rtg_ | kylem: I have 3 that all load i2c_ec, but nonoe of them oops. | 12:17 |
---|---|---|
mjg59 | On the vast majority of machines, i2c_ec will never bind | 12:17 |
mjg59 | So loading it is uninteresting | 12:17 |
maks_ | BenC we can easily write a blacklist dir in /dev/.initramfs | 12:21 |
maks_ | some m-i-t would need to pick that up | 12:22 |
=== macd [n=d@adsl-150-23-118.aby.bellsouth.net] has joined #ubuntu-kernel | ||
BenC | is i2c_ec autoloaded somehow? | 12:28 |
rtg_ | Yep. | 12:28 |
rtg_ | Well, maybe. I'm not sure. | 12:28 |
rtg_ | I think its a result of sms | 12:28 |
rtg_ | sms --> sbs | 12:30 |
=== mjg59_ [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel | ||
=== capiira [n=capiira@dslb-088-064-082-166.pools.arcor-ip.net] has left #ubuntu-kernel ["Leaving"] | ||
mjg59_ | rtg_: Uh. http://git.kernel.org/?p=linux/kernel/git/bcollins/ubuntu-feisty.git;a=commitdiff;h=82580efcd0224a29a9bef1b4ce53f7a4d4611b09 is really, really not a good idea. | 01:08 |
mjg59_ | Oh, wait, hang on. | 01:09 |
mjg59_ | Sorry, I see what you were doing there. | 01:09 |
maks_ | BenC qemu test complain about not beeing able to write to ${rootmnt}/etc/modprobe.d/initramfs as ro mounted | 01:11 |
maks_ | can post the patch if you want | 01:12 |
BenC | maks_: Ah, right, it's mounted ro until after the init scripts run fsck and such | 01:12 |
BenC | so maybe there should be a dangling symlink for /etc/modprobe.d/initramfs-blacklist -> /some/tmpfs/initramfs-blacklist | 01:13 |
BenC | initramfs could create /some/tmpfs/initramfs-blacklist | 01:13 |
mjg59_ | BenC: It looks astonishingly likely that db2f0f088a056c4ccf9054747169802db2f9ae9a is what broke the i2c_ec stuff | 01:14 |
mjg59_ | Any idea what dragged that in? | 01:14 |
maks_ | BenC currently we like to write usplash stuff into /dev/.initramfs so there could be the modprobe.d blacklist there | 01:15 |
BenC | mjg59_: right, rtg found that out too...that got dragged in either by devres stuff, or to fix some other bug, but I can't recall which now | 01:15 |
BenC | I should really edit cherry-picks as to why they got picked, especially ones like that | 01:16 |
mjg59_ | Well, just lose that line | 01:16 |
BenC | we are still seeing a bug with HPA patch, even reverted back to original ABI breaking code | 01:17 |
BenC | reverting it just covered up the ata_dev_same_device() failure | 01:17 |
mjg59_ | BenC: Well, that's not the fix that he committed | 01:17 |
pkl_ | BenC: is there a reason why c25cfc3ad3cb8cac3474febfe66cff8ee0bfba18a was applied reverting the tifm driver from 0.8 to 0.7? | 01:18 |
BenC | mjg59_: I'll look at that in a second...if it works with the i2c change reverted, I can maybe get it in | 01:18 |
BenC | I need to get HPA working | 01:18 |
mjg59_ | BenC: I can't guarantee that it works, and I recommend not reverting the entire thing | 01:18 |
BenC | [ 67.208231] ata1.00: ata_hpa_resize 1: sectors = 234441648, hpa_sectors = -1342616656 | 01:18 |
mjg59_ | Just change that one line | 01:19 |
mjg59_ | It certainly can't make anything worse | 01:19 |
BenC | mjg59_: which line? | 01:19 |
mjg59_ | + smbus->adapter.dev.parent = &device->dev; | 01:19 |
mjg59_ | Delete it | 01:19 |
BenC | for i2c_ec? | 01:19 |
mjg59_ | Yes | 01:19 |
BenC | Ok | 01:19 |
mjg59_ | You can't legally do that with the acpi in 2.6.20 | 01:19 |
BenC | mjg59_: back to HPA, there seems to be some 64-bit issue with the lba to sectors code | 01:20 |
BenC | I can see it on my xeon now | 01:20 |
BenC | I need to check the values in that function | 01:20 |
mjg59_ | Ok | 01:21 |
mjg59_ | I'm unlikely to have a 64-bit install done in time to help you | 01:21 |
BenC | mjg59_: be around for general poking if I find anything? | 01:23 |
BenC | now that I can reproduce it, I can at least find out where it's coming from, but if I get any ata questions, I'm at a loss for quick answers | 01:24 |
mjg59_ | Yup | 01:29 |
BenC | printk("%02x %02x %02x %02x %02x %02x: %016llx \n", | 01:36 |
BenC | tf->hob_lbah, tf->hob_lbam, tf->hob_lbal, | 01:36 |
BenC | tf->lbah, tf->lbam, tf->lbal, sectors); | 01:36 |
BenC | [ 67.124009] f9 4b af f9 4b af: ffffffffaff94baf | 01:37 |
BenC | prints that | 01:37 |
BenC | that can't be right | 01:37 |
BenC | this is what gets printed for another drive | 01:37 |
BenC | [ 67.112044] 00 00 00 f9 4b af: 0000000000f94baf | 01:37 |
BenC | the f94baf looks awful familiar :/ | 01:37 |
BenC | the first read is from ata_read_native_max_address_ext() | 01:46 |
BenC | the second is from the return value of ata_set_native_max_address_ext() | 01:46 |
BenC | I think I see the problem with this | 01:47 |
=== jml [n=jml@59.167.203.115] has joined #ubuntu-kernel | ||
=== Amaranth [n=travis@ubuntu/member/Amaranth] has joined #ubuntu-kernel | ||
=== johanbr [n=j@blk-224-156-151.eastlink.ca] has joined #ubuntu-kernel | ||
=== jml_ [n=jml@59.167.203.115] has joined #ubuntu-kernel | ||
cjwatson | BenC: what's the status? I see scrollback ... | 02:15 |
BenC | cjwatson: things are working, I just have a slight problem of 64-bit ness it seems | 02:30 |
BenC | cjwatson: it's a problem I can reproduce so just a matter of figuring out the thinko that is hiding in here somewhere | 02:31 |
cjwatson | ok, I'm going to doze here, | 02:32 |
cjwatson | if you phone my landline it stands some chance of waking me up | 02:32 |
BenC | cjwatson: Ok, elmo said he'd be around to process an upload, unless you prefer I contact you | 02:32 |
cjwatson | use me as a first resort and elmo as a backup, please | 02:34 |
=== mdz_ [n=mdz@host86-141-95-121.range86-141.btcentralplus.com] has joined #ubuntu-kernel | ||
BenC | cjwatson: will do | 02:44 |
=== Fujitsu [n=Fujitsu@ubuntu/member/fujitsu] has joined #ubuntu-kernel | ||
=== reitblatt [n=mark@resnet-50-180.dorm.utexas.edu] has joined #ubuntu-kernel | ||
=== tritium [n=tritium@ubuntu/member/tritium] has joined #ubuntu-kernel | ||
=== gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-kernel | ||
=== defendguin [n=supertux@cpe-72-181-7-135.houston.res.rr.com] has joined #ubuntu-kernel | ||
=== rpereira [n=rpereira@ubuntu/member/rpereira] has joined #ubuntu-kernel | ||
=== tuxmaniac [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel | ||
=== zul [n=chuck@CPE0006258ec6c1-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel | ||
jml | t. | 04:07 |
jml | (sorry, ultra-tappy touchpad) | 04:07 |
=== tritium_ [n=tritium@ubuntu/member/tritium] has joined #ubuntu-kernel | ||
=== m0rg0th [n=manugarg@219.64.120.181] has joined #ubuntu-kernel | ||
=== zul_ [n=chuck@CPE0006258ec6c1-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel | ||
defendguin | hey how do i mount the mmc device? | 06:19 |
=== defendguin [n=supertux@cpe-72-181-7-135.houston.res.rr.com] has joined #ubuntu-kernel | ||
defendguin | now my mmc card wont work in -12 and it used to and my camera doesn't connect as a mass storage device | 06:35 |
Mithrandir | defendguin: please try the absolute latest kernel, it should fix it. | 06:38 |
=== m0rg0th [n=manugarg@220.225.228.97] has joined #ubuntu-kernel | ||
defendguin | Mithrandir: what is the latest? -14? | 06:38 |
defendguin | or did they release something else with the mmc fix? | 06:38 |
fabbione | defendguin: yes. about 8 hours ago and should be in the archive.ubuntu.com as we speak | 06:40 |
fabbione | probably the mirrors didn't catch it up yet | 06:40 |
Mithrandir | 2.6.20.14.12 | 06:40 |
defendguin | ahh i see it in update | 06:40 |
defendguin | 2.6.20-14.23 | 06:40 |
defendguin | i hope that fixes the issue with the camera too | 06:41 |
=== defendguin [n=supertux@cpe-72-181-7-135.houston.res.rr.com] has joined #ubuntu-kernel | ||
defendguin | doesn't look like it is working | 06:50 |
defendguin | how can i tell that im in .23? it just says generic | 06:51 |
Mithrandir | make sure uname -a says 2.6.20-14-generic | 06:51 |
defendguin | Linux houston 2.6.20-14-generic #2 SMP Thu Apr 12 22:53:19 UTC 2007 i686 GNU/Linux | 06:52 |
Mithrandir | and neither your camera (which I presume is mass-storage) nor your mmc reader works correctly? | 06:52 |
defendguin | correct | 06:53 |
defendguin | it is most probably mass storage but i can't guarantee that | 06:55 |
defendguin | i wish i had gotten a chance to test this fix out a few days ago | 06:57 |
Mithrandir | I think it's too late to fix this now, we might be able to do it in a post-release update | 06:58 |
defendguin | Mithrandir: i really need to mount this disk | 07:02 |
defendguin | i used to remember my fstab stuff but its been so long since i needed to | 07:03 |
defendguin | what sucks is that it used to work in -12 but now that doesn't even work | 07:08 |
cjwatson | BenC: this isn't sounding good. What's up? | 07:12 |
Mithrandir | what breaks if we just blacklist that i2c-ec driver? | 07:21 |
Mithrandir | hm, the smart battery module breaks then.. | 07:23 |
Mithrandir | ugh, it's 1:30 for Ben, a bit on the late side to call him. | 07:27 |
fabbione | i don't think he will mind too much if you call him | 07:45 |
fabbione | probably the wife will mind more | 07:45 |
Mithrandir | I've sent him an SMS and will send him an email too, hopefully he'll see either if he's up | 08:09 |
=== tritium [n=tritium@ubuntu/member/tritium] has joined #ubuntu-kernel | ||
=== gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-kernel | ||
=== abogani [n=abogani@adsl203-157-083.mclink.it] has joined #ubuntu-kernel | ||
=== infinity2 [n=adconrad@cerberus.0c3.net] has joined #ubuntu-kernel | ||
=== m0rg0th_ [n=manugarg@220.225.228.97] has joined #ubuntu-kernel | ||
=== ivoks [n=ivoks@20-45.dsl.iskon.hr] has joined #ubuntu-kernel | ||
=== infinity3 [n=adconrad@cerberus.0c3.net] has joined #ubuntu-kernel | ||
=== cassidy [n=cassidy@host-213-189-171-21.brutele.be] has joined #ubuntu-kernel | ||
=== varka [n=varkatop@p54A5FF03.dip.t-dialin.net] has joined #ubuntu-kernel | ||
=== gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-kernel | ||
=== gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-kernel | ||
=== jml_ [n=jml@59.167.203.115] has joined #ubuntu-kernel | ||
=== varka [n=varkatop@p54A5FF03.dip.t-dialin.net] has left #ubuntu-kernel ["Ex-Chat"] | ||
=== Keybuk [n=scott@quest.netsplit.com] has joined #ubuntu-kernel | ||
=== fabbione [i=fabbione@gordian.fabbione.net] has joined #ubuntu-kernel | ||
=== mc44 [n=mc44@unaffiliated/mc44] has joined #ubuntu-kernel | ||
mc44 | the new kernel seems to be broken for me | 12:07 |
mc44 | should I file a bug? | 12:07 |
Mithrandir | no | 12:07 |
Mithrandir | there are about five hundred zillion dupes filed already | 12:08 |
mc44 | haha | 12:08 |
mc44 | ok sorry | 12:08 |
=== Lure [n=lure@ubuntu/member/lure] has joined #ubuntu-kernel | ||
=== Lure [n=lure@external-7.hermes.si] has joined #ubuntu-kernel | ||
=== stivani [n=stivani@d5152A75F.access.telenet.be] has joined #ubuntu-kernel | ||
=== mdz [n=mdz@yttrium.canonical.com] has joined #ubuntu-kernel | ||
=== ivoks [n=ivoks@38-155.dsl.iskon.hr] has joined #ubuntu-kernel | ||
Mithrandir | the SD/MMC reader on my x40 seems to work again, at least. | 01:38 |
Keybuk | I've yet to find a use for the SD/MMC reader | 01:53 |
Keybuk | since every single device I own has a different profile card | 01:54 |
Mithrandir | I wish I had a CF reader instead. SD is one of the few I actually don't have. | 01:54 |
zdzichuBG | so I must be lucky. card reader in my z61t is compatibile with my sistet's Minolta camera | 01:55 |
Keybuk | the 770 takes a little card that's half the size of the cards the N800 takes | 01:55 |
Mithrandir | zdzichuBG: my camera uses CF cards, since it's a tad more expensive than the cheap ones you can get. | 01:55 |
Mithrandir | at least I don't know of any prosumer DSLRs which use !CF. | 01:56 |
Keybuk | successive sony phones have left me with a small pile of Memory Stick Duo, Memory Stick Pro Duo and now Scandisk M2 cards | 01:56 |
Keybuk | none of which fit in any other device | 01:56 |
Mithrandir | yes, I have MS Pro Duo as well, for my old camera. | 01:56 |
Mithrandir | it's easier just to chuck a 250-thousand-different-kinds-of-cards reader in the bag and use that | 01:57 |
Keybuk | I just use bluetooth | 01:57 |
Keybuk | or USB | 01:57 |
=== EtienneG [n=etienne@ubuntu/member/EtienneG] has joined #ubuntu-kernel | ||
=== Lure_ [n=lure@external-7.hermes.si] has joined #ubuntu-kernel | ||
=== tuxmaniac [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel | ||
=== zul [n=chuck@CPE0006258ec6c1-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel | ||
JanC | Mithrandir: 1000-in-1 usb card readers are cheap :) | 03:01 |
Mithrandir | they cost about 2.5 cents, so yes. | 03:01 |
=== JanC just bought one | ||
JanC | I don't understand why they don't just cram one of these in laptops instead of all those things that need special drivers | 03:05 |
JanC | would probably be easier to support for manufacturers too | 03:06 |
=== ..[topic/#ubuntu-kernel:cjwatson] : Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | 2.6.20-15.24 Uploaded - If you have a BUG, it's too late to get it in for release. | ||
=== heno [n=henrik@ubuntu/member/heno] has joined #ubuntu-kernel | ||
=== jml [n=jml@59.167.203.115] has joined #ubuntu-kernel | ||
=== m0rg0th__ [n=manugarg@220.225.228.97] has joined #ubuntu-kernel | ||
=== mc44 [n=mc44@unaffiliated/mc44] has left #ubuntu-kernel ["Exit,] | ||
mjg59_ | Keybuk: The ones in the N70 (RSMMC) are electrically compatible with MMC and SD | 03:35 |
mjg59_ | There's a little adapter that clips on | 03:35 |
Keybuk | N70? | 03:36 |
Keybuk | don't have one of those | 03:36 |
Keybuk | I have a 770 and N800 | 03:36 |
mjg59_ | 770, rather | 03:36 |
Keybuk | the ones in the N800 are bigger versions of the ones in the 770, right? | 03:36 |
mjg59_ | The N800 comes with SD, not MMC | 03:39 |
mjg59_ | SD readers can read MMC, but not vice-versa | 03:39 |
BenC | zul: ping | 03:44 |
zul | BenC: pong | 03:44 |
zul | whats up | 03:44 |
BenC | zul: are you able to do anything with openvz? I sort of need to give Malc a heads up on what may or may not happen | 03:44 |
zul | yeah I have x86 kind of working kind of broken on amd64 | 03:45 |
zul | I keep timing out when Im trying to upload it with dput | 03:45 |
BenC | IMO, just apply the patches and getting it building if you can | 03:45 |
BenC | let them worry about whether it works or not | 03:45 |
zul | sure.. | 03:45 |
BenC | zul: could you reply to the email they sent? | 03:46 |
zul | sure | 03:46 |
BenC | zul: great thanks. I know you got a lot of other things going on more important, I'm just trying to follow up and make sure the communication doesn't go stale | 03:47 |
zul | not a problem, thanks | 03:47 |
zul | done | 03:48 |
zul | Ill try to upload the kernel to the archive again | 03:49 |
zul | and get the userland stuff uploaded at lunch | 03:50 |
=== m0rg0th [n=manugarg@220.225.228.97] has joined #ubuntu-kernel | ||
Keybuk | mjg59_: all I remember is that I tried to put the 770 card into my laptop | 04:19 |
Keybuk | and couldn't get it out for weeks | 04:19 |
mjg59_ | Yes. You need the clip-on adapter. | 04:21 |
kylem | wtf, could not resolve us.archive.ubuntu.com | 04:22 |
Keybuk | I don't think I got an adapter | 04:24 |
Keybuk | the M2 thing in the W880i is damned nifty though | 04:24 |
Keybuk | it's 1GB and smaller than a thumb-nail | 04:24 |
=== ivoks [n=ivoks@37-212.dsl.iskon.hr] has joined #ubuntu-kernel | ||
mjg59_ | Keybuk: The 770 ships with the card in an adapter | 04:28 |
mjg59_ | You have to unclip it before you can put it in the machine | 04:28 |
Keybuk | oh | 04:28 |
Keybuk | wonder where I put that then | 04:28 |
Keybuk | probably threw it away :) | 04:29 |
gnomefreak | Keybuk: thank you for the n-m fix it worked great :) | 04:37 |
EtienneG | hey guys ... I'm getting 403 Forbidden when trying to install linux-image-2.6.20-14-generic | 04:41 |
EtienneG | before I look for someone to bug, am I th eonly one that noticed this ? | 04:42 |
stgraber | EtienneG: this kernel is buggy, so the access is currently blocked | 04:42 |
EtienneG | ha, ok then | 04:43 |
stgraber | EtienneG: some people have crash with it another one is being built at the moment | 04:43 |
EtienneG | thanks for the info, I'll be waiting for the next upload | 04:43 |
=== tuxmaniac [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel | ||
=== ThiefOfBaghdad [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel | ||
zul | BenC: ping bu | 05:06 |
zul | er... openvz-kernel-2.6.20_2.6.20-1_source.changes is NEW | 05:06 |
=== Lure [n=lure@clj46-234.dial-up.arnes.si] has joined #ubuntu-kernel | ||
=== rpereira [n=rpereira@ubuntu/member/rpereira] has joined #ubuntu-kernel | ||
=== bleinmono [n=toffel@ppp85-141-150-48.pppoe.mtu-net.ru] has joined #ubuntu-kernel | ||
BenC | mjg59, kylem: We need to figure out the correct fix for this | 05:51 |
mjg59 | Yes. | 05:51 |
BenC | I'm leaning toward just not doing hpa if the driver is sata_nv...if it's even possible to do that check | 05:51 |
mjg59 | Not trivially. | 05:52 |
mjg59 | What other code paths are there that execute ata_exec_internal? | 05:52 |
=== pmjdebruijn [n=pmjdebru@pmjdebruijn.xs4all.nl] has joined #ubuntu-kernel | ||
mjg59 | Ah, hang on. Alan at one point suggested doing the resizing after the drive/controller timing had been set up. | 05:53 |
mjg59 | Can you move the call to... | 05:53 |
mjg59 | (Let me find this) | 05:53 |
BenC | let me check... | 05:53 |
BenC | we should do this for ATAPI, right? | 05:54 |
mjg59 | Ok, after the call to ata_set_mode() in ata_eh_recover() | 05:54 |
mjg59 | No, ATAPI isn't a problem | 05:54 |
kylem | no. | 05:54 |
mjg59 | Or, rather, I don't think ATAPI devices can implement HPAs | 05:55 |
kylem | doing HPA on ATAPI devices is illegal. | 05:55 |
BenC | they should report no hpa support though right? | 05:55 |
kylem | it won't claim ata_id_hpa_enabled | 05:55 |
kylem | right | 05:55 |
BenC | right | 05:55 |
kylem | - Mandatory when the Host Protected Area feature set and the 48-bit Address feature set are | 05:55 |
kylem | implemented. | 05:55 |
kylem | - Use prohibited when the Removable Media feature set is implemented. | 05:55 |
kylem | - Use prohibited when PACKET Command feature set is implemented. | 05:55 |
mjg59 | kylem: Can you give that a go? | 05:55 |
BenC | let me try mjg59's suggestion | 05:56 |
mjg59 | There's two codepaths where ata_set_mode() gets called, depending on whether the eh code is in use or not | 05:56 |
mjg59 | It may need adding to both | 05:56 |
mjg59 | Actually, yeah, it needs adding to both (if it works) | 05:57 |
mjg59 | But I guess we'll find out | 05:57 |
kylem | mjg59, i've moved it somewhere myself trying. | 05:57 |
BenC | mjg59: maybe put it in ata_set_mode? | 05:57 |
mjg59 | BenC: Could do, but no real need | 05:57 |
mjg59 | It's trivial to add to both | 05:58 |
kylem | trying. | 05:58 |
BenC | recompiling | 06:00 |
kylem | worked. | 06:00 |
kylem | mjg59, didn't fail that time. | 06:00 |
kylem | ah. didn't fail because it didn't execute, fun. | 06:01 |
kylem | ohhhh. hmm. maybe because hpa_id_enabled wasn't true there. | 06:02 |
mjg59 | kylem: Where did you add it? | 06:02 |
kylem | mjg59, after ata_set_mode in bus_probe. | 06:02 |
mjg59 | kylem: That won't be executed if there's an error handler | 06:03 |
mjg59 | Put it in ata_eh_recover() | 06:03 |
BenC | I've put it in both places, rebooting | 06:03 |
BenC | boots | 06:05 |
kylem | i still think we should add an avoid_all_hpa_code flag. | 06:05 |
mjg59 | Does it execute? | 06:05 |
BenC | but I want to put a printk back in there to make sure it is getting called | 06:05 |
BenC | kylem: I changed the ignore_hpa=0 to actually do that | 06:05 |
kylem | at least that way, we don't brownpaperbag the pressed cds completley and utterly. | 06:05 |
kylem | ok. | 06:06 |
mjg59 | I can sort of understand sata_nv getting upset if you're executing ata commands before actually making sure the disk and controller timings agree | 06:06 |
=== ivoks [n=ivoks@33-171.dsl.iskon.hr] has joined #ubuntu-kernel | ||
BenC | it's not getting called | 06:09 |
BenC | let me post my diff | 06:09 |
BenC | people.ubuntu.com/~bcollins/libata.diff | 06:11 |
BenC | I'm thinking about moving the checks for (ata_ignore_hpa && ata_id_hpa_enabled(dev->id) to the top of ata_hpa_resize to clean things up a bit | 06:12 |
mjg59 | Would also work | 06:13 |
kylem | ok. | 06:13 |
mjg59 | BenC: It's not obvious why it's not getting called. You rebuilt the initramfs as well, right? | 06:13 |
kylem | i have an idea. | 06:13 |
kylem | we buy everyone scsi disks. | 06:13 |
kylem | and pretend ATA doesn't exist. :) | 06:13 |
BenC | yes | 06:13 |
kylem | ergh. | 06:18 |
kylem | ok. | 06:19 |
kylem | i think ata_dev_revalidate is the right place to put it. | 06:19 |
kylem | since it's called from both the _eh and bus_probe paths. | 06:19 |
mjg59 | After set_mode? | 06:20 |
cjwatson | BenC: (is that another ABI bump, BTW?) | 06:21 |
BenC | ok, it's getting called now | 06:21 |
BenC | cjwatson: No | 06:21 |
cjwatson | ok, just checkin' | 06:21 |
cjwatson | I'm not going to say no if that's what it takes | 06:21 |
mjg59 | BenC: Getting called and breaking? | 06:22 |
BenC | mjg59: I need to check that it is actually going down the code path | 06:22 |
mjg59 | Ok | 06:23 |
BenC | but the function got called | 06:23 |
=== notts [n=eric@cpe-72-184-186-31.tampabay.res.rr.com] has joined #ubuntu-kernel | ||
notts | sorry to jump in, but there a major bug with the new updated kernel this morning | 06:25 |
notts | got error message: error loading os" | 06:26 |
mjg59 | notts: Yes, it's broken for various people | 06:26 |
mjg59 | It's being worked on | 06:26 |
kylem | hey hey. | 06:28 |
kylem | garzik to the rescue. | 06:28 |
notts | major thanks, any word on when | 06:28 |
mjg59 | When it works | 06:30 |
kylem | WOOOOOO | 06:32 |
kylem | I WIN | 06:32 |
rpereira | hi everybody. | 06:32 |
BenC | mjg59: neither of my drives says they support hpa when ata_resize_hpa() is called | 06:33 |
kylem | BenC, boot with sata_nv.adma=0 | 06:33 |
BenC | kylem: with old patch? | 06:33 |
kylem | ata4.00: ata_hpa_resize 1: sectors = 390721968, hpa_sectors = 390721968 | 06:33 |
mjg59 | BenC: Hm. Odd. | 06:33 |
kylem | BenC, yeah | 06:33 |
kylem | want my defconfig'd kernel instead of rebuilding? | 06:34 |
mjg59 | kylem: Figures. | 06:34 |
kylem | yes. | 06:34 |
kylem | fuck nvidia. | 06:34 |
kylem | right in the skull. | 06:34 |
BenC | it doesn't take me long to rebuild libata.ko | 06:34 |
kylem | ok. | 06:34 |
BenC | kylem: so should we disable adma by default? | 06:34 |
kylem | i think so. | 06:34 |
mjg59 | Fine. Dropping adma just costs us ncq | 06:34 |
mjg59 | But it's clearly broken in some respect | 06:34 |
BenC | kylem: can you do a patch for that, and I'll include a patch so that ignore_hpa steers clear of all hpa code? | 06:35 |
=== notts3 [n=eric@cpe-72-184-186-31.tampabay.res.rr.com] has joined #ubuntu-kernel | ||
BenC | email it to me and I'll test it and prepare a -15.25 | 06:36 |
kylem | 1sec, trying to figure out if we can shortcircuit adma mode so we can leave it on. | 06:36 |
BenC | kylem: just default the module param to 0 :) | 06:36 |
BenC | if people want it they can enable it manually | 06:36 |
kylem | NCQ is fairly hit or miss too. | 06:36 |
BenC | do you think making module parm adma=0 by default is enough? | 06:37 |
kylem | yes. | 06:37 |
BenC | if that's all that's needed I can do that locally | 06:37 |
kylem | ok. | 06:37 |
kylem | it's static int adma_enabled | 06:38 |
BenC | ok, I'll rebuild a kernel and see if danielk is still around to test | 06:38 |
cjwatson | any idea when sata_nv.adma was turned on? | 06:38 |
kylem | cjwatson, recently. | 06:38 |
cjwatson | nod | 06:38 |
kylem | commit 382a6652e91b34d5480cfc0ed840c196650493d4 | 06:39 |
kylem | Author: Robert Hancock <hancockr@shaw.ca> | 06:39 |
kylem | Date: Mon Feb 5 16:26:02 2007 -0800 | 06:39 |
kylem | sata_nv: use ADMA for NODATA commands | 06:39 |
kylem | turning it off wholesale is fine for now. i'll look into a real fix when we have some idle cycles. | 06:39 |
cjwatson | hmm, git-blame says 2006-10-27 | 06:39 |
cjwatson | oh, it's only one bit of adma? | 06:39 |
kylem | yeah. | 06:40 |
kylem | basically instead of transitioning from adma mode to normal mode, they decided to submit NODATA with ADMA. | 06:40 |
cjwatson | git log drivers/ata/sata_nv.c doesn't show the commit above? | 06:41 |
cjwatson | (just trying to find it) | 06:41 |
kylem | probably came in one of our imports | 06:41 |
BenC | it was a chunk patch to sync with libata | 06:41 |
cjwatson | $ git diff-tree -p 382a6652e91b34d5480cfc0ed840c196650493d4 | 06:41 |
cjwatson | fatal: bad object 382a6652e91b34d5480cfc0ed840c196650493d4 | 06:41 |
cjwatson | ah | 06:41 |
BenC | cjwatson: git-diff-tree -p --pretty SHA | 06:41 |
kylem | cjwatson, git show | 06:41 |
BenC | cjwatson: ah, you don't have upstream in there | 06:41 |
BenC | should we just revert that one patch? | 06:42 |
cjwatson | ah, I have a different tree for that | 06:42 |
kylem | BenC, testing. | 06:42 |
BenC | if we can leave adma on for default transactions we can avoid the "my driver performance sucks" crew | 06:42 |
kylem | yup. | 06:42 |
kylem | reverts cleanly, testing now. | 06:42 |
cjwatson | it does say it fixed some timeouts ... | 06:43 |
kylem | boots. | 06:43 |
cjwatson | oh, no, maybe not | 06:43 |
kylem | yeah, the language is murky. | 06:43 |
BenC | cjwatson: it fixes some previously fixed ones in a different way | 06:44 |
BenC | from the sound of it | 06:44 |
mjg59 | Well, worst case is that we revert back to edgy levels of support for that hardware | 06:44 |
cjwatson | mm | 06:44 |
cjwatson | can we bonnie++ the resulting system or something? :) | 06:44 |
mjg59 | BenC: The i2c_ec patch went in, right? | 06:44 |
kylem | NCQ isn't always a win. | 06:45 |
BenC | kylem: so reverting that patch with current code works? | 06:45 |
cjwatson | bit concerned that just booting might be a weak test there | 06:45 |
BenC | mjg59: yes | 06:45 |
mjg59 | Excellent, thanks | 06:45 |
cjwatson | (I don't care about performance, more that it doesn't fall over under load) | 06:45 |
kylem | BenC, yes in my testing tree (which is .21-rc6) building ubuntu images now | 06:45 |
cjwatson | mjg59: anything else you know about? | 06:45 |
cjwatson | of the release-critical ohmygod kind | 06:45 |
mjg59 | Nothing from me | 06:45 |
cjwatson | I thought it might be worth an explicit check :) | 06:46 |
cjwatson | thanks | 06:46 |
mjg59 | Yeah, I forgot about that one because it got downgraded to medium for some reason | 06:46 |
mjg59 | Basic problem was just that something we'd ended up pulling in expected the ACPI device semantics from 2.6.21 | 06:47 |
cjwatson | it was critical when I first saw it | 06:47 |
mjg59 | I bumped it back up yesterday | 06:47 |
cjwatson | kylem: can you arrange for as many people on the team as possible who use sata_nv to try this out? | 06:47 |
kylem | cjwatson, just ping them in #c? | 06:48 |
kylem | or check the hw db? | 06:48 |
cjwatson | #c, #distro, distro-team@ | 06:50 |
kylem | yup. | 06:50 |
BenC | I'll have kernels built with this in about 30 minutes | 06:51 |
BenC | i386 and amd64 | 06:51 |
kylem | ok | 06:52 |
kylem | BenC, this should be a suitably large hammer for now. I'm going to try and snoop the taskfile and unset ADMA for these commands and submit the patch upstream when i get a bit of free time. | 06:55 |
BenC | kylem: ok | 06:55 |
=== shawarma [n=sh@atlas.linux2go.dk] has left #ubuntu-kernel [] | ||
BenC | kylem, mjg59: thanks for the time...maybe we get a suitable kernel for release | 06:56 |
kylem | heh | 06:56 |
BenC | at least -15.24 is almost done building to free up the buildd's | 06:57 |
BenC | builds started | 06:58 |
cjwatson | let me know when it's ready (by whatever means necessary) and I'll contact Tollef if he isn't around | 06:59 |
cjwatson | so we can make sure it builds on palmer again | 06:59 |
=== tritium_ [n=tritium@ubuntu/member/tritium] has joined #ubuntu-kernel | ||
BenC | is palmer the faster machine? | 07:08 |
kylem | BenC, patch hitting your mailbox. | 07:10 |
kylem | alternate fix, reverting that commit is fine too though. | 07:10 |
BenC | I've already got the revert in here and building/preparing :) | 07:11 |
kylem | beauty. | 07:11 |
=== jpon [n=jpniane@neu67-3-82-239-80-181.fbx.proxad.net] has joined #ubuntu-kernel | ||
kylem | i fired the fix upstream since it was so simple to filter. | 07:12 |
BenC | kylem: your fix will filter into gutsy, so for feisty we'll stick with edgy type performance | 07:12 |
=== kylem nods. | ||
kylem | BenC, worst case scenario is we fix it right in -updates, with this, everyone should be able to at least boot. | 07:14 |
=== lfittl [n=lfittl@213.129.230.12] has joined #ubuntu-kernel | ||
cjwatson | sounds like the right decision | 07:16 |
kylem | with ben's ignore_hpa patch, if this crops up in another controller we aren't able to test on, we should can at least tell them to disable it to get it installed. | 07:17 |
kylem | but from a cursory glance, i don't think it will be. | 07:17 |
=== reitblatt [n=mark@resnet-50-180.dorm.utexas.edu] has joined #ubuntu-kernel | ||
lfittl | kvm, seems broken (kvm-api-9 dependency can't be fulfilled), any time when this will be fixed? | 07:31 |
cjwatson | somebody oughta fix that, I agree; is it just a rebuild? | 07:33 |
=== ivoks [n=ivoks@39-229.dsl.iskon.hr] has joined #ubuntu-kernel | ||
cjwatson | * refs/heads/Ubuntu-2.6.20-12.20: fast forward to branch 'Ubuntu-2.6.20-12.20' of git+ssh://rookery/srv/kernel-team/private/ubuntu-feisty old..new: 3ffd70d..3ffd70d | 07:39 |
cjwatson | error: Ref refs/heads/Ubuntu-2.6.20-12.20 is at 3ffd70ddba9d846eefd2d201404ad4c8a99d2899 but expected d02602542ef5ceb2af1765a179a9af6224136afe | 07:39 |
cjwatson | does anyone know what I'm supposed to do about that? | 07:39 |
rtg_ | cjwatson: I just edited .git/refs/heads/Ubuntu-2.6.20-12.20 and made it d02602542ef5ceb2af1765a179a9af6224136afe. It was happy after that. | 07:47 |
cjwatson | huh :-) | 07:48 |
cjwatson | nope, two git pulls later and it's unhappy again | 07:49 |
=== Keybuk [n=scott@quest.netsplit.com] has joined #ubuntu-kernel | ||
=== johanbr [n=j@blk-224-156-151.eastlink.ca] has joined #ubuntu-kernel | ||
rtg_ | cjwatson: Yeah, it does it to me too. Serious git training is something I'm going to extract from Ben and Kyle at UDS so I can understand some of these bizarre problems. | 07:54 |
=== lfittl_ [n=lfittl@213.129.230.12] has joined #ubuntu-kernel | ||
kylem | we could do a git bootcamp at uds if you wish? i could make slides | 07:55 |
rtg_ | Man, I would hug you for that, but it might wig you out :) | 07:55 |
kylem | haha. | 07:56 |
cjwatson | bootcamp> I'd like that, if I have any time | 07:56 |
kylem | i suspect a few other people in the company would like it too | 07:56 |
kylem | i'd like someone to give me a bzr boot-to-the-head heh | 07:56 |
kylem | cjwatson, are any of our bzr team coming? | 07:56 |
cjwatson | not sure, but plenty of distro are very capable with bz | 07:57 |
cjwatson | r | 07:57 |
=== kylem wants to bug them about the backend too, hehe | ||
zul | bug or annoy? | 08:14 |
=== lfittl_ [n=lfittl@213.129.230.12] has joined #ubuntu-kernel | ||
=== lfittl [n=lfittl@213.129.230.12] has joined #ubuntu-kernel | ||
=== joejaxx [i=joejaxx@fluxbuntu/founder/joejaxx] has joined #ubuntu-kernel | ||
=== doko_ [n=doko@dslb-088-073-111-226.pools.arcor-ip.net] has joined #ubuntu-kernel | ||
=== PhinnFort [n=martin@unaffiliated/phinnfort] has joined #ubuntu-kernel | ||
PhinnFort | where can I find a list/download the patches that are applied against the ubuntu kernel? | 08:56 |
PhinnFort | i'm especially interested in the usplash patch | 08:57 |
=== PhinnFort notices that the special thing about usplash is that it is completely userspace | ||
PhinnFort | :S | 09:08 |
cjwatson | that's one of the things the "u" stands for, yes :-) | 09:15 |
PhinnFort | :D | 09:19 |
PhinnFort | that means I won't have to worry about ck stepping on other patches' toes | 09:19 |
=== rpereira [n=rpereira@ubuntu/member/rpereira] has joined #ubuntu-kernel | ||
=== nott2 [n=eric@cpe-72-184-186-31.tampabay.res.rr.com] has joined #ubuntu-kernel | ||
nott2 | have they up loaded the new fix for kernel 2.6.20-14 ? | 10:10 |
cjwatson | 21:10 <cjwatson> -15.24 is in the archive and built everywhere | 10:25 |
cjwatson | 21:11 <cjwatson> except possibly ia64, and if not that's on its way | 10:25 |
cjwatson | 21:11 <cjwatson> -15.24 should work for everyone except people using sata_nv | 10:25 |
cjwatson | 21:11 <cjwatson> for sata_nv, -15.25 has been uploaded | 10:25 |
cjwatson | 21:11 <cjwatson> but it has not yet built | 10:25 |
cjwatson | 21:11 <cjwatson> (in progress, hamsters running as fast as they can) | 10:25 |
cjwatson | 21:12 <cjwatson> so sata_nv users should stay on -13 until -15.25 is available, and everyone else should use -15.24 | 10:25 |
cjwatson | 21:12 <cjwatson> and REPORT REGRESSIONS ASAP | 10:25 |
cjwatson | nott2: ^-- | 10:25 |
=== nott2 [n=eric@cpe-72-184-186-31.tampabay.res.rr.com] has joined #ubuntu-kernel | ||
=== nott2 [n=eric@cpe-72-184-186-31.tampabay.res.rr.com] has joined #ubuntu-kernel | ||
=== Wicker25 [n=wicker25@host-84-221-154-193.cust-adsl.tiscali.it] has joined #ubuntu-kernel | ||
nott2 | cjwatson: so the new kernel is uploaded ? | 11:12 |
nott2 | got a error "403" from the website when trying to update about hour ago.... | 11:13 |
nott2 | just try to get a update, nothing YET.... | 11:17 |
=== kane77 [n=kane@194.1.130.37] has joined #ubuntu-kernel | ||
=== fryfrog [n=fryfrog@gallery/fryfrog] has joined #ubuntu-kernel | ||
=== blueyed [n=daniel@pdpc/supporter/active/blueyed] has joined #ubuntu-kernel | ||
=== madzero [n=madzero@cpc2-cosh5-0-0-cust524.cos2.cable.ntl.com] has joined #ubuntu-kernel | ||
=== fryfrog [n=fryfrog@gallery/fryfrog] has left #ubuntu-kernel [] | ||
=== madzero [n=madzero@cpc2-cosh5-0-0-cust524.cos2.cable.ntl.com] has left #ubuntu-kernel [] |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!