[09:48] <d3xter> hey
[09:48] <d3xter> is anyone here?
[09:48] <apw> there are lots of people here
[09:49] <IntuitiveNipple> Of course not - this is a virtual world :)
[10:57] <heywood> hi all
[11:00] <heywood> trying to strip down running kernel modules on a netbook (acer aspire one, SSD version) running intrepid, 2.6.27-7-generic. having trouble understanding relationship between sdhci_pci and sdhci.
[11:01] <heywood> modprobe sdhci_pci says "WARNING: Error inserting sdhci (/lib/modules/.../sdhci.ko): Invalid module format"
[11:01] <heywood> assume sdhci_pci depends on sdhci, which modprobe tries to install first. but why would sdhci (which is unmodified from stock kernel source) give an invalid format error?
[11:18] <alex_joni> heywood: did youo check dmesg for further errors?
[11:19] <alex_joni> usually the invalid format can be generated if you compiled the module with a different compiler as the stock kernel
[11:23] <heywood> checking now. looks like one of my modules exports the same symbol as sdhci_pci, which may be causing that error. brb...
[11:28] <d3xter> i've got a problem with =>2.6.27
[11:28] <d3xter> i've got a bcm4401-b0 lan and bcm4328 wlan
[11:29] <d3xter> at startup the lan-adapter loads b44 and ssb driver
[11:29] <d3xter> and after that the propritaery broadcom driver cant use the wlan-chip, because ssb blocks it
[11:29] <heywood> @alex_joni: weird. i have an sdhci driver i've built myself (call it sdhci_extended) and i'm doing the following, in order, in a shell script (as root): rmmod all_relevant_modules [in order] --> modprobe mmc_core --> insmod sdhci_extended --> modprobe sdhci_pci
[11:30] <d3xter> so i have to manually unload ssb and b44 and load the wl driver frist
[11:30] <IntuitiveNipple> d3xter: Have you posted a bug report?
[11:30] <d3xter> IntuitiveNipple: not yet
[11:30] <IntuitiveNipple> d3xter: That would be best... generally people don't have time here to follow instant questions
[11:31] <d3xter> on launchpad?
[11:31] <IntuitiveNipple> d3xter: Yes, against the "linux" package. The post the bug reference here. It'll also be announced on the kernel-bugs mailing list
[11:32] <d3xter> ok thx
[11:33] <heywood> dumb question: in the output of lsmod, is the last column "depends on" or "used by"?
[11:34] <d3xter> used by
[11:34] <heywood> thanks
[11:36] <heywood> @alex_joni: (continuing previous comment...) when i insmod my sdhci_extended, it shows up in lsmod normally, and also in the "referred to by" list for mmc_core. so far so good.
[11:40] <heywood> now when i do modprobe sdhci_pci, it shows up in the lsmod output by itself, and also in the "referred to by" list for my sdhci_extended. both of those are as expected. but at this point i also get the error about "error inserting sdhci: invalid module format". in dmesg i see that same message, but then a bunch more from sdhci _after_ it -- which suggests sdhci is loading after all.
[11:41] <heywood> does that sound like a showstopper, or more like a nuisance?
[11:42] <heywood> (point is, i need to do many many hours of data gathering, have no easy way to check if all is OK during the run, and don't want to find out only at the end that it didn't work or dropped packets or something)
[11:51] <alex_joni> heywood: try loading your module with insmod
[11:51] <alex_joni> modprobe will do smart things and pull in dependencies, insmod won't
[11:52] <heywood> assume i should call insmod with /lib/modules/`uname -r`/... , correct? (i have several kernels on the machine)
[11:58] <alex_joni> yup
[12:04] <heywood> ok, will give it a try. thanks!
[12:07] <d3xter> ok, i've posted a bug: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/335391
[12:07] <ubot3> Malone bug 335391 in linux "Broadcom LAN Driver blocks Broadcom WLAN" [Undecided,New] 
[12:11] <heywood> @alex_joni: ran modprobe -v to check i wasn't missing any other dependencies. confirmed that it only tries to insmod sdhci (which fails), then sdhci_pci. so just doing insmod sdhci_pci should be safe and i won't miss anything else.
[12:11] <heywood> thanks for your help! :)
[12:14] <alex_joni> heywood: glad I could help
[12:26] <heywood> ?
[13:00] <heywood> can anyone help me understand what the commands in the Linplus section of "tweaks for powersaving" on https://help.ubuntu.com/community/AspireOne do?
[13:02] <heywood> trying to figure out if i can safely comment them out. don't care about power saving (running from AC adapter), but do care about minimizing system activity (interrupts, etc.) while my code is running.
[13:56] <bullgard4> [Ubuntu 8.04.2, 2.6.24-23-generic] I have doubled my RAM size from 512 to 1024 MB. Why does Ubuntu not increase the value 512*1024 in /sys/power/image_size accordingly?
[14:04] <smb_tp> bullgard4, See Documentation/ABI/testing/sysfs-power. It is just a default value not related to your ram size.
[14:09] <bullgard4> smb_tp: /usr/share/doc/linux-doc-2.6.22/Documentation/ABI/testing/sysfs-power/Rafael J. Wysocki says differently: "The /sys/power/image_size file controls the size of the image created by the suspend-to-disk mechanism."
[14:12] <smb_tp> And in the other Doc he writes, that the kernel will try to get the image within that size but will continue to hibernate if that fails.
[14:13] <smb_tp> However, if it turns out to be
[14:13] <smb_tp>                 impossible, the kernel will try to suspend anyway using the
[14:13] <smb_tp>                 smallest image possible.  In particular, if "0" is written to
[14:13] <smb_tp>                 this file, the suspend image will be as small as possible.
[14:13] <smb_tp>                 Reading from this file will display the current image size
[14:13] <smb_tp>                 limit, which is set to 500 MB by default.
[14:37] <bullgard4> So where will I find guidance how to make better use of available RAM size in order to resume or thaw faster?
[16:01] <heywood> hi all, does anyone know what the most serious consequences are of turning off sysklogd and klogd for runlevel 2? is there anything to worry about beyond not having debug info available via dmesg?
[16:43] <IntuitiveNipple> heywood: If you're not worried about missing kernel reports I don't see a problem with that
[16:43] <IntuitiveNipple> apw: ping
[16:43] <apw> hi
[16:45] <IntuitiveNipple> Hey. I'm working through all the bugs tags 'resume' right now. There's a fair few that have "High/Medium" but haven't been touched and may be fixed. Could you look at Bug #312411 and close it if appropriate please?
[16:45] <ubot3> Malone bug 312411 in linux "[Jaunty] Suspend/resume failure on ThinkPad T61 (pulseaudio-related?)" [High,In progress] https://launchpad.net/bugs/312411
[16:45] <heywood> @IN: ok, thanks.
[16:47] <apw> IntuitiveNipple, done
[16:48] <IntuitiveNipple> apw: many thanks. I want to get the launchpad report into some kind of sane order :)
[16:48] <apw> which report you looking at?
[16:48] <IntuitiveNipple> Advanced search for all tags "resume" against "linux"
[16:50] <apw> ahh i see