On Wed, Mar 9, 2016 at 2:21 AM, Dennis Dalessandro <dennis.dalessandro@xxxxxxxxx> wrote: > On Tue, Mar 08, 2016 at 10:56:48PM +0200, Or Gerlitz wrote: >> >> If indeed there's a need for such a pin down cache, it should have >> been implemented in the IB core, b/c both the problem and the solution >> are generic and have nothing to do with certain device driver. > Since this is for psm it has nothing to do with the IB core. I fail to see > what that would achieve. So your logic goes like 1. hfi is a driver which plugs into the IB core 2. OTOH hfi implements proprietary inter-connect, but we can use the IB core for fabric management so we change the IB core MAD code to support the interconnect 3. OTOH the maintainer allowed hfi to expose to user-space proprietary char-device 4. so from now on, we can patch hfi with 300 patches per release that deal with extensions to functionality exposed by the char device and we don't care to solve generic problems that has nothing to do with the HW used for the device. Really? >> The problem is that you are bypassing the IB core altogether with a >> proprietary character device in the hfi1 driver which managed to snick >> in even in the presence of reviewer comments during the submission. > Yes there were reviewer comments, and the maintainer made a decision. > I would also not call this proprietary. It is open source after all: > > See: https://github.com/01org/opa-psm2 -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html