From: Christoph Hellwig hch@infradead.org Date: Mon, 16 Nov 2020 16:27:44 +0000
On Mon, Nov 16, 2020 at 04:51:49AM -0500, Michael S. Tsirkin wrote:
On Mon, Nov 16, 2020 at 09:19:50AM +0000, Christoph Hellwig wrote:
I just noticed this showing up in Linus' tree and I'm not happy.
This whole model of the DMA subdevices in remoteproc is simply broken.
We really need to change the virtio code pass an expicit DMA device ( similar to what e.g. the USB and RDMA code does),
Could you point me at an example or two please?
Take a look at the ib_dma_* helper in include/rdma/ib_verbs.h and dma_device member in struct ib_device for the best example.
Oh, best example indeed. I did really love these helpers and kinda wish there were such for Ethernet and wireless networking. They'd allow to keep the code more readable and clean and prevent from several sorts of silly mistakes.
This could be done in e.g. 4 steps: - introduce such helpers for netdev/mac80211; - add checkpatch warnings to discourage usage of old methods like SET_NETDEV_DEV() and direct dereferencing of netdev->dev.parent; - slowly convert existing drivers to the new model; - remove the old way entirely along with checkpatch remnants.
I could take this if there'll be enough votes :)
Al