Hi James, As Tejun and Ming is suggesting to have interface close to the usage as fabric infrasture I was thinking of this approach from the inputs provided by you in the below mail. Please provide your inputs on this approach. 1) blkcg will have a new field to store driver specific information as "blkio_cg_ priv_data"(in the current patch it is app_identifier) as Tejun said he doesn’t mind cgroup data structs carrying extra bits for stuff. 2) scsi transport will provide a new interface(sysfs) as register_vm_fabric 3) As part of this interface user/deamon will provide the details of VM such as UUID,PID on VM creation to the transport . 4) With VM PID information we need to find the associated blkcg and needs to update the UUID info in blkio_cg_ priv_data. 5) Once we update the blkio_cg_ priv_data with vmid all the io’s issued from VM will have the UUID info as part of blkcg. With this we can also address the concerns raised by Tejun and the existing lpfc patches still holds good. Regards, Muneendra. -----Original Message----- From: James Smart [mailto:james.smart@xxxxxxxxxxxx] Sent: Thursday, August 6, 2020 5:08 AM To: Hannes Reinecke <hare@xxxxxxx>; Muneendra <muneendra.kumar@xxxxxxxxxxxx>; linux-block@xxxxxxxxxxxxxxx; linux-scsi@xxxxxxxxxxxxxxx Cc: pbonzini@xxxxxxxxxx; emilne@xxxxxxxxxx; mkumar@xxxxxxxxxx; Gaurav Srivastava <gaurav.srivastava@xxxxxxxxxxxx>; James Smart <jsmart2021@xxxxxxxxx> Subject: Re: [RFC 16/16] lpfc: vmid: Introducing vmid in io path. That's actually what's supposed to be happening. fastpath uses the uuid to look up a vmid tag. If no vmid tag, kick off the fabric traffic that will get one but don't wait for it to complete. Any io issued while that process is occurring will be not be vmid tagged. I'll circle back on lpfc to make sure this is happening. In the mean time - the most important patch to review is the cgroup patch - patch1. If we wanted to speed the driver's io path up, one thing to consider is adding a driver-settable value on the blkcg structure. Once the fabric traffic obtained the vmid, the driver would set the blkcg structure with the value. In this scenario though, as the vmid is destroyed as of link down, the driver needs a way, independent of an io, to reach into the blkcg struct to clear the vmid value. We also need to be sure the blkcg struct won't be on top of a multipath device or something such that the blkcg struct may be referenced by a different scsi host - I assume we're good in that area. -- james
