[00:32] <CIA-17> mbm * r4050 /branches/buildroot-ng/openwrt/target/linux/aruba-2.6/patches/002-irq.patch: cosmetic fix for irq error messages
[00:34] <CIA-17> mbm * r4051 /branches/buildroot-ng/openwrt/rules.mk: cleanup
[02:04] <CIA-17> mbm * r4052 /branches/buildroot-ng/openwrt/include/kernel.mk: set defaults for .pkginfo
[02:49] <CIA-17> mbm * r4053 /branches/buildroot-ng/openwrt/ (3 files in 2 dirs): .pkginfo cleanup #2
[10:45] <florian__> as anyone here played with the isl3893 SoC ?
[10:45] <florian__> I ignored I had one 3 years ago :)
[10:55] <florian__> Kaloz: I tried once again to boot my custo brcm63xx kernel, but it did not work, any idea why it could not ?
[10:56] <florian__> the file is in the right ELF format, I tried appending console option, but it did nothing
[10:58] <Kaloz> florian__: mmuless crap :/
[10:58] <Kaloz> florian__: well, it can have a bunch of reasons for not booting
[10:59] <Kaloz> did you add the serial driver ?
[10:59] <Kaloz> it uses it's own driver
[10:59] <florian__> I think so, if you are talking about the serial driver in arch/
[10:59] <florian__> I think I also need to check if the driver in char/serial was compiled built-in right ?
[11:02] Action: [mbm] ponders a TAINTED setting for the openwrt config
[11:02] <florian__> [mbm]: which one ?
[11:02] <[mbm]> look in #openwrt and you'll see what I mean
[11:03] <Kaloz> you mean the busybox tainted module support?
[11:04] <[mbm]> no, I mean "some idiot has screwed with the configuration, you can probably ignore his bug reports"
[11:05] <florian__> ah ok
[11:05] <florian__> I tought you were talking about tainted module support which is *slightly* different ;)
[11:09] <florian__> why would he change is default mtab filename ?
[11:10] <[mbm]> why would he even want an mtab
[11:10] <florian__> actually I don't know
[11:11] <florian__> it's probably more common to mount things at boot time, but a script could do it as well
[11:11] <[mbm]> no, that's fstab
[11:11] <[mbm]> mtab keeps track of what you've mounted
[11:11] <[mbm]> but so does /proc/mounts
[11:12] <[mbm]> which busybox uses by default
[11:12] <florian__> yep sorry
[11:17] <florian__> what are the LED options in newer kernels ?
[11:17] <[mbm]> haven't really looked into it
[17:11] <tziOm> What do I do if I want to compile something to run on rc5? Can I download some devel kit?
[17:24] <tziOm> Is libtorrent/rtorrent ported to rc5?
[17:30] <common> port yourself
[17:30] <nbd> yay... finally it works
[17:31] <nbd> writable jffs2 on first boot :)
[17:31] <nbd> (on broadcom)
[17:31] <tziOm> common, as docs say: no need to reinvent the wheel...
[17:32] <tziOm> if someone has ported it, why should I spend time on it?
[17:32] <common> check ipkg.be
[17:34] <tziOm> has to be something wrong with that one... no hits on samba
[17:43] <db90h> nice nbd
[17:44] <db90h> congrats ;)
[17:56] <h3sp4wn> tziOm: macsat's site has ctorrent on it (as I told you yesterday)
[18:03] <db90h> i heard using torrents causes birth defects anyway
[18:03] <db90h> ..sometimes, i amuse myself beyond belief.. beyond words.. beyond comprehension
[18:04] <db90h> (and I still haven't went blind)
[18:05] <common> not using torrent causes ...
[18:11] <nbd> amnesia?
[18:14] <db90h> i think he was implying it causes me ;p
[18:28] <CIA-17> nbd * r4054 /branches/buildroot-ng/openwrt/target/linux/generic-2.4/patches/227-jffs2_eofdetect.patch: jffs2 eof detect: mark erased blocks after erasing them
[18:30] <CIA-17> nbd * r4055 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/004-flash.patch: broadcom 2.4: update the trx for jffs2 directly in the flash map driver
[18:39] <andersee> nbd: morning
[18:39] <nbd> just curious... have you looked at buildroot-ng already?
[18:40] <andersee> quite a while back, but I expect it has evolved since...
[18:40] <nbd> yeah, it's really usable now
[18:40] <nbd> it'll replace the stuff in trunk/ soon
[18:57] <[mbm]> http://kerneltrap.org/node/6750
[19:01] <nbd> interesting
[19:01] <[mbm]> yep
[19:01] <nbd> btw. porting the flash map change to 2.6 now
[19:01] <nbd> 2.4 seems to work well
[19:01] <[mbm]> saw that
[19:01] <[mbm]> I think I have a working copy of the trace/verbose stuff based on MAKECMDGOALS
[19:02] <[mbm]> just doing a few more test builds
[19:02] <nbd> great
[19:02] <[mbm]> haven't decided what to do with the V=
[19:02] <[mbm]> so right no it's just hardcoded for minimum verbosity
[19:03] <nbd> as a first step we could implement three levels:
[19:03] <nbd> submake tracing, tracing + strerr, everything
[19:03] <nbd> s/strerr/stderr
[19:03] <[mbm]> right now you get trace+stderr
[19:03] <[mbm]> and full verbosity would be easy enough to add
[19:04] <[mbm]> but I haven't figured out the levels between
[19:04] <nbd> k
[19:05] <nbd> what i haven't figured out yet is how to add the jffs2 eof markers to the squashfs images
[19:05] <nbd> because the squashfs is only aligned to 1024
[19:06] <nbd> i could just put 128k of 1k-aligned end markers in there
[19:06] <nbd> but maybe there's a better way
[19:07] <nbd> i don't want to increase the squashfs alignment
[19:07] <[mbm]> I'm really surprised you didn't just hack mkfs.jffs2
[19:07] <nbd> why should i?
[19:07] <nbd> with jffs2 it's easy
[19:08] <[mbm]> anywyas, what's so tricky about squashfs?
[19:08] <nbd> i need to hit a specific spot with the end marker
[19:08] <nbd> and it needs to be inside the trx
[19:08] <nbd> and the trx utility doesn't align it to blocksize for me
[19:09] <[mbm]> you just need to pad the image you feed into the trx
[19:09] <nbd> but i don't know how much i should pad the squashfs
[19:09] <nbd> since it doesn't begin on an eraseblock
[19:09] <nbd> or at least not necessarily
[19:10] <CIA-17> nbd * r4056 /branches/buildroot-ng/openwrt/target/linux/brcm-2.6/config: update kernel config for brcm-2.6
[19:14] <CIA-17> mbm * r4057 /branches/buildroot-ng/openwrt/include/verbose.mk: new verbose/trace script for make
[19:14] <[mbm]> there's the script
[19:14] <[mbm]> I didn't add the hooks for it yet
[19:15] <[mbm]> basically just needs an include from rules.mk and the toplevel makefile
[19:15] <[mbm]> no other changes to any of the makefiles
[19:18] <CIA-17> nbd * r4058 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/004-flash.patch: add copyright statement
[19:36] <nbd> why did you add the KBUILD_VERBOSE line?
[19:36] <[mbm]> for later use
[19:39] <nbd> looks good
[19:59] <CIA-17> nbd * r4059 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/004-flash.patch: fix partition align
[20:23] <nbd> damn... something's wrong with 2.6
[20:23] <[mbm]> ?
[20:23] <nbd> it locks up when trying to erase from the jffs2 scanning function
[20:23] <nbd> but it works from the mtd map driver
[20:24] <[mbm]> did you forget an unlock?
[20:27] <nbd> i did not lock or unlock anything
[20:28] <nbd> and the deadlock seems to happen in the cfi_cmdset_0001 driver
[20:33] <nbd> hmm
[20:33] <nbd> it loops querying the chip's status
[20:33] <nbd> which is always ok
[20:41] <nbd> any ideas?
[20:48] <crazy_imp> is it possible to get ipkg working to check if it's the right version of whiterussian or kamikaze build (and here also between the kernel 2.4 & 2.6) and block the installation of a package which was compiled for another version of openwrt? (something similar to the blocker's in gentoo-portage)
[20:49] <[mbm]> crazy_imp: ipkg's dependancy handling is broken
[20:50] <crazy_imp> [mbm]: so theoreticaly this feature is allready in there but not working?
[20:53] <[mbm]> yep.
[21:00] <nbd> damn ... this mtd crap still doesn't work
[21:11] <nbd> [mbm]: oops... i know what's wrong
[21:12] <nbd> [mbm]: the chip is still locked :)
[21:19] <nbd> hmm... no
[21:19] <nbd> unlock didn't help
[21:19] <nbd> in fact unlock crashes it again
[21:21] <nbd> [mbm]: have you tested the jffs2 patch on aruba?
[21:33] <crazy_imp> nbd: i think i solved my problem with the ide support, #'ed all the lines in /etc/modules.d/20-ide and loaded them by hand, in this order: ide-core, pdc202xx_old , ide-detect, ide-disk (or is there an change in kamikaze so it works now, but a fact against this is it doesn't worked if the ide-* modules got loaded and the pdc202xx_old wasn't loaded. so i think it's only a question of the order :D) <
[21:33] <crazy_imp> - maybe someone can confirm this?
[21:34] <nbd> thanks. i'll look into it
[21:34] <crazy_imp> no problem :D now i can finish my openwrt alarm clock :)
[21:43] <nbd> [mbm]: fixed it by moving it to a different function _and_ doing the unlock
[21:44] <nbd> so it was probably a locking issue as well
[21:51] <CIA-17> nbd * r4060 /branches/buildroot-ng/openwrt/target/linux/generic-2.6/patches/204-jffs2_eofdetect.patch: fix the jffs2 eof detect patch for linux 2.6 (unlocking required for erasing blocks)
[21:53] <CIA-17> nbd * r4061 /branches/buildroot-ng/openwrt/target/linux/brcm-2.6/patches/002-flash-map.patch: port the brcm-2.4 flash map driver changes to 2.6
[22:00] <[mbm]> nbd: nope, been busy with other stuff
[22:07] <nbd> ok. really weird, that the 2.4 stuff worked without unlocking
[22:07] <nbd> i'm going to backport my 2.6 changes anyway
[22:21] <CIA-17> nbd * r4062 /branches/buildroot-ng/openwrt/target/linux/generic-2.4/patches/227-jffs2_eofdetect.patch: backport jffs2 eof detect changes from 2.6
[22:22] <[mbm]> hmm I'm noticing the kernel config being run twice
[22:23] <CIA-17> nbd * r4063 /branches/buildroot-ng/openwrt/package/base-files/ (brcm-2.4/sbin/mount_root brcm-2.6/sbin/mount_root): nuke jffs2root --move from init scripts
[22:23] <nbd> [mbm]: got any ideas on the squashfs thing?
[22:23] <[mbm]> I'll have to look
[22:23] <[mbm]> although I thought I put in an align function
[22:23] <nbd> align to 1024
[22:24] <nbd> but that's not enough. the eof marker needs to be at the beginning of an erase block
[22:24] <[mbm]> no, I mean trx -a ...
[22:25] <nbd> i'm talking about trx -a as well
[22:25] <nbd> for jffs2 it's aligned to eraseblock size and for squashfs it's aligned to 1024 via -a
[22:26] <[mbm]> that can be any value
[22:26] <nbd> sure, but why waste space?
[22:26] <[mbm]> ?
[22:26] <nbd> if we increase the alignment of the squashfs, we get more padding
[22:26] <[mbm]> all you'd do is align the end of the trx to a 64k block and then put your marker
[22:26] <nbd> -> more wasted space
[22:27] <nbd> the marker needs to be inside the trx
[22:27] <[mbm]> right
[22:27] <nbd> so we need to append it to the squashfs part
[22:27] <[mbm]> not really
[22:27] <[mbm]> you know that the -a can be used multiple times right?
[22:27] <nbd> sure
[22:28] <nbd> but how would you append the end mark to the squashfs without putting it in the next trx part?
[22:30] <[mbm]> maybe trx needs an option to just dump something to the trx without adding it to the trx header
[22:33] <[mbm]> trx -o firmware.trx loader.gz vmlinux.lzma -a <squashfs> root.squash -a 65536 -raw marker
[22:33] <[mbm]> which, amusingly would work for both squashfs and jffs2
[23:08] <nbd> haha... http://www.linuxworld.com.au/index.php/id;754084996;fp;2;fpid;1
[23:10] <[mbm]> lmao
[23:40] <[mbm]> hmm .. verbosity levels
[23:41] <[mbm]> V=0 only trace .. V=1 trace+stderr .. V=99 full
[23:41] Action: [mbm] pokes nbd
[23:41] <nbd> yep
[23:41] <[mbm]> any others?
[23:41] <nbd> can't think of any right now
[23:41] <nbd> let's use these for now
[23:44] <[mbm]> make really needs a 'less than'
[23:45] <[mbm]> default 0 or 1?
[23:46] <nbd> 0
[23:46] Action: nbd starts porting the new broadcom system code to linux 2.6
[23:47] <CIA-17> mbm * r4064 /branches/buildroot-ng/openwrt/ (5 files in 3 dirs): switch on new verbose system
[23:47] <[mbm]> btw still haven't figured out why it seems to want to run the kernel's config twice
[23:51] <[mbm]> make[5] -C target/linux/aruba-2.6 prepare .. <oldconfig> .. make[5] -C target/linux/aruba-2.6 compile .. <oldconfig>
[00:00] --- Sat Jun 24 2006
[00:34] <CIA-17> mbm * r4051 /branches/buildroot-ng/openwrt/rules.mk: cleanup
[02:04] <CIA-17> mbm * r4052 /branches/buildroot-ng/openwrt/include/kernel.mk: set defaults for .pkginfo
[02:49] <CIA-17> mbm * r4053 /branches/buildroot-ng/openwrt/ (3 files in 2 dirs): .pkginfo cleanup #2
[10:45] <florian__> as anyone here played with the isl3893 SoC ?
[10:45] <florian__> I ignored I had one 3 years ago :)
[10:55] <florian__> Kaloz: I tried once again to boot my custo brcm63xx kernel, but it did not work, any idea why it could not ?
[10:56] <florian__> the file is in the right ELF format, I tried appending console option, but it did nothing
[10:58] <Kaloz> florian__: mmuless crap :/
[10:58] <Kaloz> florian__: well, it can have a bunch of reasons for not booting
[10:59] <Kaloz> did you add the serial driver ?
[10:59] <Kaloz> it uses it's own driver
[10:59] <florian__> I think so, if you are talking about the serial driver in arch/
[10:59] <florian__> I think I also need to check if the driver in char/serial was compiled built-in right ?
[11:02] Action: [mbm] ponders a TAINTED setting for the openwrt config
[11:02] <florian__> [mbm]: which one ?
[11:02] <[mbm]> look in #openwrt and you'll see what I mean
[11:03] <Kaloz> you mean the busybox tainted module support?
[11:04] <[mbm]> no, I mean "some idiot has screwed with the configuration, you can probably ignore his bug reports"
[11:05] <florian__> ah ok
[11:05] <florian__> I tought you were talking about tainted module support which is *slightly* different ;)
[11:09] <florian__> why would he change is default mtab filename ?
[11:10] <[mbm]> why would he even want an mtab
[11:10] <florian__> actually I don't know
[11:11] <florian__> it's probably more common to mount things at boot time, but a script could do it as well
[11:11] <[mbm]> no, that's fstab
[11:11] <[mbm]> mtab keeps track of what you've mounted
[11:11] <[mbm]> but so does /proc/mounts
[11:12] <[mbm]> which busybox uses by default
[11:12] <florian__> yep sorry
[11:17] <florian__> what are the LED options in newer kernels ?
[11:17] <[mbm]> haven't really looked into it
[17:11] <tziOm> What do I do if I want to compile something to run on rc5? Can I download some devel kit?
[17:24] <tziOm> Is libtorrent/rtorrent ported to rc5?
[17:30] <common> port yourself
[17:30] <nbd> yay... finally it works
[17:31] <nbd> writable jffs2 on first boot :)
[17:31] <nbd> (on broadcom)
[17:31] <tziOm> common, as docs say: no need to reinvent the wheel...
[17:32] <tziOm> if someone has ported it, why should I spend time on it?
[17:32] <common> check ipkg.be
[17:34] <tziOm> has to be something wrong with that one... no hits on samba
[17:43] <db90h> nice nbd
[17:44] <db90h> congrats ;)
[17:56] <h3sp4wn> tziOm: macsat's site has ctorrent on it (as I told you yesterday)
[18:03] <db90h> i heard using torrents causes birth defects anyway
[18:03] <db90h> ..sometimes, i amuse myself beyond belief.. beyond words.. beyond comprehension
[18:04] <db90h> (and I still haven't went blind)
[18:05] <common> not using torrent causes ...
[18:11] <nbd> amnesia?
[18:14] <db90h> i think he was implying it causes me ;p
[18:28] <CIA-17> nbd * r4054 /branches/buildroot-ng/openwrt/target/linux/generic-2.4/patches/227-jffs2_eofdetect.patch: jffs2 eof detect: mark erased blocks after erasing them
[18:30] <CIA-17> nbd * r4055 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/004-flash.patch: broadcom 2.4: update the trx for jffs2 directly in the flash map driver
[18:39] <andersee> nbd: morning
[18:39] <nbd> just curious... have you looked at buildroot-ng already?
[18:40] <andersee> quite a while back, but I expect it has evolved since...
[18:40] <nbd> yeah, it's really usable now
[18:40] <nbd> it'll replace the stuff in trunk/ soon
[18:57] <[mbm]> http://kerneltrap.org/node/6750
[19:01] <nbd> interesting
[19:01] <[mbm]> yep
[19:01] <nbd> btw. porting the flash map change to 2.6 now
[19:01] <nbd> 2.4 seems to work well
[19:01] <[mbm]> saw that
[19:01] <[mbm]> I think I have a working copy of the trace/verbose stuff based on MAKECMDGOALS
[19:02] <[mbm]> just doing a few more test builds
[19:02] <nbd> great
[19:02] <[mbm]> haven't decided what to do with the V=
[19:02] <[mbm]> so right no it's just hardcoded for minimum verbosity
[19:03] <nbd> as a first step we could implement three levels:
[19:03] <nbd> submake tracing, tracing + strerr, everything
[19:03] <nbd> s/strerr/stderr
[19:03] <[mbm]> right now you get trace+stderr
[19:03] <[mbm]> and full verbosity would be easy enough to add
[19:04] <[mbm]> but I haven't figured out the levels between
[19:04] <nbd> k
[19:05] <nbd> what i haven't figured out yet is how to add the jffs2 eof markers to the squashfs images
[19:05] <nbd> because the squashfs is only aligned to 1024
[19:06] <nbd> i could just put 128k of 1k-aligned end markers in there
[19:06] <nbd> but maybe there's a better way
[19:07] <nbd> i don't want to increase the squashfs alignment
[19:07] <[mbm]> I'm really surprised you didn't just hack mkfs.jffs2
[19:07] <nbd> why should i?
[19:07] <nbd> with jffs2 it's easy
[19:08] <[mbm]> anywyas, what's so tricky about squashfs?
[19:08] <nbd> i need to hit a specific spot with the end marker
[19:08] <nbd> and it needs to be inside the trx
[19:08] <nbd> and the trx utility doesn't align it to blocksize for me
[19:09] <[mbm]> you just need to pad the image you feed into the trx
[19:09] <nbd> but i don't know how much i should pad the squashfs
[19:09] <nbd> since it doesn't begin on an eraseblock
[19:09] <nbd> or at least not necessarily
[19:10] <CIA-17> nbd * r4056 /branches/buildroot-ng/openwrt/target/linux/brcm-2.6/config: update kernel config for brcm-2.6
[19:14] <CIA-17> mbm * r4057 /branches/buildroot-ng/openwrt/include/verbose.mk: new verbose/trace script for make
[19:14] <[mbm]> there's the script
[19:14] <[mbm]> I didn't add the hooks for it yet
[19:15] <[mbm]> basically just needs an include from rules.mk and the toplevel makefile
[19:15] <[mbm]> no other changes to any of the makefiles
[19:18] <CIA-17> nbd * r4058 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/004-flash.patch: add copyright statement
[19:36] <nbd> why did you add the KBUILD_VERBOSE line?
[19:36] <[mbm]> for later use
[19:39] <nbd> looks good
[19:59] <CIA-17> nbd * r4059 /branches/buildroot-ng/openwrt/target/linux/brcm-2.4/patches/004-flash.patch: fix partition align
[20:23] <nbd> damn... something's wrong with 2.6
[20:23] <[mbm]> ?
[20:23] <nbd> it locks up when trying to erase from the jffs2 scanning function
[20:23] <nbd> but it works from the mtd map driver
[20:24] <[mbm]> did you forget an unlock?
[20:27] <nbd> i did not lock or unlock anything
[20:28] <nbd> and the deadlock seems to happen in the cfi_cmdset_0001 driver
[20:33] <nbd> hmm
[20:33] <nbd> it loops querying the chip's status
[20:33] <nbd> which is always ok
[20:41] <nbd> any ideas?
[20:48] <crazy_imp> is it possible to get ipkg working to check if it's the right version of whiterussian or kamikaze build (and here also between the kernel 2.4 & 2.6) and block the installation of a package which was compiled for another version of openwrt? (something similar to the blocker's in gentoo-portage)
[20:49] <[mbm]> crazy_imp: ipkg's dependancy handling is broken
[20:50] <crazy_imp> [mbm]: so theoreticaly this feature is allready in there but not working?
[20:53] <[mbm]> yep.
[21:00] <nbd> damn ... this mtd crap still doesn't work
[21:11] <nbd> [mbm]: oops... i know what's wrong
[21:12] <nbd> [mbm]: the chip is still locked :)
[21:19] <nbd> hmm... no
[21:19] <nbd> unlock didn't help
[21:19] <nbd> in fact unlock crashes it again
[21:21] <nbd> [mbm]: have you tested the jffs2 patch on aruba?
[21:33] <crazy_imp> nbd: i think i solved my problem with the ide support, #'ed all the lines in /etc/modules.d/20-ide and loaded them by hand, in this order: ide-core, pdc202xx_old , ide-detect, ide-disk (or is there an change in kamikaze so it works now, but a fact against this is it doesn't worked if the ide-* modules got loaded and the pdc202xx_old wasn't loaded. so i think it's only a question of the order :D) <
[21:33] <crazy_imp> - maybe someone can confirm this?
[21:34] <nbd> thanks. i'll look into it
[21:34] <crazy_imp> no problem :D now i can finish my openwrt alarm clock :)
[21:43] <nbd> [mbm]: fixed it by moving it to a different function _and_ doing the unlock
[21:44] <nbd> so it was probably a locking issue as well
[21:51] <CIA-17> nbd * r4060 /branches/buildroot-ng/openwrt/target/linux/generic-2.6/patches/204-jffs2_eofdetect.patch: fix the jffs2 eof detect patch for linux 2.6 (unlocking required for erasing blocks)
[21:53] <CIA-17> nbd * r4061 /branches/buildroot-ng/openwrt/target/linux/brcm-2.6/patches/002-flash-map.patch: port the brcm-2.4 flash map driver changes to 2.6
[22:00] <[mbm]> nbd: nope, been busy with other stuff
[22:07] <nbd> ok. really weird, that the 2.4 stuff worked without unlocking
[22:07] <nbd> i'm going to backport my 2.6 changes anyway
[22:21] <CIA-17> nbd * r4062 /branches/buildroot-ng/openwrt/target/linux/generic-2.4/patches/227-jffs2_eofdetect.patch: backport jffs2 eof detect changes from 2.6
[22:22] <[mbm]> hmm I'm noticing the kernel config being run twice
[22:23] <CIA-17> nbd * r4063 /branches/buildroot-ng/openwrt/package/base-files/ (brcm-2.4/sbin/mount_root brcm-2.6/sbin/mount_root): nuke jffs2root --move from init scripts
[22:23] <nbd> [mbm]: got any ideas on the squashfs thing?
[22:23] <[mbm]> I'll have to look
[22:23] <[mbm]> although I thought I put in an align function
[22:23] <nbd> align to 1024
[22:24] <nbd> but that's not enough. the eof marker needs to be at the beginning of an erase block
[22:24] <[mbm]> no, I mean trx -a ...
[22:25] <nbd> i'm talking about trx -a as well
[22:25] <nbd> for jffs2 it's aligned to eraseblock size and for squashfs it's aligned to 1024 via -a
[22:26] <[mbm]> that can be any value
[22:26] <nbd> sure, but why waste space?
[22:26] <[mbm]> ?
[22:26] <nbd> if we increase the alignment of the squashfs, we get more padding
[22:26] <[mbm]> all you'd do is align the end of the trx to a 64k block and then put your marker
[22:26] <nbd> -> more wasted space
[22:27] <nbd> the marker needs to be inside the trx
[22:27] <[mbm]> right
[22:27] <nbd> so we need to append it to the squashfs part
[22:27] <[mbm]> not really
[22:27] <[mbm]> you know that the -a can be used multiple times right?
[22:27] <nbd> sure
[22:28] <nbd> but how would you append the end mark to the squashfs without putting it in the next trx part?
[22:30] <[mbm]> maybe trx needs an option to just dump something to the trx without adding it to the trx header
[22:33] <[mbm]> trx -o firmware.trx loader.gz vmlinux.lzma -a <squashfs> root.squash -a 65536 -raw marker
[22:33] <[mbm]> which, amusingly would work for both squashfs and jffs2
[23:08] <nbd> haha... http://www.linuxworld.com.au/index.php/id;754084996;fp;2;fpid;1
[23:10] <[mbm]> lmao
[23:40] <[mbm]> hmm .. verbosity levels
[23:41] <[mbm]> V=0 only trace .. V=1 trace+stderr .. V=99 full
[23:41] Action: [mbm] pokes nbd
[23:41] <nbd> yep
[23:41] <[mbm]> any others?
[23:41] <nbd> can't think of any right now
[23:41] <nbd> let's use these for now
[23:44] <[mbm]> make really needs a 'less than'
[23:45] <[mbm]> default 0 or 1?
[23:46] <nbd> 0
[23:46] Action: nbd starts porting the new broadcom system code to linux 2.6
[23:47] <CIA-17> mbm * r4064 /branches/buildroot-ng/openwrt/ (5 files in 3 dirs): switch on new verbose system
[23:47] <[mbm]> btw still haven't figured out why it seems to want to run the kernel's config twice
[23:51] <[mbm]> make[5] -C target/linux/aruba-2.6 prepare .. <oldconfig> .. make[5] -C target/linux/aruba-2.6 compile .. <oldconfig>
[00:00] --- Sat Jun 24 2006