=== gerald is now known as Guest46363 === swordsmanz is now known as aww_ === aww_ is now known as Guest34257 === Guest34257 is now known as aww___ === aww___ is now known as swordsmanz [10:43] apw: Could you be convinced to include "s390/kernel: fix ptrace peek/poke for floating point registers" (55a423b) in the next xenial kernel? Fixes a panic we ran into on the buildds today. [10:43] (tested by building gdb; the panic was first seen in its test suite) [10:44] wgrant, yep, i see from your ppa that that is a backport? do you have the one you have tested, oculd you email it to me and i'll apply it [10:45] wgrant, and if there is a bug number wang it in here [10:46] apw: I haven't filed a bug, can do so in a bit if you need. The backport is trivial, but will find a patch. [10:49] apw: http://paste.ubuntu.com/13665866/ is the diff. [10:49] Looks identical to the git diff for me, but it didn't apply directly. Possibly line numbers too far off or something. [10:50] Except they're the same. [10:50] Weird. [11:22] ack thanks [11:24] wgrant, ok it cherry-picked for me, so ... wierd [11:26] and that ... i recon ... is a cve if "overwrites the task structure" [11:26] is correct [11:26] apw: Oh yeah, it's totally exploitable. [11:26] I probably just misapplied it, I was rushing. [11:27] given we have a bad s390x package split as well pending on tip [11:27] i think we should get that uploaded sooner rather than later [11:27] though i would like this 4.3 to go out at least [11:27] i assume this is affecting the bootstrap of main though .. right ? [11:28] I have it installed on one buildd (z13-008), and it's only affected gdb so far. [11:28] main is built, apart from various unrelated FTBFSes. [11:28] I just manually threw gdb at the one with the fixed kernel. [11:28] so you goosed gdb thorught hte fixed one, right [11:28] ok then i will let this one migrate, i should have the final testing ok in about 2 hours [11:28] Yep. [11:28] Not super urgent. [11:28] then i'll get this one uploaded [15:54] apw, yo =) [15:54] xnox, hi [15:54] totally forgot, could you please cherrypick a patch that prevents buildds blowing up, on funny PIE builds? =) [15:54] apw, unless one of the other engineers already talked to you about it [15:55] "s390/kernel: fix ptrace peek/poke for floating point registers" [15:55] xnox, ^^ that ? [15:55] apw, yes that. [15:55] xnox, is picked and will be in the next upload, i was under the impression only gdb had hit it thus far and had been handled ? [15:56] apw, yeah, wgrant quickly built a kernel and upgraded buildd8, so buildd 8 is a good one =) [15:56] but having that in the next kernel would be dandy. thank you =) [15:56] heh is that why she is offline, held for special projects [15:57] xnox, that and your split to make linux-virtual work are on deck for the next upload [15:58] apw, brilliant. [16:04] apw, I could turn the crank on that as soon as -1.10 is promoted [16:04] yep, i've dropped the tag, just waiting [16:04] otherwise it might be a few days [16:05] should be done within the hour in theory [16:05] because you know friday is a classy day to release a kernel [16:05] apw, I'll check back in a bit [16:05] especially a new kernel release [16:06] perhaps we should have waited till monday [16:12] apw, this way we'll have plenty of email to deal with first thing Monday morning [16:13] rtg, i might be able to catch it before it copies out [16:14] apw, oh, let it go. I think its pretty stable [16:14] apw, after all, its got the cking stamp of approval [16:15] yeah : [16:15] :) [16:15] who upgrades over teh weekend anyhow, thats a silly idea [16:16] release often I say [16:17] yeah i'd like it to be a lot cheaper to update but its really not [16:54] jsalisbury: hello [16:55] cristian_c, hello. I haven't had much time to look into the new issues yet. It sounds like the regression is now fixed for you? [16:56] jsalisbury: yes, as I've stated in launchpad report [16:56] cristian_c, ok. I'll make some time to look at the other issue [16:57] jsalisbury: I've got a question [16:57] cristian_c, sure [16:58] jsalisbury: last time, I've read fix went in 4.3-rc2 [16:59] cristian_c, the fix for the regression? or the fix for the new issue? [16:59] jsalisbury: but I've noticed that regression is disappeared already in 4.2.0-18 [16:59] jsalisbury: about regression [16:59] cristian_c, the fix was cc'd to upstream stable, so it's eventually going to make it's way into all the stable kernels [16:59] 4.2.0-16 is default kernel in 15.10 [17:00] after first updates, kernel goes to 4.2.0-18 version number [17:01] jsalisbury: so, where can I find the commit (if it's public) [17:01] ? [17:02] cristian_c, this is the commit: [17:02] commit 8a1513b49321e503fd6c8b6793e3b1f9a8a3285b [17:02] Author: Kyle Evans [17:02] Date: Fri Sep 11 10:40:17 2015 -0500 [17:02] hp-wmi: limit hotkey enable [17:02] ok, thank you very much [17:03] cristian_c, you can get it out of linus' tree or any of the stable trees, but it will have a different SHA1 [17:03] in the stable trees [17:03] ok [23:20] when do updates from the upstream kernel get backported to LTS kernel backport packages? [23:21] specifically, I need https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=0480334 in linux-image-generic-lts-utopic for trusty [23:26] ccope: hmm... i see. so, that commit is meant to be applied to stable kernels >= v3.18, thus it will never be applied to the corresponding stable kernel (3.16.7-ckt) [23:27] ccope: however, the utopic kernel actually backported overlayfs so that commit may actually be applicable to that kernel [23:29] ccope: my suggestion would be to open a bug against that kernel, requesting that specific commit to be included [23:29] (feel free to assign that bug to me ;-) ) [23:31] henrix, ok cool, thanks! [23:55] henrix, i'm unable to assign you but the bug is here: https://bugs.launchpad.net/ubuntu/+source/linux-lts-utopic/+bug/1523000 [23:55] Ubuntu bug 1523000 in linux-lts-utopic (Ubuntu) "writing files larger than 2GB to overlayfs fails" [Undecided,New] [23:59] ccope: ack, thanks. i'll assign it to myself later