[00:05] mahmoh: If you read the pandaboard emails, someone got that confused with the GCC-4.6 usb ehci bug. [00:06] At any rate, this bug is well documented in LP now. [00:09] mahmoh: If you have a usb stick & usb drive attatched, can you see how it performs when you have some constant i/o in the background of one while doing benchmarks to the other? I'm thinking something like "while (1);do dd if=/dev/zero of= bs=4M count=10;done" [00:44] GrueMaster: I do and I'm sure it'll work but I can try it if you're really curious to find out ;) [00:45] No, We really should get back to focusing on server tests. [00:45] agreed [03:13] mahmoh, We did discuss it. It was suggested that some changes were made that would improve things. The latest results posted seem different than earlier results, so I think the situation is improved. === zyga-afk is now known as zyga [13:18] hello, everyone ! I am trying use "dhclient" in ubuntu-arm (mini6410 board), but I have this problem http://pastebin.com/BW03ucUJ [13:18] Someone Can help me? [14:07] arcaico, Which kernel are you using? [14:10] persia, Linux FriendlyARM 2.6.28.6-FriendlyARM #2 Sat Jun 26 13:24:08 CST 2010 armv6l GNU/Linux [14:20] ppisati: I ran three pandas overnight with 3.0.0-1 with "heavy" IO and have seen no problems, I did see one problem before that on one of the boards, unsure if it's related or not, http://paste.ubuntu.com/647721/ http://paste.ubuntu.com/647768/ [14:42] arcaico: We don't support armv6. Ubuntu-arm shifted to armv7 with Lucid. [14:48] arcaico, That's probably the issue then: that doesn't look like an Ubuntu kernel. Are you sure the options mentioned in the messsage you posted are enabled? [14:49] mahmoh: looks like you lost / === prpplague is now known as prpplague^2 [14:59] ppisati: hm? should I write that one up or ? I have logs too [15:08] mahmoh: i already pushed a new kernel on my tree on zinc [15:08] mahmoh: care to try that one? [15:08] point me to the deb and I'm all over it as soon as my IO testing is complete (2 hrs or so) [15:08] ppisati: do you think it'll fix my problem above? [15:13] mahmoh: don't think so [15:14] mahmoh: anyway, 2 things: 1) can you try with the actual oneiric/natty omap4 kernel? 2) how did you get it? is it reproducible? [15:22] ppisati: I hit other problems with the actual kernel, io-usb related [15:22] I can share those with you [19:39] persia, nudge, where was that pandaboard bug? [19:42] this one? https://bugs.launchpad.net/linux-linaro/+bug/709245 [19:42] Ubuntu bug 709245 in linux-ti-omap4 "panda: USB disk IO slow" [High,Confirmed] [19:43] Neko: techincally it isn't just panda [19:43] Neko: i've been able to replicate the problem on my tegra2 board [19:44] what does the tegra have between usb host and the port on the board? [19:44] smsc hub [19:44] > [19:45] if it's trimslice it's a USB2514 which is the same thing all told as the one in the Panda.. just the Panda has the one with the built-in ethernet controller [19:45] I really think it is down to enhanced tt scheduling.. [19:47] for trimslice the ethernet is on a completely different bus so messing with USB shouldn't change any of the behavior of the ethernet stuff which is more scary indeed.. but, for now, I can't find the stupid bug number [19:48] are you using the usb-to-sata onboard? [19:49] therefore hammering usb sata makes other usb stuff go faster, or moving the mouse makes usb-sata faster? I'd love to know the effect === jkridner_ is now known as jkridner === Jack87|Away is now known as Jack87 === Jack87 is now known as Jack87|Away === Jack87|Away is now known as Jack87 === rsalveti` is now known as rsalveti === Jack87 is now known as Jack87|Away [23:46] anyone know what that awesome yet very unfriendly to users way is to basically roll back a git repository to a certain commit on a branch, and force that branch up so that commits are removed and not just HEAD moves around?