build: add low-latency GPIF buffer size build options
Add build options for low-latency GPIF buffer configuration to support
latency-sensitive applications. When enabled, both FX3 firmware and FPGA
report legacy versions (< 2.5.0 / 0.16.0) to trigger libbladeRF's
automatic buffer size detection.
FX3 Firmware (cmake -DLOW_LATENCY_GPIF_BUFFERS=ON):
- Set version to 2.4.0 (triggers legacy mode in libbladeRF)
- Configure GPIF DMA: size*2, count=22 buffers
- Generate VERSION string with correct 2.4.0 display
FPGA (build_bladerf.sh -L):
- Set version to 0.15.3 (includes critical metadata/timestamp fixes)
- Swap fx3_gpif_p.vhd with fx3_gpif_p_lowlatency.vhd during build
- Create fx3_gpif_p_lowlatency.vhd package: HS=256, SS=512 samples
Low-latency buffer sizes (pre-2.5.0/0.16.0 behavior):
- FPGA: HS=256, SS=512 samples (vs HS=1024, SS=2048)
- FX3: size*2, count=22 DMA buffers (vs size*8, count=11)
- Host: USB_MSG_SIZE_*_LEGACY used by libbladeRF
libbladeRF 2.6.0+ automatically detects FW < 2.5.0 AND FPGA < 0.16.0
and uses legacy buffer sizes for backward compatibility.