GNOME Bugzilla – Bug 794991
msdk: plugin fails to load
Last modified: 2018-04-05 00:37:17 UTC
$ gst-inspect-1.0 msdk 0:00:00.192163330 26590 0x1204b20 ERROR msdk msdk.c:167:msdk_open_session: Intel Media SDK not available (undeveloped feature) No such element or plugin 'msdk' libva (master) heads/master-0-g3f3e70542a4f https://github.com/01org/libva msdk (master) heads/master-0-gc56e75c1beee https://github.com/Intel-Media-SDK/MediaSDK gstreamer (master) heads/master-0-g22fac0b21200 https://anongit.freedesktop.org/git/gstreamer/gstreamer gst-plugins-base (master) heads/master-0-gc64d2cf3969f https://anongit.freedesktop.org/git/gstreamer/gst-plugins-base gst-plugins-good (master) heads/master-0-g0f1770f2bdfe https://anongit.freedesktop.org/git/gstreamer/gst-plugins-good gst-plugins-bad (master) heads/master-0-gfc989ce544c5 https://anongit.freedesktop.org/git/gstreamer/gst-plugins-bad gstreamer-vaapi (master) heads/master-0-g966d4d0476a9 https://anongit.freedesktop.org/git/gstreamer/gstreamer-vaapi
Created attachment 370534 [details] [review] patch
Review of attachment 370534 [details] [review]: There is a property called "hardware" in decoder, vpp and encoder elements to choose between hardware and software implementations. This is a useful feature to debug issues in hardware acceleration but unfortunately software implementation is only supported in windows. I believe this issue should be fixed in mediasdk side. WDT?
Review of attachment 370534 [details] [review]: Actually, I was wrong about MFX_IMPL_HARDWARE_ANY... it seems to work fine. So we just need a way for msdk_is_available to invoke msdk_open_session with an "AUTO" to check for SW and/or HW availability.
(In reply to U. Artie Eoff from comment #3) > Review of attachment 370534 [details] [review] [review]: > > Actually, I was wrong about MFX_IMPL_HARDWARE_ANY... it seems to work fine. > So we just need a way for msdk_is_available to invoke msdk_open_session with > an "AUTO" to check for SW and/or HW availability. Exactly :)
Created attachment 370535 [details] [review] patch v2
Review of attachment 370535 [details] [review]: git apply failed...Can you please rebase it on master? Also remove the trailing whitespaces in 28, 29, 42, 52, 61
Created attachment 370538 [details] [review] patch v3 This version is a direct upload of the file from `git format-patch`... my previous patch was a copy/paste text upload (which I guess introduced the ws artifacts). Patch is already based on latest master fc989ce544c5. Let me know if you still have `git am/apply` issues.
(In reply to U. Artie Eoff from comment #7) > Created attachment 370538 [details] [review] [review] > patch v3 > > This version is a direct upload of the file from `git format-patch`... my > previous patch was a copy/paste text upload (which I guess introduced the ws > artifacts). Patch is already based on latest master fc989ce544c5. Let me > know if you still have `git am/apply` issues. There is a slight indentation issue, but I will fix it locally. Next time onwards please run gst-indent-1.0 on all *.c files before git format patch.
(In reply to sreerenj from comment #8) > (In reply to U. Artie Eoff from comment #7) > > Created attachment 370538 [details] [review] [review] [review] > > patch v3 > > > > This version is a direct upload of the file from `git format-patch`... my > > previous patch was a copy/paste text upload (which I guess introduced the ws > > artifacts). Patch is already based on latest master fc989ce544c5. Let me > > know if you still have `git am/apply` issues. > > There is a slight indentation issue, but I will fix it locally. > Next time onwards please run gst-indent-1.0 on all *.c files before git > format patch. Thanks. Good to know and will comply going forward.
(In reply to sreerenj from comment #8) > (In reply to U. Artie Eoff from comment #7) > > Created attachment 370538 [details] [review] [review] [review] > > patch v3 > > > > This version is a direct upload of the file from `git format-patch`... my > > previous patch was a copy/paste text upload (which I guess introduced the ws > > artifacts). Patch is already based on latest master fc989ce544c5. Let me > > know if you still have `git am/apply` issues. > > There is a slight indentation issue, but I will fix it locally. > Next time onwards please run gst-indent-1.0 on all *.c files before git > format patch. What provides gst-indent-1.0?
(In reply to U. Artie Eoff from comment #10) > (In reply to sreerenj from comment #8) > > (In reply to U. Artie Eoff from comment #7) > > > Created attachment 370538 [details] [review] [review] [review] [review] > > > patch v3 > > > > > > This version is a direct upload of the file from `git format-patch`... my > > > previous patch was a copy/paste text upload (which I guess introduced the ws > > > artifacts). Patch is already based on latest master fc989ce544c5. Let me > > > know if you still have `git am/apply` issues. > > > > There is a slight indentation issue, but I will fix it locally. > > Next time onwards please run gst-indent-1.0 on all *.c files before git > > format patch. > > What provides gst-indent-1.0? It comes up with gstreamer core: https://cgit.freedesktop.org/gstreamer/gstreamer/tree/tools
Review of attachment 370538 [details] [review]: Pushed as commit 275d754156d6113568bc7f74e184a3b41393e49a in master Also cherry-picked to 1.14.