Thanks for the review. Please see my response inline.
> > + select DMA_CMA
>
> Thіs needs to be
>
> select DMA_CMA if HAVE_DMA_CONTIGUOUS
>
> > +#include <linux/dma-contiguous.h>
>
> > + /* Allocate from CMA */
> > + // request_pages = (request_size >> PAGE_SHIFT) + 1;
> > + request_pages = (round_up(request_size, PAGE_SIZE) >> PAGE_SHIFT);
> > + page = dma_alloc_from_contiguous(NULL, request_pages, 0, false);
>
> dma_alloc_from_contiguous is an internal helper, you must use it
> through dma_alloc_coherent and pass a struct device to that function.
>
Can I directly use cma_alloc() and cma_release() in this case? The contiguous
memory allocated is just for virtual framebuffer device, not for any DMA
operation. I think using dma_alloc_coherent() might be a bit of overkill.
> > + if (!gen2vm) {
> > + pdev = pci_get_device(PCI_VENDOR_ID_MICROSOFT,
> > + PCI_DEVICE_ID_HYPERV_VIDEO, NULL);
> > + if (!pdev) {
> > + pr_err("Unable to find PCI Hyper-V video\n");
> > + return -ENODEV;
> > + }
> > + }
>
> Please actually implement a pci_driver instead of hacks like this.
>
I don't quite follow this. What do you mean implementing a pci_driver
in this case?
> > + par->need_docopy = false;
> > + goto getmem1;
> > + } else {
>
> No need for an else after a goto.
Thanks. Will do.
Wei
