Am Fri, 1 Sep 2017 05:36:31 -0700 schrieb Constantin Petra <[email protected]>:
> Hi, > > I have a question related to this above driver implementation, in an > Ultrascale+ environment. > > This was taken from > https://github.com/henning-schild/ivshmem-guest-code/blob/master/kernel_module/uio/uio_ivshmem.c, > in an attempt to set up IVSHMEM connection (not ivshmem-net) with a > different inmate OS (not Linux). > > here, in ivshmem_pci_probe(): > info->mem[1].addr = pci_resource_start(dev, 2); You will have to check out the "jailhouse" branch from the above github repository. In jailhouse the location and size of the shared memory are not in the BAR but somewhere in custom config space. The next thing you will run into might be problems with legacy interrupts. A patch for that is in "jailhouse-next" but i never tested it myself or received any feedback on it. If you can confirm that it works i will merge it into the jailhouse-branch. @Jonas, can you confirm that the patch in jailhouse-next made the interrupts work, or did you have to change ivshmem-guest-code further? Henning > the returned "addr" is zero, causing the probe to return -ENODEV when > ivshmem_pci_probe() is called by the OS. > > It's hard for me to follow the logic without a through pci driver > knownledge, do you have any idea what could cause this or would you > mind sharing some tips for debugging this? > > Thank you, > Constantin > -- 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]. For more options, visit https://groups.google.com/d/optout.
