[01:06] <dja> Hey kernel people! I've been using perf, and I'm finding that there's a man page for perf-archive, but every time I try to run perf archive I see: perf: 'archive' is not a perf-command. See 'perf --help'. 
[01:07] <dja> I haven't found a bug report for this - is it known/expected?
[01:11] <Guest83662> @dja have you tried sudo for both record and archive?
[01:12] <dja> Guest83662, yes; I will just double check
[01:12] <dja> dja@linkitivity ~/d/l/linux> sudo perf record -a -- sleep 10
[01:12] <dja> [ perf record: Woken up 16 times to write data ]
[01:12] <dja> [ perf record: Captured and wrote 5.906 MB perf.data (94138 samples) ]
[01:12] <dja> dja@linkitivity ~/d/l/linux> sudo perf archive
[01:12] <dja> perf: 'archive' is not a perf-command. See 'perf --help'.
[04:33] <apw> dja, i do not know of that issue, but perf is a kernel locked command and as such a manual process to package.  so it might need a tweek.  could you file a bug against the kernel please and drop the number in here
[04:33] <dja> apw, yeah, will do
[04:34] <dja> apw, might take a bit of time, in the middle of something
[04:34] <dja> but I will let you know when I do
[04:34] <dja> I suspect we're missing perf-archive.sh from the kernel source
[04:34] <dja> but I haven't checked
[04:35] <apw> dja, yes something like that i am sure, but it is probabally the tip of the iceberg
[04:35] <apw> dja, no rush, i have plenty of other things to do
[04:35] <dja> apw, don't we all :)
[10:01] <nouveau_GTX_750_> Hi, is there anyone around for nouveau related status questions ?
=== zyga_ is now known as zyga
[10:14] <apw> nouveau_GTX_750_, you should just ask
[10:16] <nouveau_GTX_750_> OK, here goes. It will include 2 kernel traces so estimated number of lines is 100-200 lines
[10:16] <nouveau_GTX_750_> Is it OK ?
[10:17] <nouveau_GTX_750_> Subject: [nouveau/17.04] "RH Bug 1435000 - system freeze when screensaver puts monitor on standby" presumably reproduced at my 17.04 Ubuntu desktop + a 2nd possible nouveau issue
[10:17] <ubot5`> Error: Launchpad bug 1435000 could not be found
[10:25] <nouveau_GTX_750_> There is a limit to the number of chars per message so I will send it line by line
[10:25] <nouveau_GTX_750_> Is Canonical considering a merge of RH_1435000 fix into 17.04 (Zesty) Ubuntu release?
[10:26] <nouveau_GTX_750_> From my /var/log/syslog after reboot:
[10:27] <nouveau_GTX_750_> May  1 19:53:58 MS-7971 kernel: [ 7877.771144] general protection fault: 0000 [#1] SMP
[10:28] <nouveau_GTX_750_> May  1 19:53:58 MS-7971 kernel: [ 7877.771328] CPU: 7 PID: 1168 Comm: Xorg Not tainted 4.10.0-20-generic #22-Ubuntu May  1 19:53:58 MS-7971 kernel: [ 7877.771337] Hardware name: MSI MS-7971/Z170A PC MATE (MS-7971), BIOS A.20 07/17/2015
[10:29] <nouveau_GTX_750_> May  1 19:53:58 MS-7971 kernel: [ 7877.771362] RIP: 0010:kmem_cache_alloc_trace+0x7b/0x190
[10:29] <nouveau_GTX_750_> May  1 19:53:58 MS-7971 kernel: [ 7877.771469] Call Trace:
[10:30] <nouveau_GTX_750_> May  1 19:53:58 MS-7971 kernel: [ 7877.771546] nouveau_fence_new+0x42/0xc0 [nouveau] May  1 19:53:58 MS-7971 kernel: [ 7877.771618] nouveau_gem_ioctl_pushbuf+0xe80/0x1610 [nouveau] May  1 19:53:58 MS-7971 kernel: [ 7877.771647] drm_ioctl+0x21b/0x4c0 [drm]
[10:30] <nouveau_GTX_750_> May  1 19:53:58 MS-7971 kernel: [ 7877.771714]  ? nouveau_gem_ioctl_new+0x150/0x150 [nouveau] May  1 19:53:58 MS-7971 kernel: [ 7877.771781] nouveau_drm_ioctl+0x74/0xc0 [nouveau] May  1 19:53:58 MS-7971 kernel: [ 7877.771793] do_vfs_ioctl+0xa3/0x610
[10:31] <nouveau_GTX_750_> $ lspci | grep -i nvidia 01:00.0 VGA compatible controller: NVIDIA Corporation GM107 [GeForce GTX 750 Ti] (rev a2)
[10:32] <nouveau_GTX_750_> RH Bug 1435000 details:  $ curl https://bugzilla.redhat.com/show_bug.cgi?id=1435000 > RH_1435000
[10:32] <ubot5`> bugzilla.redhat.com bug 1435000 in xorg-x11-drv-nouveau "system freeze when screensaver puts monitor on standby" [High,New]
[10:32] <ubot5`> Error: Launchpad bug 1435000 could not be found
[10:32] <nouveau_GTX_750_> 2389 Apr 08 21:03:24 max.kudr.me kernel:  nouveau_fence_new+0x42/0xc0 [nouveau]
[10:44] <nouveau_GTX_750_> 3770 &gt; You need to grab kernel 4.10.12.</span > 3771 3772 Awesome!!! It works! I can finally go for a coffee break again (without a reboot at least).
[10:45] <nouveau_GTX_750_> vim /usr/share/doc/linux-image-4.10.0-20-generic/changelog.Debian.gz  ...   69   * Zesty update to v4.10.8 stable release (LP: #1678930) ...   87     - Linux 4.10.8
[10:45] <ubot5`> Launchpad bug 1678930 in linux (Ubuntu Zesty) "Zesty update to v4.10.8 stable release" [Undecided,Fix released] https://launchpad.net/bugs/1678930
[10:45] <nouveau_GTX_750_> A 2nd trace I am also experiencing:
[10:46] <nouveau_GTX_750_> May  2 16:08:04 MS-7971 kernel: [ 8773.740485] BUG: unable to handle kernel NULL pointer dereference at 0000000000000020 May  2 16:08:04 MS-7971 kernel: [ 8773.741395] IP: dma_fence_wait_timeout+0x36/0xf0 May  2 16:08:04 MS-7971 kernel: [ 8773.742286] PGD 0 May  2 16:08:04 MS-7971 kernel: [ 8773.742287] May  2 16:08:04 MS-7971 kernel: [ 8773.744044] Oops: 0000 [#1] SMP
[10:46] <nouveau_GTX_750_> May  2 16:08:04 MS-7971 kernel: [ 8773.748223] CPU: 6 PID: 1265 Comm: InputThread Not tainted 4.10.0-20-generic #22-Ubuntu May  2 16:08:04 MS-7971 kernel: [ 8773.748724] Hardware name: MSI MS-7971/Z170A PC MATE (MS-7971), BIOS A.20 07/17/2015
[10:47] <nouveau_GTX_750_> May  2 16:08:04 MS-7971 kernel: [ 8773.755429] Call Trace: May  2 16:08:04 MS-7971 kernel: [ 8773.755884] drm_atomic_helper_wait_for_fences+0x48/0x120 [drm_kms_helper] May  2 16:08:04 MS-7971 kernel: [ 8773.756350] nv50_disp_atomic_commit+0x19c/0x2a0 [nouveau] May  2 16:08:04 MS-7971 kernel: [ 8773.756784] drm_atomic_commit+0x4b/0x50 [drm]
[10:47] <smb> nouveau_GTX_750_, if the fix is in upstream 4.10.12 (which I think was said somewhere this lot of messages), then it is pending on the stack for the next update after this update cycle
[10:47]  * ogra_ points gently to http://paste.ubuntu.com
[10:47] <nouveau_GTX_750_> yes, at least to the issue I am about to write
[10:48] <nouveau_GTX_750_> Any ETA to 4.10.12 landing ?
[10:48] <smb> nouveau_GTX_750_, if things go smoothly (which they rarely do), then about 5 weeks from now on
[10:50] <smb> nouveau_GTX_750_, 3 weeks maybe if you take kernel packages from the proposed pocket
[10:50] <nouveau_GTX_750_> OK, I will be patient. Do you need help in covering nouveau + "NVIDIA Corporation GM107 [GeForce GTX 750 Ti] (rev a2)" ?
[10:50] <nouveau_GTX_750_> I am paranoid so I I use official packages only
[10:53] <smb> nouveau_GTX_750_, if what is needed is in https://lists.ubuntu.com/archives/kernel-team/2017-May/083889.html it should be there. I think we should also get to include 4.10.13 at least
[11:01] <nouveau_GTX_750_> "drm/nouveau/kms/nv50: fix double dma_fence_put() when destroying plane state" should fix my 2nd issue at least
[11:02] <nouveau_GTX_750_> I offer testing facilities - Canonical should be able to reproduce "double dma_fence_put" issue using current 17.04 release
[11:24] <nouveau_GTX_750_> I will wait for 6 more minutes and logout. IMHO, this nouveau issue is a nice metric for nvidia's open source driver coverage using induced sleep state. My interest in adding it to the test plan - knowing that chance of getting the same issue when using Ubuntu's latest release is minimal.
=== mamarley is now known as Guest96638
=== mamarley_ is now known as mamarley
=== JanC is now known as Guest52042
=== JanC_ is now known as JanC
=== DavidDuffey is now known as dduffey