[10:26] <fairuz> Hi, Can I use request_irq in user space?
[11:36] <apw> fairuz, don't believe there is an interface for that no
[11:59] <fairuz> apw: so the irq number, how can I know it exactly? i assume that it's different for each arch?
[12:04] <mjg59> fairuz: the irq number for what?
[12:04] <fairuz> for PL310, a L2 cache controller
[12:05] <fairuz> it says in the TRM, the irq for this device is on MA_IRQ_0
[12:05] <fairuz> just wondering does it mean irq number = 0 in linux?
[12:23] <fairuz> apw: is there a way to know the irq number for each irq source?
[12:35] <mjg59> fairuz: Depends on your architecture
[12:36] <mjg59> fairuz: There's no real mechanism for delivering IRQs to userspace - you'll need something in kernel anyway
[12:36] <fairuz> mjg59: yes, i'm writing in kernel space right now
[12:36] <fairuz> just dont really understand the irq number that I should put
[12:38] <mjg59> fairuz: You'll need to work out how your architecture wires stuff up
[12:38] <fairuz> mjg59: just need a confirmation, does irq number 0 in linux means the hard wired 0 in interrupt controller?
[12:39] <mjg59> No, there can be offsets due to the architecture
[12:39] <fairuz> mjg59: o
[12:39] <fairuz> mjg59: ok
[12:40] <mjg59> fairuz: There's already some amount of PL310-related code in the omap and tegra trees
[12:41] <fairuz> mjg59: I dont found any that touches the interrupt part (or maybe it's me who's not look hard enough)
[12:41] <mjg59> fairuz: What hardware are you using?
[12:41] <fairuz> omap4430
[12:41] <mjg59> #define OMAP44XX_IRQ_PL310                      (0 + OMAP44XX_IRQ_GIC_START)
[12:41] <mjg59> Where OMAP44XX_IRQ_GIC_START is 32
[12:42] <mjg59> So that's probably the IRQ you want!
[12:42] <fairuz> yes!!!!
[12:42] <mjg59> Look at arch/arm/plat-omap/include/plat/irqs-44xx.h
[12:42] <fairuz> i doubt it, I found SPIs start at ID32 and PL310 hard wired to IRQ 0
[12:42] <fairuz> mjg59: Thanks
[12:43] <fairuz> mjg59: I'm still familiarise myself with all those searching in the kernel tree -.-
[12:45] <fairuz> mjg59: one stupid question, in the kernel tree, any file in any include folders can be included right? e.g #include <plat/irqs-44xx.h> ?
[12:45] <fairuz> regardless where the include folder is
[12:45] <fairuz> :D
[12:52] <mjg59> fairuz: Yes, the paths get set appropriately
[13:18]  * ogasawara waves
[13:21] <JFo> o/
[13:22] <tgardner> ogasawara, hey there. kinda early for you, isn't it :)
[13:22] <ogasawara> tgardner: according to the lil guy it's not :)
[13:23] <ogasawara> tgardner: so I figure I'll start my days about now from now on
[13:23] <tgardner> ogasawara, hence the smiley.
[13:23] <tgardner> ogasawara, does your mumble client still work?
[13:23] <ogasawara> tgardner: yep, just gotta dig for my headphones
[13:47] <cooloney> ogasawara: long time no see, welcome
[13:47] <ogasawara> cooloney: thanks
[13:49] <cooloney> ogasawara: how's going? don't work too hard, -:)
[14:04] <JFo> apw, care to have a look at bug 715330 ?
[14:04] <ubot2> Launchpad bug 715330 in xserver-xorg-driver-ati "Freeze after login with KMS enabled on Radeon HD6310" [Medium,Confirmed] https://launchpad.net/bugs/715330
[14:04] <JFo> you may already have a patch req for it.
[14:24] <apw> JFo, looks like we already have that coming, will handle
[14:41] <apw> JFo, abou ?
[14:41] <apw> about ?
[14:41]  * apw slaps his '' key
[14:41] <apw> grrrr
[14:49] <JFo> heh
[15:13]  * ogasawara back in 20min
[16:02] <JFo> <-grabbing lunch
[16:34] <tseliot> mjg59: are you around?
[16:35] <mjg59> tseliot: Hi
[16:36] <tseliot> mjg59: hi. Do you know how to debug the lack of a backlight device after backlight_device_register() succeeds? This is with 2.6.32
[16:37] <mjg59> No
[16:37] <mjg59> If that's succeeded then you should have a device
[16:38] <mjg59> When you say it's succeeded, you know it returns ERR_PTR and not NULL on error, right?
[16:38] <tseliot> mjg59: that's what I was about to ask
[16:39] <mjg59> So if you're just checking against NULL then you'll think you've succeeded when in fact you've failed
[16:39] <tseliot> mjg59: so, shall I test it with "if (IS_ERR(backlight_device))" ?
[16:39] <mjg59> Yes
[16:40] <tseliot> mjg59: well, that's the way the poulsbo driver checks it if(backlight_device), that is
[16:40] <tseliot> I guess it's ok, as it's psb...
[16:40] <tseliot> ;)
[16:41] <tseliot> mjg59: so my next question would be, how would you debug a failure there?
[16:44] <tseliot> mjg59: is there any documentation I can have a look at?
[16:44] <mjg59> tseliot: drivers/video/backlight/backlight.c
[16:45] <mjg59> You should be able to work it out from the failure cases there
[16:45] <tseliot> mjg59: right. Thanks a lot
[16:47] <apw> tgardner, indeed ... finger trouble
[16:57] <tgardner> apw, do you remember if we decided mpt's macbook had a drive hardware issue?
[16:57] <apw> i think we decided it was possible, but we had no evidence
[16:57] <apw> we asked him to collect something when it happened again
[16:58] <tgardner> apw, k
[16:58] <apw> tgardner, though waht i now forget
[16:58] <tgardner> maybe he'll remember
[16:59] <Kano> hi, why is CONFIG_HZ_250 for 32 bit and CONFIG_HZ_100 for 64 bit?
[16:59] <mjg59> I think a better question is arguably why they're not both HZ_1000
[17:00] <apw> mjg59, cause that costs you 10% of performance
[17:00] <Kano> sure,best would be 1000 hz for both
[17:00] <apw> actaully 10-17% depending on load
[17:00] <mjg59> apw: It's very dependent on what your definition of performance is :)
[17:00] <apw> mjg59, indeed, overall throughput in this context
[17:01] <Kano> but at least it should be the same for 32+64 bit or not?
[17:01] <apw> Kano, no it was a deliberate split
[17:02] <Kano> how to change the config to update the files in debian.master?
[17:03] <apw> Kano, in normally use vi to change the option i want, then use updateconfigs target to propogate
[17:04] <Kano> ah
[17:04] <apw> Kano, the source is all included in the tree
[17:05] <Kano> will check it out
[17:09] <davek> hello kernel folk
[17:10] <davek> does anyone know the estimated release of ubuntu using 2.6.38 stable?
[17:11] <apw> davek, we are waiting on final from linus, w
[17:11] <Kano> there is no .38 stable kernel yet
[17:11] <apw> though i would expect it would be in the next day or two
[17:11] <davek> okay
[17:11] <davek> how long would it take after release?
[17:11] <apw> thoguh we are struggling with compiler issues
[17:11] <davek> for ubuntu to add it?  (im kindof a noob here)
[17:12] <davek> i recently purchased a very new motherboard and am looking forward to support for drivers within 2.6.38
[17:13] <apw> davek, if they arn't in the current kernels based on the 38-rcN releases you are likely not goting to get it in any furhter kernel we upload
[17:13] <davek> from what i read there should be Zacate support in there
[17:14] <davek> so would one run Natty to get this kernel, or could i do a trick with Maverick to install it?
[17:14] <davek> (in the future)
[17:14] <apw> davek, well they arn't adding any new features to .38 everything is bug fixes
[17:14] <apw> so if a ntty kernel doesn't have what you want now, its not likely to later
[17:14] <davek> okay
[17:15] <davek> im running Maverick 10.10 presently
[17:15] <davek> will that pick up 2.6.38 too? 
[17:16] <apw> we do backports for lucid, but you can hand install them on maverick.  they won't update on maverick tho.
[17:16] <tgardner> davek, you can test also natty by adding this PPA: https://launchpad.net/~kernel-ppa/+archive/pre-proposed?field.series_filter=natty
[17:16] <davek> okay..  wow thanks for the response guys
[17:16] <davek> you rock
[17:16] <davek> thanks apw and tgardner
[18:10]  * tgardner --> lunch
[18:13] <Kano> bye
[19:16]  * jjohansen -> lunch
[20:47] <rupert_millard> Hi. It's my first time ever on IRC! On my new laptop, I am getting an intermittent kernel panic within about a minute of powering up. I installed "linux-crashdump" so I can give you a proper bug report, but although it seems to dump a load of data on the panic, I can't find it anywhere when I boot up. So I've got two problems - what (is/am I doing) wrong with linux-crashdump, and how do I give you the best bug report I can? I do have 
[20:47] <rupert_millard> a photograph of the screen showing the panic.	
[20:55] <charlie-tca> rupert_millard: You can file a bug by opening a terminal and typing "ubuntu-bug linux" without the quotes
[20:55] <rupert_millard> yes but is that very useful to you?
[20:55] <rupert_millard> i really wanted to include a dump, but linux-crashdump doesn't seem to work for me
[20:57] <charlie-tca> ubuntu-bug provides most of the needed logs, and is a very useful method of filing bugs
[20:57] <rupert_millard> ok
[20:58] <rupert_millard> charlie-tca: what's the point of linux-crashdump then?
[20:58] <charlie-tca> You can look in /var/log for the linux-crashdump, but I don't know if it will be there
[20:58] <rupert_millard> charlie-tca: should I file a bug against it too?
[20:58] <charlie-tca> rupert_millard: no idea, I don't know what it is, myself
[20:58] <charlie-tca> Is there anything in /usr/share/docs that explain it?
[20:59] <rupert_millard> charlie-tca: https://wiki.ubuntu.com/Kernel/CrashdumpRecipe
[21:00] <charlie-tca> The LKCD utility is not designed to gather helpful information in the case of a hardware caused panic or a segment violation
[21:01] <rupert_millard> what constitutes a hardware caused panic?
[21:01] <charlie-tca> ubuntu-bug should pick it up
[21:01] <rupert_millard> well my computer completely locked up - i had to power cycle it
[21:01] <charlie-tca> hm, maybe there is a crash log in /var/crash ?
[21:02] <rupert_millard> there's no /var/crash directory
[21:04] <charlie-tca> then there is no crash report available. 
[21:04] <Zelozelos> anyone wanna help out w gwinwrap, im trying to get the videos to play again. the -wid option is causing an error 
[21:04] <charlie-tca> Perhaps it is panicing before it can save it
[21:06] <rupert_millard> well since installing linux-crash, there is a load of activity - the second kernel seems to be doing something
[21:07] <rupert_millard> i think init might be starting swap up before kdump has a chance to inspect that partition for a dump
[23:31]  * jjohansen heads out for a bit, back on later