symbols now | 17:52 |
| ogra_ | thats wried | 17:52 |
| ogra_ | hmm, unbelivable | 17:53 |
| ogra_ | ok, so its definitely fbcon that kills it | 18:06 |
| geotec | arm-linux-ld: unrecognized option '-Wl,-Bsymbolic-functions' .. this is tricky :S | 18:26 |
| TypoNAM | tried removing the comma , I don't think its suppose to be there | 18:36 |
| geotec | i think is a bug cauz ld shouldnt be passed -Wl | 18:38 |
| geotec | gcc is | 18:38 |
| TypoNAM | hmmm, based off of some googling it should be -Wl,-B,symbolic-functions so needs another comma ;) | 18:42 |
| ogra_ | hah, it helps to also disable android network crap isf you want network :P | 19:35 |
| ogra_ | silly stuff | 19:36 |
| ogra_ | grumble | 19:49 |
| ogra_ | so my /e/n/i network config doesnt get picked up and NM doesnt work in lightdm | 19:50 |
| highvoltage | still on the zatab ogra_? | 19:52 |
| ogra_ | ogra@zatab:~$ uname -a | 19:53 |
| ogra_ | Linux zatab 3.0.38+ #29 PREEMPT Thu Sep 20 21:11:35 CEST 2012 armv7l armv7l armv7l GNU/Linux | 19:53 |
| ogra_ | ogra@zatab:~$ | 19:53 |
| ogra_ | ;) | 19:54 |
| ogra_ | got ssh :) | 19:54 |
| ajmitch | ogra_: so you're making progress with yours? I'd love to see what's different & work out why mine booted so easily :) | 20:06 |
| ogra_ | stgraber, hmm, i thought the a10 was a dual core | 20:12 |
| ogra_ | i dont see any second core here | 20:12 |
| ogra_ | oh, and i only have 400M | 20:14 |
| ogra_ | hmm, no, 325M | 20:14 |
| ogra_ | and a constant system load of 3.00 | 20:15 |
| ogra_ | uh | 20:15 |
| ogra_ | bad plymouth | 20:15 |
| stgraber | ogra_: nope, a10 isn't dual-core, it's a single core Cortex A8 | 20:16 |
| ogra_ | wow, its pretty speedy for that | 20:16 |
| stgraber | yeah, it's surprisingly good for an a8 | 20:16 |
| infinity | Is it remarkably speedier than, say, an mx53 (also a single-core A8)? | 20:27 |
| ogra_ | dunno | 20:27 |
| ogra_ | i have never touched an mx53 | 20:28 |
| infinity | Ahh. Well, they're pretty speedy for what they are. :P | 20:28 |
| ogra_ | it feels a lot speedier than an XM | 20:28 |
| infinity | And my guess is that the A10 is just a stock design from ARM, I somehow doubt that, at that pricepoint, they're employing brilliant engineers to tweak it. | 20:28 |
| infinity | (Or it could be a relicensed part from another ARM licensee, like Freescale, just to confuse matters more) | 20:30 |
| NekoXP | infinity, it's not speedier than an MX53.. except it's easy to find an A10 1.2GHz, you have to special order 1.2GHz MX53 | 20:30 |
| ogra_ | well, no idea, it comes from deepeds china | 20:30 |
| infinity | NekoXP: Indeed, getting parts out of Freescale is vaguely like pulling teeth. | 20:31 |
| ogra_ | might have stolen designs at fsl or developed something themselves | 20:31 |
| NekoXP | it's also definitely not even close to i.MX or OMAP. An ARM reference design with some IP cores they generated themselves seems like the best bet, but a really fast turnaround and very little "optimization" | 20:31 |
| NekoXP | as far as the docs I've seen it's about as far from i.MX as you can get, either they randomly changed every register or the IP is just different | 20:32 |
| infinity | Heh. | 20:32 |
| infinity | Yeah, I've not looked at one at all. | 20:32 |
| ogra_ | ogra@zatab:~$ cat /proc/meminfo |grep MemTotal | 20:32 |
| ogra_ | MemTotal: 331520 kB | 20:32 |
| ogra_ | thats definitely not right | 20:32 |
| infinity | Though, I did say "relicensed from some licensee 'like Freescale'", I didn't imply it would have been Freescale. Tons of people design A8-alike parts. | 20:32 |
| NekoXP | a lot of IP comes out of China though. There's probably some dirty little video codec supplier around there that gave them a video decoder.. the display unit is a very reasonable and boring design. | 20:33 |
| infinity | Just seems unlikely that, if pushing for volume and low cost, they'd employ any in-house engineering to reinvent any wheels. | 20:33 |
| ogra_ | infinity, its china, why would they relicense, they just take it :) | 20:33 |
| infinity | ogra_: That only works as long as they don't want to sell into IP-whiney markets (like Europe and North America). | 20:34 |
| ogra_ | lol | 20:34 |
| infinity | ogra_: If their goal was world domination, they have to play by our bizarre rules. | 20:34 |
| ogra_ | well, go to fairs in europe and find all the gucci and rolex made in china | 20:34 |
| infinity | Yeahp, and Gucci tried to shut them down as they pop up. | 20:35 |
| ogra_ | didnt they even copy a complete BMW once ? | 20:35 |
| NekoXP | Gucci probably makes all their stuff in china anyway | 20:35 |
| infinity | But most of them are much harder to find than AllWinner, who's made themselves a large target. | 20:35 |
| infinity | So, here's hoping they're playing nicely. | 20:35 |
| ogra_ | yeah, i'm indeed not serious here :) | 20:36 |
| NekoXP | it's no wonder all that stuff gets Shanzhai treatment | 20:36 |
| ogra_ | anyway, enough tabletting for a day ... | 20:36 |
| * ogra_ wonders off to the TV | 20:36 | |
| infinity | Any day with tabletting is a bad day, IMO. | 20:36 |
| highvoltage | thanks for doing it ogra_ | 20:36 |
| infinity | I'm still so not the target market for those things. | 20:36 |
| highvoltage | ogra_: so does booting from USB make a difference? | 20:36 |
| ogra_ | highvoltage, well, now that i have it, i want to use it :) | 20:37 |
| * infinity goes back to arguing with clang. | 20:37 | |
| NekoXP | it could be that the IP in use is some standard FPGA library stuff. Some of it looks a lot like the things you'd find on opensource websites for Altera or Xilinx FPGAs | 20:37 |
| ogra_ | highvoltage, yeah, still cant boot from MMC | 20:37 |
| infinity | I swear, I develop a new brain bleed every time I dive into this codebase. | 20:37 |
| infinity | And it boggled the mind every time someone else claims it's "clean and well-written" or, my favourite "has great ARM support". | 20:37 |
| ajmitch | ogra_: I didn't think mine was too different (got it from zareason at UDS), as it boots from MMC just fine | 20:37 |
| infinity | s/boggled/boggles/ | 20:37 |
| ajmitch | infinity: their definition of 'great' might be that it compiles sometimes | 20:38 |
| highvoltage | ajmitch: yeah there's definitely some difference between the two | 20:38 |
| ogra_ | ajmitch, well, might be pre-production | 20:38 |
| infinity | ajmitch: Yeah, well. There are some things about the llvm architecture that are pretty neat, and if we could go back in time a few decades, GCC could learn some lessons. | 20:38 |
| infinity | ajmitch: But overall, the project is painful to work with, especially with their haphazard support for multiple architectures. | 20:38 |
| infinity | And every small mistake llvm makes is magnified about four thousand times by clang's clunky interface on top. | 20:39 |
| infinity | BUT I'M NOT BITTER. | 20:39 |
| infinity | llvm as a backend to gcc might actually be easier to support than clang is. | 20:40 |
| ajmitch | gcc's architecture has been driven by the fsf philosophy, not always in a good way | 20:41 |
| infinity | Nope, GCC has a lot of warts, and a lot of things that are almost literally unfixable at this point. | 20:41 |
| infinity | But llvm's hardly a utopia of software design either. | 20:41 |
| XorA|gone | you should know by now open source devs just jump from cool eTLA to cool eTLA :-D | 20:43 |
| infinity | Whatever. Perl 4 lyfe, yo. | 20:45 |
| TypoNAM | lol | 20:46 |
| infinity | Anyhow. Let's see if this hacked-up clang actually (A) builds, and (B) produces binaries that look kinda like armhf ones. | 20:46 |
| * infinity crosses his fingers. | 20:46 | |
| micahg | infinity: abiword keeps aborting on armel | 21:42 |
| infinity | micahg: Yes, this is known. Stop giving it back. :P | 21:44 |
| micahg | infinity: I figured I wasn't running into the wall fast enough :P | 21:46 |
| micahg | any idea when things might return to normalish? | 21:47 |
| infinity | Not sure. It's a catch-22 where hacking around this breaks other things. | 21:47 |
| infinity | And fixing it properly is real effort. | 21:47 |
| David7 | Has anyone here built the kernel.org kernel using Ubuntu arm-linux-gnueabihf-gcc? | 23:39 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!