Hi all,
I'm now using the EVM6678L via PCIe connection with PC under Desktop-linux-sdk, as far as running the demo application (sending and backing test file to and from DSP), I'd like to dig a little more from the SDK's perspective.
Now the demo application works by first loading a pre-build firmware into the DSP to reset the DSP and configure the communication path with PC, then at PC side the demo application sends test file batches to DSP, host uses the mailBox to inform the DSP to make a copy of that file and send back to the host. The support modules (mailBox, cmem, bufmgr, pciedrv) are compiled together with the demo application.
What I'd like to do is to separate the demo application and support modules, making the latter as a static library which can be called by multiple applications. My questions are:
1) is it possible to do such a thing? If not, why?
2) It seems the pre-build firmware for DSP cannot be built by SDK user unless some extra configuration, what should I do to re-build that firmware?
3) As I could see, user can specify how many cores to configure and use for some applications, to do this I think the DSP should be configured before running the application? Can the number of cores be dynamically scheduled during runtime?
Thanks,
Jie
