[03:33] <BenC> ah sweet stability
[03:33] <BenC> I think the last bcm43xx update fixed my dropped packets that were causing me to reconnect so often
[03:38] <mjg59> Rock
[06:12] <fabbione> morning guys
[01:02] <fabbione> BenC: please pull from my archive when you have time. RedHat Cluster Suite update.
[01:02] <fabbione> BenC: i would like to get it in for the next upload
[01:03] <fabbione> since it has some nice bug fixes
[01:21] <BenC> sure thing
[01:21] <BenC> is it going to break ABI?
[01:33] <fabbione> i don't think so
[01:33] <fabbione> you forgot to add the ABI file for amd64-server
[01:33] <fabbione> and that's the one i was building
[01:33] <fabbione> so diff complained there were no files to diff
[01:56] <fabbione> BenC: this kernel is rocking.. 
[01:56] <fabbione> on all the test arches/machines i only have one regressions that i need to reverify with the latest build
[02:12] <jbailey> BenC: You working today?
[02:13] <fabbione> jbailey: that's the kind of question that will never get a proper answer :)
[02:14] <fabbione> jbailey: considering Ben git commits all the way trough xmas and new year :)
[02:14] <jbailey> fabbione: Right.  Sort of why I'm not asking him if he's hacking or not. =)
[02:14] <fabbione> hehehe
[02:14] <jbailey> fabbione: Working is a someone related state, but not the same. =)
[02:14] <fabbione> BenC: ah one important thing.. i am going to take my webcam in London, we need to get it working :))))
[02:15] <fabbione> i have the code from some external crap
[02:15] <fabbione> but it's not exactly working.. anymore.. used to in .12
[02:15] <fabbione> and the code is a fork of the $cantrememberwhatdriver that we do have
[02:15] <fabbione> so perhaps we could do some magic on it
[02:19] <jbailey> BenC: Whenever you are back, I wasn't able to find someone over the break who can talk about the asm-ppc/asm-ppc64 integration.
[02:19] <jbailey> BenC: If you can help me find someone that would be lovely.  It's a very simple problem to fix, but will probably break all sorts of things.
[02:19] <zul> gday
[02:19] <jbailey> fabbione: FWIW, with the -10 kernel I haven't seen a crash in a couple of days.
[02:20] <jbailey> fabbione: But usually it crashes shortly after I say that.
[02:20] <fabbione> jbailey: ask benh? ;)
[02:20] <jbailey> fabbione: He wasn't in #ppc64 much over the break.
[02:21] <fabbione> jbailey: #kernel
[02:21] <jbailey> I'll get beaten for talking about using those headers. =)
[02:21] <fabbione> jbailey: ask in /msg ;)
[02:22] <zul> he'll get slaughtered then
[02:22] <fabbione> nah
[02:22] <fabbione> not from benh
[02:23] <jbailey> I'll /notify on him here for now
[02:29] <zul> brb...new grub
[02:30] <zul> oops...wrong machine
[02:34] <fabbione> BenC: i think i need a newer firmware
[02:34] <fabbione> BenC: i do the iwconfig dance
[02:34] <fabbione> ifconfig eth1 up
[02:34] <fabbione> and it switches again to 54M
[02:35] <fabbione> but if i do again iwconfig rate it works just fine
[02:44] <BenC> weird
[02:44] <BenC> so it's working?
[02:45] <BenC> jbailey: ok, just let me know what you need done
[02:45] <fabbione> BenC: more or less yes
[02:45] <fabbione> one thing to avoid
[02:45] <BenC> fabbione: can you send me specs on the webcam?
[02:46] <BenC> lsusb out and such
[02:46] <fabbione> ifdown foo if your rate is 54Mbit
[02:46] <jbailey> BenC: asm-powerpc/signal.h relies on BITS_PER_LONG, which is only defined if __kernel__.
[02:46] <fabbione> BenC: i can send you the code that was working on .12
[02:46] <fabbione> BenC: and all the other infos
[02:46] <fabbione> i don't have a URL to the specs
[02:46] <jbailey> BenC: asm-s390 just seems to pick an arbitrary value for public consumption in their signal.h
[02:47] <BenC> 32 is a good value for ppc I think
[02:47] <BenC> since if it isn't __kernel__, then it's userspace and probably 32-bit
[02:48] <BenC> fabbione: ok, just lsusb output and vendor/model if you can
[02:48] <fabbione> BenC: yup
[02:48] <jbailey> BenC: Debian has a true 64bit ppc64 port.
[02:48] <jbailey> So that doesn't seem actually true. =(
[02:48] <BenC> I wonder if I get mol running on my wife's G5 if she will even notice that I installed ubuntu on it
[02:49] <jbailey> BenC: I think it's just leaving buffer space in structs for things, though.  In which case the 64 bit one should be fine.
[02:49] <jbailey> BenC: mol?
[02:49] <BenC> jbailey: then probably a quick macro test for __powerpc64__ since the compiler should define that
[02:49] <BenC> mac-on-linux
[02:49] <jbailey> Oh!  I didn't know that existed.
[02:50] <BenC> shit, mol is the bomb
[02:50] <jbailey> So far the only mac thing I want is a game that I don't remember what it's called that ran on m68k., =)
[02:50] <BenC> it's not an emulator
[02:50] <BenC> it's a virtual machine (native mostly)
[02:50] <jbailey> Ah, cool.
[02:50] <BenC> runs OSX 10.4 easily
[02:51] <BenC> runs in a framebuffer window, and is almost native speed (no graphics accel, but that's ok)
[02:51] <BenC> #ifndef __KERNEL__
[02:51] <BenC> #ifdef __powerpc64__
[02:52] <BenC> #define BITS_PER_LONG 64
[02:52] <BenC> #else
[02:52] <BenC> #define BITS_PER_LONG 32
[02:52] <BenC> #endif
[02:52] <BenC> #endif
[02:52] <BenC> make that first line "#if defined(__KERNEL__) && !defined(BITS_PER_LONG)
[02:52] <BenC> damnit
[02:52] <BenC> #if !defined(__KERNEL__) && !defined(BITS_PER_LONG)
[02:53] <BenC> yeah, that will work
[02:54] <BenC> gcc -m32 doesn't define __powerpc64__
[02:54] <BenC> but gcc -m64 does
[02:54] <BenC> jbailey: would you be able to add that to your build, or do I need to put it in the actual headers?
[02:55] <jbailey> BenC: It should be in the actual headers, otherwise anything that uses signal.h will break.
[02:55] <BenC> ok, adding it now for 10.16
[02:55] <jbailey> Thanks.
[02:57] <zul> BenC: new crack for you
[03:55] <zul> vi brb
[04:31] <zul> heylo
[04:37] <fabbione> BenC: 
[04:37] <fabbione> -Toolchain package versions: libc6-dev_2.3.5-1ubuntu12 dpkg-dev_1.13.10ubuntu4 linux-kernel-headers_2.6.11.2-0ubuntu13 gcc-4.0_4.0.1-4ubuntu9 g++-4.0_4.0.1-
[04:37] <fabbione> 4ubuntu9 binutils_2.16.1-2ubuntu5 libstdc++6-4.0-dev_4.0.1-4ubuntu9 libstdc++6_4.0.1-4ubuntu9
[04:37] <fabbione> +Toolchain package versions: libc6-dev_2.3.5-1ubuntu12 dpkg-dev_1.13.10ubuntu4 linux-kernel-headers_2.6.11.2-0ubuntu13 gcc-4.0_4.0.1-4ubuntu9 g++-4.0_4.0.1-
[04:37] <fabbione> 4ubuntu9 binutils_2.16.1-2ubuntu6 libstdc++6-4.0-dev_4.0.1-4ubuntu9 libstdc++6_4.0.1-4ubuntu9
[04:37] <fabbione> spot the difference...
[04:41] <BenC> binutils?
[04:41] <fabbione> and now:
[04:41] <fabbione>  Checking correctness of source dependencies...
[04:41] <fabbione> -Toolchain package versions: libc6-dev_2.3.5-1ubuntu11 dpkg-dev_1.13.10ubuntu4 linux-kernel-headers_2.6.11.2-0ubuntu12 gcc-4.0_4.0.1-4ubuntu9 g++-4.0_4.0.1-
[04:41] <fabbione> 4ubuntu9 binutils_2.16.1-2ubuntu4 libstdc++6-4.0-dev_4.0.1-4ubuntu9 libstdc++6_4.0.1-4ubuntu9
[04:41] <fabbione> +Toolchain package versions: libc6-dev_2.3.5-1ubuntu12 dpkg-dev_1.13.10ubuntu4 linux-kernel-headers_2.6.11.2-0ubuntu13 gcc-4.0_4.0.1-4ubuntu9 g++-4.0_4.0.1-
[04:41] <fabbione> 4ubuntu9 binutils_2.16.1-2ubuntu5 libstdc++6-4.0-dev_4.0.1-4ubuntu9 libstdc++6_4.0.1-4ubuntu9
[04:42] <fabbione> i am digging into 20771
[04:42] <zul> i love battling with grub
[04:42] <BenC> so you think it's a binutils issue?
[04:42] <fabbione> the only thing changed between 9.23 and 10.24 is binutils
[04:42] <fabbione> now
[04:42] <fabbione> binutils changes are trivial
[04:42] <fabbione> only packaging stuff
[04:42] <fabbione> the second diff is the toolchain info on which binutils has been built
[04:43] <fabbione> you might notice clearly that there is more that did change
[04:43] <BenC> is binutils built self contained like gcc (IOW, it gets compiled twice, the second time by itself)?
[04:43] <fabbione> i have no idea
[04:43] <fabbione> that's doko/elmo question
[04:43] <BenC> maybe binutils compiled itself wrong the second time :)
[04:44] <fabbione> it's a possibility that we can't exclude
[04:44] <fabbione> ubuntu4 was hppa/sparc specific change
[04:44] <fabbione> assuming it was done correctly (and i think so
[04:44] <BenC> -5 was built with -4, and -6 with -5, so it could be that -5 was broken
[04:45] <fabbione> ubuntu3 is the first hit with some i386/x64_64 changes
[04:45] <fabbione> it's a possibility
[04:45] <fabbione> i don't have enough experience to understand binutils
[04:46] <fabbione> but there are no differences in the test suite
[04:46] <fabbione> at least that i can see with diff -y 
[04:46] <fabbione> diff -u
[04:47] <doko> BenC: no, it's the normal build, and then the multiarch build
[04:52] <fabbione> BenC: it's the only thing i will have to check when elmo is back
[04:52] <fabbione> get old binutils and try to rebuild the kernel
[04:52] <fabbione> and see what happens
[04:55] <AcidPils> rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr^x^x^x^rxx9~
[04:55] <AcidPils> upss
[04:56] <BenC> damn this sucks, mol wont run on G5(64-bit), and development on it seems pretty dead for over a year now
[04:57] <BenC> will a G5 run 32-bit kernel?
[04:58] <fabbione> BenC: yes, but it's not very stable and upstream doesn't support it anymore
[04:58] <fabbione> that's why we had the changes between hoary and breezy with power3/power4 into powerpc64-smp
[05:00] <BenC> that reminds me, I want to make yaboot choose the correct kernel based on 64-bit/32-bit
[05:00] <zul> 2
[05:00] <zul> oops
[05:01] <BenC> amazing how much yaboot code is like silo
[05:03] <fabbione> BenC: ahhaha did they copy it?
[05:04] <mjg59> BenC: I've got the SD driver running fairly stable now (though still stupidly slow), so I'll probably submit that to you soon
[05:06] <BenC> fabbione: yeah, yaboot was based on silo
[05:06] <BenC> mjg59: ok, thanks
[05:06] <fabbione> ehhehe
[05:21] <mjg59> Anyone intimitely familiar with x86 around?
[05:42] <jbailey> mjg59: I cuddled my new laptop about 6 years ago.  About 15 years ago I remember cuddling my first 486.
[05:42] <jbailey> mjg59: It's a long and sordid history.  I don't know if she's told me all her secrets or not, though.
[06:34] <jbailey> Is suspend to ram on a desktop supposed to work?  For fun I suspended my wife's machine and it doesn't resume.
[06:34] <fabbione> my suspend too
[06:34] <fabbione> it dies on resume :)
[06:35] <jbailey> Ah, okay.
[06:35] <jbailey> Just thinking how nice it would be to have instant-on for my desktop machines. =)
[06:36] <mjg59> It's "supposed" to work
[06:36] <mjg59> But will tend to fail for much the same reasons as on laptops
[06:55] <jbailey> mjg59: Ah, cool.
[06:56] <jbailey> mjg59: Do the tweaks you do to acpi for laptops generally make things better all around, or are they often specific deviant modes for specific models?
[06:58] <mjg59> They generally make things better all around
[06:58] <mjg59> But it would be nice to know how it fails on desktops
[06:59] <jbailey> Is there a way I can get more information out of the box other than "solid red light, doesn't take the monitor out of green mode"?
[07:00] <jbailey> s/red/harddrive/
[07:02] <mjg59> Heh
[07:02] <mjg59> Serial console?
[07:15] <jbailey> 'k. I'll try that next time.
[07:15] <jbailey> See if I get anything off of it at all.
[07:16] <mjg59> You might need to hack wakeup.S to reprogram the serial port
[08:01] <BenC> anyone with a G3?
[08:01] <BenC> I need to see "cat /proc/device-tree/compatible"
[08:02] <cjb> Yeah, hang on.
[08:08] <jbailey> mjg59: But otherwise does it already send status messages there?
[08:13] <mjg59> jbailey: No, you need console=ttyS0,9600n1 (or something)
[08:13] <jbailey> Oh right, so it's just the usual kernel messages.
[08:13] <jbailey> Is there any other special debugging worth turning on?
[08:14] <mjg59> CONFIG_PM_DEBUG, possibly
[08:14] <jbailey> 'k
[08:14] <jbailey> I think Angie's away next weekend, so her machine will be free to break then.
[08:17] <cjb> BenC: Mine contains "PowerBook2,1^@MacRISC^@Power Macintosh^@".
[08:17] <BenC> cjb: cool, thanks
[08:37] <cjb> Ack.
[08:39] <zul> heh...my wife and i at were at each others throats..
[08:40] <zul> presents are better than strees no doubt
[08:41] <cjb> I mean, whether having presents makes up for the holiday stress and makes Christmas worthwhile rather than annoying.  :)
[08:41] <zul> heh..brb...new grub
[08:45] <zul> bleah..
[09:01] <zul> oh goody the muppet show is on