Now the kernel code branch I use is v5.4.39 from https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git linux-kernel. This driver comes from https://github.com/fixstars/cRTOS-drivers.git.The original ivshmem-net source code is contributed by Mans Rullgard .
在2021年12月30日星期四 UTC+8 00:38:13<Bezdeka, Florian> 写道： > On Wed, 2021-12-29 at 08:13 -0800, jiajun huang wrote: > > Hi, > > Unfortunately, after I tried almost all methods, the network still > > cannot be connected. But I can be sure that the ivshmem-net device > > did not receive the interrupt, but the shadow device that also uses > > the ivshmem interface can receive the interrupt normally, and the > > state change of another virtual device is detected. > > I want to know if this is due to the kernel version. Are there any > > kernel branches that you maintain yourself? > > Seems I overlooked your linux-net-driver.c. Where does that come from? > I'm normally using Linux from [1] or [2]. > > [1] https://github.com/siemens/linux/tree/jailhouse-enabling/5.10 > [2] https://github.com/siemens/linux/tree/jailhouse-enabling/5.10-rt > > I would expect build failures if there is none of the jailhouse- > enabling branches in use, but that might have changed. > > Regards, > Florian > > > Best wish > > 在2021年12月27日星期一 UTC+8 19:49:29<Florian Bezdeka> 写道： > > > On 27.12.21 12:34, Florian Bezdeka wrote: > > > > On 26.12.21 17:21, jiajun huang wrote: > > > > > Happy New Year， > > > > > > > > > > I successfully created root cell and inmate cell on QEMU, and > > > > > nuttx > > > > > running in the inmate cell. At the same time, I added the > > > > > ivshmem-net > > > > > device to the root cell and the intimate cell, and loaded the > > > > > NIC > > > > > driver. Finally, I configured ip 172.16.0.1 and 172.16.0.2 for > > > > > the > > > > > network card. But when I execute ping 172.16.0.2 in the root > > > > > cell, the > > > > > error "From 172.16.0.1 icmp_seq=1 Destination Host Unreachable" > > > > > appears. > > > > > The attachment is the driver I used in linux and nuttx > > > > > respectively.The > > > > > network driver uses the virtio interface. I tried to log in the > > > > > driver, > > > > > but I found that the control flow did not enter > > > > > ndo_start_xmit(). > > > > > > > > Looking at the hypervisor logs, it seems that the memory > > > > configuration > > > > of your virtual interfaces is correct now. The shared memory > > > > between the > > > > cells is detected, the device probing seems successful. > > > > > > > > I guess you don't receive interrupts for your virtual devices. > > > > Could you > > > > verify that by looking into /proc/interrupts? > > > > > > > > Are we still running on qemu? All your devices have .iommu = 0 > > > > set, but > > > > the .irqchip configuration looks different (root cell as well as > > > > inmate) > > > > compared to the qemu example that we have. > > > > > > > > > > As you have configured MSI-X interrupts for the virtual devices, > > > the > > > documentation says: > > > > > > You may also need to set the iommu field to match the IOMMU unit > > > that > > > the guest expects based on the bdf value. Try 1 if MSI-X interrupts > > > do > > > not make it when using 0. > > > > > > Just give it a try... > > -- You received this message because you are subscribed to the Google Groups "Jailhouse" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/jailhouse-dev/91742023-c4b3-4280-809a-5fa78cb7311bn%40googlegroups.com.
