diff options
| author | Dan Williams <dan.j.williams@intel.com> | 2022-12-02 01:03:35 +0300 |
|---|---|---|
| committer | Dan Williams <dan.j.williams@intel.com> | 2022-12-03 10:52:32 +0300 |
| commit | dc370b28c8425669e7ed5af4c01540645cfb00ec (patch) | |
| tree | 37e58358cc8be8750eca80a2342e7459d25c9464 /drivers/nvdimm/region.c | |
| parent | 07cb5f705b4fe9e1386a610da4cb3c063267714f (diff) | |
| download | linux-dc370b28c8425669e7ed5af4c01540645cfb00ec.tar.xz | |
nvdimm/region: Move cache management to the region driver
Now that cpu_cache_invalidate_memregion() is generically available, use
it to centralize CPU cache management in the nvdimm region driver.
This trades off removing redundant per-dimm CPU cache flushing with an
opportunistic flush on every region disable event to cover the case of
sensitive dirty data in the cache being written back to media after a
secure erase / overwrite event.
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/166993221550.1995348.16843505129579060258.stgit@dwillia2-xfh.jf.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Diffstat (limited to 'drivers/nvdimm/region.c')
| -rw-r--r-- | drivers/nvdimm/region.c | 11 |
1 files changed, 11 insertions, 0 deletions
diff --git a/drivers/nvdimm/region.c b/drivers/nvdimm/region.c index 390123d293ea..88dc062af5f8 100644 --- a/drivers/nvdimm/region.c +++ b/drivers/nvdimm/region.c @@ -2,6 +2,7 @@ /* * Copyright(c) 2013-2015 Intel Corporation. All rights reserved. */ +#include <linux/memregion.h> #include <linux/cpumask.h> #include <linux/module.h> #include <linux/device.h> @@ -100,6 +101,16 @@ static void nd_region_remove(struct device *dev) */ sysfs_put(nd_region->bb_state); nd_region->bb_state = NULL; + + /* + * Try to flush caches here since a disabled region may be subject to + * secure erase while disabled, and previous dirty data should not be + * written back to a new instance of the region. This only matters on + * bare metal where security commands are available, so silent failure + * here is ok. + */ + if (cpu_cache_has_invalidate_memregion()) + cpu_cache_invalidate_memregion(IORES_DESC_PERSISTENT_MEMORY); } static int child_notify(struct device *dev, void *data) |