On 1/28/25 11:19 AM, Jerome Forissier wrote:
Use the coroutines framework to scan USB buses in parallel for better
performance. Tested on arm64 QEMU on a somewhat contrived example
(4 USB buses, each with one audio device, one keyboard, one mouse and
one tablet).
$ make qemu_arm64_defconfig
$ make -j$(nproc) CROSS_COMPILE="ccache aarch64-linux-gnu-"
$ qemu-system-aarch64 -M virt -nographic -cpu max -bios u-boot.bin \
$(for i in {1..4}; do echo -device qemu-xhci,id=xhci$i \
-device\ usb-{audio,kbd,mouse,tablet},bus=xhci$i.0; \
done)
The time spent in usb_init() is reported on the console and shows a
significant improvement with COROUTINES enabled.
Have you considered using the cyclic framework ( cyclic_register() and
co. ) to fully offload USB operations away from the main thread, i.e. to
make the U-Boot shell and e.g. 'usb start' run fully in parallel ? That
could then be extended to block transfers, which could run in
background, and ... we would also get a concept of shell pipes to move
data around like we do in Linux.
