linux-toradex.git/drivers/nvdimm/region_devs.c, branch v4.20-rc6

libnvdimm, pfn: Pad pfn namespaces relative to other regions

2018-12-05T22:16:12+00:00

Commit cfe30b872058 "libnvdimm, pmem: adjust for section collisions with
'System RAM'" enabled Linux to workaround occasions where platform
firmware arranges for "System RAM" and "Persistent Memory" to collide
within a single section boundary. Unfortunately, as reported in this
issue [1], platform firmware can inflict the same collision between
persistent memory regions.

The approach of interrogating iomem_resource does not work in this
case because platform firmware may merge multiple regions into a single
iomem_resource range. Instead provide a method to interrogate regions
that share the same parent bus.

This is a stop-gap until the core-MM can grow support for hotplug on
sub-section boundaries.

[1]: https://github.com/pmem/ndctl/issues/76

Fixes: cfe30b872058 ("libnvdimm, pmem: adjust for section collisions with...")
Cc: 
Reported-by: Patrick Geary 
Tested-by: Patrick Geary 
Reviewed-by: Vishal Verma 
Signed-off-by: Dan Williams

libnvdimm, region: Fail badblocks listing for inactive regions

2018-09-28T18:12:34+00:00

While experimenting with region driver loading the following backtrace
was triggered:

 INFO: trying to register non-static key.
 the code is fine but needs lockdep annotation.
 turning off the locking correctness validator.
 [..]
 Call Trace:
  dump_stack+0x85/0xcb
  register_lock_class+0x571/0x580
  ? __lock_acquire+0x2ba/0x1310
  ? kernfs_seq_start+0x2a/0x80
  __lock_acquire+0xd4/0x1310
  ? dev_attr_show+0x1c/0x50
  ? __lock_acquire+0x2ba/0x1310
  ? kernfs_seq_start+0x2a/0x80
  ? lock_acquire+0x9e/0x1a0
  lock_acquire+0x9e/0x1a0
  ? dev_attr_show+0x1c/0x50
  badblocks_show+0x70/0x190
  ? dev_attr_show+0x1c/0x50
  dev_attr_show+0x1c/0x50

This results from a missing successful call to devm_init_badblocks()
from nd_region_probe(). Block attempts to show badblocks while the
region is not enabled.

Fixes: 6a6bef90425e ("libnvdimm: add mechanism to publish badblocks...")
Cc: 
Reviewed-by: Johannes Thumshirn 
Reviewed-by: Dave Jiang 
Signed-off-by: Dan Williams

libnvdimm: Export max available extent

2018-07-25T21:11:43+00:00

The 'available_size' attribute showing the combined total of all
unallocated space isn't always useful to know how large of a namespace
a user may be able to allocate if the region is fragmented. This patch
will export the largest extent of unallocated space that may be allocated
to create a new namespace.

Signed-off-by: Keith Busch 
Reviewed-by: Vishal Verma 
Signed-off-by: Dave Jiang

libnvdimm: Use max contiguous area for namespace size

2018-07-25T21:11:09+00:00

This patch will find the max contiguous area to determine the largest
pmem namespace size that can be created. If the requested size exceeds
the largest available, ENOSPC error will be returned.

This fixes the allocation underrun error and wrong error return code
that have otherwise been observed as the following kernel warning:

  WARNING: CPU:  PID:  at drivers/nvdimm/namespace_devs.c:913 size_store

Fixes: a1f3e4d6a0c3 ("libnvdimm, region: update nd_region_available_dpa() for multi-pmem support")
Cc: 
Signed-off-by: Keith Busch 
Reviewed-by: Vishal Verma 
Signed-off-by: Dave Jiang

libnvdimm, pmem: Do not flush power-fail protected CPU caches

2018-06-06T18:02:32+00:00

This commit:

5fdf8e5ba566 ("libnvdimm: re-enable deep flush for pmem devices via fsync()")

intended to make sure that deep flush was always available even on
platforms which support a power-fail protected CPU cache.  An unintended
side effect of this change was that we also lost the ability to skip
flushing CPU caches on those power-fail protected CPU cache.

Fix this by skipping the low level cache flushing in dax_flush() if we have
CPU caches which are power-fail protected.  The user can still override this
behavior by manually setting the write_cache state of a namespace.  See
libndctl's ndctl_namespace_write_cache_is_enabled(),
ndctl_namespace_enable_write_cache() and
ndctl_namespace_disable_write_cache() functions.

Cc: 
Fixes: 5fdf8e5ba566 ("libnvdimm: re-enable deep flush for pmem devices via fsync()")
Signed-off-by: Ross Zwisler 
Signed-off-by: Dan Williams

Merge branch 'for-4.17/libnvdimm' into libnvdimm-for-next

2018-04-09T17:50:08+00:00

libnvdimm: Add of_node to region and bus descriptors

2018-04-07T14:53:23+00:00

We want to be able to cross reference the region and bus devices
with the device tree node that they were spawned from. libNVDIMM
handles creating the actual devices for these internally, so we
need to pass in a pointer to the relevant node in the descriptor.

Signed-off-by: Oliver O'Halloran 
Acked-by: Dan Williams 
Acked-by: Balbir Singh 
Signed-off-by: Dan Williams

libnvdimm: add an api to cast a 'struct nd_region' to its 'struct device'

2018-04-03T18:51:42+00:00

For debug, it is useful for bus providers to be able to retrieve the
'struct device' associated with an nd_region instance that it
registered. We already have to_nd_region() to perform the reverse cast
operation, in fact its duplicate declaration can be removed from the
private drivers/nvdimm/nd.h header.

Reviewed-by: Dave Jiang 
Signed-off-by: Dan Williams

libnvdimm, nfit: fix persistence domain reporting

2018-03-21T22:12:07+00:00

The persistence domain is a point in the platform where once writes
reach that destination the platform claims it will make them persistent
relative to power loss. In the ACPI NFIT this is currently communicated
as 2 bits in the "NFIT - Platform Capabilities Structure". The bits
comprise a hierarchy, i.e. bit0 "CPU Cache Flush to NVDIMM Durability on
Power Loss Capable" implies bit1 "Memory Controller Flush to NVDIMM
Durability on Power Loss Capable".

Commit 96c3a239054a "libnvdimm: expose platform persistence attr..."
shows the persistence domain as flags, but it's really an enumerated
hierarchy.

Fix this newly introduced user ABI to show the closest available
persistence domain before userspace develops dependencies on seeing, or
needing to develop code to tolerate, the raw NFIT flags communicated
through the libnvdimm-generic region attribute.

Fixes: 96c3a239054a ("libnvdimm: expose platform persistence attr...")
Reviewed-by: Dave Jiang 
Cc: "Rafael J. Wysocki" 
Cc: Ross Zwisler 
Signed-off-by: Dan Williams

libnvdimm, region: hide persistence_domain when unknown

2018-03-21T21:06:23+00:00

Similar to other region attributes, do not emit the persistence_domain
attribute if its contents are empty.

Fixes: 96c3a239054a ("libnvdimm: expose platform persistence attr...")
Cc: Dave Jiang 
Reviewed-by: Ross Zwisler 
Signed-off-by: Dan Williams