Hi all, On 3/10/25 6:04 PM, Greg Kroah-Hartman wrote:
6.6-stable review patch. If anyone has any objections, please let me know. ------------------ From: Farouk Bouabid <farouk.bouabid@xxxxxxxxxxxxxxxxxxxxx> [ Upstream commit 5963d97aa780619ffb66cf4886c0ca1175ccbd3e ] A hardware switch can set the rs485 transceiver into half or full duplex mode. Switching to the half-duplex mode requires the user to enable em485 on uart5 using ioctl, DE/RE are both connected to GPIO0_B5 which is the RTS signal for uart0. Implement GPIO0_B5 as rts-gpios with RTS_ON_SEND option enabled (default) so that driver mode gets enabled while sending (RTS high) and receiver mode gets enabled while not sending (RTS low). In full-duplex mode (em485 is disabled), DE is connected to GPIO0_B5 and RE is grounded (enabled). Since GPIO0_B5 is implemented as rts-gpios, the driver mode gets enabled whenever we want to send something and RE is not affected (always enabled) in this case by the state of RTS. Signed-off-by: Farouk Bouabid <farouk.bouabid@xxxxxxxxxxxxxxxxxxxxx> Link: https://lore.kernel.org/r/20240208-dev-rx-enable-v6-2-39e68e17a339@xxxxxxxxxxxxxxxxxxxxx Signed-off-by: Heiko Stuebner <heiko@xxxxxxxxx> Stable-dep-of: 5ae4dca718ea ("arm64: dts: rockchip: Disable DMA for uart5 on px30-ringneck")
I don't mind the backport of this 005/145 patch, but this Stable-dep-of commit has already been merged in 6.6 with a conflict resolution, see bcc87e2e01637a90d67c66ce6c2eb28a78bf79f2.
Cheers, Quentin
