[00:10] jsalisbury: e63a94f12b5fc67b2b92a89d4058e7a9021e900e is good [00:12] jsalisbury: so that's a merge commit, and 'git bisect' stops there for me? [00:14] slangasek, Yeah, your right. The bisect usually goes down into a merge, but it isn't in this case. [00:14] slangasek, I'll look at that merge and see if there are commits in it that stuck out and could be the fix [00:15] slangasek, If nothting sticks out, I can manually bisect in that merge I suppose. [00:15] jsalisbury: k. it's "only" 105 commits on the tty stack, 4kloc :) [00:15] slangasek, heh, yeah [00:17] slangasek, I hate when bisects don't have a easy and happy ending. I'll let you know when a kernel is ready. May not be until the morning though. [00:18] jsalisbury: no problem [00:19] 979990c6284814617d8f2179d197f72ff62b5d85 looks vaguely interesting [00:20] by and large, the merge commit's description of being "mostly serial" is supported by the commit log [00:22] slangasek, that was cc'd to stable, so it's on its way to us anyway. I could build a test kenel with it just for the heck of it. [01:22] slangasek, if you want to test it out, I built an Artful kernel with a pick of commit 979990c6284814617d8f2179d197f72ff62b5d85 [01:23] slangasek, if that's not it, I'll dig further into that merge [03:09] jsalisbury: that one didn't do the trick either :/ [11:15] slangasek, I'll look at that merge closeer and manually bisect it if needed. [11:18] jsalisbury, is there any reason you canont just start a new bisct between the first commit on the branch and the git mergebase of the two parents of the merge ? [11:55] apw, I'll give that a try. I'm not sure whey the bisect didn't go down into the merge. [11:56] no indeed that is idd [11:59] maybe it was an octopus merge (3 or more branches at once) ? [12:02] TJ-, maybe. apw, it looks like your method shoudl work [12:11] slangasek I started a new "Reverse" bisect of just the merge using: [12:11] # good: [1a3b85ea36d38d5732fdd86b321b10bcaeb53512] Merge tag 'usb-4.14-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb [12:11] # bad: [3840ed9548f778717aaab5eab744da798c3ea055] tty: goldfish: Implement support for kernel 'earlycon' parameter [12:11] I built the first test kernel, up to the following commit: [12:11] ef954844c7ace62f773f4f23e28d2d915adc419f === cachio is now known as cachio_lunch === cachio_lunch is now known as cachio [17:57] jsalisbury: ef954844c7ace62f773f4f23e28d2d915adc419f bad [18:00] slangasek, thanks, I'll build the next one. [18:29] slangasek, kernel up to commit ec085c5a51b768947ca481f90b66653e36b3c566 is ready [18:44] jsalisbury: ec085c5a51b768947ca481f90b66653e36b3c566 bad [18:44] slangasek, ack. I'll build the next one [18:54] slangasek, test kernel up to commit cdb939456f81878cda1f31ba0ef1a3b1d7be07e8 is ready. [19:57] jsalisbury: cdb939456f81878cda1f31ba0ef1a3b1d7be07e8 bad [20:08] slangasek, ack, building the next one [20:18] slangasek, commit 2ce8008711e4837c11e99a94df55406085d0d098 is ready. [21:14] jsalisbury: 2ce8008711e4837c11e99a94df55406085d0d098 bad [21:14] slangasek, ack, I'll build the next one [21:25] slangasek, commit d01c3289e7d68162e32bc08c2b65dd1a216a7ef8 is ready. [21:29] jsalisbury: not visible on http://kernel.ubuntu.com/~jsalisbury/lp1724911/ [21:44] slangasek, sorry, should be there now. [21:47] jsalisbury: looks good, thanks [22:14] jsalisbury: d01c3289e7d68162e32bc08c2b65dd1a216a7ef8 bad [22:15] slangasek, ack [22:38] slangasek, commit ae28d7402a7e7cf3dd0da1f2eb6a52b337873e08 is ready [23:26] jsalisbury: eventually one of these kernels is going to be good again, right? :) ae28d7402a7e7cf3dd0da1f2eb6a52b337873e08 bad [23:28] jsalisbury: so I went back to the bug log and noticed you listed your bisect as between 1a3b85ea36d38d5732fdd86b321b10bcaeb53512 and 3840ed9548f778717aaab5eab744da798c3ea055 but both of those commits were bad [23:51] slangasek, hmm, I must have gotten the merge commits wrong from e63a94f12b. I'll investigate a bit and ask apw if I can't figure it out.