[11:31] <florian__> ola
[12:05] <CIA-17> nico * r4308 /packages/libs/pwlib/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port pwlib to buildroot-ng
[12:08] <CIA-17> nico * r4309 /packages/libs/openh323/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port openh323 to buildroot-ng
[13:36] <florian__> {Nico}: stil had not time to test au1000 ?
[13:58] <florian__> nbd: did you make some tests on the wrt54gx ?
[13:58] <nbd> not yet
[13:59] <florian__> ok, looks like people can have it running, only the wireless driver was not working, probably because there was not mimo enabled wl.o at this thime
[14:01] <florian__> http://forum.openwrt.org/viewtopic.php?pid=30880#p30880
[14:14] <[mbm]> http://code.google.com/
[14:14] <[mbm]> google is now doing opensource project hosting
[14:14] <[mbm]> interesting but the whole thing lacks the typical google polished ui
[14:15] <[mbm]> the 'browse source' is just a link to the svn server itself which is horrible for browsing
[14:16] <[mbm]> example project - http://code.google.com/p/nautilus-scripts/
[14:17] <[mbm]> the really funny thing is that some of the people behind it were also the people that were working for va linux at the time sourceforge was launched
[14:17] <florian__> ah ah :)
[14:18] <florian__> {Nico}: are your porting quagga ? vrrp ? openvpn ?
[14:18] <florian__> because if, no, I will do it
[16:53] <CIA-17> nico * r4310 /packages/libs/bluez-libs/ (. Config.in Makefile ipkg/): port bluez-libs to buildroot-ng
[17:04] <CIA-17> nico * r4311 /packages/utils/bluez-utils/ (. Config.in Makefile ipkg/): port bluez-utils to buildroot-ng
[17:10] <CIA-17> nico * r4312 /packages/libs/bluez-libs/Makefile: remove dependency on libdaemon
[17:52] <nbd> groz: ping
[18:02] <CIA-17> nico * r4313 /packages/libs/libamsel/ (. Makefile): port libamsel to buildroot-ng
[18:05] <CIA-17> nico * r4314 /branches/whiterussian/openwrt/toolchain/uClibc/patches/140-ipv6_missing_sockopts.patch: backport [1729] to WhiteRussian
[18:13] <CIA-17> nico * r4315 /packages/libs/libamsel/ (Config.in ipkg/): oops, forgot to remove unneeded files for -ng from my last commit (again)
[18:15] <CIA-17> nico * r4316 /packages/libs/serdisplib/ (. Config.in Makefile ipkg/): port serdisplib to buildroot-ng
[18:34] <florian__> re
[19:53] <CIA-17> mbm * r4317 /branches/buildroot-ng/openwrt/include/package.mk: fix build issues with SMP compiles
[20:18] <florian__> malbon: ping ?
[20:30] <[mbm]> should probably consider adding distcc or ccache to ng
[20:30] <malbon> florian__: yo
[20:31] <malbon> distcc would be pretty nifty
[20:32] <[mbm]> well, to clarify I meant for building the firmware, not as a package
[20:32] <Bartman007> [mbm]: mmm, that would be cool.
[20:32] <malbon> yeah I thought so.
[20:32] <Bartman007> I could have 3x amd64 3700s building for me.
[20:33] <[mbm]> (a whole army of wrt's might have the same horse power as the old pentium which is being used as an nfs server to host the wrt's gcc ... :P)
[20:33] <Bartman007> gcc on emulated fpu, yey!
[20:33] <[mbm]> Bartman007: I use quad xeons
[20:34] <Bartman007> so compiling openwrt consumes enough power to run a wrt for a month?
[20:34] <[mbm]> a month? let's be realistic here
[20:35] <[mbm]> we're talking 3 months at least
[20:36] <florian__> malbon: do you remember which mtd driver you told me about for redboot parsing reference ?
[20:37] <florian__> [mbm]: did not you make the brilliant performance test on the wiki ;)
[20:37] <[mbm]> florian__: no, I was the one that flamed the hell out of them for doing so
[20:37] <florian__> [mbm]: yes I remember, it's far from being relevant
[20:38] <florian__> but maybe we could try running the spec.org test on the wrt \o/
[20:38] <[mbm]> if you want a useless benchmark, go brag about bogomips
[20:38] <florian__> nah, it does not take enough time !:p
[20:39] <florian__> I want something like 3dmark or another shit
[20:39] <[mbm]> wait .. what about distributed.net or folding@home ?
[20:39] <florian__> this is a good idea
[20:39] <db90h> the question is, are u both being sarcastic, or just one of you
[20:40] <db90h> u underestimate the power of these routers, i've been making love to my overclocked wrt for weeks now
[20:40] <[mbm]> db90h: if it were me I'd be setting up my own locate distributed.net server and sending out aes packets to be decrypted ;)
[20:40] <florian__> we are sarcastic, we are misanthropist at the same time, yay !
[20:41] <db90h> don't use words like misanthropist, that waste considerable time since u know i gotta go look itup
[20:41] <Bartman007> florian__: I like you Europeans, you use worlds that confuse most Americans but are perfectly valid English.
[20:42] <db90h> it would be kinda cool to have an army of wrt's working on a distributed computing project, just for amusement
[20:42] <florian__> Bartman007: lol :)
[20:42] <[mbm]> db90h: is that a complaint against large words, or large words not relating to onanism?
[20:42] <db90h> umm
[20:43] <db90h> hehe
[20:43] <db90h> both it turns out
[20:43] <[mbm]> wondered if you'd look that up
[20:43] <db90h> i was almost too lazy
[20:43] <db90h> but glad i did
[20:44] <db90h> i'll use this often from now on ;p
[20:44] <Bartman007> I had a history teacher in 7th grade who read the dictionary cover to cover, you could tell what letter he was on each day...
[20:45] <[mbm]> lmao
[20:45] <db90h> hehe
[20:45] Action: [mbm] used to have a few 'obscure word of the day' things
[20:46] <[mbm]> tend not to use them because even the english speakers don't know most of them
[20:49] <malbon> florian__: use the sa1100-flash one, it's got a static map and redboot in that one.
[20:49] <florian__> malbon: thanks again, in fact, I found the necessary modifications in the inventel sources :)
[20:49] <malbon> florian__: sorry, I was watching firefly.
[20:50] <florian__> I really wonder why english is so popular and widespread,apart from the fact that they had the biggest empire/colonies ever
[20:50] <[mbm]> good show
[20:50] <florian__> malbon: no pb
[20:50] <malbon> I like it. :)
[20:50] <[mbm]> florian__: that's exactly it
[20:50] <[mbm]> english as a language is one of the most backwards languages
[20:50] <malbon> but its a lot easier to learn than chinese.
[20:51] <florian__> right, only few twice as less as french as far as I remember
[20:51] <[mbm]> I suppose it's easier in the regard that most words aren't gender specific
[20:52] <Bartman007> [mbm]: if rap slang like hyphy is working it's way into the english vernacular, you should use obscure words.
[20:52] <florian__> Bartman007: lol, I don't understand a word of your sentence :)
[20:52] <[mbm]> Bartman007: nah, then I'd look just as elitist as those people that type using 'j00' and 'u' as words
[20:53] <Bartman007> [mbm]: y, r u n0t r33t enough to understand 1337ness?
[20:55] <Bartman007> bah, I can't even stick with it if I try, I missed to==2, that crap is so annoying.
[20:56] <florian__> if you can type quickly you dramatically improve your orthograph
[20:59] <Bartman007> I believe that if you are speaking of spelling alone, definitely doesn't decrease the number of typos
[21:07] <[mbm]> hmm .. looks like there's still some smp issues
[21:07] <[mbm]> make spends forever in the packages directory even when there's nothing to be done
[21:09] <[mbm]> probably have to dig through make and try to reduce the number of targets
[21:12] <florian__> ah ah better use a wrt for cross compiling [mbm] !
[21:12] <florian__> woah I am glad to be in weekend, I become to have packet loss inside my brain
[21:13] <[mbm]> latency is pretty high too
[21:30] <db90h> i'm a little teacup
[21:30] <Bartman007> that's all 2MB of flash is good for.
[21:31] <db90h> speaking of 2mb flash, i just built a compressed CFE
[21:31] <db90h> which ends up at 113KB by default
[21:31] <db90h> giving us 128KB more flash space
[21:32] <Bartman007> wouldn't mind 128KB extra space on his 4MB devices also
[21:32] <Bartman007> +/me
[21:33] <db90h> i used the belkin 7230-4 sources, which include the CFE source.. that unit uses a compressed CFE
[21:35] <florian__> ah with which algo ?
[21:35] <db90h> it uses gzip
[21:35] <db90h> by default
[21:35] <florian__> ok
[21:35] <db90h> i haven't done any tweaking to it, i just got it built after mucking with the belkin sources for a while
[21:36] <florian__> gg
[21:36] <db90h> i'd like to repackage the CFE source so we can all create custom, optionally compressed, CFE image more easily
[21:37] <florian__> yes, but I wonder if we can do it with the actual CFE licensing ?
[21:37] <db90h> shhh
[21:37] <florian__> ah, yes of course, we can do it ;)
[21:37] <db90h> I think i'm in violation already with my vxworks killer, of both Broadcom and Wind River
[21:38] <db90h> but not positive on that
[21:39] <db90h> (Wind River part being in the reversion firmware, where I distribute the VxWorks boot loader)
[21:41] <Bartman007> db90h: would it be possible to just distribute your code, and a script to strip the bootloader out of existing vxworks firmware?
[21:43] <db90h> well, there is technically a way, but it's not trivial... you'd have to write a firmware compatible with the VxWorks boot loader that does the job of reading and storing the VxWorks boot loader
[21:43] <Bartman007> oh hmmmm, I was thinking that the Vx bootloader was included in firmware updates, dunno why....
[21:44] <db90h> no, not that i know of.. though since that capability exists in the firmware format, it's not inconcievable that some vendor somewhere has, or will, do that
[21:46] <florian__> damn, I just found out so many things regarding the bootloader type in the inventel sources !
[21:46] <db90h> The VxWorks boot loader is fucking huge compared to the CFE.. it takes up 384KB of ROM space compressed
[21:46] <florian__> there was complete sources on how handling board initialization for the 2 bootloaders
[21:46] <florian__> that's quite huge
[21:46] <florian__> but maybe there are some code that is not the kernel task then
[21:47] <db90h> well, actually its 383KB, last KB is the BOOTP block (which i've just documented and wrote a tool for, if anybody cares)
[21:47] <db90h> it's got more functionality than CFEs
[21:47] <db90h> it has a built in httpd server
[21:47] <db90h> full tcp/ip stack
[21:47] <db90h> and dhcpd/dhcpc
[21:48] <florian__> ah that's the reason why
[21:48] <db90h> it's a pretty capable OS in its own right ;p
[21:49] <florian__> well, actually it is one of the best hard real time OS ever
[21:50] <db90h> u mean VxWorks itself? yea, i was just joking about its boot loader as an OS
[21:50] <db90h> I don't know much about VxWorks, but i appreciate the fact that it seems to do its job well in less resources than linux
[21:51] <florian__> yes
[21:51] <db90h> though as mbm i think educated me a while back, these aren't technically RTOSes.. maybe VxWorks can be tho
[21:52] <florian__> and actually, apart QNX and VxWorks, I don't know any other hard real time OSes
[21:52] <florian__> it is quite easy to find people to work under VxWorks, QNX, and Linux, for others, I really doubt it is as easy
[21:53] <florian__> I really need to improve my english
[21:55] <db90h> me too
[21:55] <db90h> i'd like to be an embedded developer by profession
[21:56] <db90h> i might start puring that.. it's just hard to afford things like Wind River's SDKs, and any specialized hardware i might need
[21:56] <db90h> err pursuing
[21:57] <db90h> i noticed btw an awesome Broadcom doc on these platforms
[21:57] <florian__> actually I am following telecommunication studies to become engineer (master executive I guess is the equivalent of what I follow in france) but specializing in embedded systems would be great !
[21:57] <db90h> it instructed me, amongst other things, on how to set up gdb to remotely debug the OS
[21:58] <db90h> i'm sure u guys have seen it
[21:58] <db90h> the one in the Belkin 7230-4 sources, and also in others i'm sure
[21:59] <florian__> I did not look at it
[21:59] <db90h> here, let me throw it up on web real quick
[21:59] <db90h> for anyone who hasn't read it.. it's awesome reading
[22:00] <db90h> http://www.bitsum.com/files/BCM47XX_Resource_Guide.pdf
[22:02] <db90h> there's also one for the CFE
[22:03] <florian__> malbon: in parse_redboot_partitions, how do I use the fis_origin ?
[22:05] <florian__> I guess it is the address of the fis directory ?
[22:31] <florian__> ..
[22:54] <malbon> yes, usually they put the last block offset iirc.
[22:58] <florian__> malbon: thanks
[22:58] <florian__> malbon: do you know why parse_redboot_partitions is defined returns static int, instead of int ?
[22:58] <malbon> florian__: I don't quite see what you mean actually, the redboot partition map is *usually* in the last erase block of the mtd device.
[22:59] <florian__> malbon: sorry I changed of question quickly
[22:59] <malbon> florian__: ok no prob.
[22:59] <florian__> in fact the parse_redboot_partitions symbol is defined to return static int, thus preventing other file from using it as extern symbol
[23:00] <malbon> florian__: my version of redboot.c defines parse_redboot_partitions as int.
[23:00] <florian__> it's just strange, in the linux-2.4 tree, it was not like that
[23:00] <florian__> malbon: ah ok
[23:01] <malbon> you'll probably need to add a define somewhere in redboot.c to make that change where the fis directory is in flash.
[23:03] <florian__> right
[23:03] <florian__> it's ok, I modified the file, just wondered why it was like that
[23:08] <florian__> it's getting real exciting to hack a hardware in fact
[23:09] <malbon> hehe, you should do some redboot ;)
[23:09] <nbd> .
[23:10] <nbd> florian__: the 'static' does not say anything about the return type
[23:10] <nbd> florian__: it just means the function is not exported outside of the c file that it's in
[23:10] <florian__> nbd: sure, but what I was interested in is that it cannot be accessed outside the redboot.c, which I needed because I use in another file
[23:11] <[mbm]> nbd: think we might need to clean up some of the make targets .. seems to take longer for a smp build to remake the firmware (ie no files changed, nothing to do) than it does for a normal build
[23:13] Action: [mbm] hasn't had time to dig through the make debug to figure out why
[23:33] <[mbm]> hmm.. I can already spot some problematic rules
[23:37] <db90h> i wish u guys would set u a visual studio solution for openwrt
[23:37] <db90h> ;p
[23:40] <groz> get me a visual stuido running on linux
[23:40] <groz> and i will
[23:41] <groz> but, couple of things, none of my linux boxes have montiors or keyboards
[23:41] <groz> so it should be something i can use with ssh too
[23:43] <[mbm]> groz: wine
[23:44] <groz> mbm, no x
[23:44] <groz> i dont have it installed
[23:45] <[mbm]> well, with no x it's hard to do "visual" but you can still do mono
[23:45] <groz> but, the little project i'm starting later this afternoon will probably suffice
[23:45] <groz> i'm gonna get a colinux variant of openwrt going
[23:46] <groz> with that up, if one plays judicously with cygwin
[23:46] <groz> should be able to integrate the whole thing into visual studio
[23:49] <[mbm]> lmao
[23:49] <[mbm]> not sure one would want to
[23:49] <groz> i dont, but somebody might
[23:49] <[mbm]> nbd: doh.. the prereq check doesn't check for ncurses
[23:58] <db90h> plz God don't make ur prereq checks as aggressive as freewrt's
[23:59] <db90h> they put in shit released like a couple weeks ago, so instead of using ubuntu packages, i gotta go install them manually
[23:59] <db90h> and there's no need to require these newest versions, afaik
[00:00] --- Sat Jul 29 2006
[12:05] <CIA-17> nico * r4308 /packages/libs/pwlib/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port pwlib to buildroot-ng
[12:08] <CIA-17> nico * r4309 /packages/libs/openh323/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port openh323 to buildroot-ng
[13:36] <florian__> {Nico}: stil had not time to test au1000 ?
[13:58] <florian__> nbd: did you make some tests on the wrt54gx ?
[13:58] <nbd> not yet
[13:59] <florian__> ok, looks like people can have it running, only the wireless driver was not working, probably because there was not mimo enabled wl.o at this thime
[14:01] <florian__> http://forum.openwrt.org/viewtopic.php?pid=30880#p30880
[14:14] <[mbm]> http://code.google.com/
[14:14] <[mbm]> google is now doing opensource project hosting
[14:14] <[mbm]> interesting but the whole thing lacks the typical google polished ui
[14:15] <[mbm]> the 'browse source' is just a link to the svn server itself which is horrible for browsing
[14:16] <[mbm]> example project - http://code.google.com/p/nautilus-scripts/
[14:17] <[mbm]> the really funny thing is that some of the people behind it were also the people that were working for va linux at the time sourceforge was launched
[14:17] <florian__> ah ah :)
[14:18] <florian__> {Nico}: are your porting quagga ? vrrp ? openvpn ?
[14:18] <florian__> because if, no, I will do it
[16:53] <CIA-17> nico * r4310 /packages/libs/bluez-libs/ (. Config.in Makefile ipkg/): port bluez-libs to buildroot-ng
[17:04] <CIA-17> nico * r4311 /packages/utils/bluez-utils/ (. Config.in Makefile ipkg/): port bluez-utils to buildroot-ng
[17:10] <CIA-17> nico * r4312 /packages/libs/bluez-libs/Makefile: remove dependency on libdaemon
[17:52] <nbd> groz: ping
[18:02] <CIA-17> nico * r4313 /packages/libs/libamsel/ (. Makefile): port libamsel to buildroot-ng
[18:05] <CIA-17> nico * r4314 /branches/whiterussian/openwrt/toolchain/uClibc/patches/140-ipv6_missing_sockopts.patch: backport [1729] to WhiteRussian
[18:13] <CIA-17> nico * r4315 /packages/libs/libamsel/ (Config.in ipkg/): oops, forgot to remove unneeded files for -ng from my last commit (again)
[18:15] <CIA-17> nico * r4316 /packages/libs/serdisplib/ (. Config.in Makefile ipkg/): port serdisplib to buildroot-ng
[18:34] <florian__> re
[19:53] <CIA-17> mbm * r4317 /branches/buildroot-ng/openwrt/include/package.mk: fix build issues with SMP compiles
[20:18] <florian__> malbon: ping ?
[20:30] <[mbm]> should probably consider adding distcc or ccache to ng
[20:30] <malbon> florian__: yo
[20:31] <malbon> distcc would be pretty nifty
[20:32] <[mbm]> well, to clarify I meant for building the firmware, not as a package
[20:32] <Bartman007> [mbm]: mmm, that would be cool.
[20:32] <malbon> yeah I thought so.
[20:32] <Bartman007> I could have 3x amd64 3700s building for me.
[20:33] <[mbm]> (a whole army of wrt's might have the same horse power as the old pentium which is being used as an nfs server to host the wrt's gcc ... :P)
[20:33] <Bartman007> gcc on emulated fpu, yey!
[20:33] <[mbm]> Bartman007: I use quad xeons
[20:34] <Bartman007> so compiling openwrt consumes enough power to run a wrt for a month?
[20:34] <[mbm]> a month? let's be realistic here
[20:35] <[mbm]> we're talking 3 months at least
[20:36] <florian__> malbon: do you remember which mtd driver you told me about for redboot parsing reference ?
[20:37] <florian__> [mbm]: did not you make the brilliant performance test on the wiki ;)
[20:37] <[mbm]> florian__: no, I was the one that flamed the hell out of them for doing so
[20:37] <florian__> [mbm]: yes I remember, it's far from being relevant
[20:38] <florian__> but maybe we could try running the spec.org test on the wrt \o/
[20:38] <[mbm]> if you want a useless benchmark, go brag about bogomips
[20:38] <florian__> nah, it does not take enough time !:p
[20:39] <florian__> I want something like 3dmark or another shit
[20:39] <[mbm]> wait .. what about distributed.net or folding@home ?
[20:39] <florian__> this is a good idea
[20:39] <db90h> the question is, are u both being sarcastic, or just one of you
[20:40] <db90h> u underestimate the power of these routers, i've been making love to my overclocked wrt for weeks now
[20:40] <[mbm]> db90h: if it were me I'd be setting up my own locate distributed.net server and sending out aes packets to be decrypted ;)
[20:40] <florian__> we are sarcastic, we are misanthropist at the same time, yay !
[20:41] <db90h> don't use words like misanthropist, that waste considerable time since u know i gotta go look itup
[20:41] <Bartman007> florian__: I like you Europeans, you use worlds that confuse most Americans but are perfectly valid English.
[20:42] <db90h> it would be kinda cool to have an army of wrt's working on a distributed computing project, just for amusement
[20:42] <florian__> Bartman007: lol :)
[20:42] <[mbm]> db90h: is that a complaint against large words, or large words not relating to onanism?
[20:42] <db90h> umm
[20:43] <db90h> hehe
[20:43] <db90h> both it turns out
[20:43] <[mbm]> wondered if you'd look that up
[20:43] <db90h> i was almost too lazy
[20:43] <db90h> but glad i did
[20:44] <db90h> i'll use this often from now on ;p
[20:44] <Bartman007> I had a history teacher in 7th grade who read the dictionary cover to cover, you could tell what letter he was on each day...
[20:45] <[mbm]> lmao
[20:45] <db90h> hehe
[20:45] Action: [mbm] used to have a few 'obscure word of the day' things
[20:46] <[mbm]> tend not to use them because even the english speakers don't know most of them
[20:49] <malbon> florian__: use the sa1100-flash one, it's got a static map and redboot in that one.
[20:49] <florian__> malbon: thanks again, in fact, I found the necessary modifications in the inventel sources :)
[20:49] <malbon> florian__: sorry, I was watching firefly.
[20:50] <florian__> I really wonder why english is so popular and widespread,apart from the fact that they had the biggest empire/colonies ever
[20:50] <[mbm]> good show
[20:50] <florian__> malbon: no pb
[20:50] <malbon> I like it. :)
[20:50] <[mbm]> florian__: that's exactly it
[20:50] <[mbm]> english as a language is one of the most backwards languages
[20:50] <malbon> but its a lot easier to learn than chinese.
[20:51] <florian__> right, only few twice as less as french as far as I remember
[20:51] <[mbm]> I suppose it's easier in the regard that most words aren't gender specific
[20:52] <Bartman007> [mbm]: if rap slang like hyphy is working it's way into the english vernacular, you should use obscure words.
[20:52] <florian__> Bartman007: lol, I don't understand a word of your sentence :)
[20:52] <[mbm]> Bartman007: nah, then I'd look just as elitist as those people that type using 'j00' and 'u' as words
[20:53] <Bartman007> [mbm]: y, r u n0t r33t enough to understand 1337ness?
[20:55] <Bartman007> bah, I can't even stick with it if I try, I missed to==2, that crap is so annoying.
[20:56] <florian__> if you can type quickly you dramatically improve your orthograph
[20:59] <Bartman007> I believe that if you are speaking of spelling alone, definitely doesn't decrease the number of typos
[21:07] <[mbm]> hmm .. looks like there's still some smp issues
[21:07] <[mbm]> make spends forever in the packages directory even when there's nothing to be done
[21:09] <[mbm]> probably have to dig through make and try to reduce the number of targets
[21:12] <florian__> ah ah better use a wrt for cross compiling [mbm] !
[21:12] <florian__> woah I am glad to be in weekend, I become to have packet loss inside my brain
[21:13] <[mbm]> latency is pretty high too
[21:30] <db90h> i'm a little teacup
[21:30] <Bartman007> that's all 2MB of flash is good for.
[21:31] <db90h> speaking of 2mb flash, i just built a compressed CFE
[21:31] <db90h> which ends up at 113KB by default
[21:31] <db90h> giving us 128KB more flash space
[21:32] <Bartman007> wouldn't mind 128KB extra space on his 4MB devices also
[21:32] <Bartman007> +/me
[21:33] <db90h> i used the belkin 7230-4 sources, which include the CFE source.. that unit uses a compressed CFE
[21:35] <florian__> ah with which algo ?
[21:35] <db90h> it uses gzip
[21:35] <db90h> by default
[21:35] <florian__> ok
[21:35] <db90h> i haven't done any tweaking to it, i just got it built after mucking with the belkin sources for a while
[21:36] <florian__> gg
[21:36] <db90h> i'd like to repackage the CFE source so we can all create custom, optionally compressed, CFE image more easily
[21:37] <florian__> yes, but I wonder if we can do it with the actual CFE licensing ?
[21:37] <db90h> shhh
[21:37] <florian__> ah, yes of course, we can do it ;)
[21:37] <db90h> I think i'm in violation already with my vxworks killer, of both Broadcom and Wind River
[21:38] <db90h> but not positive on that
[21:39] <db90h> (Wind River part being in the reversion firmware, where I distribute the VxWorks boot loader)
[21:41] <Bartman007> db90h: would it be possible to just distribute your code, and a script to strip the bootloader out of existing vxworks firmware?
[21:43] <db90h> well, there is technically a way, but it's not trivial... you'd have to write a firmware compatible with the VxWorks boot loader that does the job of reading and storing the VxWorks boot loader
[21:43] <Bartman007> oh hmmmm, I was thinking that the Vx bootloader was included in firmware updates, dunno why....
[21:44] <db90h> no, not that i know of.. though since that capability exists in the firmware format, it's not inconcievable that some vendor somewhere has, or will, do that
[21:46] <florian__> damn, I just found out so many things regarding the bootloader type in the inventel sources !
[21:46] <db90h> The VxWorks boot loader is fucking huge compared to the CFE.. it takes up 384KB of ROM space compressed
[21:46] <florian__> there was complete sources on how handling board initialization for the 2 bootloaders
[21:46] <florian__> that's quite huge
[21:46] <florian__> but maybe there are some code that is not the kernel task then
[21:47] <db90h> well, actually its 383KB, last KB is the BOOTP block (which i've just documented and wrote a tool for, if anybody cares)
[21:47] <db90h> it's got more functionality than CFEs
[21:47] <db90h> it has a built in httpd server
[21:47] <db90h> full tcp/ip stack
[21:47] <db90h> and dhcpd/dhcpc
[21:48] <florian__> ah that's the reason why
[21:48] <db90h> it's a pretty capable OS in its own right ;p
[21:49] <florian__> well, actually it is one of the best hard real time OS ever
[21:50] <db90h> u mean VxWorks itself? yea, i was just joking about its boot loader as an OS
[21:50] <db90h> I don't know much about VxWorks, but i appreciate the fact that it seems to do its job well in less resources than linux
[21:51] <florian__> yes
[21:51] <db90h> though as mbm i think educated me a while back, these aren't technically RTOSes.. maybe VxWorks can be tho
[21:52] <florian__> and actually, apart QNX and VxWorks, I don't know any other hard real time OSes
[21:52] <florian__> it is quite easy to find people to work under VxWorks, QNX, and Linux, for others, I really doubt it is as easy
[21:53] <florian__> I really need to improve my english
[21:55] <db90h> me too
[21:55] <db90h> i'd like to be an embedded developer by profession
[21:56] <db90h> i might start puring that.. it's just hard to afford things like Wind River's SDKs, and any specialized hardware i might need
[21:56] <db90h> err pursuing
[21:57] <db90h> i noticed btw an awesome Broadcom doc on these platforms
[21:57] <florian__> actually I am following telecommunication studies to become engineer (master executive I guess is the equivalent of what I follow in france) but specializing in embedded systems would be great !
[21:57] <db90h> it instructed me, amongst other things, on how to set up gdb to remotely debug the OS
[21:58] <db90h> i'm sure u guys have seen it
[21:58] <db90h> the one in the Belkin 7230-4 sources, and also in others i'm sure
[21:59] <florian__> I did not look at it
[21:59] <db90h> here, let me throw it up on web real quick
[21:59] <db90h> for anyone who hasn't read it.. it's awesome reading
[22:00] <db90h> http://www.bitsum.com/files/BCM47XX_Resource_Guide.pdf
[22:02] <db90h> there's also one for the CFE
[22:03] <florian__> malbon: in parse_redboot_partitions, how do I use the fis_origin ?
[22:05] <florian__> I guess it is the address of the fis directory ?
[22:31] <florian__> ..
[22:54] <malbon> yes, usually they put the last block offset iirc.
[22:58] <florian__> malbon: thanks
[22:58] <florian__> malbon: do you know why parse_redboot_partitions is defined returns static int, instead of int ?
[22:58] <malbon> florian__: I don't quite see what you mean actually, the redboot partition map is *usually* in the last erase block of the mtd device.
[22:59] <florian__> malbon: sorry I changed of question quickly
[22:59] <malbon> florian__: ok no prob.
[22:59] <florian__> in fact the parse_redboot_partitions symbol is defined to return static int, thus preventing other file from using it as extern symbol
[23:00] <malbon> florian__: my version of redboot.c defines parse_redboot_partitions as int.
[23:00] <florian__> it's just strange, in the linux-2.4 tree, it was not like that
[23:00] <florian__> malbon: ah ok
[23:01] <malbon> you'll probably need to add a define somewhere in redboot.c to make that change where the fis directory is in flash.
[23:03] <florian__> right
[23:03] <florian__> it's ok, I modified the file, just wondered why it was like that
[23:08] <florian__> it's getting real exciting to hack a hardware in fact
[23:09] <malbon> hehe, you should do some redboot ;)
[23:09] <nbd> .
[23:10] <nbd> florian__: the 'static' does not say anything about the return type
[23:10] <nbd> florian__: it just means the function is not exported outside of the c file that it's in
[23:10] <florian__> nbd: sure, but what I was interested in is that it cannot be accessed outside the redboot.c, which I needed because I use in another file
[23:11] <[mbm]> nbd: think we might need to clean up some of the make targets .. seems to take longer for a smp build to remake the firmware (ie no files changed, nothing to do) than it does for a normal build
[23:13] Action: [mbm] hasn't had time to dig through the make debug to figure out why
[23:33] <[mbm]> hmm.. I can already spot some problematic rules
[23:37] <db90h> i wish u guys would set u a visual studio solution for openwrt
[23:37] <db90h> ;p
[23:40] <groz> get me a visual stuido running on linux
[23:40] <groz> and i will
[23:41] <groz> but, couple of things, none of my linux boxes have montiors or keyboards
[23:41] <groz> so it should be something i can use with ssh too
[23:43] <[mbm]> groz: wine
[23:44] <groz> mbm, no x
[23:44] <groz> i dont have it installed
[23:45] <[mbm]> well, with no x it's hard to do "visual" but you can still do mono
[23:45] <groz> but, the little project i'm starting later this afternoon will probably suffice
[23:45] <groz> i'm gonna get a colinux variant of openwrt going
[23:46] <groz> with that up, if one plays judicously with cygwin
[23:46] <groz> should be able to integrate the whole thing into visual studio
[23:49] <[mbm]> lmao
[23:49] <[mbm]> not sure one would want to
[23:49] <groz> i dont, but somebody might
[23:49] <[mbm]> nbd: doh.. the prereq check doesn't check for ncurses
[23:58] <db90h> plz God don't make ur prereq checks as aggressive as freewrt's
[23:59] <db90h> they put in shit released like a couple weeks ago, so instead of using ubuntu packages, i gotta go install them manually
[23:59] <db90h> and there's no need to require these newest versions, afaik
[00:00] --- Sat Jul 29 2006