<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> <meta name="Generator" content="Microsoft Word 15 (filtered medium)"> <style></style> </head> <body lang="EN-US" link="#0563C1" vlink="#954F72"> <div class="WordSection1"> For vaCreateContext, I think render_targets/num_render_targets are optional (it depends on driver implementation), and the surface of vaBeginPicture can a brand-new surface<o:p></o:p> VAStatus vaCreateContext (<o:p></o:p> VADisplay dpy,<o:p></o:p> VAConfigID config_id,<o:p></o:p> int picture_width,<o:p></o:p> int picture_height,<o:p></o:p> int flag,<o:p></o:p> VASurfaceID *render_targets,<o:p></o:p> int num_render_targets,<o:p></o:p> VAContextID *context /* out */<o:p></o:p> );<o:p></o:p> <o:p> </o:p> DMA buffer can be imported through VA_SURFACE_ATTRIB_MEM_TYPE_DRM_PRIME, e.g.<o:p></o:p> VASurfaceAttribExternalBuffers buf;<o:p></o:p> VASurfaceAttribType type,<o:p></o:p> VASurfaceAttrib attrib_list[2]<o:p></o:p> unsigned long prime_fd; /* an input variable, it is the imported prime fd from other process */<o:p></o:p> <o:p> </o:p> int stride = frame_width_aligned, height_aligned = frame_height_aligned;<o:p></o:p> <o:p> </o:p> if (stride == 0)<o:p></o:p> stride = frame_width_mbaligned;<o:p></o:p> if (height_aligned == 0)<o:p></o:p> height_aligned = frame_height_mbaligned;<o:p></o:p> <o:p></o:p> buf.pixel_format = VA_FOURCC_NV12;<o:p></o:p> buf.width = frame_width_mbaligned;<o:p></o:p> buf.height = frame_height_mbaligned;<o:p></o:p> buf.data_size = stride * height_aligned * 1.5;<o:p></o:p> buf.num_buffers = 1; /* import only 1 buffer */<o:p></o:p> buf.num_planes = 3;<o:p></o:p> buf.pitches[0] = stride;<o:p></o:p> buf.pitches[1] = stride;<o:p></o:p> buf.pitches[2] = stride;<o:p></o:p> buf.pitches[3] = 0;<o:p></o:p> buf.offsets[0] = 0;<o:p></o:p> buf.offsets[1] = stride * height_aligned;<o:p></o:p> buf.offsets[2] = buf.offsets[1];<o:p></o:p> buf.offsets[3] = 0;<o:p></o:p> buf.buffers = &prime_fd;<o:p></o:p> buf.flags = 0;<o:p></o:p> buf.private_data = NULL;<o:p></o:p> <o:p> </o:p> attrib_list[0].type = (VASurfaceAttribType)VASurfaceAttribMemoryType;<o:p></o:p> attrib_list[0].flags = VA_SURFACE_ATTRIB_SETTABLE;<o:p></o:p> attrib_list[0].value.type = VAGenericValueTypeInteger;<o:p></o:p> attrib_list[0].value.value.i = VA_SURFACE_ATTRIB_MEM_TYPE_DRM_PRIME<o:p></o:p> <o:p> </o:p> attrib_list[1].type = (VASurfaceAttribType)VASurfaceAttribExternalBufferDescriptor;<o:p></o:p> attrib_list[1].flags = VA_SURFACE_ATTRIB_SETTABLE;<o:p></o:p> attrib_list[1].value.type = VAGenericValueTypePointer;<o:p></o:p> attrib_list[1].value.value.p = (void *)buf;<o:p></o:p> vaCreateSurfaces(<o:p></o:p> va_dpy,<o:p></o:p> VA_RT_FORMAT_YUV420,<o:p></o:p> frame_width_mbaligned, frame_height_mbaligned,<o:p></o:p> &surface_id, 1,<o:p></o:p> &attrib_list, 2);<o:p></o:p> <o:p> </o:p> <a name="_MailEndCompose"><o:p> </o:p></a> <a name="_____replyseparator"></a>From: Libva [mailto:libva-bounces@lists.freedesktop.org] On Behalf Of Anon Sent: Wednesday, May 25, 2016 2:23 PM To: libva@lists.freedesktop.org Subject: [Libva] encode surfaces created from prime fd (backed by OpenGL 2D texture)<o:p></o:p> <o:p> </o:p> <div> Hi All,<o:p></o:p> <div> <o:p> </o:p> </div> <div> I want to implement the following use case on Intel Haswell CPU with HD graphics:<o:p></o:p> </div> <div> <o:p> </o:p> </div> <div> offscreen OpenGL rendering to OpenGL 2D texture -> use EGL_MESA_image_dma_buf_export to export the texture as prime fd -> vaCreateSurfaces from the prime fd -> use vaapi to hw acclerate h264 encoding<o:p></o:p> </div> <div> <o:p> </o:p> </div> <div> Is this supported by the latest va-api and libva-intel-driver release (i.e., 1.7.0)? If so, is there any example I can start with?<o:p></o:p> </div> <div> <o:p> </o:p> </div> <div> I did read the sample application h264encode.c and thought it might be a good starting point. However, the first major issue I encountered is when creating context through vaCreateContext, a number of pre-allocated va surfaces are required and thus statically associated with the new va context. However, with prime fd, the fd will change so a new va surface is created with every new prime fd. I don't see any API can be used to dynamically add/remove va surfaces after the context is created. Can you please give me some suggestions?<o:p></o:p> </div> <div> <o:p> </o:p> </div> <div> Best,<o:p></o:p> </div> <div> Kristine<o:p></o:p> </div> </div> </div> </body> </html>