=== gerald is now known as Guest36633 [07:16] hi. has anyone seen the 3.18 kernel hang bug? [07:16] the one that wedges everything and disables everything except alt+sysrq+b [07:16] * hyperair thinks it's the same as the hpet bug === gerald is now known as Guest11798 [09:08] hyperair, not seen any particular mention of that no [09:08] damn [09:08] it keeps hitting me, and for some reason nobody else sees it [09:08] * hyperair sighs [09:08] hyperair, does it happen with the v3.19 in the CKT PPA? as that will be our kernel shortly anyhow [09:09] apw: dunno, i just compiled my 3.19 kernel. haven't rebooted yet [09:09] hello [09:09] apw, Have you read the paste? [09:36] cristian_c, i have no indication in my backscroll that you told me where it was after you said you'd do it; but i have been travelling [09:37] apw, http://pastebin.com/TVmxXzUZ [09:39] cristian_c, that looks comprehensive to me [09:39] ok [10:00] apw, ok, also, I've to open an upstream report for a regression [10:00] related to this bug [10:01] in total, two reports [11:01] I must collect info in kernel format for an upstream bug report [11:02] the upstream bug report should be about a regression [11:02] How can I locate the specific commit, exactly? [11:02] Any ideas? [11:18] cristian_c, the specific commit which introduced the regression ? [11:20] apw, yes [11:20] correct [11:20] I think they request the exact kernel version [11:23] the exact kernel version is easier, you would use the mainline builds to narrow in on it, in a bisection style [11:24] apw, ok, so have I to try many kernel until I find the nearest previous version and the nearest next version? [11:25] to the kernel version has introduced the regression [11:26] cristian_c, right exactly that, but you do it in divide-and-conquor manner, so [11:26] ok [11:26] say you know it worked in 3.2 and not 3.16 you'd pick like 3.9 first and then you have gotten rid of half either result [11:27] cristian_c, i would use the prebuilt test kernels in the ubuntu mainline archive, so you don't have to make any [11:27] 3.8.0 is not affected by the regresssion [11:28] apw, can I use the prebuilt kernels even if I've to report in the upstrem? [11:28] *to the [11:28] to locate the regression version [11:29] apw, usually I download the kernel version from mainline kernel archive [11:29] http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D [11:30] and I always install three packages: headers, headers-generic, linux-image [11:50] cristian_c, yes, those kernels are mainline code exactly, there is no ubuntu patches on those [11:52] ok, usually I install the mainline .deb packages [11:52] for the prebuilt is the same , I think [11:52] *prebuilt kernels [11:52] i believe we are talking about the same ones yes, those are suitible for a bisect, and for reporting exact versions to upstream [11:52] as they are unmodified upstream code, built on ubuntu for debugging [12:08] apw, anyway, I've to identifying the mantainer [12:09] If the bug is not in USB, the first step in reporting a bug upstream is to find the maintainer of the driver for the bug from the MAINTAINERS list. Here are some typical scenarios and suggested mailing list and procedures: [12:09] in the first case [12:09] If this is a regression, CC the submitter of the regression commit. [12:09] in the second case [12:10] http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/MAINTAINERS [12:10] if you can find the before and after mainline commit, we can ask joe to do a proper bisect for you to get the actual commit for it [12:12] I'll try === ara is now known as Guest94549 [14:44] hi guys, i noticed the automatic DEB build of the 3.19 kernel on kernel.ubuntu.org failed, is this a known issue? [14:47] sarnex, failed how so ? [14:47] apw: BUILD.LOG.amd64 and BUILD.LOG.i386 say "Access not authroized" and there are no debs in the filder [14:48] apw: see here http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.19-vivid/ [14:54] sarnex, thanks, sorting [14:54] apw: awesome thanks === FreezingDroid is now known as FreezingCold === pgraner is now known as pgraner-afk === mwenning is now known as mwenning-rr5