On 16.08.21 16:25, David Hildenbrand wrote:
Let's add the basic infrastructure to exclude some physical memory regions completely from /dev/mem access, on any architecture and under any system configuration (independent of CONFIG_STRICT_DEVMEM and independent of "iomem="). Use it for virtio-mem, to disallow mapping any virtio-mem memory via /dev/mem to user space after the virtio-mem driver was loaded: there is no sane use case to access the device-managed memory region via /dev/mem once the driver is actively (un)plugging memory within that region and we want to make sure that nobody will accidentially access unplugged memory in a sane environment. Details can be found in patch #1. v1 -> v2: - "/dev/mem: disallow access to explicitly excluded system RAM regions" -- Introduce and use for_each_resource() and next_resource_skip_children() -- s/iomem_range_contains_excluded/iomem_range_contains_excluded_devmem/ - "kernel/resource: cleanup and optimize iomem_is_exclusive()" -- Use for_each_resource() Cc: Arnd Bergmann <arnd@xxxxxxxx> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> Cc: "Michael S. Tsirkin" <mst@xxxxxxxxxx> Cc: Jason Wang <jasowang@xxxxxxxxxx> Cc: "Rafael J. Wysocki" <rafael.j.wysocki@xxxxxxxxx> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Cc: Dan Williams <dan.j.williams@xxxxxxxxx> Cc: Hanjun Guo <guohanjun@xxxxxxxxxx> Cc: Andy Shevchenko <andy.shevchenko@xxxxxxxxx> Cc: virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx Cc: linux-mm@xxxxxxxxx David Hildenbrand (3): /dev/mem: disallow access to explicitly excluded system RAM regions virtio-mem: disallow mapping virtio-mem memory via /dev/mem kernel/resource: cleanup and optimize iomem_is_exclusive() drivers/char/mem.c | 22 ++++++-------- drivers/virtio/virtio_mem.c | 4 ++- include/linux/ioport.h | 1 + kernel/resource.c | 60 +++++++++++++++++++++++++++++++++---- lib/Kconfig.debug | 4 ++- 5 files changed, 71 insertions(+), 20 deletions(-) base-commit: 7c60610d476766e128cc4284bb6349732cbd6606
More review welcome; I'd suggest this should go via the -mm tree, and not via the vhost tree.
-- Thanks, David / dhildenb
