Leveraging LX2160a capabilities
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I've been using the Honeycomb LX2160acex7 board provided by solidrun for over a year now, slowly learning the board, it's capabilties, and what I've been missing.
Link to the product: (https://www.solid-run.com/arm-servers-networking-platforms/honeycomb-lx2/)
I also developed a C++ application that uses DPDK 22.11 that's been compiled straight from the DPDK website, using my aarch64 toolchain provided from buildroot.
I feel for a while that I have not been fully taken advantage of the SoC's capabilities, either by using its QDMA support, its QBman support, or any other capabilities.
Also, I only lately found out that NXP provide a repository, forked from DPDK, that supposed to be fully supported on DPAA2 used platforms, and I've been trying to understand how to compile it the right way and intergrate it on my platform.
I have a few questions I'd like to ask, so I could know how to fully take advantage of the platform:
1. How can I take advantage the SoC's capabilities when it comes to linux kernel?
2. Is there a provided modified linux kernel (or with unapplied patches) either by NXP or solidrun (the company I got the board from)?
3. Are there any kernel patches needed to be applied to support LX2160a?
4. Is the NXP provided DPDK repository better supported and more optimized for LX2160a when it comes to performance?
5. What's the difference between qorlq-lf and qorlq branches in NXP's dpdk repository (https://github.com/nxp-qoriq/dpdk)?
6. Is NXP's provided DPDK repository already uses DPAA2 functions (qdma for example) by default, or do I need to use it myself in my C++ application?
7. Is there a right and different way to compile NXP's provided DPDK repository (qorlq branch for example) so it would work on my platform and support DPAA2 architecture?
