On 14/11/17 01:29 PM, Tom St Denis wrote: > Hi Harry, > > At display/dc/dcn10/dcn10_hw_sequencer.c:2140 > >     if (num_planes > 0) { >        struct dc_stream_state *stream_for_cursor; > >        program_all_pipe_in_tree(dc, top_pipe_to_program, context); > >        for (i = 0; i < dc->res_pool->pipe_count; i++) { >            struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; > >            if (stream == pipe_ctx->stream) { >                stream_for_cursor = pipe_ctx->stream; >                break; >            } >        } > >        /* TODO: this is a hack w/a for switching from mpo to pipe split */ >        if (stream_for_cursor->cursor_attributes.address.quad_part != 0) { >            struct dc_cursor_position position = { 0 }; > >            dc_stream_set_cursor_position(stream_for_cursor, &position); >            dc_stream_set_cursor_attributes(stream_for_cursor, >                &stream_for_cursor->cursor_attributes); >        } >     } > > The variable "stream_for_cursor" could be used uninitialized in the TODO > block. The pointer should be probably initialized to NULL at the top. BTW, the compiler points this out: drivers/gpu/drm//amd/amdgpu/../display/dc/dcn10/dcn10_hw_sequencer.c: In function â??dcn10_apply_ctx_for_surfaceâ??: drivers/gpu/drm//amd/amdgpu/../display/dc/dcn10/dcn10_hw_sequencer.c:2146:5: warning: â??stream_for_cursorâ?? may be used uninitialized in this function [-Wmaybe-uninitialized] &stream_for_cursor->cursor_attributes); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Mesa and X developer