On 8/23/2016 2:43 PM, Leon Romanovsky wrote: > On Tue, Aug 23, 2016 at 06:13:42PM +0300, Yuval Shaia wrote: >> On Wed, Aug 03, 2016 at 04:27:33PM -0700, Adit Ranadive wrote: >>> + */ >>> +int pvrdma_query_port(struct ib_device *ibdev, u8 port, >>> + struct ib_port_attr *props) >>> +{ >>> + struct pvrdma_dev *dev = to_vdev(ibdev); >>> + union pvrdma_cmd_req req; >>> + union pvrdma_cmd_resp rsp; >>> + struct pvrdma_cmd_query_port *cmd = &req.query_port; >>> + struct pvrdma_cmd_query_port_resp *resp = &rsp.query_port_resp; >>> + int err; >>> + >>> + memset(cmd, 0, sizeof(*cmd)); >>> + cmd->hdr.cmd = PVRDMA_CMD_QUERY_PORT; >>> + cmd->port_num = port; >>> + >>> + err = pvrdma_cmd_post(dev, &req, &rsp); >> >> Patch order again, this guy is added in patch 7/15 ("IB/pvrdma: Add device >> command support"). >> Maybe it is only me having trouble to read it so do not want to be a nagger >> here but don't we have some (hidden) agreement that each patch is atomic >> and i should be able to compile the kernel after each one? > > IMHO, Doug will squash all these patches to one a second before adding > it into his tree. The patches divided here for easy review. > > Thanks > Indeed. When adding a new driver, the separation for review is handy, but bisectability is best maintained by a squashed commit. I like to have a good cover letter page for using as the squashed commit's log message when I do that. -- Doug Ledford <dledford@xxxxxxxxxx> GPG Key ID: 0E572FDD
Attachment:
signature.asc
Description: OpenPGP digital signature