en: MT7601U driver performance note; VStarcam Hi3518EV200+GC2023 device page - #481
en: MT7601U driver performance note; VStarcam Hi3518EV200+GC2023 device page#481phedoreanu wants to merge 1 commit into
Conversation
…ce page adding-wifi-driver.md: document that the mainline mt7601u and the vendor-lineage mt7601sta (BR2_PACKAGE_MT7601U_OPENIPC) drivers are not equivalent - measured 0.4-1.0 Mbps vs 15-17 Mbps transmit on the same Hi3518EV200 camera, dongle and AP, which is the difference between an RTSP stream that lags by seconds and one that keeps up. New device page for VStarcam pan/tilt cameras (Hi3518EV200 + GC2023 MIPI + MT7601U): install layout that keeps the vendor U-Boot, the missing-tftpdstp trap, GC2023 MIPI sensor notes including the digital gain 0xb1=0 black-frame failure mode, PTZ stepper pins recovered from the vendor motogpio.ko, and the IR-cut pinmux that sysfs alone cannot reach.
PR Summary by QodoDocument MT7601U driver differences and add VStarcam Hi3518EV200 setup page
AI Description
Diagram
High-Level Assessment
Files changed (3)
|
Code Review by Qodo🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)
Great, no issues found!Qodo reviewed your code and found no material issues that require reviewTip of the day💡 Did you know, you can ask Qodo to dismiss a finding you disagree with, with your reason on record |
Two additions from converting a VStarcam Hi3518EV200 pan/tilt camera:
adding-wifi-driver.md— a note under MediaTek that the mainlinemt7601uand the vendor-lineagemt7601sta(whatBR2_PACKAGE_MT7601U_OPENIPCbuilds) are not equivalent. Measured on the same camera, dongle, AP and channel: 0.4–1.0 Mbps transmit on mainline (withmt7601u_mcu_wait_resp retryingin dmesg) vs 15–17 Mbps on the STA driver — the difference between an RTSP stream that lags by seconds and one that keeps its send queue empty. Also warns against swapping the two drivers on a live system, which can wedge the dongle MCU until a power cycle.New device page for VStarcam pan/tilt cameras (Hi3518EV200 + GC2023 MIPI + MT7601U): an install layout that keeps the vendor U-Boot (its environment at
0x80000sits inside the kernel region of the standard layout, so standard offsets let a latersaveenvcorrupt the kernel), the missing-tftpdstptrap in the vendor U-Boot, GC2023 MIPI sensor notes including the digital-gain0xb1=0black-frame failure mode, PTZ stepper GPIOs recovered from the vendor'smotogpio.ko, and the IR-cut pinmux that sysfs alone cannot reach. All verified on real hardware.