Hi, I am using OpenGL to render camera image into a wayland surface, and I use IVI functions to set the position and size of the surface. So, I do as described in IVI quick guide: LayerManagerControl create layer 1 1024 600 LayerManagerControl set screen 0 render order 1 LayerManagerControl set layer 1 render order 5242880 LayerManagerControl set layer 1 visibility 1 LayerManagerControl set surface 5242880 source region 0 0 300 200 LayerManagerControl set surface 5242880 destination region 20 20 600 400 LayerManagerControl set surface 5242880 visibility 1 And the surface is shown fine. If I do exactly the same from my C++ program, it works only if I put sleeps between ILM functions. Something like this: ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); ilm_commitChanges(); std::this_thread::sleep( 100ms ); and if the sleep is not long enough, it will not work. The "LayerManagerControl get scene" will return output as if all functions were called properly, but what is shown is not what "LayerManagerControl get scene" returns. For example, somethimes the source region what is rendered is different from what should be. Even with sleeps, sometimes it happens that the scene get wrong numbers. Also, I see next lines in weston.log: [07:29:02.248] ivi-shell: source rectangle is not yet set by ivi_layout_surface_set_source_rectangle [07:29:02.885] ivi-shell: source rectangle is not yet set by ivi_layout_surface_set_source_rectangle [07:29:03.525] ivi-shell: source rectangle is not yet set by ivi_layout_surface_set_source_rectangle [07:29:03.178] ivi-shell: source rectangle is not yet set by ivi_layout_surface_set_source_rectangle Versions that I use are: weston: 6.0 IVI: revision 9acdaa1 Wayland: 1.17 The question is, is there a way to somehow bump the ivi shell? I thought that " ilm_commitChanges();" is supposed to do that, but somehow it doesn't. Also, what I noticed is that order of functions somehow matters. For example, if I set the visibility at end, then the image is not shown. By the way, I set some things from one program (add a surface to a layer) and other things in another program (set source region). Mit freundlichen Grüßen i. A. Vladimir Jovic ifm ecomatic gmbh

Hi Vladimir, I would guess that the issue is coming because the calls between different process are not synchronized. You see the error because visibility is set before surface got source region assigned to it, so compositor doesn't know how to render it. The implementation of the ivi-controller doesn't forbid you to use the ivi-wm protocol (ilm api's are wrapper around it) from different processes, there are a few use-cases but in general the should be only one, it should be application controller which will know where and how to render the surface. The minimalistic example of it is here: https://github.com/GENIVI/wayland-ivi-extension/tree/master/ivi-layermanagement-examples/layer-add-surfaces you will see that there are callback available to get to know when the rendering client provides a buffer to a ivi-surface and after this source/destination regions and afterwards visibility can be set. Best regards Eugen Friedrich Engineering Software Multimedia 1 (ADITG/ESM1) Tel. +49 5121 49 6921 > -----Original Message----- > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > Sent: Dienstag, 21. Mai 2019 16:50 > To: genivi-ivi-layer-management@lists.genivi.org > Subject: ivi functions not executed immediately? > > Hi, > > I am using OpenGL to render camera image into a wayland surface, and I use > IVI functions to set the position and size of the surface. > So, I do as described in IVI quick guide: > LayerManagerControl create layer 1 1024 600 > LayerManagerControl set screen 0 render order 1 > LayerManagerControl set layer 1 render order 5242880 > LayerManagerControl set layer 1 visibility 1 > LayerManagerControl set surface 5242880 source region 0 0 300 200 > LayerManagerControl set surface 5242880 destination region 20 20 600 400 > LayerManagerControl set surface 5242880 visibility 1 > > And the surface is shown fine. > If I do exactly the same from my C++ program, it works only if I put sleeps > between ILM functions. Something like this: > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > ilm_commitChanges(); > std::this_thread::sleep( 100ms ); > and if the sleep is not long enough, it will not work. > The "LayerManagerControl get scene" will return output as if all functions > were called properly, but what is shown is not what "LayerManagerControl > get scene" returns. For example, somethimes the source region what is > rendered is different from what should be. > Even with sleeps, sometimes it happens that the scene get wrong numbers. > > Also, I see next lines in weston.log: > [07:29:02.248] ivi-shell: source rectangle is not yet set by > ivi_layout_surface_set_source_rectangle > [07:29:02.885] ivi-shell: source rectangle is not yet set by > ivi_layout_surface_set_source_rectangle > [07:29:03.525] ivi-shell: source rectangle is not yet set by > ivi_layout_surface_set_source_rectangle > [07:29:03.178] ivi-shell: source rectangle is not yet set by > ivi_layout_surface_set_source_rectangle > > Versions that I use are: > weston: 6.0 > IVI: revision 9acdaa1 > Wayland: 1.17 > > The question is, is there a way to somehow bump the ivi shell? I thought that > " ilm_commitChanges();" is supposed to do that, but somehow it doesn't. > > Also, what I noticed is that order of functions somehow matters. For > example, if I set the visibility at end, then the image is not shown. > > By the way, I set some things from one program (add a surface to a layer) > and other things in another program (set source region). > > Mit freundlichen Grüßen > > i. A. Vladimir Jovic > > ifm ecomatic gmbh > > > > _______________________________________________ > genivi-ivi-layer-management mailing list > genivi-ivi-layer-management@lists.genivi.org > http://lists.genivi.org/mailman/listinfo/genivi-ivi-layer- > management_lists.genivi.org

Hi, Thanks for the info, and example. I tried to do all IVI function calls in one program, so called "windows manager", in following order: 1. when it starts, create a layer, and add all existing surfaces to it 2. from another program, create a surface, and when it becomes visible, start opengl rendering 3. set visibility in window manager 4. set source region in window manager 5. set destination region in window manager When I do "LayerManagerControl get scene", I get exactly numbers I used: root@pdm3:/home# LayerManagerControl get scene screen 0 (0x0) --------------------------------------- - connector name: Unknown-1 - resolution: x=1280, y=480 - layer render order: 1(0x1), layer 1 (0x1) --------------------------------------- - destination region: x=0, y=0, w=1280, h=480 - source region: x=0, y=0, w=1280, h=480 - opacity: 1 - visibility: 1 - surface render order: 5242880(0x500000), - on screen: 0(0x0) surface 5242880 (0x500000) --------------------------------------- - created by pid: 2757 - original size: x=720, y=576 - destination region: x=20, y=20, w=300, h=200 - source region: x=0, y=0, w=719, h=575 - opacity: 1 - visibility: 1 - frame counter: 7731 - on layer: 1(0x1) however, the image is shown as if the source and destination are 100x100. That is the initial surface size, but later I resize it using If I modify destination or source region, then the image is shown normally: LayerManagerControl set surface 5242880 destination region 20 20 300 201 Sometimes, it works normally, and I do not have to do anything. The only difference from your example, is that I do "ilm_commitChanges()" one after another command. Can I do that when setting destination, before setting source region and visibility? ************************ I tried to use ivi-wm, but it doesn't seem to work properly. But you said that ILM is just a wrapper. I generated the source and header with wayland-scanner public-code < wm-ivi.xml > wm-ivi.c wayland-scanner client-header < wm-ivi.xml > wm-ivi.h I compiled it into a shared library, and linked to my program. Then in the callback for "wl_registry_add_listener", when it tries to use "ivi_wm_add_listener()" it just seg fault somewhere. Mit freundlichen Grüßen i. A. Vladimir Jovic ifm ecomatic gmbh > -----Ursprüngliche Nachricht----- > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > Gesendet: Mittwoch, 22. Mai 2019 09:05 > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > management@lists.genivi.org > Betreff: RE: ivi functions not executed immediately? > > Hi Vladimir, > > I would guess that the issue is coming because the calls between different process > are not synchronized. > You see the error because visibility is set before surface got source region > assigned to it, so compositor doesn't know how to render it. > > The implementation of the ivi-controller doesn't forbid you to use the ivi-wm > protocol (ilm api's are wrapper around it) from different processes, there are a few > use-cases but in general the should be only one, it should be application controller > which will know where and how to render the surface. > > The minimalistic example of it is here: > https://urldefense.proofpoint.com/v2/url?u=https- > 3A__github.com_GENIVI_wayland-2Divi-2Dextension_tree_master_ivi- > 2Dlayermanagement-2Dexamples_layer-2Dadd- > 2Dsurfaces&d=DwIFAw&c=riR7jviByh3sGm7GIiSlHkFN0_aSATB6A8x0nHa2EM0&r > =8wfdsMUxnRudpdLb_Cg4rnnnKI- > gQZ4ykr3tjUrejXg&m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s > =nccLBZJ7E0vToNaw1lyLmvQUFywU-_p7FMp03yIxPLM&e= > > you will see that there are callback available to get to know when the rendering > client provides a buffer to a ivi-surface and after this source/destination regions > and afterwards visibility can be set. > > Best regards > > Eugen Friedrich > Engineering Software Multimedia 1 (ADITG/ESM1) > > Tel. +49 5121 49 6921 > > > -----Original Message----- > > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > > Sent: Dienstag, 21. Mai 2019 16:50 > > To: genivi-ivi-layer-management@lists.genivi.org > > Subject: ivi functions not executed immediately? > > > > Hi, > > > > I am using OpenGL to render camera image into a wayland surface, and I > > use IVI functions to set the position and size of the surface. > > So, I do as described in IVI quick guide: > > LayerManagerControl create layer 1 1024 600 > > LayerManagerControl set screen 0 render order 1 > > LayerManagerControl set layer 1 render order 5242880 > > LayerManagerControl set layer 1 visibility 1 > > LayerManagerControl set surface 5242880 source region 0 0 300 200 > > LayerManagerControl set surface 5242880 destination region 20 20 600 400 > > LayerManagerControl set surface 5242880 visibility 1 > > > > And the surface is shown fine. > > If I do exactly the same from my C++ program, it works only if I put > > sleeps between ILM functions. Something like this: > > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > > ilm_commitChanges(); > > std::this_thread::sleep( 100ms ); > > and if the sleep is not long enough, it will not work. > > The "LayerManagerControl get scene" will return output as if all > > functions were called properly, but what is shown is not what > > "LayerManagerControl get scene" returns. For example, somethimes the > > source region what is rendered is different from what should be. > > Even with sleeps, sometimes it happens that the scene get wrong numbers. > > > > Also, I see next lines in weston.log: > > [07:29:02.248] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > [07:29:02.885] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > [07:29:03.525] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > [07:29:03.178] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > > > Versions that I use are: > > weston: 6.0 > > IVI: revision 9acdaa1 > > Wayland: 1.17 > > > > The question is, is there a way to somehow bump the ivi shell? I > > thought that " ilm_commitChanges();" is supposed to do that, but somehow it > doesn't. > > > > Also, what I noticed is that order of functions somehow matters. For > > example, if I set the visibility at end, then the image is not shown. > > > > By the way, I set some things from one program (add a surface to a > > layer) and other things in another program (set source region). > > > > Mit freundlichen Grüßen > > > > i. A. Vladimir Jovic > > > > ifm ecomatic gmbh > > > > > > > > _______________________________________________ > > genivi-ivi-layer-management mailing list > > genivi-ivi-layer-management@lists.genivi.org > > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.genivi.org_m > > ailman_listinfo_genivi-2Divi-2Dlayer-2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > gQZ4ykr3tjUrejXg > > > &m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s=WyP3ZHmMEVz > wrB_isOPX7 > > FQxE8kMpAytfQhoigO7PY8&e= > > management_lists.genivi.org

Hi, In my "windows manager", I do this: void SurfaceCreatedNotificationFunc( ilmObjectType object, t_ilm_uint id, t_ilm_bool created, void* user_data ) { if ( ( ILM_SURFACE == object ) && ( ILM_TRUE == created ) ) { if ( ILM_SUCCESS != ilm_layerAddSurface( 1, id ) ) { std::cout << __PRETTY_FUNCTION__ << " ilm_layerAddSurface() failed after trying to add a surface to layer\n"; } ilm_commitChanges(); } } void AutoAddSurfaces() { ilm_registerNotification( SurfaceCreatedNotificationFunc, 0 ); } and that is going to add every created surface to the layer with id 1. Now, I created a script with this content: EGLWLMockNavigation & #sleep 0.005 LayerManagerControl set surface 10 source region 0 0 600 400 LayerManagerControl set surface 10 destination region 50 40 300 250 LayerManagerControl set surface 10 visibility 1 Without the sleep, sometimes the shown surface will be at random position with random size, and the " LayerManagerControl get scene" returns this: root@pdm3:/home# LayerManagerControl get scene screen 0 (0x0) --------------------------------------- - connector name: Unknown-1 - resolution: x=1280, y=480 - layer render order: 1(0x1), layer 1 (0x1) --------------------------------------- - destination region: x=0, y=0, w=1280, h=480 - source region: x=0, y=0, w=1280, h=480 - opacity: 1 - visibility: 1 - surface render order: 10(0xa), - on screen: 0(0x0) surface 10 (0xa) --------------------------------------- - created by pid: 5820 - original size: x=800, y=480 - destination region: x=50, y=40, w=300, h=250 - source region: x=0, y=0, w=0, h=0 - opacity: 1 - visibility: 1 - frame counter: 169 - on layer: 1(0x1) When my "windows manager" is not running, and I do " LayerManagerControl add surface 10 to layer 1" from the script, instead of sleep, then it works every time. So, it looks like there is some time after creating a surface, that it needs to somehow "settle down" and for the Weston to catch up. BTW how do you find some consulting services, when it is so new? Mit freundlichen Grüßen i. A. Vladimir Jovic ifm ecomatic gmbh > -----Ursprüngliche Nachricht----- > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > Gesendet: Mittwoch, 22. Mai 2019 09:05 > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > management@lists.genivi.org > Betreff: RE: ivi functions not executed immediately? > > Hi Vladimir, > > I would guess that the issue is coming because the calls between different process > are not synchronized. > You see the error because visibility is set before surface got source region > assigned to it, so compositor doesn't know how to render it. > > The implementation of the ivi-controller doesn't forbid you to use the ivi-wm > protocol (ilm api's are wrapper around it) from different processes, there are a few > use-cases but in general the should be only one, it should be application controller > which will know where and how to render the surface. > > The minimalistic example of it is here: > https://urldefense.proofpoint.com/v2/url?u=https- > 3A__github.com_GENIVI_wayland-2Divi-2Dextension_tree_master_ivi- > 2Dlayermanagement-2Dexamples_layer-2Dadd- > 2Dsurfaces&d=DwIFAw&c=riR7jviByh3sGm7GIiSlHkFN0_aSATB6A8x0nHa2EM0&r > =8wfdsMUxnRudpdLb_Cg4rnnnKI- > gQZ4ykr3tjUrejXg&m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s > =nccLBZJ7E0vToNaw1lyLmvQUFywU-_p7FMp03yIxPLM&e= > > you will see that there are callback available to get to know when the rendering > client provides a buffer to a ivi-surface and after this source/destination regions > and afterwards visibility can be set. > > Best regards > > Eugen Friedrich > Engineering Software Multimedia 1 (ADITG/ESM1) > > Tel. +49 5121 49 6921 > > > -----Original Message----- > > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > > Sent: Dienstag, 21. Mai 2019 16:50 > > To: genivi-ivi-layer-management@lists.genivi.org > > Subject: ivi functions not executed immediately? > > > > Hi, > > > > I am using OpenGL to render camera image into a wayland surface, and I > > use IVI functions to set the position and size of the surface. > > So, I do as described in IVI quick guide: > > LayerManagerControl create layer 1 1024 600 > > LayerManagerControl set screen 0 render order 1 > > LayerManagerControl set layer 1 render order 5242880 > > LayerManagerControl set layer 1 visibility 1 > > LayerManagerControl set surface 5242880 source region 0 0 300 200 > > LayerManagerControl set surface 5242880 destination region 20 20 600 400 > > LayerManagerControl set surface 5242880 visibility 1 > > > > And the surface is shown fine. > > If I do exactly the same from my C++ program, it works only if I put > > sleeps between ILM functions. Something like this: > > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > > ilm_commitChanges(); > > std::this_thread::sleep( 100ms ); > > and if the sleep is not long enough, it will not work. > > The "LayerManagerControl get scene" will return output as if all > > functions were called properly, but what is shown is not what > > "LayerManagerControl get scene" returns. For example, somethimes the > > source region what is rendered is different from what should be. > > Even with sleeps, sometimes it happens that the scene get wrong numbers. > > > > Also, I see next lines in weston.log: > > [07:29:02.248] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > [07:29:02.885] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > [07:29:03.525] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > [07:29:03.178] ivi-shell: source rectangle is not yet set by > > ivi_layout_surface_set_source_rectangle > > > > Versions that I use are: > > weston: 6.0 > > IVI: revision 9acdaa1 > > Wayland: 1.17 > > > > The question is, is there a way to somehow bump the ivi shell? I > > thought that " ilm_commitChanges();" is supposed to do that, but somehow it > doesn't. > > > > Also, what I noticed is that order of functions somehow matters. For > > example, if I set the visibility at end, then the image is not shown. > > > > By the way, I set some things from one program (add a surface to a > > layer) and other things in another program (set source region). > > > > Mit freundlichen Grüßen > > > > i. A. Vladimir Jovic > > > > ifm ecomatic gmbh > > > > > > > > _______________________________________________ > > genivi-ivi-layer-management mailing list > > genivi-ivi-layer-management@lists.genivi.org > > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.genivi.org_m > > ailman_listinfo_genivi-2Divi-2Dlayer-2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > gQZ4ykr3tjUrejXg > > > &m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s=WyP3ZHmMEVz > wrB_isOPX7 > > FQxE8kMpAytfQhoigO7PY8&e= > > management_lists.genivi.org

 LayerManagerControl set surface 5242880 visibility 1

Hi Vladimir, If you would follow the example application you will see that there is additional callback installed to find out the original dimensions of the surface. After creation is doesn't have any, only after application will really attach a buffer and send this buffer to the compositor (wl_surface_attach , wl-surface_damage, wl_surface_comit, wl_display_flush sequence, should be called from you EGL driver initiated by eglSwapBuffer call). This would be the synchronization point which you miss sometimes: setting the source region for the surface which doesn't have any size. I'm surprised that the position and size are random. Looks like a issue in the ivi-shell or ivi-conroller: actually you should never set source or destination without having the original size not sure if we documented this somewhere properly, will take a look... Best regards Eugen Friedrich Engineering Software Multimedia 1 (ADITG/ESM1) Tel. +49 5121 49 6921 > -----Original Message----- > From: Jovic, Vladimir <vladimir.jovic@ifm.com> > Sent: Donnerstag, 23. Mai 2019 15:55 > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; genivi-ivi- > layer-management@lists.genivi.org > Subject: AW: ivi functions not executed immediately? > > Hi, > > In my "windows manager", I do this: > void SurfaceCreatedNotificationFunc( ilmObjectType object, > t_ilm_uint id, > t_ilm_bool created, > void* user_data ) > { > if ( ( ILM_SURFACE == object ) && ( ILM_TRUE == created ) ) > { > if ( ILM_SUCCESS != ilm_layerAddSurface( 1, id ) ) > { > std::cout << __PRETTY_FUNCTION__ << " ilm_layerAddSurface() > failed after trying to add a surface to layer\n"; > } > ilm_commitChanges(); > } > } > void AutoAddSurfaces() > { > ilm_registerNotification( SurfaceCreatedNotificationFunc, 0 ); > } > > and that is going to add every created surface to the layer with id 1. > > Now, I created a script with this content: > > EGLWLMockNavigation & > #sleep 0.005 > LayerManagerControl set surface 10 source region 0 0 600 400 > LayerManagerControl set surface 10 destination region 50 40 300 250 > LayerManagerControl set surface 10 visibility 1 > > Without the sleep, sometimes the shown surface will be at random position > with random size, and the " LayerManagerControl get scene" returns this: > root@pdm3:/home# LayerManagerControl get scene > screen 0 (0x0) > --------------------------------------- > - connector name: Unknown-1 > - resolution: x=1280, y=480 > - layer render order: 1(0x1), > > layer 1 (0x1) > --------------------------------------- > - destination region: x=0, y=0, w=1280, h=480 > - source region: x=0, y=0, w=1280, h=480 > - opacity: 1 > - visibility: 1 > - surface render order: 10(0xa), > - on screen: 0(0x0) > > surface 10 (0xa) > --------------------------------------- > - created by pid: 5820 > - original size: x=800, y=480 > - destination region: x=50, y=40, w=300, h=250 > - source region: x=0, y=0, w=0, h=0 > - opacity: 1 > - visibility: 1 > - frame counter: 169 > - on layer: 1(0x1) > > > When my "windows manager" is not running, and I do " > LayerManagerControl add surface 10 to layer 1" from the script, instead of > sleep, then it works every time. > So, it looks like there is some time after creating a surface, that it needs to > somehow "settle down" and for the Weston to catch up. > > BTW how do you find some consulting services, when it is so new? Just ask this mailing list :-) > > Mit freundlichen Grüßen > > i. A. Vladimir Jovic > ifm ecomatic gmbh > > > > > -----Ursprüngliche Nachricht----- > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > Gesendet: Mittwoch, 22. Mai 2019 09:05 > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > management@lists.genivi.org > > Betreff: RE: ivi functions not executed immediately? > > > > Hi Vladimir, > > > > I would guess that the issue is coming because the calls between different > process > > are not synchronized. > > You see the error because visibility is set before surface got source region > > assigned to it, so compositor doesn't know how to render it. > > > > The implementation of the ivi-controller doesn't forbid you to use the ivi- > wm > > protocol (ilm api's are wrapper around it) from different processes, there > are a few > > use-cases but in general the should be only one, it should be application > controller > > which will know where and how to render the surface. > > > > The minimalistic example of it is here: > > https://urldefense.proofpoint.com/v2/url?u=https- > > 3A__github.com_GENIVI_wayland-2Divi-2Dextension_tree_master_ivi- > > 2Dlayermanagement-2Dexamples_layer-2Dadd- > > > 2Dsurfaces&d=DwIFAw&c=riR7jviByh3sGm7GIiSlHkFN0_aSATB6A8x0nHa2E > M0&r > > =8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > gQZ4ykr3tjUrejXg&m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs > &s > > =nccLBZJ7E0vToNaw1lyLmvQUFywU-_p7FMp03yIxPLM&e= > > > > you will see that there are callback available to get to know when the > rendering > > client provides a buffer to a ivi-surface and after this source/destination > regions > > and afterwards visibility can be set. > > > > Best regards > > > > Eugen Friedrich > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > Tel. +49 5121 49 6921 > > > > > -----Original Message----- > > > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > > > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > > > Sent: Dienstag, 21. Mai 2019 16:50 > > > To: genivi-ivi-layer-management@lists.genivi.org > > > Subject: ivi functions not executed immediately? > > > > > > Hi, > > > > > > I am using OpenGL to render camera image into a wayland surface, and I > > > use IVI functions to set the position and size of the surface. > > > So, I do as described in IVI quick guide: > > > LayerManagerControl create layer 1 1024 600 > > > LayerManagerControl set screen 0 render order 1 > > > LayerManagerControl set layer 1 render order 5242880 > > > LayerManagerControl set layer 1 visibility 1 > > > LayerManagerControl set surface 5242880 source region 0 0 300 200 > > > LayerManagerControl set surface 5242880 destination region 20 20 600 > 400 > > > LayerManagerControl set surface 5242880 visibility 1 > > > > > > And the surface is shown fine. > > > If I do exactly the same from my C++ program, it works only if I put > > > sleeps between ILM functions. Something like this: > > > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > > > ilm_commitChanges(); > > > std::this_thread::sleep( 100ms ); > > > and if the sleep is not long enough, it will not work. > > > The "LayerManagerControl get scene" will return output as if all > > > functions were called properly, but what is shown is not what > > > "LayerManagerControl get scene" returns. For example, somethimes the > > > source region what is rendered is different from what should be. > > > Even with sleeps, sometimes it happens that the scene get wrong > numbers. > > > > > > Also, I see next lines in weston.log: > > > [07:29:02.248] ivi-shell: source rectangle is not yet set by > > > ivi_layout_surface_set_source_rectangle > > > [07:29:02.885] ivi-shell: source rectangle is not yet set by > > > ivi_layout_surface_set_source_rectangle > > > [07:29:03.525] ivi-shell: source rectangle is not yet set by > > > ivi_layout_surface_set_source_rectangle > > > [07:29:03.178] ivi-shell: source rectangle is not yet set by > > > ivi_layout_surface_set_source_rectangle > > > > > > Versions that I use are: > > > weston: 6.0 > > > IVI: revision 9acdaa1 > > > Wayland: 1.17 > > > > > > The question is, is there a way to somehow bump the ivi shell? I > > > thought that " ilm_commitChanges();" is supposed to do that, but > somehow it > > doesn't. > > > > > > Also, what I noticed is that order of functions somehow matters. For > > > example, if I set the visibility at end, then the image is not shown. > > > > > > By the way, I set some things from one program (add a surface to a > > > layer) and other things in another program (set source region). > > > > > > Mit freundlichen Grüßen > > > > > > i. A. Vladimir Jovic > > > > > > ifm ecomatic gmbh > > > > > > > > > > > > _______________________________________________ > > > genivi-ivi-layer-management mailing list > > > genivi-ivi-layer-management@lists.genivi.org > > > https://urldefense.proofpoint.com/v2/url?u=http- > 3A__lists.genivi.org_m > > > ailman_listinfo_genivi-2Divi-2Dlayer- > 2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > > gQZ4ykr3tjUrejXg > > > > > > &m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s=WyP3ZHmME > Vz > > wrB_isOPX7 > > > FQxE8kMpAytfQhoigO7PY8&e= > > > management_lists.genivi.org >

 LayerManagerControl set surface 5242880 visibility 1

Hi Actually, I was wondering what that "origin size" really is, since I couldn't find any info on that. I do register to that callback, to get the visibility flag. Because if the visibility is 0, then opengl functions are blocking the process. I do set the window size with "wl_egl_window_resize()", and I see these numbers with "LayerManagerControl get scene". But I am not sure if I have to call "wl_display_dispatch" or " wl_display_flush" or "sl_surface_commit" or anything else after calling "wl_egl_window_resize()". Because, the "original size" is 0x0 - I need to set visibility first, then source and then destination - and only then the origin is set. Do I need to start rendering to set the origin? Or is calling the "eglSwapBuffer " enough? Do you have a document explaining all this (link or something)? Is there a book describing how exactly to program in ivi and wayland? I couldn't find anything describing IVI in details. Mit freundlichen Grüßen i. A. Vladimir Jovic ifm ecomatic gmbh > -----Ursprüngliche Nachricht----- > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > Gesendet: Donnerstag, 23. Mai 2019 18:48 > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > management@lists.genivi.org > Betreff: RE: ivi functions not executed immediately? > > Hi Vladimir, > > If you would follow the example application you will see that there is additional > callback installed to find out the original dimensions of the surface. > After creation is doesn't have any, only after application will really attach a buffer > and send this buffer to the compositor (wl_surface_attach , wl-surface_damage, > wl_surface_comit, wl_display_flush sequence, should be called from you EGL > driver initiated by eglSwapBuffer call). > > This would be the synchronization point which you miss sometimes: setting the > source region for the surface which doesn't have any size. > I'm surprised that the position and size are random. Looks like a issue in the ivi- > shell or ivi-conroller: actually you should never set source or destination without > having the original size not sure if we documented this somewhere properly, will > take a look... > > Best regards > > Eugen Friedrich > Engineering Software Multimedia 1 (ADITG/ESM1) > > Tel. +49 5121 49 6921 > > -----Original Message----- > > From: Jovic, Vladimir <vladimir.jovic@ifm.com> > > Sent: Donnerstag, 23. Mai 2019 15:55 > > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; > > genivi-ivi- layer-management@lists.genivi.org > > Subject: AW: ivi functions not executed immediately? > > > > Hi, > > > > In my "windows manager", I do this: > > void SurfaceCreatedNotificationFunc( ilmObjectType object, > > t_ilm_uint id, > > t_ilm_bool created, > > void* user_data ) { > > if ( ( ILM_SURFACE == object ) && ( ILM_TRUE == created ) ) > > { > > if ( ILM_SUCCESS != ilm_layerAddSurface( 1, id ) ) > > { > > std::cout << __PRETTY_FUNCTION__ << " ilm_layerAddSurface() > > failed after trying to add a surface to layer\n"; > > } > > ilm_commitChanges(); > > } > > } > > void AutoAddSurfaces() > > { > > ilm_registerNotification( SurfaceCreatedNotificationFunc, 0 ); } > > > > and that is going to add every created surface to the layer with id 1. > > > > Now, I created a script with this content: > > > > EGLWLMockNavigation & > > #sleep 0.005 > > LayerManagerControl set surface 10 source region 0 0 600 400 > > LayerManagerControl set surface 10 destination region 50 40 300 250 > > LayerManagerControl set surface 10 visibility 1 > > > > Without the sleep, sometimes the shown surface will be at random > > position with random size, and the " LayerManagerControl get scene" returns > this: > > root@pdm3:/home# LayerManagerControl get scene screen 0 (0x0) > > --------------------------------------- > > - connector name: Unknown-1 > > - resolution: x=1280, y=480 > > - layer render order: 1(0x1), > > > > layer 1 (0x1) > > --------------------------------------- > > - destination region: x=0, y=0, w=1280, h=480 > > - source region: x=0, y=0, w=1280, h=480 > > - opacity: 1 > > - visibility: 1 > > - surface render order: 10(0xa), > > - on screen: 0(0x0) > > > > surface 10 (0xa) > > --------------------------------------- > > - created by pid: 5820 > > - original size: x=800, y=480 > > - destination region: x=50, y=40, w=300, h=250 > > - source region: x=0, y=0, w=0, h=0 > > - opacity: 1 > > - visibility: 1 > > - frame counter: 169 > > - on layer: 1(0x1) > > > > > > When my "windows manager" is not running, and I do " > > LayerManagerControl add surface 10 to layer 1" from the script, > > instead of sleep, then it works every time. > > So, it looks like there is some time after creating a surface, that it > > needs to somehow "settle down" and for the Weston to catch up. > > > > BTW how do you find some consulting services, when it is so new? > Just ask this mailing list :-) > > > > > Mit freundlichen Grüßen > > > > i. A. Vladimir Jovic > > ifm ecomatic gmbh > > > > > > > > > -----Ursprüngliche Nachricht----- > > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > > Gesendet: Mittwoch, 22. Mai 2019 09:05 > > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > > management@lists.genivi.org > > > Betreff: RE: ivi functions not executed immediately? > > > > > > Hi Vladimir, > > > > > > I would guess that the issue is coming because the calls between > > > different > > process > > > are not synchronized. > > > You see the error because visibility is set before surface got > > > source region assigned to it, so compositor doesn't know how to render it. > > > > > > The implementation of the ivi-controller doesn't forbid you to use > > > the ivi- > > wm > > > protocol (ilm api's are wrapper around it) from different processes, > > > there > > are a few > > > use-cases but in general the should be only one, it should be > > > application > > controller > > > which will know where and how to render the surface. > > > > > > The minimalistic example of it is here: > > > https://urldefense.proofpoint.com/v2/url?u=https- > > > 3A__github.com_GENIVI_wayland-2Divi-2Dextension_tree_master_ivi- > > > 2Dlayermanagement-2Dexamples_layer-2Dadd- > > > > > 2Dsurfaces&d=DwIFAw&c=riR7jviByh3sGm7GIiSlHkFN0_aSATB6A8x0nHa2E > > M0&r > > > =8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > > > gQZ4ykr3tjUrejXg&m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs > > &s > > > =nccLBZJ7E0vToNaw1lyLmvQUFywU-_p7FMp03yIxPLM&e= > > > > > > you will see that there are callback available to get to know when > > > the > > rendering > > > client provides a buffer to a ivi-surface and after this > > > source/destination > > regions > > > and afterwards visibility can be set. > > > > > > Best regards > > > > > > Eugen Friedrich > > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > > > Tel. +49 5121 49 6921 > > > > > > > -----Original Message----- > > > > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > > > > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > > > > Sent: Dienstag, 21. Mai 2019 16:50 > > > > To: genivi-ivi-layer-management@lists.genivi.org > > > > Subject: ivi functions not executed immediately? > > > > > > > > Hi, > > > > > > > > I am using OpenGL to render camera image into a wayland surface, > > > > and I use IVI functions to set the position and size of the surface. > > > > So, I do as described in IVI quick guide: > > > > LayerManagerControl create layer 1 1024 600 > > > > LayerManagerControl set screen 0 render order 1 > > > > LayerManagerControl set layer 1 render order 5242880 > > > > LayerManagerControl set layer 1 visibility 1 > > > > LayerManagerControl set surface 5242880 source region 0 0 300 200 > > > > LayerManagerControl set surface 5242880 destination region 20 > > > > 20 600 > > 400 > > > > LayerManagerControl set surface 5242880 visibility 1 > > > > > > > > And the surface is shown fine. > > > > If I do exactly the same from my C++ program, it works only if I > > > > put sleeps between ILM functions. Something like this: > > > > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > > > > ilm_commitChanges(); > > > > std::this_thread::sleep( 100ms ); and if the sleep is not long > > > > enough, it will not work. > > > > The "LayerManagerControl get scene" will return output as if all > > > > functions were called properly, but what is shown is not what > > > > "LayerManagerControl get scene" returns. For example, somethimes > > > > the source region what is rendered is different from what should be. > > > > Even with sleeps, sometimes it happens that the scene get wrong > > numbers. > > > > > > > > Also, I see next lines in weston.log: > > > > [07:29:02.248] ivi-shell: source rectangle is not yet set by > > > > ivi_layout_surface_set_source_rectangle > > > > [07:29:02.885] ivi-shell: source rectangle is not yet set by > > > > ivi_layout_surface_set_source_rectangle > > > > [07:29:03.525] ivi-shell: source rectangle is not yet set by > > > > ivi_layout_surface_set_source_rectangle > > > > [07:29:03.178] ivi-shell: source rectangle is not yet set by > > > > ivi_layout_surface_set_source_rectangle > > > > > > > > Versions that I use are: > > > > weston: 6.0 > > > > IVI: revision 9acdaa1 > > > > Wayland: 1.17 > > > > > > > > The question is, is there a way to somehow bump the ivi shell? I > > > > thought that " ilm_commitChanges();" is supposed to do that, but > > somehow it > > > doesn't. > > > > > > > > Also, what I noticed is that order of functions somehow matters. > > > > For example, if I set the visibility at end, then the image is not shown. > > > > > > > > By the way, I set some things from one program (add a surface to a > > > > layer) and other things in another program (set source region). > > > > > > > > Mit freundlichen Grüßen > > > > > > > > i. A. Vladimir Jovic > > > > > > > > ifm ecomatic gmbh > > > > > > > > > > > > > > > > _______________________________________________ > > > > genivi-ivi-layer-management mailing list > > > > genivi-ivi-layer-management@lists.genivi.org > > > > https://urldefense.proofpoint.com/v2/url?u=http- > > 3A__lists.genivi.org_m > > > > ailman_listinfo_genivi-2Divi-2Dlayer- > > 2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > gQZ4ykr3tjUrejXg > > > > > > > > > &m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s=WyP3ZHmME > > Vz > > > wrB_isOPX7 > > > > FQxE8kMpAytfQhoigO7PY8&e= > > > > management_lists.genivi.org > > >

 LayerManagerControl set surface 5242880 visibility 1

Hi, Origin size is the size of the buffer of the surface. Therefore, you have to start to render to set the origin size. Best regards Emre Ucan Engineering Software Multimedia (ADITG/ESM) Tel. +49 5121 49 6937 > -----Original Message----- > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > Sent: Freitag, 24. Mai 2019 09:06 > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; genivi-ivi- > layer-management@lists.genivi.org > Subject: AW: ivi functions not executed immediately? > > Hi > > Actually, I was wondering what that "origin size" really is, since I couldn't find > any info on that. I do register to that callback, to get the visibility flag. > Because if the visibility is 0, then opengl functions are blocking the process. > > I do set the window size with "wl_egl_window_resize()", and I see these > numbers with "LayerManagerControl get scene". But I am not sure if I have > to call "wl_display_dispatch" or " wl_display_flush" or "sl_surface_commit" > or anything else after calling "wl_egl_window_resize()". Because, the > "original size" is 0x0 - I need to set visibility first, then source and then > destination - and only then the origin is set. Do I need to start rendering to > set the origin? Or is calling the "eglSwapBuffer " enough? > > Do you have a document explaining all this (link or something)? Is there a > book describing how exactly to program in ivi and wayland? I couldn't find > anything describing IVI in details. > > > > Mit freundlichen Grüßen > > i. A. Vladimir Jovic > ifm ecomatic gmbh > > > > > -----Ursprüngliche Nachricht----- > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > Gesendet: Donnerstag, 23. Mai 2019 18:48 > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > management@lists.genivi.org > > Betreff: RE: ivi functions not executed immediately? > > > > Hi Vladimir, > > > > If you would follow the example application you will see that there is > additional > > callback installed to find out the original dimensions of the surface. > > After creation is doesn't have any, only after application will really attach a > buffer > > and send this buffer to the compositor (wl_surface_attach , wl- > surface_damage, > > wl_surface_comit, wl_display_flush sequence, should be called from you > EGL > > driver initiated by eglSwapBuffer call). > > > > This would be the synchronization point which you miss sometimes: setting > the > > source region for the surface which doesn't have any size. > > I'm surprised that the position and size are random. Looks like a issue in the > ivi- > > shell or ivi-conroller: actually you should never set source or destination > without > > having the original size not sure if we documented this somewhere > properly, will > > take a look... > > > > Best regards > > > > Eugen Friedrich > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > Tel. +49 5121 49 6921 > > > -----Original Message----- > > > From: Jovic, Vladimir <vladimir.jovic@ifm.com> > > > Sent: Donnerstag, 23. Mai 2019 15:55 > > > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; > > > genivi-ivi- layer-management@lists.genivi.org > > > Subject: AW: ivi functions not executed immediately? > > > > > > Hi, > > > > > > In my "windows manager", I do this: > > > void SurfaceCreatedNotificationFunc( ilmObjectType object, > > > t_ilm_uint id, > > > t_ilm_bool created, > > > void* user_data ) { > > > if ( ( ILM_SURFACE == object ) && ( ILM_TRUE == created ) ) > > > { > > > if ( ILM_SUCCESS != ilm_layerAddSurface( 1, id ) ) > > > { > > > std::cout << __PRETTY_FUNCTION__ << " ilm_layerAddSurface() > > > failed after trying to add a surface to layer\n"; > > > } > > > ilm_commitChanges(); > > > } > > > } > > > void AutoAddSurfaces() > > > { > > > ilm_registerNotification( SurfaceCreatedNotificationFunc, 0 ); } > > > > > > and that is going to add every created surface to the layer with id 1. > > > > > > Now, I created a script with this content: > > > > > > EGLWLMockNavigation & > > > #sleep 0.005 > > > LayerManagerControl set surface 10 source region 0 0 600 400 > > > LayerManagerControl set surface 10 destination region 50 40 300 250 > > > LayerManagerControl set surface 10 visibility 1 > > > > > > Without the sleep, sometimes the shown surface will be at random > > > position with random size, and the " LayerManagerControl get scene" > returns > > this: > > > root@pdm3:/home# LayerManagerControl get scene screen 0 (0x0) > > > --------------------------------------- > > > - connector name: Unknown-1 > > > - resolution: x=1280, y=480 > > > - layer render order: 1(0x1), > > > > > > layer 1 (0x1) > > > --------------------------------------- > > > - destination region: x=0, y=0, w=1280, h=480 > > > - source region: x=0, y=0, w=1280, h=480 > > > - opacity: 1 > > > - visibility: 1 > > > - surface render order: 10(0xa), > > > - on screen: 0(0x0) > > > > > > surface 10 (0xa) > > > --------------------------------------- > > > - created by pid: 5820 > > > - original size: x=800, y=480 > > > - destination region: x=50, y=40, w=300, h=250 > > > - source region: x=0, y=0, w=0, h=0 > > > - opacity: 1 > > > - visibility: 1 > > > - frame counter: 169 > > > - on layer: 1(0x1) > > > > > > > > > When my "windows manager" is not running, and I do " > > > LayerManagerControl add surface 10 to layer 1" from the script, > > > instead of sleep, then it works every time. > > > So, it looks like there is some time after creating a surface, that it > > > needs to somehow "settle down" and for the Weston to catch up. > > > > > > BTW how do you find some consulting services, when it is so new? > > Just ask this mailing list :-) > > > > > > > > Mit freundlichen Grüßen > > > > > > i. A. Vladimir Jovic > > > ifm ecomatic gmbh > > > > > > > > > > > > > -----Ursprüngliche Nachricht----- > > > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > > > Gesendet: Mittwoch, 22. Mai 2019 09:05 > > > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > > > management@lists.genivi.org > > > > Betreff: RE: ivi functions not executed immediately? > > > > > > > > Hi Vladimir, > > > > > > > > I would guess that the issue is coming because the calls between > > > > different > > > process > > > > are not synchronized. > > > > You see the error because visibility is set before surface got > > > > source region assigned to it, so compositor doesn't know how to render > it. > > > > > > > > The implementation of the ivi-controller doesn't forbid you to use > > > > the ivi- > > > wm > > > > protocol (ilm api's are wrapper around it) from different processes, > > > > there > > > are a few > > > > use-cases but in general the should be only one, it should be > > > > application > > > controller > > > > which will know where and how to render the surface. > > > > > > > > The minimalistic example of it is here: > > > > https://urldefense.proofpoint.com/v2/url?u=https- > > > > 3A__github.com_GENIVI_wayland-2Divi- > 2Dextension_tree_master_ivi- > > > > 2Dlayermanagement-2Dexamples_layer-2Dadd- > > > > > > > > 2Dsurfaces&d=DwIFAw&c=riR7jviByh3sGm7GIiSlHkFN0_aSATB6A8x0nHa2E > > > M0&r > > > > =8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > > > > > > gQZ4ykr3tjUrejXg&m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs > > > &s > > > > =nccLBZJ7E0vToNaw1lyLmvQUFywU-_p7FMp03yIxPLM&e= > > > > > > > > you will see that there are callback available to get to know when > > > > the > > > rendering > > > > client provides a buffer to a ivi-surface and after this > > > > source/destination > > > regions > > > > and afterwards visibility can be set. > > > > > > > > Best regards > > > > > > > > Eugen Friedrich > > > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > > > > > Tel. +49 5121 49 6921 > > > > > > > > > -----Original Message----- > > > > > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > > > > > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > > > > > Sent: Dienstag, 21. Mai 2019 16:50 > > > > > To: genivi-ivi-layer-management@lists.genivi.org > > > > > Subject: ivi functions not executed immediately? > > > > > > > > > > Hi, > > > > > > > > > > I am using OpenGL to render camera image into a wayland surface, > > > > > and I use IVI functions to set the position and size of the surface. > > > > > So, I do as described in IVI quick guide: > > > > > LayerManagerControl create layer 1 1024 600 > > > > > LayerManagerControl set screen 0 render order 1 > > > > > LayerManagerControl set layer 1 render order 5242880 > > > > > LayerManagerControl set layer 1 visibility 1 > > > > > LayerManagerControl set surface 5242880 source region 0 0 300 200 > > > > > LayerManagerControl set surface 5242880 destination region 20 > > > > > 20 600 > > > 400 > > > > > LayerManagerControl set surface 5242880 visibility 1 > > > > > > > > > > And the surface is shown fine. > > > > > If I do exactly the same from my C++ program, it works only if I > > > > > put sleeps between ILM functions. Something like this: > > > > > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > > > > > ilm_commitChanges(); > > > > > std::this_thread::sleep( 100ms ); and if the sleep is not long > > > > > enough, it will not work. > > > > > The "LayerManagerControl get scene" will return output as if all > > > > > functions were called properly, but what is shown is not what > > > > > "LayerManagerControl get scene" returns. For example, somethimes > > > > > the source region what is rendered is different from what should be. > > > > > Even with sleeps, sometimes it happens that the scene get wrong > > > numbers. > > > > > > > > > > Also, I see next lines in weston.log: > > > > > [07:29:02.248] ivi-shell: source rectangle is not yet set by > > > > > ivi_layout_surface_set_source_rectangle > > > > > [07:29:02.885] ivi-shell: source rectangle is not yet set by > > > > > ivi_layout_surface_set_source_rectangle > > > > > [07:29:03.525] ivi-shell: source rectangle is not yet set by > > > > > ivi_layout_surface_set_source_rectangle > > > > > [07:29:03.178] ivi-shell: source rectangle is not yet set by > > > > > ivi_layout_surface_set_source_rectangle > > > > > > > > > > Versions that I use are: > > > > > weston: 6.0 > > > > > IVI: revision 9acdaa1 > > > > > Wayland: 1.17 > > > > > > > > > > The question is, is there a way to somehow bump the ivi shell? I > > > > > thought that " ilm_commitChanges();" is supposed to do that, but > > > somehow it > > > > doesn't. > > > > > > > > > > Also, what I noticed is that order of functions somehow matters. > > > > > For example, if I set the visibility at end, then the image is not shown. > > > > > > > > > > By the way, I set some things from one program (add a surface to a > > > > > layer) and other things in another program (set source region). > > > > > > > > > > Mit freundlichen Grüßen > > > > > > > > > > i. A. Vladimir Jovic > > > > > > > > > > ifm ecomatic gmbh > > > > > > > > > > > > > > > > > > > > _______________________________________________ > > > > > genivi-ivi-layer-management mailing list > > > > > genivi-ivi-layer-management@lists.genivi.org > > > > > https://urldefense.proofpoint.com/v2/url?u=http- > > > 3A__lists.genivi.org_m > > > > > ailman_listinfo_genivi-2Divi-2Dlayer- > > > 2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > > > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > > gQZ4ykr3tjUrejXg > > > > > > > > > > > > > &m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s=WyP3ZHmME > > > Vz > > > > wrB_isOPX7 > > > > > FQxE8kMpAytfQhoigO7PY8&e= > > > > > management_lists.genivi.org > > > > > > > > > _______________________________________________ > genivi-ivi-layer-management mailing list > genivi-ivi-layer-management@lists.genivi.org > http://lists.genivi.org/mailman/listinfo/genivi-ivi-layer- > management_lists.genivi.org

 LayerManagerControl set surface 5242880 visibility 1

Hi, So, in order to set the "origin size", I need to start rendering. But to start rendering, I need to set the visibility, otherwise opengl blocks. To set the visibility, I need to set both source and destination regions. Source and destination can't be set before origin is set. That means that something bellow, what I was told, to do is wrong. Mit freundlichen Grüßen i. A. Vladimir Jovic ifm ecomatic gmbh > -----Ursprüngliche Nachricht----- > Von: Ucan, Emre (ADITG/ESM1) <eucan@de.adit-jv.com> > Gesendet: Freitag, 24. Mai 2019 10:14 > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; Friedrich, Eugen (ADITG/ESM1) > <efriedrich@de.adit-jv.com>; genivi-ivi-layer-management@lists.genivi.org > Betreff: RE: ivi functions not executed immediately? > > Hi, > > Origin size is the size of the buffer of the surface. Therefore, you have to start to > render to set the origin size. > > Best regards > > Emre Ucan > Engineering Software Multimedia (ADITG/ESM) > > Tel. +49 5121 49 6937 > > -----Original Message----- > > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > > Sent: Freitag, 24. Mai 2019 09:06 > > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; > > genivi-ivi- layer-management@lists.genivi.org > > Subject: AW: ivi functions not executed immediately? > > > > Hi > > > > Actually, I was wondering what that "origin size" really is, since I > > couldn't find any info on that. I do register to that callback, to get the visibility > flag. > > Because if the visibility is 0, then opengl functions are blocking the process. > > > > I do set the window size with "wl_egl_window_resize()", and I see > > these numbers with "LayerManagerControl get scene". But I am not sure > > if I have to call "wl_display_dispatch" or " wl_display_flush" or > "sl_surface_commit" > > or anything else after calling "wl_egl_window_resize()". Because, the > > "original size" is 0x0 - I need to set visibility first, then source > > and then destination - and only then the origin is set. Do I need to > > start rendering to set the origin? Or is calling the "eglSwapBuffer " enough? > > > > Do you have a document explaining all this (link or something)? Is > > there a book describing how exactly to program in ivi and wayland? I > > couldn't find anything describing IVI in details. > > > > > > > > Mit freundlichen Grüßen > > > > i. A. Vladimir Jovic > > ifm ecomatic gmbh > > > > > > > > > -----Ursprüngliche Nachricht----- > > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > > Gesendet: Donnerstag, 23. Mai 2019 18:48 > > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > > management@lists.genivi.org > > > Betreff: RE: ivi functions not executed immediately? > > > > > > Hi Vladimir, > > > > > > If you would follow the example application you will see that there > > > is > > additional > > > callback installed to find out the original dimensions of the surface. > > > After creation is doesn't have any, only after application will > > > really attach a > > buffer > > > and send this buffer to the compositor (wl_surface_attach , wl- > > surface_damage, > > > wl_surface_comit, wl_display_flush sequence, should be called from > > > you > > EGL > > > driver initiated by eglSwapBuffer call). > > > > > > This would be the synchronization point which you miss sometimes: > > > setting > > the > > > source region for the surface which doesn't have any size. > > > I'm surprised that the position and size are random. Looks like a > > > issue in the > > ivi- > > > shell or ivi-conroller: actually you should never set source or > > > destination > > without > > > having the original size not sure if we documented this somewhere > > properly, will > > > take a look... > > > > > > Best regards > > > > > > Eugen Friedrich > > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > > > Tel. +49 5121 49 6921 > > > > -----Original Message----- > > > > From: Jovic, Vladimir <vladimir.jovic@ifm.com> > > > > Sent: Donnerstag, 23. Mai 2019 15:55 > > > > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; > > > > genivi-ivi- layer-management@lists.genivi.org > > > > Subject: AW: ivi functions not executed immediately? > > > > > > > > Hi, > > > > > > > > In my "windows manager", I do this: > > > > void SurfaceCreatedNotificationFunc( ilmObjectType object, > > > > t_ilm_uint id, > > > > t_ilm_bool created, > > > > void* user_data ) { > > > > if ( ( ILM_SURFACE == object ) && ( ILM_TRUE == created ) ) > > > > { > > > > if ( ILM_SUCCESS != ilm_layerAddSurface( 1, id ) ) > > > > { > > > > std::cout << __PRETTY_FUNCTION__ << " ilm_layerAddSurface() > > > > failed after trying to add a surface to layer\n"; > > > > } > > > > ilm_commitChanges(); > > > > } > > > > } > > > > void AutoAddSurfaces() > > > > { > > > > ilm_registerNotification( SurfaceCreatedNotificationFunc, 0 ); > > > > } > > > > > > > > and that is going to add every created surface to the layer with id 1. > > > > > > > > Now, I created a script with this content: > > > > > > > > EGLWLMockNavigation & > > > > #sleep 0.005 > > > > LayerManagerControl set surface 10 source region 0 0 600 400 > > > > LayerManagerControl set surface 10 destination region 50 40 300 > > > > 250 LayerManagerControl set surface 10 visibility 1 > > > > > > > > Without the sleep, sometimes the shown surface will be at random > > > > position with random size, and the " LayerManagerControl get scene" > > returns > > > this: > > > > root@pdm3:/home# LayerManagerControl get scene screen 0 (0x0) > > > > --------------------------------------- > > > > - connector name: Unknown-1 > > > > - resolution: x=1280, y=480 > > > > - layer render order: 1(0x1), > > > > > > > > layer 1 (0x1) > > > > --------------------------------------- > > > > - destination region: x=0, y=0, w=1280, h=480 > > > > - source region: x=0, y=0, w=1280, h=480 > > > > - opacity: 1 > > > > - visibility: 1 > > > > - surface render order: 10(0xa), > > > > - on screen: 0(0x0) > > > > > > > > surface 10 (0xa) > > > > --------------------------------------- > > > > - created by pid: 5820 > > > > - original size: x=800, y=480 > > > > - destination region: x=50, y=40, w=300, h=250 > > > > - source region: x=0, y=0, w=0, h=0 > > > > - opacity: 1 > > > > - visibility: 1 > > > > - frame counter: 169 > > > > - on layer: 1(0x1) > > > > > > > > > > > > When my "windows manager" is not running, and I do " > > > > LayerManagerControl add surface 10 to layer 1" from the script, > > > > instead of sleep, then it works every time. > > > > So, it looks like there is some time after creating a surface, > > > > that it needs to somehow "settle down" and for the Weston to catch up. > > > > > > > > BTW how do you find some consulting services, when it is so new? > > > Just ask this mailing list :-) > > > > > > > > > > > Mit freundlichen Grüßen > > > > > > > > i. A. Vladimir Jovic > > > > ifm ecomatic gmbh > > > > > > > > > > > > > > > > > -----Ursprüngliche Nachricht----- > > > > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > > > > Gesendet: Mittwoch, 22. Mai 2019 09:05 > > > > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > > > > management@lists.genivi.org > > > > > Betreff: RE: ivi functions not executed immediately? > > > > > > > > > > Hi Vladimir, > > > > > > > > > > I would guess that the issue is coming because the calls between > > > > > different > > > > process > > > > > are not synchronized. > > > > > You see the error because visibility is set before surface got > > > > > source region assigned to it, so compositor doesn't know how to > > > > > render > > it. > > > > > > > > > > The implementation of the ivi-controller doesn't forbid you to > > > > > use the ivi- > > > > wm > > > > > protocol (ilm api's are wrapper around it) from different > > > > > processes, there > > > > are a few > > > > > use-cases but in general the should be only one, it should be > > > > > application > > > > controller > > > > > which will know where and how to render the surface. > > > > > > > > > > The minimalistic example of it is here: > > > > > https://urldefense.proofpoint.com/v2/url?u=https- > > > > > 3A__github.com_GENIVI_wayland-2Divi- > > 2Dextension_tree_master_ivi- > > > > > 2Dlayermanagement-2Dexamples_layer-2Dadd- > > > > > > > > > > > 2Dsurfaces&d=DwIFAw&c=riR7jviByh3sGm7GIiSlHkFN0_aSATB6A8x0nHa2E > > > > M0&r > > > > > =8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > > > > > > > > > gQZ4ykr3tjUrejXg&m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs > > > > &s > > > > > =nccLBZJ7E0vToNaw1lyLmvQUFywU-_p7FMp03yIxPLM&e= > > > > > > > > > > you will see that there are callback available to get to know > > > > > when the > > > > rendering > > > > > client provides a buffer to a ivi-surface and after this > > > > > source/destination > > > > regions > > > > > and afterwards visibility can be set. > > > > > > > > > > Best regards > > > > > > > > > > Eugen Friedrich > > > > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > > > > > > > Tel. +49 5121 49 6921 > > > > > > > > > > > -----Original Message----- > > > > > > From: genivi-ivi-layer-management > > > > > > <genivi-ivi-layer-management- bounces@lists.genivi.org> On > > > > > > Behalf Of Jovic, Vladimir > > > > > > Sent: Dienstag, 21. Mai 2019 16:50 > > > > > > To: genivi-ivi-layer-management@lists.genivi.org > > > > > > Subject: ivi functions not executed immediately? > > > > > > > > > > > > Hi, > > > > > > > > > > > > I am using OpenGL to render camera image into a wayland > > > > > > surface, and I use IVI functions to set the position and size of the > surface. > > > > > > So, I do as described in IVI quick guide: > > > > > > LayerManagerControl create layer 1 1024 600 > > > > > > LayerManagerControl set screen 0 render order 1 > > > > > > LayerManagerControl set layer 1 render order 5242880 > > > > > > LayerManagerControl set layer 1 visibility 1 > > > > > > LayerManagerControl set surface 5242880 source region 0 0 300 200 > > > > > > LayerManagerControl set surface 5242880 destination region > > > > > > 20 > > > > > > 20 600 > > > > 400 > > > > > > LayerManagerControl set surface 5242880 visibility 1 > > > > > > > > > > > > And the surface is shown fine. > > > > > > If I do exactly the same from my C++ program, it works only if > > > > > > I put sleeps between ILM functions. Something like this: > > > > > > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > > > > > > ilm_commitChanges(); > > > > > > std::this_thread::sleep( 100ms ); and if the sleep is not > > > > > > long enough, it will not work. > > > > > > The "LayerManagerControl get scene" will return output as if > > > > > > all functions were called properly, but what is shown is not > > > > > > what "LayerManagerControl get scene" returns. For example, > > > > > > somethimes the source region what is rendered is different from what > should be. > > > > > > Even with sleeps, sometimes it happens that the scene get > > > > > > wrong > > > > numbers. > > > > > > > > > > > > Also, I see next lines in weston.log: > > > > > > [07:29:02.248] ivi-shell: source rectangle is not yet set by > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > [07:29:02.885] ivi-shell: source rectangle is not yet set by > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > [07:29:03.525] ivi-shell: source rectangle is not yet set by > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > [07:29:03.178] ivi-shell: source rectangle is not yet set by > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > > > > > > > Versions that I use are: > > > > > > weston: 6.0 > > > > > > IVI: revision 9acdaa1 > > > > > > Wayland: 1.17 > > > > > > > > > > > > The question is, is there a way to somehow bump the ivi shell? > > > > > > I thought that " ilm_commitChanges();" is supposed to do that, > > > > > > but > > > > somehow it > > > > > doesn't. > > > > > > > > > > > > Also, what I noticed is that order of functions somehow matters. > > > > > > For example, if I set the visibility at end, then the image is not shown. > > > > > > > > > > > > By the way, I set some things from one program (add a surface > > > > > > to a > > > > > > layer) and other things in another program (set source region). > > > > > > > > > > > > Mit freundlichen Grüßen > > > > > > > > > > > > i. A. Vladimir Jovic > > > > > > > > > > > > ifm ecomatic gmbh > > > > > > > > > > > > > > > > > > > > > > > > _______________________________________________ > > > > > > genivi-ivi-layer-management mailing list > > > > > > genivi-ivi-layer-management@lists.genivi.org > > > > > > https://urldefense.proofpoint.com/v2/url?u=http- > > > > 3A__lists.genivi.org_m > > > > > > ailman_listinfo_genivi-2Divi-2Dlayer- > > > > 2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > > > > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > > > gQZ4ykr3tjUrejXg > > > > > > > > > > > > > > > > > &m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s=WyP3ZHmME > > > > Vz > > > > > wrB_isOPX7 > > > > > > FQxE8kMpAytfQhoigO7PY8&e= > > > > > > management_lists.genivi.org > > > > > > > > > > > > > > > _______________________________________________ > > genivi-ivi-layer-management mailing list > > genivi-ivi-layer-management@lists.genivi.org > > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.genivi.org_m > > ailman_listinfo_genivi-2Divi-2Dlayer-2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > gQZ4ykr3tjUrejXg > > > &m=3uigexTynM1KYbjVAwQ0MZeGfnSZUU32vWqG_ytgYVQ&s=UV4fR6P3ndYPu > Wx__t3vS > > s1_u-jAFWe4gkRHSLE6MZ4&e= > > management_lists.genivi.org

 LayerManagerControl set surface 5242880 destination region

 LayerManagerControl set surface 5242880 visibility 1

Hi Vladimir, The EGL implementation will block only after sending a first buffer it blocks on the second eglSwapBuffer call(you can maybe verify this with your driver), also you can run it in "unblocking mode", for testing, just set the eglSwapInterval to 0. EGL is blocking waiting for frame sync (default behavior which is controlled with the swap interval, default is 1 so EGL will wait till the buffer is used for the composition in the compositor). After the first frame is rendered Weston will get the original size and now you can setup the source /destination regions and set corresponding surface visible. Best regards Eugen Friedrich Engineering Software Multimedia 1 (ADITG/ESM1) Tel. +49 5121 49 6921 > -----Original Message----- > From: Jovic, Vladimir <vladimir.jovic@ifm.com> > Sent: Freitag, 24. Mai 2019 10:34 > To: Ucan, Emre (ADITG/ESM1) <eucan@de.adit-jv.com>; Friedrich, Eugen > (ADITG/ESM1) <efriedrich@de.adit-jv.com>; genivi-ivi-layer- > management@lists.genivi.org > Subject: AW: ivi functions not executed immediately? > > Hi, > > So, in order to set the "origin size", I need to start rendering. But to start > rendering, I need to set the visibility, otherwise opengl blocks. To set the > visibility, I need to set both source and destination regions. Source and > destination can't be set before origin is set. > > That means that something bellow, what I was told, to do is wrong. > > > Mit freundlichen Grüßen > > i. A. Vladimir Jovic > ifm ecomatic gmbh > > > > > > -----Ursprüngliche Nachricht----- > > Von: Ucan, Emre (ADITG/ESM1) <eucan@de.adit-jv.com> > > Gesendet: Freitag, 24. Mai 2019 10:14 > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; Friedrich, Eugen > (ADITG/ESM1) > > <efriedrich@de.adit-jv.com>; genivi-ivi-layer- > management@lists.genivi.org > > Betreff: RE: ivi functions not executed immediately? > > > > Hi, > > > > Origin size is the size of the buffer of the surface. Therefore, you have to > start to > > render to set the origin size. > > > > Best regards > > > > Emre Ucan > > Engineering Software Multimedia (ADITG/ESM) > > > > Tel. +49 5121 49 6937 > > > -----Original Message----- > > > From: genivi-ivi-layer-management <genivi-ivi-layer-management- > > > bounces@lists.genivi.org> On Behalf Of Jovic, Vladimir > > > Sent: Freitag, 24. Mai 2019 09:06 > > > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; > > > genivi-ivi- layer-management@lists.genivi.org > > > Subject: AW: ivi functions not executed immediately? > > > > > > Hi > > > > > > Actually, I was wondering what that "origin size" really is, since I > > > couldn't find any info on that. I do register to that callback, to get the > visibility > > flag. > > > Because if the visibility is 0, then opengl functions are blocking the > process. > > > > > > I do set the window size with "wl_egl_window_resize()", and I see > > > these numbers with "LayerManagerControl get scene". But I am not sure > > > if I have to call "wl_display_dispatch" or " wl_display_flush" or > > "sl_surface_commit" > > > or anything else after calling "wl_egl_window_resize()". Because, the > > > "original size" is 0x0 - I need to set visibility first, then source > > > and then destination - and only then the origin is set. Do I need to > > > start rendering to set the origin? Or is calling the "eglSwapBuffer " > enough? > > > > > > Do you have a document explaining all this (link or something)? Is > > > there a book describing how exactly to program in ivi and wayland? I > > > couldn't find anything describing IVI in details. > > > > > > > > > > > > Mit freundlichen Grüßen > > > > > > i. A. Vladimir Jovic > > > ifm ecomatic gmbh > > > > > > > > > > > > > -----Ursprüngliche Nachricht----- > > > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > > > Gesendet: Donnerstag, 23. Mai 2019 18:48 > > > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > > > management@lists.genivi.org > > > > Betreff: RE: ivi functions not executed immediately? > > > > > > > > Hi Vladimir, > > > > > > > > If you would follow the example application you will see that there > > > > is > > > additional > > > > callback installed to find out the original dimensions of the surface. > > > > After creation is doesn't have any, only after application will > > > > really attach a > > > buffer > > > > and send this buffer to the compositor (wl_surface_attach , wl- > > > surface_damage, > > > > wl_surface_comit, wl_display_flush sequence, should be called from > > > > you > > > EGL > > > > driver initiated by eglSwapBuffer call). > > > > > > > > This would be the synchronization point which you miss sometimes: > > > > setting > > > the > > > > source region for the surface which doesn't have any size. > > > > I'm surprised that the position and size are random. Looks like a > > > > issue in the > > > ivi- > > > > shell or ivi-conroller: actually you should never set source or > > > > destination > > > without > > > > having the original size not sure if we documented this somewhere > > > properly, will > > > > take a look... > > > > > > > > Best regards > > > > > > > > Eugen Friedrich > > > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > > > > > Tel. +49 5121 49 6921 > > > > > -----Original Message----- > > > > > From: Jovic, Vladimir <vladimir.jovic@ifm.com> > > > > > Sent: Donnerstag, 23. Mai 2019 15:55 > > > > > To: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com>; > > > > > genivi-ivi- layer-management@lists.genivi.org > > > > > Subject: AW: ivi functions not executed immediately? > > > > > > > > > > Hi, > > > > > > > > > > In my "windows manager", I do this: > > > > > void SurfaceCreatedNotificationFunc( ilmObjectType object, > > > > > t_ilm_uint id, > > > > > t_ilm_bool created, > > > > > void* user_data ) { > > > > > if ( ( ILM_SURFACE == object ) && ( ILM_TRUE == created ) ) > > > > > { > > > > > if ( ILM_SUCCESS != ilm_layerAddSurface( 1, id ) ) > > > > > { > > > > > std::cout << __PRETTY_FUNCTION__ << " > ilm_layerAddSurface() > > > > > failed after trying to add a surface to layer\n"; > > > > > } > > > > > ilm_commitChanges(); > > > > > } > > > > > } > > > > > void AutoAddSurfaces() > > > > > { > > > > > ilm_registerNotification( SurfaceCreatedNotificationFunc, 0 ); > > > > > } > > > > > > > > > > and that is going to add every created surface to the layer with id 1. > > > > > > > > > > Now, I created a script with this content: > > > > > > > > > > EGLWLMockNavigation & > > > > > #sleep 0.005 > > > > > LayerManagerControl set surface 10 source region 0 0 600 400 > > > > > LayerManagerControl set surface 10 destination region 50 40 300 > > > > > 250 LayerManagerControl set surface 10 visibility 1 > > > > > > > > > > Without the sleep, sometimes the shown surface will be at random > > > > > position with random size, and the " LayerManagerControl get scene" > > > returns > > > > this: > > > > > root@pdm3:/home# LayerManagerControl get scene screen 0 (0x0) > > > > > --------------------------------------- > > > > > - connector name: Unknown-1 > > > > > - resolution: x=1280, y=480 > > > > > - layer render order: 1(0x1), > > > > > > > > > > layer 1 (0x1) > > > > > --------------------------------------- > > > > > - destination region: x=0, y=0, w=1280, h=480 > > > > > - source region: x=0, y=0, w=1280, h=480 > > > > > - opacity: 1 > > > > > - visibility: 1 > > > > > - surface render order: 10(0xa), > > > > > - on screen: 0(0x0) > > > > > > > > > > surface 10 (0xa) > > > > > --------------------------------------- > > > > > - created by pid: 5820 > > > > > - original size: x=800, y=480 > > > > > - destination region: x=50, y=40, w=300, h=250 > > > > > - source region: x=0, y=0, w=0, h=0 > > > > > - opacity: 1 > > > > > - visibility: 1 > > > > > - frame counter: 169 > > > > > - on layer: 1(0x1) > > > > > > > > > > > > > > > When my "windows manager" is not running, and I do " > > > > > LayerManagerControl add surface 10 to layer 1" from the script, > > > > > instead of sleep, then it works every time. > > > > > So, it looks like there is some time after creating a surface, > > > > > that it needs to somehow "settle down" and for the Weston to catch > up. > > > > > > > > > > BTW how do you find some consulting services, when it is so new? > > > > Just ask this mailing list :-) > > > > > > > > > > > > > > Mit freundlichen Grüßen > > > > > > > > > > i. A. Vladimir Jovic > > > > > ifm ecomatic gmbh > > > > > > > > > > > > > > > > > > > > > -----Ursprüngliche Nachricht----- > > > > > > Von: Friedrich, Eugen (ADITG/ESM1) <efriedrich@de.adit-jv.com> > > > > > > Gesendet: Mittwoch, 22. Mai 2019 09:05 > > > > > > An: Jovic, Vladimir <vladimir.jovic@ifm.com>; genivi-ivi-layer- > > > > > > management@lists.genivi.org > > > > > > Betreff: RE: ivi functions not executed immediately? > > > > > > > > > > > > Hi Vladimir, > > > > > > > > > > > > I would guess that the issue is coming because the calls between > > > > > > different > > > > > process > > > > > > are not synchronized. > > > > > > You see the error because visibility is set before surface got > > > > > > source region assigned to it, so compositor doesn't know how to > > > > > > render > > > it. > > > > > > > > > > > > The implementation of the ivi-controller doesn't forbid you to > > > > > > use the ivi- > > > > > wm > > > > > > protocol (ilm api's are wrapper around it) from different > > > > > > processes, there > > > > > are a few > > > > > > use-cases but in general the should be only one, it should be > > > > > > application > > > > > controller > > > > > > which will know where and how to render the surface. > > > > > > > > > > > > The minimalistic example of it is here: > > > > > > https://urldefense.proofpoint.com/v2/url?u=https- > > > > > > 3A__github.com_GENIVI_wayland-2Divi- > > > 2Dextension_tree_master_ivi- > > > > > > 2Dlayermanagement-2Dexamples_layer-2Dadd- > > > > > > > > > > > > > > > 2Dsurfaces&d=DwIFAw&c=riR7jviByh3sGm7GIiSlHkFN0_aSATB6A8x0nHa2E > > > > > M0&r > > > > > > =8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > > > > > > > > > > > > > gQZ4ykr3tjUrejXg&m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs > > > > > &s > > > > > > =nccLBZJ7E0vToNaw1lyLmvQUFywU-_p7FMp03yIxPLM&e= > > > > > > > > > > > > you will see that there are callback available to get to know > > > > > > when the > > > > > rendering > > > > > > client provides a buffer to a ivi-surface and after this > > > > > > source/destination > > > > > regions > > > > > > and afterwards visibility can be set. > > > > > > > > > > > > Best regards > > > > > > > > > > > > Eugen Friedrich > > > > > > Engineering Software Multimedia 1 (ADITG/ESM1) > > > > > > > > > > > > Tel. +49 5121 49 6921 > > > > > > > > > > > > > -----Original Message----- > > > > > > > From: genivi-ivi-layer-management > > > > > > > <genivi-ivi-layer-management- bounces@lists.genivi.org> On > > > > > > > Behalf Of Jovic, Vladimir > > > > > > > Sent: Dienstag, 21. Mai 2019 16:50 > > > > > > > To: genivi-ivi-layer-management@lists.genivi.org > > > > > > > Subject: ivi functions not executed immediately? > > > > > > > > > > > > > > Hi, > > > > > > > > > > > > > > I am using OpenGL to render camera image into a wayland > > > > > > > surface, and I use IVI functions to set the position and size of the > > surface. > > > > > > > So, I do as described in IVI quick guide: > > > > > > > LayerManagerControl create layer 1 1024 600 > > > > > > > LayerManagerControl set screen 0 render order 1 > > > > > > > LayerManagerControl set layer 1 render order 5242880 > > > > > > > LayerManagerControl set layer 1 visibility 1 > > > > > > > LayerManagerControl set surface 5242880 source region 0 0 300 > 200 > > > > > > > LayerManagerControl set surface 5242880 destination region > > > > > > > 20 > > > > > > > 20 600 > > > > > 400 > > > > > > > LayerManagerControl set surface 5242880 visibility 1 > > > > > > > > > > > > > > And the surface is shown fine. > > > > > > > If I do exactly the same from my C++ program, it works only if > > > > > > > I put sleeps between ILM functions. Something like this: > > > > > > > ilm_surfaceSetSourceRectangle( surfaceId, 0, 0, 300, 200 ); > > > > > > > ilm_commitChanges(); > > > > > > > std::this_thread::sleep( 100ms ); and if the sleep is not > > > > > > > long enough, it will not work. > > > > > > > The "LayerManagerControl get scene" will return output as if > > > > > > > all functions were called properly, but what is shown is not > > > > > > > what "LayerManagerControl get scene" returns. For example, > > > > > > > somethimes the source region what is rendered is different from > what > > should be. > > > > > > > Even with sleeps, sometimes it happens that the scene get > > > > > > > wrong > > > > > numbers. > > > > > > > > > > > > > > Also, I see next lines in weston.log: > > > > > > > [07:29:02.248] ivi-shell: source rectangle is not yet set by > > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > > [07:29:02.885] ivi-shell: source rectangle is not yet set by > > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > > [07:29:03.525] ivi-shell: source rectangle is not yet set by > > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > > [07:29:03.178] ivi-shell: source rectangle is not yet set by > > > > > > > ivi_layout_surface_set_source_rectangle > > > > > > > > > > > > > > Versions that I use are: > > > > > > > weston: 6.0 > > > > > > > IVI: revision 9acdaa1 > > > > > > > Wayland: 1.17 > > > > > > > > > > > > > > The question is, is there a way to somehow bump the ivi shell? > > > > > > > I thought that " ilm_commitChanges();" is supposed to do that, > > > > > > > but > > > > > somehow it > > > > > > doesn't. > > > > > > > > > > > > > > Also, what I noticed is that order of functions somehow matters. > > > > > > > For example, if I set the visibility at end, then the image is not > shown. > > > > > > > > > > > > > > By the way, I set some things from one program (add a surface > > > > > > > to a > > > > > > > layer) and other things in another program (set source region). > > > > > > > > > > > > > > Mit freundlichen Grüßen > > > > > > > > > > > > > > i. A. Vladimir Jovic > > > > > > > > > > > > > > ifm ecomatic gmbh > > > > > > > > > > > > > > > > > > > > > > > > > > > > _______________________________________________ > > > > > > > genivi-ivi-layer-management mailing list > > > > > > > genivi-ivi-layer-management@lists.genivi.org > > > > > > > https://urldefense.proofpoint.com/v2/url?u=http- > > > > > 3A__lists.genivi.org_m > > > > > > > ailman_listinfo_genivi-2Divi-2Dlayer- > > > > > 2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > > > > > > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > > > > > > gQZ4ykr3tjUrejXg > > > > > > > > > > > > > > > > > > > > > > &m=YrRXzuqcsVwn7R2MF9iARhCLYWgZfYp1do7BrPdPCHs&s=WyP3ZHmME > > > > > Vz > > > > > > wrB_isOPX7 > > > > > > > FQxE8kMpAytfQhoigO7PY8&e= > > > > > > > management_lists.genivi.org > > > > > > > > > > > > > > > > > > > > > _______________________________________________ > > > genivi-ivi-layer-management mailing list > > > genivi-ivi-layer-management@lists.genivi.org > > > https://urldefense.proofpoint.com/v2/url?u=http- > 3A__lists.genivi.org_m > > > ailman_listinfo_genivi-2Divi-2Dlayer- > 2D&d=DwIFAw&c=riR7jviByh3sGm7GIiS > > > lHkFN0_aSATB6A8x0nHa2EM0&r=8wfdsMUxnRudpdLb_Cg4rnnnKI- > > gQZ4ykr3tjUrejXg > > > > > > &m=3uigexTynM1KYbjVAwQ0MZeGfnSZUU32vWqG_ytgYVQ&s=UV4fR6P3n > dYPu > > Wx__t3vS > > > s1_u-jAFWe4gkRHSLE6MZ4&e= > > > management_lists.genivi.org >