|
1 iMaker-ConE integration |
|
2 ======================= |
|
3 |
|
4 iMaker tool is able to create images from CPF file. This mechanism is used for example in Carbide.v |
|
5 where iMaker plugin (Eclipse plugin) is calling iMaker to create flash images for modified CPF file. |
|
6 The following picture shows how the calling hierarchy goes. |
|
7 |
|
8 .. image:: images/imaker_cone.jpg |
|
9 |
|
10 #. Carbide.v's iMaker plugin instantiates call to iMaker where CPF and configuration inside the CPF is given as parameter. |
|
11 Also other information like report template and report output filename is given. Imaker plugin is responsible of providing |
|
12 progress information to Carbide.v user so that he/she knows that generation is progressing. |
|
13 #. iMaker calls ConE to generate makefile that contains image creation configuration for this CPF. |
|
14 #. ConE returns makefile to iMaker |
|
15 #. iMaker calls itself using the makefile. Default target in the makefile is create_selected and in Carbide.v use cases it |
|
16 contains ROFS3 and UDA targets. |
|
17 #. iMaker calls ConE to generate content for ROFS3 (parameter --impl-tag=target:rofs3) and gives output directory as parameter. |
|
18 #. ConE filters implementations as defined by iMaker and calls each plugin separately to create output files to output folder. |
|
19 #. Data is generated to output folder. |
|
20 #. iMaker creates ROFS3 image from that data using normal iMaker variant build step. |
|
21 #. iMaker calls ConE to generate content for UDA (parameter --impl-tag=target:uda) and gives output directory as parameter. |
|
22 #. ConE filters implementations as defined by iMaker and calls each plugin separately to create output files to output folder. |
|
23 #. Data is generated to output folder. |
|
24 #. iMaker creates UDA image from that data using normal iMaker variant build step. |
|
25 #. iMaker performs data package copy step which basically uses predefined dcp, vpl and signature files and copies flash images |
|
26 based on variant configuration to one folder with certain filenames. This way there is no need to resign the package anymore. |
|
27 Downside of this approach is that data package definition and information cannot be changed but in case of operator variant |
|
28 verification it is not that critical. Of course same data package cannot be used in production. |
|
29 #. Data package and generation reports are located in places that Carbide.v specified in the first call. |
|
30 |