=== Ming is now known as Guest62359 === BenC__ is now known as BenC [04:20] Here it goes. Please excuse me if I offend you. It is not my intent. [04:21] For a long time kernel devs have been ignoring users. [04:21] I believe they simply don't care. If for a year there [04:21] has been an OPEN SOURCE drivers for the LARGEST laptop [04:21] manufacturer with millions of computers sold with this [04:21] particular hardware and at least 20 versions of kernel [04:21] since, where does that leave us? [04:21] Here it goes: [04:21] Here are the OS drivers that are available: [04:21] http://www.ralinktech.com/en/04_support/support.php?sn=501 [04:21] The drivers I mean are: [04:21] RT539x PCIe 09/26/2011 2.5.0.3 [04:21] Our lspci -k: [04:21] Network controller: Ralink corp. RT5390 [802.11 b/g/n 1T1R G-band PCI Express Single Chip] [04:21] Kernel driver in use: rt2800pci [04:21] Kernel modules: rt2800pci [04:21] These ARE NOT GOOD. They work, but are designed for other Ralink hardware. [04:21] They have bugs. [04:21] For the above module it should be the OS drivers from the above website. [04:21] They are OPEN SOURCE and AVAILABLE! [04:21] The RT5390 PCI is in shit loads of HPs laptops. Shit loads of! [04:21] And the hardware is great! The drivers for Windows are awesome! [04:21] uh, no [04:21] And the drivers for Linux have been available for a long long time! [04:21] And yet WE ARE IGNORED for shit load of kernel versions, shit loads [04:21] of versions! WTF?! [04:21] PS. You have to probably blacklist some other shitty drivers like: [04:21] gksudo leafpad /etc/modprobe.d/blacklist.conf [04:21] # Blacklist conflicting RaLink driver modules: [04:21] blacklist rt2800pci [04:21] blacklist rt2800lib [04:21] blacklist rt2x00usb [04:21] blacklist rt2x00pci [04:21] blacklist rt2x00lib [04:21] blacklist rt2860sta [04:22] oh, no? [04:22] this isn't even the right place for that, if there was a place to flood a screed [04:22] do you think the linux wireless people don't know of, or have a real reason to not include those drivers? [04:22] oh yeah not the place, just another FU user! You are nice as those kernel devs who care more about google and android than users! [04:23] you don't know what you are talking about [04:23] if you presume to put yourself in the place of a person who can make that decision, decide to use the driver yourself and forget about the actual people who are [04:24] there has been at least 20 versions and they ignored us. I don't simply understand why android is more important than us. [04:25] They always cry that there are no drivers, but there are and even Open Source. In a moth there will be a YEAR! [04:26] if you post that every time, i'm not surprised you are ignored; you are not ignored [04:26] you aren't even asking the right people, in this case [04:26] ? [04:26] further, you assume the people who decide these things have no other reason to not do it, besides ignoring you [04:27] I only care about resaults. The resault is NO DRIVER. [04:27] if you only care about results then you can't factually say that you have been ignored, as you have made no obvious attempt at actually meeting your goals [06:48] moin [07:47] morning [07:48] brb [07:48] moin Stefan :) [07:50] ppisati, Morning Paolo. Right back within one second. Not bad. :) === emma is now known as em === JanC_ is now known as JanC [10:08] * ppisati goes out for a bit [10:24] hello, on my samsung laptop, xubuntu 12.04 the screen-brightnes-control with fn keys was working with kernel 3.2.0-23 and after update to 3.2.0-27 there is only low light and bright, nothing in-between and the notification (with the sun) doesnt appear anymore [10:24] what can i do? [10:24] to fix this? [10:29] the grub entries are the same, just the kernel changed, semms to be a bug of 3.2.0-27 === yofel_ is now known as yofel [12:41] oumpa, if you were still here i would suggest fileing a bug and testing the kernels in between from launchpad to see which broke it === amitk is now known as amitk-afk [13:29] * rtg delivers vehicle for maintenance === rtg is now known as rtg-afk === tjaalton_ is now known as tjaalton === kentb-out is now known as kentb === rtg-afk is now known as rtg === kengyu_ is now known as kengyu === ante_ is now known as ivoks [14:46] * ogasawara back in 20 [15:51] I'd love to have the AF_BUS patch added to the kernel for the i.MX53 QSB. [15:51] Should I submit that to Linaro? Or is there somewhere in Ubuntu I should ask? [15:52] linaro ... [15:52] Okay, thanks [15:52] since the kernel package for mx5 is maintained by them [16:18] ogasawara, thx for the update ! [16:19] ogra_: np [16:45] kees, finally! http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=800179c9b8a1e796e441674776d11cd4c05d61d7 [16:56] * ppisati takes a break... [17:07] tyhicks, re 3.6-rc1 'The write-through cache changes are being reverted due to bugs...' - should we be backporting those patches ? [17:10] rtg: Probably so [17:10] rtg: I didn't feel comfortable marking them for the stable kernel just yet because I haven't had a chance to look at the backport [17:10] tyhicks, ack [17:11] rtg: I was going to speak with cking about it because he is mostly familiar with the bugs caused by the reverted code [17:11] tyhicks, I think he's on vacation for a couple of weeks [17:11] (he wrote a test case for one of them) [17:11] oh [17:12] tyhicks, prolly for the empty files case ? [17:13] rtg: Nope. The biggest problem is that ENOSPC can't be detected at the correct time to return from write(). [17:13] ah [17:14] rtg: So userspace thinks everything is fine and may keep writing but eCryptfs can't detect a full lower filesystem until the page writeback is performed. [17:15] tyhicks, wouldn't an unlayered FS have the same problem ? [17:15] rtg: No, because they are able to do internal tracking of available space [17:33] * smb -> EOD [17:42] * rtg -> lunch [18:00] * ppisati goes out for some chinese food [19:36] ogasawara, pushed quantal master-next to fix 'II: Checking modules for generic... [19:36] reading new modules...read 3474 modules. [19:36] reading old modules... [19:36] MISS: aufs [19:36] read 3475 modules : new(0) missing(1) [19:36] EE: Missing modules (start begging for mercy) [19:36] ma' [19:36] rtg: ah, thanks [19:37] ogasawara, am working through recent ecryptfs updates to see what might be SRU worthy. [19:37] rtg: ack [19:40] rtg: Sorry, I'm now seeing that I forgot to set the "Cc: stable@..." tag on any of the patches in my last pull request [19:40] rtg: Let me get you a list of what I think is SRU-worthy [19:40] tyhicks, that would be great, thanks. [19:40] 821f749 eCryptfs: Revert to a writethrough cache model [19:40] e3ccaa9 eCryptfs: Initialize empty lower files when opening them [19:41] 8bc2d3c eCryptfs: Unlink lower inode when ecryptfs_create() fails [19:41] 069ddcd eCryptfs: Copy up POSIX ACL and read-only flags from lower mount [19:41] rtg: ^ I see that you've already got the ball rolling on that last one [19:41] tyhicks, yep, it looked like a good one [19:42] rtg: The empty lower files fixes should solve one of the most nagging usability problems [19:43] tyhicks, I'm also looking at 2.6.32.y to see if it would benefit from a backport of 069ddcd. All the other kernel versions are a clean cherry-pick [19:43] tyhicks, agreed about lower files fixes... [19:49] rtg: re 2.6.32.y> I would say so [19:49] tyhicks, k, I'll look into it [19:50] rtg: I see the if "(... !IS_POSIXACL(inode)) { mode &= ~current_umask(); }" pattern several places in fs/namei.c [19:51] tyhicks, I saw we had on precise/oneiric commit 4a26620 "eCryptfs: Improve statfs reporting" and forwarded to stable. It seemed worth to have on stables, cking was in favour too. But by Ben's response, not sure he will pick. [19:51] herton: Yeah, I was about to respond to that. I think Ben is right. :/ [19:52] herton: I hadn't considered his use case [21:00] * rtg -> EOD [22:55] i'm running ubuntu 12.04 LTS in a KVM VM on Ubuntu 12.04 LTS, if I bidirection iperf my VM from the host node the NIC will lock up requiring if up/down after a few hours [22:56] i've tried to reproduce this with centos5 VM on the same host (its the only other readily available template right now) and I was unable to reproduce the bug. I'm currently testing 10.04 but it will be a few hours until I know if its working [22:56] has anyone seen this kind of issue before?