| 00:08.39 | *** join/#edev purl (~purl@rikers.org) |
| 00:08.39 | *** topic/#edev is Discussion of embedded systems (Hardware/Software/Linux/WinCE/Commercial Device Repurposing) : Busybox and Buildroot : http://www.elinux.org |
| 00:53.31 | *** join/#edev file (~file@asterisk/developer-and-muffin-lover/file) |
| 00:56.32 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 01:36.53 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 01:51.46 | *** join/#edev prpplague (~prpplague@ppp-70-242-119-115.dsl.rcsntx.swbell.net) |
| 01:55.03 | *** join/#edev risca (~risca@wi-secure-1813.cc.umanitoba.ca) |
| 01:59.08 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 01:59.08 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 01:59.08 | *** join/#edev uwe_ (~uwe_@dslb-088-066-186-052.pools.arcor-ip.net) |
| 01:59.08 | *** join/#edev RushPL (~quassel@89-69-171-30.dynamic.chello.pl) |
| 01:59.16 | *** join/#edev RushPL (~quassel@89-69-171-30.dynamic.chello.pl) |
| 02:08.28 | *** join/#edev av500 (~av500@p50996ded.dip0.t-ipconnect.de) |
| 02:08.28 | *** join/#edev NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) |
| 02:08.28 | *** join/#edev solar (~solar@smtp.gentoo.org) |
| 02:08.28 | *** join/#edev n17ikh (~peter@c-174-56-150-44.hsd1.sc.comcast.net) |
| 02:08.28 | *** join/#edev madsy (~madsy@fu/coder/madsy) |
| 02:08.28 | *** join/#edev alex4nder- (~alexander@ns2.lexterieur.net) |
| 02:12.44 | *** join/#edev roxfan (dunno@91.180.21.99) |
| 02:14.56 | *** join/#edev slashbeast (~slashbeas@funtoo/core/slashbeast) |
| 02:16.57 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 02:16.57 | *** join/#edev sterNiX (~LessIsMor@unaffiliated/nu253r/x-0655220) |
| 02:16.57 | *** join/#edev htns (~htns@175.139.208.57) |
| 02:16.57 | *** join/#edev ricky26 (~quassel@80.83.125.94) |
| 02:16.57 | *** join/#edev skynet-2000 (SkyNet-200@gateway/shell/trekweb.org/x-jtbslgsopdmxqotx) |
| 02:16.57 | *** mode/#edev [+o GPSFan] by verne.freenode.net |
| 02:19.21 | *** join/#edev skynet-2000 (SkyNet-200@gateway/shell/trekweb.org/x-jtbslgsopdmxqotx) |
| 02:19.21 | *** join/#edev ricky26 (~quassel@80.83.125.94) |
| 02:19.21 | *** join/#edev htns (~htns@175.139.208.57) |
| 02:19.22 | *** join/#edev sterNiX (~LessIsMor@unaffiliated/nu253r/x-0655220) |
| 02:19.22 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 02:19.22 | *** join/#edev slashbeast (~slashbeas@funtoo/core/slashbeast) |
| 02:19.22 | *** join/#edev roxfan (dunno@91.180.21.99) |
| 02:19.22 | *** join/#edev alex4nder- (~alexander@ns2.lexterieur.net) |
| 02:19.22 | *** join/#edev madsy (~madsy@fu/coder/madsy) |
| 02:19.22 | *** join/#edev n17ikh (~peter@c-174-56-150-44.hsd1.sc.comcast.net) |
| 02:19.22 | *** join/#edev solar (~solar@smtp.gentoo.org) |
| 02:19.22 | *** join/#edev NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) |
| 02:19.22 | *** join/#edev av500 (~av500@p50996ded.dip0.t-ipconnect.de) |
| 02:19.22 | *** join/#edev RushPL (~quassel@89-69-171-30.dynamic.chello.pl) |
| 02:19.22 | *** join/#edev uwe_ (~uwe_@dslb-088-066-186-052.pools.arcor-ip.net) |
| 02:19.22 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 02:19.22 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 02:19.44 | *** join/#edev prpplague (~prpplague@ppp-70-242-119-115.dsl.rcsntx.swbell.net) |
| 02:19.44 | *** join/#edev file (~file@asterisk/developer-and-muffin-lover/file) |
| 02:19.44 | *** join/#edev dfarnsworth (~dfarnswor@nat/montavista/x-razkvcadwbtxqsxt) |
| 02:19.44 | *** join/#edev ohsix (ohsix@208.100.175.111) |
| 02:19.44 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 02:19.44 | *** join/#edev nullpuppy (~nullpuppy@freematrix/staff/nullpuppy) |
| 02:19.44 | *** join/#edev calculus (~calculus@gentoo/user/calculus) |
| 02:19.44 | *** join/#edev devilsadvocate (~quassel@109.200.19.188) |
| 02:19.44 | *** join/#edev MonMotha (~monmotha@2001:470:c5f2:1:224:7eff:fedc:893e) |
| 02:19.44 | *** join/#edev jevin (~jevin@napalm.jevinskie.com) |
| 02:19.44 | *** join/#edev Spida (~timo@ns1.spinnennetz.org) |
| 02:19.44 | *** join/#edev jonmasters (~jcm@edison.jonmasters.org) |
| 02:19.44 | *** join/#edev grummund (~user@unaffiliated/grummund) |
| 02:19.44 | *** join/#edev sunrunner20 (~jamesdoe@pool-71-123-194-101.dllstx.fios.verizon.net) |
| 02:19.44 | *** join/#edev bsdfox (~h36sa@unaffiliated/bsdfox) |
| 02:19.44 | *** join/#edev kergoth (~kergoth@covenant.kergoth.com) |
| 02:19.44 | *** join/#edev ccole (~cole@coledd.com) |
| 02:19.44 | *** join/#edev SpeedEvil (~user@tor/regular/SpeedEvil) |
| 02:19.44 | *** join/#edev pytey (~pytey@howler.blatero.com) |
| 02:19.44 | *** join/#edev wmat (wmat@wallace.mixdown.ca) |
| 02:19.44 | *** join/#edev MuscleNerd (~MuscleNer@99-65-242-37.lightspeed.irvnca.sbcglobal.net) |
| 02:19.44 | *** join/#edev xxiao (~xxiao@li41-126.members.linode.com) |
| 02:19.44 | *** join/#edev samxor (simoirc@206.217.201.48) |
| 02:19.44 | *** join/#edev yoann_ (~yoann@archos.rain.fr) |
| 02:19.45 | *** join/#edev mpk (mp@aggr.com) |
| 02:19.45 | *** join/#edev keesj (~keesj@ip49-193-210-87.adsl2.static.versatel.nl) |
| 02:19.45 | *** join/#edev mnemoc (~amery@shell.opensde.net) |
| 02:19.45 | *** join/#edev Crofton|work (~balister@pool-96-240-160-175.ronkva.east.verizon.net) |
| 02:19.45 | *** join/#edev uwe_mobile (~uwe@static.88-198-8-117.clients.your-server.de) |
| 02:19.45 | *** join/#edev ChanServ (ChanServ@services.) |
| 02:19.45 | *** mode/#edev [+oo GPSFan ChanServ] by verne.freenode.net |
| 02:20.37 | *** join/#edev landley (~landley@cpe-24-27-31-176.austin.res.rr.com) |
| 02:20.37 | *** mode/#edev [+o landley] by ChanServ |
| 02:20.57 | landley | Oh _now_ southwest decides to offer $70 each way airfares on sale. |
| 02:21.11 | landley | Too late, I put down an offer on a house, can't afford a trip to CELF just now. |
| 02:21.25 | prpplague | landley: doh |
| 02:21.39 | landley | Eh, it's a really nice house. |
| 02:21.39 | SpeedEvil | :) |
| 02:22.07 | landley | Dunno if we'll get it. Hard to afford it. But dude, mortage rates at _3%_ right now. |
| 02:22.20 | landley | If I'm _ever_ going to be able to afford to trade up from this condo, now would be the time... |
| 02:22.52 | landley | (Yeah, between taxes and insurance the place is $7k/year even without the mortgage or home maintenance, but still.) |
| 02:23.17 | landley | Mirell: 45th and airport. |
| 02:23.25 | landley | North campus instead of west campus. |
| 02:23.27 | landley | if we get it. |
| 02:23.32 | Mirell | Ah. |
| 02:23.57 | landley | Basically moving from "surrounded by undergrads" to "surrounded by grad students". |
| 02:24.07 | Mirell | Hyde-Parkish sortof |
| 02:25.03 | landley | Yup. |
| 02:25.14 | landley | Within _easy_walking_ distance of the 24 hour HEB. |
| 02:41.50 | prpplague | landley: well i wish i could wish you good luck, but i hate owning a house and would not wish it on anyone, let alone someone i call a friend... |
| 02:42.15 | landley | prpplague: I own a condominium right now. |
| 02:42.20 | landley | It's across the street from a fraternity. |
| 02:42.52 | prpplague | hehe |
| 02:43.36 | prpplague | landley: *cough* your boss lady isn't in nesting mode? |
| 02:43.47 | landley | prpplague: nope, I'm driving this one. |
| 02:44.04 | landley | I keep trying to have guests over from out of town and the couch is too small. |
| 02:44.07 | landley | Kitchen's cramped. |
| 02:44.10 | prpplague | hehe |
| 02:44.10 | landley | FOUR CATS. |
| 02:44.43 | landley | (Queue Jean-Luc: THERE! ARE! FOUR! CATS!) |
| 02:46.35 | prpplague | landley: hehe |
| 02:46.38 | prpplague | landley: i get that |
| 02:48.06 | prpplague | landley: you know...... |
| 02:50.29 | prpplague | landley: every time i see "Fade" in your notes i hear in my mind's eye "Feyd"......... from Dune...... |
| 02:51.05 | prpplague | s/eye/ear ?? |
| 03:11.59 | MonMotha | THERE! ARE! CAT! FIVE! eeeeeeeeee! |
| 03:14.10 | prpplague | hehe |
| 03:24.59 | Mirell | ...what |
| 03:26.11 | prpplague | Mirell: which one? |
| 03:26.17 | Mirell | CAT FIVE |
| 03:29.29 | MonMotha | prpplague: so apparently SD cards are rather inconsistent in their pre-init behavior and very picky about the init sequence |
| 03:31.19 | *** join/#edev NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) |
| 03:57.27 | *** join/#edev RushPL_ (~quassel@89-69-171-30.dynamic.chello.pl) |
| 04:03.12 | landley | Mortgages are unpleasant paperwork. |
| 04:33.55 | *** join/#edev mbuf (user@nat/redhat/x-qucyykyudzntnvwp) |
| 05:26.16 | *** join/#edev file (~file@asterisk/developer-and-muffin-lover/file) |
| 05:26.16 | *** join/#edev ohsix (ohsix@208.100.175.111) |
| 05:35.09 | *** join/#edev file (~file@asterisk/developer-and-muffin-lover/file) |
| 05:35.09 | *** join/#edev ohsix (ohsix@208.100.175.111) |
| 05:36.38 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 05:37.34 | *** join/#edev uwe__ (~uwe_@dslb-088-066-186-052.pools.arcor-ip.net) |
| 05:38.53 | *** join/#edev djerome_ (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 05:39.21 | *** join/#edev djerome (~djerome@ip68-2-20-108.ph.ph.cox.net) |
| 05:46.42 | *** join/#edev kergoth (~kergoth@covenant.kergoth.com) |
| 05:48.10 | *** join/#edev clio (~andrej@85.159.109.222) |
| 07:13.16 | *** join/#edev devilsadvocate (~quassel@109.200.19.188) |
| 07:50.39 | *** join/#edev amarsman (~marsman@90-145-17-249.wxdsl.nl) |
| 07:51.20 | *** join/#edev devilsadvocate (~quassel@109.200.19.188) |
| 08:00.36 | *** join/#edev devilsadvocate (~quassel@109.200.19.188) |
| 08:17.53 | *** join/#edev uwe_mobile (~uwe@static.88-198-8-117.clients.your-server.de) |
| 09:27.43 | *** join/#edev mazert (~mazert@vpn2.freedom-ip.com) |
| 09:42.09 | *** join/#edev lfelipe (~lfelipe@enlightenment/developer/lfelipe) |
| 09:44.26 | mazert | plop |
| 09:45.36 | mazert | some1? :) |
| 09:50.02 | av500 | you? |
| 09:51.16 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 10:18.16 | *** join/#edev lfelipe (~lfelipe@enlightenment/developer/lfelipe) |
| 11:16.29 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 13:16.41 | *** join/#edev troth (~troth@nat/hp/x-lemhjzjpktstwqpd) |
| 14:52.36 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 14:52.39 | *** mode/#edev [+o GPSFan] by ChanServ |
| 15:02.25 | *** join/#edev landley (~landley@cpe-24-27-31-176.austin.res.rr.com) |
| 15:25.02 | lfelipe | landley: btw, have you seen newlib? it might be useful as a replacement libc that is not (l)gpl |
| 15:25.50 | landley | I've heard of newlib used in combination with libgloss to port busybox to bare metal (circa 2005 or so), but haven't heard anything about it since. |
| 15:26.39 | landley | Dmitry Tochansky on the toybox list is writing his own BSD licensed libc, and just started looking at bionic. He'd probably be the guy to ask. |
| 15:26.39 | lfelipe | not sure how well maintained is the project right now, we used it quite a bit back in my university days |
| 15:26.56 | lfelipe | yeah, I asked because of that discussion in fact :) |
| 15:27.09 | landley | I vaguely recall it was part of the cygnus stuff red hat acquired, and then red hat moved away from the emebedded space when Enterprise happened. |
| 15:27.21 | landley | Ask on the list and I'll tell you that story. Right now I've got to go commute... |
| 15:27.37 | av500 | have fun :) |
| 15:45.31 | *** join/#edev risca (~risca@wi-secure-1813.cc.umanitoba.ca) |
| 16:00.01 | *** join/#edev purl (~purl@rikers.org) |
| 16:00.02 | *** topic/#edev is Discussion of embedded systems (Hardware/Software/Linux/WinCE/Commercial Device Repurposing) : Busybox and Buildroot : http://www.elinux.org |
| 16:01.58 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-fef8de00-126.dhcp.inet.fi) |
| 16:17.48 | *** join/#edev MonMotha (~monmotha@2001:470:c5f2:1:224:7eff:fedc:893e) |
| 16:52.06 | *** join/#edev landley (~landley@140.242.26.2) |
| 16:56.37 | *** join/#edev prpplague (~danders@nat/ti/x-exbniyuiahwezbbz) |
| 16:57.59 | *** join/#edev risca (~risca@wi-secure-1813.cc.umanitoba.ca) |
| 17:19.51 | *** join/#edev risca (~risca@wi-secure-1813.cc.umanitoba.ca) |
| 17:35.32 | *** join/#edev sterNiX (~LessIsMor@2.176.204.40) |
| 17:35.32 | *** join/#edev sterNiX (~LessIsMor@unaffiliated/nu253r/x-0655220) |
| 18:04.50 | *** join/#edev florian (~fuchs@sign-4db6ad45.pool.mediaWays.net) |
| 18:04.50 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 18:35.45 | landley | Morning. |
| 18:43.15 | *** join/#edev troth (~troth@nat/hp/x-lemhjzjpktstwqpd) |
| 18:43.58 | *** mode/#edev [+o landley] by ChanServ |
| 19:02.21 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 19:48.22 | *** join/#edev T0mW (~Tom@70.15.162.118.res-cmts.t132.ptd.net) |
| 20:12.20 | *** join/#edev calculus (~calculus@99.48.50.183) |
| 20:12.20 | *** join/#edev calculus (~calculus@gentoo/user/calculus) |
| 20:37.58 | *** join/#edev risca (~risca@wi-secure-6257.cc.umanitoba.ca) |
| 20:57.25 | *** join/#edev lfelipe (~lfelipe@186.214.73.132) |
| 20:57.37 | *** join/#edev lfelipe (~lfelipe@enlightenment/developer/lfelipe) |
| 21:11.09 | *** join/#edev prpplague (~danders@nat/ti/x-fiomaqhggbhmhidk) |
| 21:24.19 | *** join/#edev sjhill (~sjhill@63.228.164.32) |
| 21:24.19 | *** mode/#edev [+o sjhill] by ChanServ |
| 22:17.17 | landley | It's amazing how much I get done as long as i have something specific to avoid working on. |
| 22:27.12 | roxfan | procrastination keeps you busy! |
| 22:27.52 | kergoth | meh, I can't concentrate worth a damn today |
| 22:29.02 | prpplague | kergoth: me either |
| 22:29.43 | landley | Anybody else have this deep, buried rage that they work in an industry where most of the people involved aren't very good at it? |
| 22:29.51 | landley | wants to hurt TI. |
| 22:30.06 | landley | Reading the comments about how feature sets can't coexist due to space limitations. |
| 22:30.07 | kergoth | I suspect that's true of most industries, to some extent |
| 22:30.31 | landley | 2/3 of the space (256k on-chip memory, usable before the DRAM init happens) was used by 3 instances of a 64k FAT buffer. |
| 22:30.41 | landley | So you can load files from a FAT partition on flash. |
| 22:31.19 | landley | The default FAT sector size is 512 bytes. The partition format can go up to 64k sectors, which means you're using a multi-gigabyte disk. |
| 22:32.05 | landley | With 4k sectors, fat16 can go up to a 268 megabyte disk. The boot partition is 7 megabytes. |
| 22:32.51 | landley | So 2/3 of this 256k on-chip buffer is just _wasted_, yet they have the engineering bandwidth to document what sets of features they can fit into the space they have left. |
| 22:33.22 | roxfan | can't you get away with smaller buffer? |
| 22:37.11 | landley | I went from 64k to 4k. We're actually _using_ 512 bytes. |
| 22:37.14 | landley | So yes, very much so. |
| 22:37.38 | landley | And now that I've freed up 2/3 of the space, I no longer need MLO to load u-boot to load uImage, I can rip out a step there. |
| 22:38.19 | landley | (Their rom loads a uboot version called MLO from flash into the on-chip memory, then there's the DRAM init, then that first uboot loads a second uboot with presumably more features, and then that second uboot loads the kernel.) |
| 22:38.34 | landley | No need for the second uboot now, but there's no reason _I_ had to do this. |
| 22:38.47 | landley | Apparently they never ran objdump -x against their binary, or some such... |
| 22:39.54 | landley | prpplague: you want to step in and defend TI? :) |
| 22:40.01 | landley | Or plead the 5th? |
| 22:40.23 | kergoth | that's pretty sad. being able to deal with resource constraints should be one of the defining characteristics of anyone working in embedded |
| 22:40.25 | roxfan | "objdump? wazzat?" |
| 22:41.01 | *** join/#edev risca (~risca@wi-secure-6257.cc.umanitoba.ca) |
| 22:41.05 | prpplague | reads the log |
| 22:41.48 | kergoth | knew they used 3 steps, but never looked into the reasoning |
| 22:45.23 | prpplague | landley: the 64k limitation for x-loader is due to the fact x-loader is used across a wide range of chips where some only have 64k available. most of x-loader has been deprecated in that the new versions of u-boot generate a MLO files that is far more robust due to the fact it is based off the optomize u-boot code rather than a bit rot x-loader |
| 22:45.49 | landley | I'm not using x-loader, this is MLO. |
| 22:45.58 | landley | In MLO, there are three FAT buffers which are 64k each. |
| 22:46.10 | landley | They fill up 2/3 of the Netra's 256k on chip memory. |
| 22:46.57 | prpplague | landley: right MLO is just the name it looks for, MLO can be anything executable |
| 22:47.10 | landley | I shrunk them down to 4k to get that memory _back_ so I didn't need both MLO and uboot. |
| 22:47.40 | prpplague | landley: right, very easy to do |
| 22:47.53 | prpplague | landley: what code are you using to generate the MLO? |
| 22:48.09 | prpplague | landley: sorry didn't read properly |
| 22:48.57 | prpplague | landley: yes there are a bunch of tweaks that still need to happen to the MLO generated from u-boot |
| 22:49.07 | prpplague | landley: the support in u-boot to generate a MLO is very new |
| 22:49.45 | prpplague | landley: i prefer actually to use barebox which is small enough to operate from the 64k base so no second step is needed |
| 22:51.44 | prpplague | landley: trust me, the overall condition of x-loader/u-boot is much much better now than it was a year ago |
| 22:52.53 | landley | prpplague: we're using "psp-14", whatever that is. |
| 22:53.05 | landley | Apparently we downloaded some code from a playstation portable. |
| 22:53.41 | roxfan | PSP can also mean Program Segment Prefix \o/ |
| 22:53.46 | landley | Sigh. Need energy drink... |
| 22:55.27 | prpplague | landley: yea that is some team builds inside of the ASP teams, but it is based on mainline u-boot |
| 22:55.45 | prpplague | landley: there are still about 50+ patches queued up for u-boot and SPL support |
| 22:55.47 | landley | It's based a u-boot from almost 2 years ago, anyway. |
| 22:56.06 | prpplague | landley: oh, then you are most likely using the x-loader code |
| 22:56.08 | landley | Eh, I'm just irritable becaue my caffeine's run out. |
| 22:56.20 | prpplague | landley: if it is that old |
| 22:56.50 | landley | It's not a TI release from 2 years ago. It's a current release, based on TI having forked the upstream project. |
| 22:58.07 | prpplague | landley: hmm, sorry i guess i did not understand your previous statement |
| 22:58.20 | landley | needs a nap. |
| 22:58.48 | landley | It's not turning 40, it's sitting in a cubicle all day with no exercise... |
| 22:58.50 | prpplague | landley: so yes, to answer your question, the whole boot method stuff is a mess |
| 22:59.39 | landley | I'm reminded of the one and only episode with the 8th doctor, where he takes a tiny little piece out of a giant atomic clock since that's the bit that does the actual work. |
| 22:59.58 | landley | I feel the need to do that to this code, but sadly I'm not here to clean it up. |
| 23:00.18 | prpplague | landley: hehe indeed |
| 23:00.38 | prpplague | landley: the mainline u-boot is getting pretty clean and robust |
| 23:00.52 | prpplague | landley: its only when TI does their hacks for internal releases that it gets nasty again |
| 23:01.57 | landley | Let's see, what does a bootloader need: |
| 23:02.10 | landley | 1) Entry point, possibly setting up interrupt vectors with stubs. |
| 23:02.26 | landley | 2) polling serial I/O so you can see it do stuff. |
| 23:03.04 | landley | 3) Mini-stack before DRAM init (possibly the stack-in-cache trick with a hand-hacked TLB entry or two). |
| 23:03.09 | landley | 4) DRAM init code. |
| 23:03.30 | landley | 5) load-a-kernel function, possibly with FAT parsing. |
| 23:04.12 | prpplague | landley: hehe |
| 23:04.23 | landley | 6) jump to kernel with "supplemental data" (kernel command line, device tree, possibly external initramfs archive). |
| 23:04.33 | prpplague | landley: now, add on top of that the 400 other requests from TI customers.... |
| 23:04.36 | landley | Altogether we're talking _maybe_ 10k. |
| 23:04.57 | *** join/#edev madsy (~madsy@fu/coder/madsy) |
| 23:05.00 | landley | Screw the 400 other requests. Load linux, do it there. |
| 23:05.05 | landley | If you can't load linux, you have a jtag. |
| 23:05.34 | prpplague | landley: hehe, yea problem is that under a biz environment it doesn't always work that way, hehe |
| 23:05.36 | prpplague | landley: http://www.elinux.org/APEX |
| 23:05.42 | kergoth`meh | maintainers need to get better at saying no |
| 23:05.45 | kergoth`meh | heh |
| 23:05.45 | prpplague | landley: my preferred bootloader for arm devices |
| 23:05.48 | landley | I'm not talking about TI, I'm talking about the open source project upstream. |
| 23:06.43 | prpplague | landley: i agree in your concept, but groups that maintain stuff like u-boot always are pushing for new features and such, basically becomes bloated |
| 23:07.02 | prpplague | landley: i am not a big fan of u-boot, but it is the big standard these days for many embedded systems |
| 23:07.04 | landley | The maintainer of an embedded thing needs to be able to push back. |
| 23:07.22 | prpplague | landley: in this case the maintainers are the ones wanting the features |
| 23:07.24 | landley | "This project exists to do X and do it well. That is not X." |
| 23:08.23 | prpplague | landley: yea, u-boot says "we want to be a super feature rich bootloader, if you want small and optimized, use barebox(aka u-boot-v2)" |
| 23:08.45 | prpplague | landley: have you tried barebox? |
| 23:08.51 | landley | not yet. |
| 23:08.55 | landley | what license is APEX under? |
| 23:08.56 | roxfan | ...until that one becomes big and bloated :) |
| 23:09.03 | landley | Ah, MIT. |
| 23:09.17 | landley | roxfan: if the project's goal is to do X and only X, it doesn't have to become bloated. |
| 23:09.32 | landley | It can reach a point where development naturally STOPS, because it does what it set out to do. |
| 23:09.52 | prpplague | landley: http://www.elinux.org/Barebox |
| 23:09.54 | landley | After that you have adjustments for new C standards, and to fight off amazingly stupid gcc warnings, and so on. |
| 23:10.25 | prpplague | landley: barebox fits more into the idea of toybox and other minimal implementations |
| 23:10.34 | prpplague | landley: everything is configurable for small size |
| 23:10.51 | prpplague | landley: i have some barebox builds for panda that are less that 8k with a number of features |
| 23:11.03 | landley | Cool. |
| 23:11.15 | landley | I'm kinda full up right now, but I'll bookmark it. |
| 23:11.19 | roxfan | afair zipit had quite a small bootloader |
| 23:12.40 | roxfan | or not zipit... some other thingy |
| 23:12.50 | prpplague | there are tons of small bootloaders out there |
| 23:13.03 | prpplague | problem is getting them to a point that they are easily modified for other platforms |
| 23:13.15 | prpplague | apex is pretty robust and is pretty extensible |
| 23:16.45 | prpplague | apex is now up on gitorious |
| 23:20.07 | calculus | sometimes people want bootloaders to play sound or display an image |
| 23:21.05 | prpplague | or have backup firmware loading options |
| 23:21.27 | prpplague | or have special sleep/suspend functions |
| 23:22.13 | calculus | once you throw nfs in there, you start needing network support, dhcp, etc |
| 23:23.16 | prpplague | and people want to support network via usb which entails adding OHCI and EHCI support |
| 23:23.17 | calculus | and sometimes usb support when the network device is on usb (hi panda and beagleboardS) |
| 23:23.33 | prpplague | which then people want to be able to boot from usb storage devices |
| 23:25.04 | calculus | to be fair to uboot, grub is also a multistage bootloader |
| 23:28.36 | prpplague | calculus: well also the issue is not a u-boot item, it is the implemetation of the bootrom from TI that allows it load the bootloader from the sd card that really puts the requirements on u-boot |
| 23:36.04 | landley | If you want a bootloader to do arbitrary things, make it with a Great Big Hook to call a board-specific function, and then having said function is not the bootloader's problem. |
| 23:36.13 | landley | "We link in crap.o, which you supply yourself." |
| 23:36.31 | landley | calculus: grub is horrid. |
| 23:55.21 | calculus | landley: I haven't looked at the code, but it served me well on my x86 machines |