[12:12] <BenC> lamont: the patch we have is one I did, but it may be similar
[05:27] <fabbione> morning
[08:02] <airlied> do you guys have pre-built kernels for Intel Pentium D + SMP?
[08:20] <infinity> airlied: In theory, the -686 kernel should be the right thing, there.
[08:20] <infinity> airlied: Our 686 kernels do SMP/UP switching on the fly, through some dirty trickery.
[08:22] <infinity> airlied: "apt-get install linux-686" for the minimum fuss (this is on dapper, of course... On breezy or earlier, you probably want linux-686-smp)
[08:23] <airlied> do I need to go to dapper? or does breezy work?
[08:23] <airlied> ah thx..
[08:23] <airlied> nope that doesn't work on breeszy I might just upgrade to dapper anyways..
[08:24] <infinity> It will only work if you have restricted in your sources.list...
[08:24] <infinity> If you don't give a hoot about resitrcted modules, then "apt-get install linux-image-686-smp" instead (which should come from "main")
[08:25] <infinity> (linux-FLAVOR depends on linux-image-FLAVOUR and linux-restricted-modules-FLAVOUR to make sure you always have both in sync, should you actually be using the latter)
[08:25] <infinity> LRM includes stuff like madwifi, nvidia and fglrx kernel components, etc.
[08:26] <airlied> shouldn't need those then.. 
[08:26] <infinity> Or just upgrade to dapper anyway, cause you're a l33t freedesktop hacker, and you need the latest and greatest anyhow. ;)
[08:26] <infinity> (and kernel 2.6.15 will set much better with you than 2.6.12, I suspect)
[08:26] <infinity> s/set/sit/
[08:27] <airlied> well at least I might have ethernet with 2.6.15
[08:27] <infinity> That would be a good thing.
[08:27] <Mithrandir> ethernet's supposedly good.
[08:27] <airlied> yes I've got a USB Ethernet at the moment.. its mostly not good..
[02:01] <davekempe> anyone got dapper to install on a sun 4100 or 4200?
[02:01] <davekempe> the megaraid_sas driver seems to dump a stack trace and not detect any drives for me
[02:01] <fabbione> davekempe: url to the hw please?
[02:02] <davekempe> http://www.sun.com/servers/entry/x4100/
[02:03] <davekempe> http://www.sun.com/servers/entry/x4200/
[02:03] <davekempe> btw, breezy works fine on the next one up, the v40z
[02:04] <davekempe> but its just megaraid_mm or some older megaraid
[02:04] <davekempe> not sas
[02:04] <fabbione> please file a bug on lauchpad.net/malone component linux-source-2.6.15 and add: lspci -vvv to the bug
[02:05] <fabbione> include also the 2 URL to the hw
[02:05] <davekempe> no worries.
[02:05] <fabbione> the megaraid did change a lot between .12 and .15
[02:06] <fabbione> but i thing that machine actually uses mptsas
[02:06] <davekempe> I will try out dapper flight 4 amd64 tomorrow and try get it working
[02:06] <fabbione> and not megaraid
[02:06] <fabbione> ok
[02:06] <fabbione> thanks
[02:06] <davekempe> i really would like hoary on there though. I will have to make a different installer
[02:07] <infinity> ... hoary?
[02:07] <fabbione> why hoary???
[02:07] <davekempe> so i don't have to resolve the libc PT_GNU_STACK mess and can use grsec in the kernel
[02:07] <davekempe> its a server, not much i need has changed between hoary and dapper
[02:08] <davekempe> (that I can't backport if really really needed)
[02:08] <Mithrandir> dapper's 9-10 months into the support period, though, so you'll be stuck without security fixes in less than a year.
[02:08] <mjg59> Hoary, not dapper
[02:09] <Mithrandir> uh, true.
[02:09] <davekempe> yeah, so we have that long to figure out the libc mess one way or another
[02:10] <davekempe> i really would prefer to not have an executable stack for my purposes
[02:12] <davekempe> this is what im talking about: http://forums.grsecurity.net/viewtopic.php?t=1152&highlight=debian
[02:14] <davekempe> its all redhat's fault
[02:18] <zul> okie dokie
[02:27] <zul> whats the program thats a front gui for a tcpdump
[02:27] <zul> i forget what it is called
[02:28] <Keybuk> ethereal
[02:29] <zul> thanks...
[04:06] <mjg59> BenC: Are you sure about http://kernel.org/git/?p=linux/kernel/git/bcollins/ubuntu-2.6.git;a=blobdiff;h=99c0ce2a87645bea35ad161d0552b35307d6af0f;hp=f26d819c9ae13c403b904c97e0be34866a32213e;hb=14de3561b1eab87faf8591b667e14e8e93523dac;f=arch/i386/Kconfig ?
[04:11] <BenC> mjg59: Pretty sure. The default is 1G of user address space, with anything over that as virtual, and ~190Megs is used for kernel related things to manage that, which gives us the ~870 megs of non-highmem we have by default
[04:11] <BenC> atleast, that's the way I understand it
[04:12] <BenC> it could be backwards, because the patch I used for it is really confusing
[04:12] <mjg59> BenC: No, that's 1G of physical address space, surely?
[04:12] <BenC> yeah, I think you're right
[04:12] <mjg59> The more physical memory the kernel can access, the less address space available to userspace
[04:12] <BenC> Should change the description
[04:13] <mjg59> Yeah
[04:15] <Keybuk> Linus and Andrew both ganging up on kay and gregkh for breaking sysfs/uevent/etc. every week
[04:19] <BenC> goodness
[05:19] <fabbione> hey guys
[05:20] <fabbione> BenC: please pull from me again.. more |Viagara love ;)
[05:20] <BenC> ok :)
[05:20] <fabbione> BenC: we need silo.. do you think you can make it today?
[05:20] <fabbione> .10 + the 2 patches
[05:21] <BenC> Yeah, I can try
[05:21] <fabbione> thanks
[05:22] <fabbione> BenC: both .15 UP and SMP have been tested.
[05:22] <fabbione> BenC: so we are "certified" :)
[05:24] <BenC> gotta love that :)
[05:26] <fabbione> oh yeah
[05:27] <fabbione> that's when we found out a few extra bugs with silo (second patch) and an error in trap/syscall table not being properly alligned
[05:29] <infinity> BenC: Has anyone whined to you yet about a bunch of ISAPNP (I think it was?) printk's flashing by on boot before usplash kicks in?
[05:29] <BenC> yeah, I see them too
[05:29] <BenC> I'm bugging mjg59 about them, since they are his fault :)
[05:29] <infinity> Yeah, can we get the severity lowered, so they get shut up by "quiet"?
[05:30] <infinity> Unless they shouldn't be happening at all.
[05:30] <BenC> they seem to be harmless, so I'll make them quiet
[05:31] <infinity> Also, has anyone yet requested that we bump up the default size of our ringbuffer a tad?  I find it painfully irritating that by the time I boot, "dmesg" doesn't go back to the beginning anymore, cause our boot is so noisy.
[06:25] <Keybuk> BenC: how are you on the kernel filesystem layer?
[06:46] <BenC> Keybuk: so so
[06:46] <Keybuk> I'm having issues with the FIBMAP ioctl
[06:48] <Keybuk> which may be just me ... :)
[06:48] <Keybuk> if I pass 0 rather than "uninitialised local variable" it works
[07:01] <fabbione> hmmmm
[07:02] <fabbione> there has been hell of a lot of activity on OCFS2
[07:55] <BenC> Keybuk: looking at the code, you are supposed to pass a block number
[07:59] <BenC> mjg59: your last email contained no patch
[09:43] <mjg59> BenC: Argh. Hang on.