On Aug 30 08:54, Cédric Le Goater wrote: > Hello Byron, > > On 8/30/23 06:05, www wrote: > > Dear Sir， > > > > I have a few questions for you to ask. > > 1. According to my data collection, MCTP function should not be implemented > > in qemu. > > I would like to ask you how to simulate MCTP Over SMBus devices？Or do we > > have a device program with similar functions for reference? > > (The biggest problem with simulating MCTP Over SMBus devices is that there > > is a master-slave switch between request and response. > > It requires the device to actively respond to the requestor, that is, the > > device initiates the reply.) > > > > 2. Among the BMC functions, the communication between the BMC and the OS is > > a very important and basic function. > > Is there a way to simulate the communication between BMC and OS to test > > device drivers and applications? > > If we want to implement this feature, how do we go about it? > > > > I am looking forward to your reply. > > Byron > > Initial support for MCTP over I2C is being discussed here : > > > https://lore.kernel.org/qemu-devel/20230823-nmi-i2c-v4-0-2b0f86e5b...@samsung.com/ > > This is not my domain of expertise. So I am adding a few people who could > help. >
I think the already upstreamed support for asynchronous send (i2c_send_async + i2c_ack) along with the above series should get you started. The series referenced by Cédric implements an abstract mctp device that other devices can derive from. This model implements MCTP control message handling as well as handling the actual I2C transport (packetization). Thanks, Klaus
signature.asc
Description: PGP signature
