**** BEGIN LOGGING AT Sun Jul 18 02:59:56 2010
Jul 18 07:29:27 <cwillu_at_work> rcn-ee, back
Jul 18 07:31:41 <JaMa> cwillu_at_work, rcn-ee: can you please point me to patches used to fix "unknown rellocations" in 2.6.35-rc*?
Jul 18 07:32:12 <cwillu_at_work> JaMa, you might have missed him by about 5\ hours
Jul 18 07:32:56 <JaMa> I can find out, if you point me to repo where it's fixed..
Jul 18 07:33:18 <cwillu_at_work> later 2.6.35 builds from http://rcn-ee.net/deb/lucid/
Jul 18 07:33:42 <cwillu_at_work> it'll either be in the patch file (among other things), or it's already in mainline git
Jul 18 07:34:29 <JaMa> thanks
Jul 18 07:40:13 <JaMa> ok found it
Jul 18 11:06:13 <cwillu_at_work> rcn-ee, building from your 2.6.35-devel with the micrel patches disabled still results in the null pointer
Jul 18 11:06:27 <cwillu_at_work> going to try again against 2.6.34
Jul 18 12:28:04 <cwillu_at_work> rcn-ee, looks like that null pointer thing is purely a 2.6.35 thing, of which you may or may not be aware
Jul 18 12:28:17 <cwillu_at_work> checking for the memory leak under 2.6.34 right now, with micrel patches disabled
Jul 18 12:29:03 <cwillu_at_work> rcn-ee, ... and the verdict is nack
Jul 18 12:29:28 <cwillu_at_work> I'll run it over night (need to go to bed anyway :p)
Jul 18 12:29:54 <cwillu_at_work> slabtop -s s is showing a consistent increase in allocations to the size-2048 pool, and the usual rate
Jul 18 12:32:10 <cwillu_at_work> rcn-ee, send me your id_rsa.pub, I'll set you up so you can log into the beagle images I'm making if you want
Jul 18 12:32:30 <cwillu_at_work> cwillu@cwillu.com
Jul 18 12:33:31 <cwillu_at_work> hmm, actually, there's a flicker of hope
Jul 18 12:33:40 <cwillu_at_work> I guess I should give the box a couple minutes to calm down
Jul 18 12:33:51 <cwillu_at_work> it _seems_ to be holding at around 230-250
Jul 18 12:34:31 <cwillu_at_work> so I'll let you know in a few hours (i.e., when I wake up and check :p) if it's increased from that
Jul 18 12:34:47 <cwillu_at_work> I was seeing values in the 20,000's,
Jul 18 12:35:31 <cwillu_at_work> so if it stays under a thousand for a day or so, I'll be satisfied that disabling the micrel patchset eliminates the behaviour
Jul 18 12:35:43 <cwillu_at_work> might even be able to convince me to bisect it :p
Jul 18 12:40:25 <cwillu_at_work> ...
Jul 18 12:41:09 <cwillu_at_work> (I'll still check tomorrow, but it appears to still be climbing, maybe a little less consistently than before, or I might not have noticed the bumping around before)
Jul 18 12:41:24 <cwillu_at_work> it's up at 450 now
Jul 18 21:12:59 <cwillu> rcn-ee, koen on #beagle mentioned that there was some patches on the linux-omap mailing list for a micrel memory leak
Jul 18 23:00:05 <cwillu> rcn-ee, null pointer patch -> http://marc.info/?l=linux-netdev&m=127490516731906&w=2
Jul 18 23:00:54 <cwillu> going to try to run with that, and check if the memory leak was fixed elsewhere in 2.6.35
Jul 19 01:38:10 <cwillu> rcn-ee, okay, that patch fixes the null pointer at least, just got it booting
Jul 19 01:39:07 <cwillu> currently 36 objects in slab size-2048
Jul 19 01:39:20 <cwillu> ... and no increases yet
Jul 19 01:39:27 <cwillu> I'll ping back in an hour
Jul 19 01:41:48 <cwillu> rcn-ee, note that this is _with_ the micrel patches, plus an additional one which I'll email you
Jul 19 01:42:27 <cwillu> (er, it's actually the one from the link, so I'll only email you if you want a known-to-be-working copy of it :p)
Jul 19 01:57:12 <rcn-ee> cool cwillu, was playing with the patch.. http://marc.info/?l=linux-netdev&m=127490516731906&w=2 needed for 2.6.34 and 2.6.35 ? ;)
Jul 19 01:58:53 <cwillu> rcn-ee, haven't tested it under 2.6.34
Jul 19 01:59:15 <cwillu> rcn-ee, I'm suspicious that the null-pointer was a resulting from freeing the memory (i.e., what wasn't happening in .34 and earlier)
Jul 19 01:59:32 <cwillu> you're referring to the 2 -> 4 line?
Jul 19 01:59:35 <rcn-ee> okay, will queue up for 2.6.35 and get a build out, will test 2.6.34 tomorrow too just for kicks..
Jul 19 01:59:42 <cwillu> k
Jul 19 02:00:00 <cwillu> re, leak test, it's still holding at 36 objects
Jul 19 02:00:12 <cwillu> I've seen it go up to 100 or so, but it came back down
Jul 19 02:00:19 <cwillu> so, it seems to be working properly
Jul 19 02:00:41 <rcn-ee> yeah, it's more stablized then the previous version which shot up too fast..
Jul 19 02:04:25 <cwillu> going for a walk for an hour
Jul 19 02:04:32 <cwillu> I might dance a little while I'm walking :p
Jul 19 02:05:03 <rcn-ee> you should... too much stress from the ks8851 driver for you this weekend..
Jul 19 02:05:03 <cwillu> (road trip to replace / upgrade a bunch of hardware with bb + zippy's, and that wasn't going to work out to well if the max uptime of a zippy is a day :p)
Jul 19 02:05:23 <cwillu> now I can actually write features to pack in before I leave wednesday :) :D \o/
Jul 19 02:05:57 <rcn-ee> yeah that would have sucked big time, nothing like hardware that refuses to stay up..
Jul 19 02:07:23 <cwillu> yep
Jul 19 02:07:34 <cwillu> I knew I had this problem, but I thought I remembered that 2.6.33 didn't have it
Jul 19 02:07:42 <cwillu> as it turns out, I just never used the ethernet under 2.6.33 :)
Jul 19 02:08:01 <cwillu> I mean, I had a couple crappy workarounds, but I didn't want to have to explain their need to my boss :p
Jul 19 02:08:13 <cwillu> hardware that stays up is better
Jul 19 02:09:11 <rcn-ee> yeap... :) (crap replaced device 1 with 2, migrate to 2, shutdown 1, device 1 no longer repsonding..)
Jul 19 02:09:23 <rcn-ee> (device 2 no longer responding)
Jul 19 02:09:47 <cwillu> we get enough lightning strikes that I already have enough issues :p
Jul 19 02:10:02 <cwillu> had a customer call me up saying my system was rebooting every few seconds
Jul 19 02:10:13 <cwillu> he didn't feel the need to mention that his desktop was also doing the same
Jul 19 02:10:52 <rcn-ee> yeah, you need good ups in the summer.. (or two or three in line..)
Jul 19 02:13:01 <cwillu> in this case, if the power is out, they're not doing anything anyway
Jul 19 02:13:11 <cwillu> and I've gone through the sd testing routine and so forth
Jul 19 02:13:40 <cwillu> I'm actually considering reimplementing reboot as "sync; dontechodangerous b > /proc/sys-rq-trigger"
Jul 19 02:13:51 <cwillu> nice and quick :)
**** ENDING LOGGING AT Mon Jul 19 02:59:56 2010