Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-09 Thread Ulrich Pegelow
Thanks. Works here. Let's see if his also fixes the SegFault reported by the TO. Ulrich Am 09.04.2017 um 16:49 schrieb Roman Lebedev: And, pushed. ___ darktable developer mailing list to unsubscribe send a mail to darkta

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-09 Thread Roman Lebedev
On Sun, Apr 9, 2017 at 3:43 PM, Ulrich Pegelow wrote: > Am 09.04.2017 um 09:31 schrieb Roman Lebedev: >> >> On Sun, Apr 9, 2017 at 9:59 AM, Ulrich Pegelow >> wrote: >>> >>> Am 08.04.2017 um 20:04 schrieb Roman Lebedev: Well, that is *very* strange indeed. If it *reliably* happ

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-09 Thread Roman Lebedev
On Sun, Apr 9, 2017 at 3:43 PM, Ulrich Pegelow wrote: > Am 09.04.2017 um 09:31 schrieb Roman Lebedev: >> >> On Sun, Apr 9, 2017 at 9:59 AM, Ulrich Pegelow >> wrote: >>> >>> Am 08.04.2017 um 20:04 schrieb Roman Lebedev: Well, that is *very* strange indeed. If it *reliably* happ

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-09 Thread Ulrich Pegelow
Am 09.04.2017 um 09:31 schrieb Roman Lebedev: On Sun, Apr 9, 2017 at 9:59 AM, Ulrich Pegelow wrote: Am 08.04.2017 um 20:04 schrieb Roman Lebedev: Well, that is *very* strange indeed. If it *reliably* happens for you, then maybe you could also bisect this within the submodule itself? Very c

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-09 Thread Ulrich Pegelow
Am 09.04.2017 um 09:31 schrieb Roman Lebedev: On Sun, Apr 9, 2017 at 9:59 AM, Ulrich Pegelow wrote: Am 08.04.2017 um 20:04 schrieb Roman Lebedev: Well, that is *very* strange indeed. If it *reliably* happens for you, then maybe you could also bisect this within the submodule itself? Very c

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-09 Thread Roman Lebedev
On Sun, Apr 9, 2017 at 9:59 AM, Ulrich Pegelow wrote: > Am 08.04.2017 um 20:04 schrieb Roman Lebedev: >> >> Well, that is *very* strange indeed. >> If it *reliably* happens for you, then maybe you could also bisect this >> within the submodule itself? >> > > Very clear result: Aha, now that makes

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-09 Thread Ulrich Pegelow
Am 08.04.2017 um 20:04 schrieb Roman Lebedev: Well, that is *very* strange indeed. If it *reliably* happens for you, then maybe you could also bisect this within the submodule itself? Very clear result: 7f087325d09e2b6d4ecc392f7aee44dd29fafe62 is the first bad commit commit 7f087325d09e2b6d4e

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-08 Thread Ulrich Pegelow
I'll try to have a closer look tomorrow. Ulrich Am 08.04.2017 um 20:04 schrieb Roman Lebedev: On Sat, Apr 8, 2017 at 6:23 PM, Ulrich Pegelow wrote: Strange things also happen here. All of a sudden darktable would hang during startup in the OpenCL init phase of my AMD card (catalyst driver).

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-08 Thread Roman Lebedev
On Sat, Apr 8, 2017 at 6:23 PM, Ulrich Pegelow wrote: > Strange things also happen here. All of a sudden darktable would hang during > startup in the OpenCL init phase of my AMD card (catalyst driver). > > I bisected the issue and found this: > 97085123a6c29bed86fc7795d527c95bd50a is the firs

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-08 Thread Mark Heieis
I found that clinfo also segfaulted. Further investigation showed that the amdgpu libs had been changed. Reinstalling the few required libs solved my issue. So I suspect that an OS upgrade clobbered them. I placed them in /lib64 rather than in /opt

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-08 Thread Ulrich Pegelow
Strange things also happen here. All of a sudden darktable would hang during startup in the OpenCL init phase of my AMD card (catalyst driver). I bisected the issue and found this: 97085123a6c29bed86fc7795d527c95bd50a is the first bad commit commit 97085123a6c29bed86fc7795d527c95bd50a A

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-05 Thread Germano Massullo
Il 05/04/2017 22:36, Mark Heieis ha scritto: > > Yep, darktable opens fine with opencl disabled, so clearly an (amd > driver/lib?) opencl issue. > Then you should fill a bugreport on AMD support website, so that they can fix it __

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-05 Thread Mark Heieis
Yep, darktable opens fine with opencl disabled, so clearly an (amd driver/lib?) opencl issue. On 2017-04-05 13:01, Mark Heieis wrote: Yes, I found that clinfo seg faults as well. So not sure if it's with the AMD

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-05 Thread Mark Heieis
Yes, I found that clinfo seg faults as well. So not sure if it's with the AMD libs, the open source driver piece or a kernel update that's making this happen. On 2017-04-05 12:42, Roman Lebedev wrote: On Wed, Apr 5, 2017 a

Re: [darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-05 Thread Roman Lebedev
On Wed, Apr 5, 2017 at 10:29 PM, Mark Heieis wrote: > Hi There, > > Today, for some reason, darktable seg faults with the following. I used it > yesterday and it worked fine. I may have updated the OS after my last use > and could roll-back the update to check. Any thoughts? > > Linux copernicus 4

[darktable-dev] All of a sudden, darktable Thread Seg faults

2017-04-05 Thread Mark Heieis
Hi There, Today, for some reason, darktable seg faults with the following. I used it yesterday and it worked fine. I may have updated the OS after my last use and could roll-back the update to check. Any thoughts? Linux copernicus 4.10.8-20