[12:08] <CIA-17> florian * r4120 /branches/buildroot-ng/openwrt/include/target.mk: Add missing brcm63xx-2.6 target to the target list
[12:19] <florian__> Kaloz: ping ?
[12:25] <florian__> db90h: when you constated the power output differences, how did you measured them ?
[13:38] <florian__> malbon: do you have any brcm63xx based device using CFE ?
[14:33] <florian__> ..
[15:07] <florian__> nbd: ping ?
[16:36] <x-db90h> ok, i've extended mksquashfs-lzma and the zlib lzma wrapper so that the lzma compression parameters can be specified on the mksquashfs-lzma command line. Regarding the patch, it is ok to send one that patches the post-openwrt-patched files (final merge in ./build_*)?
[16:40] <florian__> x-db90h: yep
[16:41] <x-db90h> k, thx
[17:17] <florian__> Kaloz: ping ?
[17:17] <Kaloz> pong
[17:18] <florian__> ok, I got few questions for you ;)
[17:18] <florian__> how do I mark files as binary with svn ?
[17:18] <Kaloz> same here, let me know if you have an idea about this - http://pastebin.ca/83245
[17:19] <Kaloz> i'm pretty sure it's some minor typo or something, but can't find why does it wants that damn file
[17:19] <florian__> Kaloz: I think that in your arch/mips/ar531x/Makefile
[17:19] <florian__> you must have something like lib-y instead of obj-y, or just the opposite
[17:21] <florian__> let me check with brcm63xx for instace
[17:21] <florian__> ok, I think it should be obj-y if not already the case
[17:22] <florian__> Kaloz: so no idea for marking files as binary ?
[17:22] <Kaloz> http://svnbook.red-bean.com/en/1.1/apas08.html
[17:23] <Kaloz> svn:mime-type seems to be the answer
[17:23] <Kaloz> florian__: i get that error with obj-y :p
[17:23] <florian__> Kaloz: then try with lib-y ;) ?
[17:26] <florian__> Kaloz: do you remember where you found the sources for the brcm63xx enet driver ?
[17:27] <florian__> I remember it was for 2.4
[17:27] <Kaloz> it compiles with that :P (no idea if this is politically correct btw, everything uses obj-y)
[17:27] <Kaloz> iirc i didn't find those
[17:27] <Kaloz> just some header files
[17:27] <Kaloz> but i'm not sure
[17:27] <florian__> ok
[17:27] <florian__> I thought we had complete sources
[17:28] <florian__> maybe b44 will work as well ?
[17:30] <Kaloz> no idea
[17:32] <malbon> florian__: not currently, but I will do sometime this week.
[17:33] <florian__> ok
[17:34] <florian__> Kaloz: I do not understand if I can put files as binary now they have been added ?
[17:38] <florian__> ok I think I got it
[17:40] <CIA-17> florian * r4121 /docs/rmll2006/ (presentation_en.pdf presentation_fr.pdf): Change file properties for the PDF files, mark them as binary
[17:40] <florian__> Kaloz: is it ok ?
[17:58] <florian__> Kaloz: are you dead ?
[18:10] <malbon> I would take that as a temporary yes. :)
[18:13] <florian__> lol
[18:31] <florian__> anyone knows if the realtek 865x devices are using patent free mips instructions ?
[18:35] <nbd> they are
[18:44] <florian__> ok, so you have to patch the toolchain with the non-mips patches right ?
[18:48] <florian__> is there any interesting work at rtl8180 we can use for the rtl865x port ?
[18:53] <x-db90h> i am sick of pastebin being so damn slow
[18:53] Action: x-db90h yells.. come the fuck on!!
[18:54] <x-db90h> sorry, i hate waiting on stuff
[18:54] <florian__> x-db90h: use pastebin.ca
[18:56] <malbon> or pastebin.co.uk
[18:58] <x-db90h> thx
[18:59] <x-db90h> it added some blank lines for some reason, some paste problem.. http://pastebin.ca/83306
[19:00] <x-db90h> despite the blank lines, that seem ok? that's for white russian rc5 tho.. need to make one for kamikaze still, but gotta expand this damn virtual disk first
[19:02] <x-db90h> i guess i'll need to throw it in the patch folder and see if it works ok too.. this diff procedure has taken 4x longer than the code ;p
[19:04] <x-db90h> i see some minor issues i need to correct too.. gotta recreate it, whopee
[19:06] <florian__> nbd: are there any chances that a patent-free mips toolchain creates valid binaries ?
[19:06] <x-db90h> good luck finding anything that's patent free ;p
[19:07] <x-db90h> damn sw patents are sickening
[19:09] <florian__> http://nonmips.sourceforge.net
[19:12] <x-db90h> glad ppl are making such efforts
[19:13] <x-db90h> in the end, tho something has to be done about this sw patent issue.. as we all know. Else no matter what you do, even a 'hello world' app, you must simply hope nobody chooses to try to enforce some insane patent that your software violates
[19:19] <db90h> ok, here's a better formatted patch with that newline paste issue gone: http://pastebin.ca/83331
[19:19] <db90h> tho the syntax highlting for C++ doesn't do so well with a diff ;p
[19:41] <jr-> florian. i think i saw some bcmenet sources few weeks ago. just can't remember where and was unable to locate them on my disk either. of course it's possible that i remember incorrectly.
[19:42] <florian__> jr-: I remember reading that too ;)
[19:42] <florian__> if you get your hand on it, I take :D
[20:20] <florian__> is it possible to make the calculation of floating value with ash ?
[20:20] <florian__> $((my_value / my_divider)) seems to work only for integers
[20:21] <db90h> i dunno, but you can always multiply away the decimal points ;)
[20:21] <florian__> well I wanted to avoid that, but I think you are right :)
[20:30] <db90h> i tell u what, i'm about starting to like linux well enough to run Windows in a virtual machine instead of linux
[20:30] <db90h> but i wouldn't want to have to go through all the pain and agony of driver issues
[20:37] <malbon> db90h: why don't you run *BSD?
[20:39] <malbon> can mips machines read from anywhere in memory, or does it need to be word aligned?
[20:41] <db90h> never tried a bsd OS.. i suppose i should some time, though I would have expected as many driver and missing application issues as with linux..
[20:42] <malbon> so what's the problem with linux that makes you want to go back to windows?
[20:42] <db90h> i never went to linux to start with, i've always ran it in a virtual machine (vmware)
[20:42] <malbon> on top of?
[20:42] <db90h> windows
[20:43] <db90h> XP, x32
[20:43] <malbon> eugh.
[20:43] <db90h> or x64
[20:43] <db90h> yea, well i'm a low-level Windows develop for many years, I don't want to get into an OS debate, I was merely commenting that I'm liking linux more and more
[20:43] <db90h> err developer
[20:43] <malbon> yes I guessed. :)
[20:44] <malbon> I find Xen quite nifty, so if your machine is quite modern you could try running windows within that.
[20:45] <florian__> Xen is not really dedicated to developpers
[20:45] <florian__> qemu gives better performance for common hardware
[20:45] <florian__> if you have hardware virtualisation then, Xen is good for you
[20:45] <malbon> nope, but as a virtualisation platform it's pretty good.
[20:46] <db90h> i wish i did have a CPU /w virtualization extensions.. maybe someday
[20:46] <malbon> yes, Pentium D900's and AMD am2 socket processors have hardware-foo
[20:47] <florian__> a Xen developper told me it was better to use Xen on hardware virtualisation ready machines rather than the hardware virtualisation itself
[20:47] <florian__> don't know why
[20:53] <malbon> db90h: your bsp hack, can that execute programs rather than just blowing away the existing boot code?
[21:06] <db90h> depends on what you mean.. by BSP hack you mean what? All I did was reverse engineer the firmware image format.. and the format has the capabibility to replace the BSP itself, so I used that to replace the CFE
[21:07] <db90h> err to flash the CFE over the BSP
[21:07] <malbon> ok, so you had the CFE image in there and the BSP took care of the rest?
[21:07] <db90h> i wrote a tool to manipulate firmware images.. so you can use it to flash new firmware, if that's your question
[21:08] <db90h> (without wiping out the BSP)
[21:08] <malbon> I'm thinking about non-broadcom uses of it.
[21:09] <malbon> so on atheros it'd be nice to load a redboot which you can telnet to before burning an image that isn't quite right...
[21:09] <db90h> if the firmare image format is consistent with that I encountered on the WRT54G/GS v5-v6 then you could definitely use it to slap a new boot loader on there..
[21:11] <malbon> it's a nifty tool for non-linux friendly devices especially if it is platform indipendent. I assume there is a linux command line version? ;)
[21:12] <db90h> i wrote it to be linux-ready, tho never built it for linux ;p.. probably require a few minor tweaks, but nothing major
[21:12] <malbon> I'll test it against some vxworks devices I have here if you can make it friendly
[21:12] <malbon> no rush though.
[21:12] <db90h> it is friendly<g>.. oh, you mean linux friendly ;p
[21:13] <db90h> yea, i've been meaning to do that.. it'll probably take all of 10 mins to fix up for linux, so maybe I'll do that later today
[21:13] <malbon> I don't have a win32 platform (at all) so running it would be a bit of a task for me.
[21:13] <db90h> i would love to know just how standard this format is.. if other vxworks devices use the same format, then that's awesome
[21:14] <db90h> i'll fix it up in just a minute, just for you ;)
[21:14] <db90h> but if it takes more than 30 minutes of my time, then I will curse you like you've never been cursed<g>.. fortunately for you, you won't be able to hear me ;)
[21:16] <malbon> I'm sure there will be no curse you'll be able to think up I haven't ben called before.
[21:16] <db90h> hehe, i am pretty creative ;p
[21:16] <db90h> you got any specific devices in mind here?
[21:17] <malbon> yes, the wrt55ag Atheros AP30 design.
[21:17] <malbon> the wgt614v7 Atheros ap61 design
[21:17] <malbon> I am pretty sure I have a few more in my box o goodies.
[21:18] <malbon> wgu634 which is another ap30 design.
[21:18] <florian__> hu you have ugly devices malbon :)
[21:19] <db90h> looking at a wrt55ag v2 firmware image, it is definitely not the same format
[21:19] <db90h> well, i'll let you figure out if any are.. please let me know if you discover some
[21:19] <malbon> I like netgear ones, don't ask me why.
[21:20] <malbon> certainly will.
[21:20] <malbon> florian__: why ugly?
[21:21] <florian__> because it's atheror :p
[21:21] <florian__> atheros sorry
[21:21] <malbon> but the radios are good in atheros gear.
[21:21] <florian__> even if it could be understood as a joke
[21:26] <florian__> yep radios are good, but SoC are far from being easily supported
[21:26] <malbon> :)
[21:27] <db90h> what's the issue with supporting Atheros chips?
[21:27] <db90h> i mean, i knwo there is some support.. like madwifi, but u infer there's problems..
[21:28] <db90h> err imply
[21:28] <malbon> there are problems with madwifi, the main problem is that they use ancient bin utils to make that hal, which in turn on mips makes it incompatible with modern compilers
[21:29] <db90h> ah, ic
[21:29] <h3sp4wn> Madwifi is rock solid on my wgt634u
[21:30] <h3sp4wn> better than nas ever was (with hostapd)
[21:30] <malbon> yeah, it is on my wgt634u too, but that's not a atheros soc, it is an atheros pci card.
[21:30] <db90h> wgt634u has a usb port doesn't it? I always thought those were very economical, if its the model i'm thinking of
[21:30] <malbon> yes they do have usb2 port, but I have never used it.
[21:31] <db90h> my next router will be one with a usb port for sure
[21:31] <db90h> i may never use it either, but it'd be nice to have
[21:31] <malbon> getting few and far between with usb now :(
[21:32] <db90h> problem is rationalizing buying another router when i have no real use for one
[21:32] <malbon> I got a Compex NetPassage 28G which has got 4xusb2 on it.
[21:32] <florian__> I need to find a device with broadcom cryptonet device :)
[21:35] <db90h> yea, me too.. nbd gave me a job to reverse engineer the broadcom nas encryption module, but i've slacked on it (mostly because I have no device to 'really' use it with)
[21:36] <db90h> the WRTSLG54 (or whatever) I'd like to have, but can't afford it at all
[21:36] <db90h> WRTSL54GS
[21:37] <db90h> then there's the 802.11n issue.. whether to wait for it or not
[21:38] <db90h> (or buy a pre-802.11n, etc..).. i wish i had more money, but will never have any cash if I keep mucking around with you guys ;p
[21:40] <malbon> for accellerated encryption or just to work out how nas communicated with the driver?
[21:40] <db90h> to figure out the encryption algorithm and other specifics
[21:41] <malbon> hmm, I thought nas only did the hostapd and wpa_supplicant job, and the encryption was done by the wmac like on most other wireless cards.
[21:41] <db90h> hehe
[21:41] <db90h> i meant NAS == network accessible storage
[21:41] <db90h> not NAS wifi driver, sorry ;)
[21:42] <malbon> ah
[21:43] <florian__> malbon: to port the BSD driver to linux and make it use the crypto API
[21:43] <malbon> acronym soup. :)
[21:43] <db90h> hehe, yea, i bet you were really scratching ur head on that one ;p
[21:43] <florian__> lol
[21:44] <florian__> well my main problem is that my wl500-deluxe is bricked, I mean electrically
[21:44] <malbon> florian__: I think that'd be pretty cool, 75Mbs of IP sec of a wgt634u would just be uber-cool.
[21:44] <florian__> and is not longer sold in europe
[21:44] <malbon> jtag?
[21:44] <florian__> malbon: well basically there are not issues with low frequency device
[21:44] <florian__> but for instance, if you have a P4
[21:45] <florian__> it is sometimes better to make calculation inside the proc
[21:45] <florian__> rather than generating interrupts in order to use the Crypto card
[21:45] <florian__> so the drive should be "frequency" adaptative I think
[21:45] <malbon> dma :)
[21:46] <florian__> that's it
[21:47] <florian__> unless we can directly make the network card "speak" with the "crypto" card
[21:47] <florian__> which I highly doubt
[21:48] <florian__> in fact, there are dma issues with low latency devices anyway :)
[21:49] <florian__> and all the more as IPsec is giving fixed size packets to encrypt
[21:49] <florian__> so it would only be interesting for ssh stuff, or VPN SSL maybe
[21:50] <malbon> there are always issues, but it should be quicker than polled or interrupt driven transfers as it should be able to continue and allow other interrupts to be processed.
[21:50] <florian__> right
[21:53] <florian__> well I am tempted to say, without driver, how can we now ?
[21:53] <florian__> so I say it
[21:54] <malbon> very true, but then the function isn't used at present on the commercial offerings, so it really would be an added benefit.
[21:55] <florian__> rigth
[21:55] <florian__> damn, I type like a french cow
[21:55] <malbon> not watching the football?
[21:56] <db90h> french cows can type?! holy shit.. man, this should be slashdot'd right away!
[21:57] <db90h> if you guys only knew how much i amused myself
[21:57] <malbon> there so advanced now, that they produce cheese straight from the udder. :P
[21:57] <db90h> bought time they started integrating that into the cow ;)
[21:58] <florian__> lol
[21:58] <malbon> hehe, the unions wouldn't allow it ;)
[21:58] <florian__> malbon: well no I listen to the football match
[21:58] <florian__> I only care for the result
[21:58] <malbon> which is?
[21:58] <florian__> and as I am 25% italian, I am divided :)
[21:59] <malbon> difficult, you must be fighting with yourself.
[21:59] <florian__> yep
[21:59] <florian__> sometimes my left hand hits me, sometimes the right one
[22:00] <malbon> you could take the third option?
[22:00] <florian__> not sure
[22:00] <malbon> not support either side?
[22:00] <florian__> sure I do not care really
[22:00] <malbon> I was getting that impression.
[22:01] <florian__> 0-0 for the moment as far as I have understood
[22:02] <malbon> hmm, endianness does my head in. :(
[22:03] <florian__> what's the link with football ;) ?
[22:03] <malbon> none.
[22:05] <florian__> ah sorry 1-1 :)
[23:24] <db90h> buildroot-ng is nice
[23:24] <db90h> first time i've taken a look at it
[00:00] --- Mon Jul 10 2006
[12:19] <florian__> Kaloz: ping ?
[12:25] <florian__> db90h: when you constated the power output differences, how did you measured them ?
[13:38] <florian__> malbon: do you have any brcm63xx based device using CFE ?
[14:33] <florian__> ..
[15:07] <florian__> nbd: ping ?
[16:36] <x-db90h> ok, i've extended mksquashfs-lzma and the zlib lzma wrapper so that the lzma compression parameters can be specified on the mksquashfs-lzma command line. Regarding the patch, it is ok to send one that patches the post-openwrt-patched files (final merge in ./build_*)?
[16:40] <florian__> x-db90h: yep
[16:41] <x-db90h> k, thx
[17:17] <florian__> Kaloz: ping ?
[17:17] <Kaloz> pong
[17:18] <florian__> ok, I got few questions for you ;)
[17:18] <florian__> how do I mark files as binary with svn ?
[17:18] <Kaloz> same here, let me know if you have an idea about this - http://pastebin.ca/83245
[17:19] <Kaloz> i'm pretty sure it's some minor typo or something, but can't find why does it wants that damn file
[17:19] <florian__> Kaloz: I think that in your arch/mips/ar531x/Makefile
[17:19] <florian__> you must have something like lib-y instead of obj-y, or just the opposite
[17:21] <florian__> let me check with brcm63xx for instace
[17:21] <florian__> ok, I think it should be obj-y if not already the case
[17:22] <florian__> Kaloz: so no idea for marking files as binary ?
[17:22] <Kaloz> http://svnbook.red-bean.com/en/1.1/apas08.html
[17:23] <Kaloz> svn:mime-type seems to be the answer
[17:23] <Kaloz> florian__: i get that error with obj-y :p
[17:23] <florian__> Kaloz: then try with lib-y ;) ?
[17:26] <florian__> Kaloz: do you remember where you found the sources for the brcm63xx enet driver ?
[17:27] <florian__> I remember it was for 2.4
[17:27] <Kaloz> it compiles with that :P (no idea if this is politically correct btw, everything uses obj-y)
[17:27] <Kaloz> iirc i didn't find those
[17:27] <Kaloz> just some header files
[17:27] <Kaloz> but i'm not sure
[17:27] <florian__> ok
[17:27] <florian__> I thought we had complete sources
[17:28] <florian__> maybe b44 will work as well ?
[17:30] <Kaloz> no idea
[17:32] <malbon> florian__: not currently, but I will do sometime this week.
[17:33] <florian__> ok
[17:34] <florian__> Kaloz: I do not understand if I can put files as binary now they have been added ?
[17:38] <florian__> ok I think I got it
[17:40] <CIA-17> florian * r4121 /docs/rmll2006/ (presentation_en.pdf presentation_fr.pdf): Change file properties for the PDF files, mark them as binary
[17:40] <florian__> Kaloz: is it ok ?
[17:58] <florian__> Kaloz: are you dead ?
[18:10] <malbon> I would take that as a temporary yes. :)
[18:13] <florian__> lol
[18:31] <florian__> anyone knows if the realtek 865x devices are using patent free mips instructions ?
[18:35] <nbd> they are
[18:44] <florian__> ok, so you have to patch the toolchain with the non-mips patches right ?
[18:48] <florian__> is there any interesting work at rtl8180 we can use for the rtl865x port ?
[18:53] <x-db90h> i am sick of pastebin being so damn slow
[18:53] Action: x-db90h yells.. come the fuck on!!
[18:54] <x-db90h> sorry, i hate waiting on stuff
[18:54] <florian__> x-db90h: use pastebin.ca
[18:56] <malbon> or pastebin.co.uk
[18:58] <x-db90h> thx
[18:59] <x-db90h> it added some blank lines for some reason, some paste problem.. http://pastebin.ca/83306
[19:00] <x-db90h> despite the blank lines, that seem ok? that's for white russian rc5 tho.. need to make one for kamikaze still, but gotta expand this damn virtual disk first
[19:02] <x-db90h> i guess i'll need to throw it in the patch folder and see if it works ok too.. this diff procedure has taken 4x longer than the code ;p
[19:04] <x-db90h> i see some minor issues i need to correct too.. gotta recreate it, whopee
[19:06] <florian__> nbd: are there any chances that a patent-free mips toolchain creates valid binaries ?
[19:06] <x-db90h> good luck finding anything that's patent free ;p
[19:07] <x-db90h> damn sw patents are sickening
[19:09] <florian__> http://nonmips.sourceforge.net
[19:12] <x-db90h> glad ppl are making such efforts
[19:13] <x-db90h> in the end, tho something has to be done about this sw patent issue.. as we all know. Else no matter what you do, even a 'hello world' app, you must simply hope nobody chooses to try to enforce some insane patent that your software violates
[19:19] <db90h> ok, here's a better formatted patch with that newline paste issue gone: http://pastebin.ca/83331
[19:19] <db90h> tho the syntax highlting for C++ doesn't do so well with a diff ;p
[19:41] <jr-> florian. i think i saw some bcmenet sources few weeks ago. just can't remember where and was unable to locate them on my disk either. of course it's possible that i remember incorrectly.
[19:42] <florian__> jr-: I remember reading that too ;)
[19:42] <florian__> if you get your hand on it, I take :D
[20:20] <florian__> is it possible to make the calculation of floating value with ash ?
[20:20] <florian__> $((my_value / my_divider)) seems to work only for integers
[20:21] <db90h> i dunno, but you can always multiply away the decimal points ;)
[20:21] <florian__> well I wanted to avoid that, but I think you are right :)
[20:30] <db90h> i tell u what, i'm about starting to like linux well enough to run Windows in a virtual machine instead of linux
[20:30] <db90h> but i wouldn't want to have to go through all the pain and agony of driver issues
[20:37] <malbon> db90h: why don't you run *BSD?
[20:39] <malbon> can mips machines read from anywhere in memory, or does it need to be word aligned?
[20:41] <db90h> never tried a bsd OS.. i suppose i should some time, though I would have expected as many driver and missing application issues as with linux..
[20:42] <malbon> so what's the problem with linux that makes you want to go back to windows?
[20:42] <db90h> i never went to linux to start with, i've always ran it in a virtual machine (vmware)
[20:42] <malbon> on top of?
[20:42] <db90h> windows
[20:43] <db90h> XP, x32
[20:43] <malbon> eugh.
[20:43] <db90h> or x64
[20:43] <db90h> yea, well i'm a low-level Windows develop for many years, I don't want to get into an OS debate, I was merely commenting that I'm liking linux more and more
[20:43] <db90h> err developer
[20:43] <malbon> yes I guessed. :)
[20:44] <malbon> I find Xen quite nifty, so if your machine is quite modern you could try running windows within that.
[20:45] <florian__> Xen is not really dedicated to developpers
[20:45] <florian__> qemu gives better performance for common hardware
[20:45] <florian__> if you have hardware virtualisation then, Xen is good for you
[20:45] <malbon> nope, but as a virtualisation platform it's pretty good.
[20:46] <db90h> i wish i did have a CPU /w virtualization extensions.. maybe someday
[20:46] <malbon> yes, Pentium D900's and AMD am2 socket processors have hardware-foo
[20:47] <florian__> a Xen developper told me it was better to use Xen on hardware virtualisation ready machines rather than the hardware virtualisation itself
[20:47] <florian__> don't know why
[20:53] <malbon> db90h: your bsp hack, can that execute programs rather than just blowing away the existing boot code?
[21:06] <db90h> depends on what you mean.. by BSP hack you mean what? All I did was reverse engineer the firmware image format.. and the format has the capabibility to replace the BSP itself, so I used that to replace the CFE
[21:07] <db90h> err to flash the CFE over the BSP
[21:07] <malbon> ok, so you had the CFE image in there and the BSP took care of the rest?
[21:07] <db90h> i wrote a tool to manipulate firmware images.. so you can use it to flash new firmware, if that's your question
[21:08] <db90h> (without wiping out the BSP)
[21:08] <malbon> I'm thinking about non-broadcom uses of it.
[21:09] <malbon> so on atheros it'd be nice to load a redboot which you can telnet to before burning an image that isn't quite right...
[21:09] <db90h> if the firmare image format is consistent with that I encountered on the WRT54G/GS v5-v6 then you could definitely use it to slap a new boot loader on there..
[21:11] <malbon> it's a nifty tool for non-linux friendly devices especially if it is platform indipendent. I assume there is a linux command line version? ;)
[21:12] <db90h> i wrote it to be linux-ready, tho never built it for linux ;p.. probably require a few minor tweaks, but nothing major
[21:12] <malbon> I'll test it against some vxworks devices I have here if you can make it friendly
[21:12] <malbon> no rush though.
[21:12] <db90h> it is friendly<g>.. oh, you mean linux friendly ;p
[21:13] <db90h> yea, i've been meaning to do that.. it'll probably take all of 10 mins to fix up for linux, so maybe I'll do that later today
[21:13] <malbon> I don't have a win32 platform (at all) so running it would be a bit of a task for me.
[21:13] <db90h> i would love to know just how standard this format is.. if other vxworks devices use the same format, then that's awesome
[21:14] <db90h> i'll fix it up in just a minute, just for you ;)
[21:14] <db90h> but if it takes more than 30 minutes of my time, then I will curse you like you've never been cursed<g>.. fortunately for you, you won't be able to hear me ;)
[21:16] <malbon> I'm sure there will be no curse you'll be able to think up I haven't ben called before.
[21:16] <db90h> hehe, i am pretty creative ;p
[21:16] <db90h> you got any specific devices in mind here?
[21:17] <malbon> yes, the wrt55ag Atheros AP30 design.
[21:17] <malbon> the wgt614v7 Atheros ap61 design
[21:17] <malbon> I am pretty sure I have a few more in my box o goodies.
[21:18] <malbon> wgu634 which is another ap30 design.
[21:18] <florian__> hu you have ugly devices malbon :)
[21:19] <db90h> looking at a wrt55ag v2 firmware image, it is definitely not the same format
[21:19] <db90h> well, i'll let you figure out if any are.. please let me know if you discover some
[21:19] <malbon> I like netgear ones, don't ask me why.
[21:20] <malbon> certainly will.
[21:20] <malbon> florian__: why ugly?
[21:21] <florian__> because it's atheror :p
[21:21] <florian__> atheros sorry
[21:21] <malbon> but the radios are good in atheros gear.
[21:21] <florian__> even if it could be understood as a joke
[21:26] <florian__> yep radios are good, but SoC are far from being easily supported
[21:26] <malbon> :)
[21:27] <db90h> what's the issue with supporting Atheros chips?
[21:27] <db90h> i mean, i knwo there is some support.. like madwifi, but u infer there's problems..
[21:28] <db90h> err imply
[21:28] <malbon> there are problems with madwifi, the main problem is that they use ancient bin utils to make that hal, which in turn on mips makes it incompatible with modern compilers
[21:29] <db90h> ah, ic
[21:29] <h3sp4wn> Madwifi is rock solid on my wgt634u
[21:30] <h3sp4wn> better than nas ever was (with hostapd)
[21:30] <malbon> yeah, it is on my wgt634u too, but that's not a atheros soc, it is an atheros pci card.
[21:30] <db90h> wgt634u has a usb port doesn't it? I always thought those were very economical, if its the model i'm thinking of
[21:30] <malbon> yes they do have usb2 port, but I have never used it.
[21:31] <db90h> my next router will be one with a usb port for sure
[21:31] <db90h> i may never use it either, but it'd be nice to have
[21:31] <malbon> getting few and far between with usb now :(
[21:32] <db90h> problem is rationalizing buying another router when i have no real use for one
[21:32] <malbon> I got a Compex NetPassage 28G which has got 4xusb2 on it.
[21:32] <florian__> I need to find a device with broadcom cryptonet device :)
[21:35] <db90h> yea, me too.. nbd gave me a job to reverse engineer the broadcom nas encryption module, but i've slacked on it (mostly because I have no device to 'really' use it with)
[21:36] <db90h> the WRTSLG54 (or whatever) I'd like to have, but can't afford it at all
[21:36] <db90h> WRTSL54GS
[21:37] <db90h> then there's the 802.11n issue.. whether to wait for it or not
[21:38] <db90h> (or buy a pre-802.11n, etc..).. i wish i had more money, but will never have any cash if I keep mucking around with you guys ;p
[21:40] <malbon> for accellerated encryption or just to work out how nas communicated with the driver?
[21:40] <db90h> to figure out the encryption algorithm and other specifics
[21:41] <malbon> hmm, I thought nas only did the hostapd and wpa_supplicant job, and the encryption was done by the wmac like on most other wireless cards.
[21:41] <db90h> hehe
[21:41] <db90h> i meant NAS == network accessible storage
[21:41] <db90h> not NAS wifi driver, sorry ;)
[21:42] <malbon> ah
[21:43] <florian__> malbon: to port the BSD driver to linux and make it use the crypto API
[21:43] <malbon> acronym soup. :)
[21:43] <db90h> hehe, yea, i bet you were really scratching ur head on that one ;p
[21:43] <florian__> lol
[21:44] <florian__> well my main problem is that my wl500-deluxe is bricked, I mean electrically
[21:44] <malbon> florian__: I think that'd be pretty cool, 75Mbs of IP sec of a wgt634u would just be uber-cool.
[21:44] <florian__> and is not longer sold in europe
[21:44] <malbon> jtag?
[21:44] <florian__> malbon: well basically there are not issues with low frequency device
[21:44] <florian__> but for instance, if you have a P4
[21:45] <florian__> it is sometimes better to make calculation inside the proc
[21:45] <florian__> rather than generating interrupts in order to use the Crypto card
[21:45] <florian__> so the drive should be "frequency" adaptative I think
[21:45] <malbon> dma :)
[21:46] <florian__> that's it
[21:47] <florian__> unless we can directly make the network card "speak" with the "crypto" card
[21:47] <florian__> which I highly doubt
[21:48] <florian__> in fact, there are dma issues with low latency devices anyway :)
[21:49] <florian__> and all the more as IPsec is giving fixed size packets to encrypt
[21:49] <florian__> so it would only be interesting for ssh stuff, or VPN SSL maybe
[21:50] <malbon> there are always issues, but it should be quicker than polled or interrupt driven transfers as it should be able to continue and allow other interrupts to be processed.
[21:50] <florian__> right
[21:53] <florian__> well I am tempted to say, without driver, how can we now ?
[21:53] <florian__> so I say it
[21:54] <malbon> very true, but then the function isn't used at present on the commercial offerings, so it really would be an added benefit.
[21:55] <florian__> rigth
[21:55] <florian__> damn, I type like a french cow
[21:55] <malbon> not watching the football?
[21:56] <db90h> french cows can type?! holy shit.. man, this should be slashdot'd right away!
[21:57] <db90h> if you guys only knew how much i amused myself
[21:57] <malbon> there so advanced now, that they produce cheese straight from the udder. :P
[21:57] <db90h> bought time they started integrating that into the cow ;)
[21:58] <florian__> lol
[21:58] <malbon> hehe, the unions wouldn't allow it ;)
[21:58] <florian__> malbon: well no I listen to the football match
[21:58] <florian__> I only care for the result
[21:58] <malbon> which is?
[21:58] <florian__> and as I am 25% italian, I am divided :)
[21:59] <malbon> difficult, you must be fighting with yourself.
[21:59] <florian__> yep
[21:59] <florian__> sometimes my left hand hits me, sometimes the right one
[22:00] <malbon> you could take the third option?
[22:00] <florian__> not sure
[22:00] <malbon> not support either side?
[22:00] <florian__> sure I do not care really
[22:00] <malbon> I was getting that impression.
[22:01] <florian__> 0-0 for the moment as far as I have understood
[22:02] <malbon> hmm, endianness does my head in. :(
[22:03] <florian__> what's the link with football ;) ?
[22:03] <malbon> none.
[22:05] <florian__> ah sorry 1-1 :)
[23:24] <db90h> buildroot-ng is nice
[23:24] <db90h> first time i've taken a look at it
[00:00] --- Mon Jul 10 2006