BMC/Intel-BMC/linux.git/drivers/nvdimm/namespace_devs.c, branch dev-4.6

libnvdimm, pmem: adjust for section collisions with 'System RAM'

2016-03-05T20:25:45+00:00

On a platform where 'Persistent Memory' and 'System RAM' are mixed within a given sparsemem section, trim the namespace and notify about the sub-optimal alignment. Cc: Toshi Kani Cc: Ross Zwisler Signed-off-by: Dan Williams

libnvdimm: fix mode determination for e820 devices

2016-01-26T17:40:32+00:00

Correctly display "safe" mode when a btt is established on a e820/memmap defined pmem namespace. Signed-off-by: Dan Williams

libnvdimm: fix namespace object confusion in is_uuid_busy()

2016-01-06T02:37:23+00:00

When btt devices were re-worked to be child devices of regions this routine was overlooked. It mistakenly attempts to_nd_namespace_pmem() or to_nd_namespace_blk() conversions on btt and pfn devices. By luck to date we have happened to be hitting valid memory leading to a uuid miscompare, but a recent change to struct nd_namespace_common causes: BUG: unable to handle kernel NULL pointer dereference at 0000000000000001 IP: [] memcmp+0xc/0x40 [..] Call Trace: [] is_uuid_busy+0xc1/0x2a0 [libnvdimm] [] ? to_nd_blk_region+0x50/0x50 [libnvdimm] [] device_for_each_child+0x50/0x90 Cc: Signed-off-by: Dan Williams

libnvdimm, pfn: move 'memory mode' indication to sysfs

2015-12-24T20:20:20+00:00

'Memory mode' is defined as the capability of a DAX mapping to be the source/target of DMA and other "direct I/O" scenarios. While it currently requires allocating 'struct page' for each page frame of persistent memory in the namespace it will not always be the case. Work continues on reducing the kernel's dependency on 'struct page'. Let's not maintain a suffix that is expected to lose meaning over time. In other words a future 'raw mode' pmem namespace may be as capable as today's 'memory mode' namespace. Undo the encoding of the mode in the device name and leave it to other tooling to determine the mode of the namespace from its attributes. Reported-by: Matthew Wilcox Signed-off-by: Dan Williams

libnvdimm, pfn: fix pfn seed creation

2015-12-13T19:41:36+00:00

Similar to btt, plant a new pfn seed when the existing one is activated. Signed-off-by: Dan Williams

nvdimm: improve diagnosibility of namespaces

2015-12-09T00:27:30+00:00

In order to bind namespace to the driver user must first set all mandatory attributes in the following order: - uuid - size - sector_size (for blk namespace only) If the order is wrong, then user either won't be able to set the attribute or bind the namespace. This simple patch improves diagnosibility of common operations with namespaces by printing some details about the error instead of failing silently. Below are examples of error messages (assuming dyndbg is enabled for nvdimms): [/]# echo 4194304 > /sys/bus/nd/devices/region5/namespace5.0/size [ 288.372612] nd namespace5.0: __size_store: uuid not set [ 288.374839] nd namespace5.0: size_store: 400000 fail (-6) sh: write error: No such device or address [/]# [/]# echo namespace5.0 > /sys/bus/nd/drivers/nd_blk/bind [ 554.671648] nd_blk namespace5.0: nvdimm_namespace_common_probe: sector size not set [ 554.674688] ndbus1: nd_blk.probe(namespace5.0) = -19 sh: write error: No such device [/]# Signed-off-by: Dmitry V. Krivenok Signed-off-by: Dan Williams

libnvdimm, pmem: direct map legacy pmem by default

2015-08-29T03:40:05+00:00

The expectation is that the legacy / non-standard pmem discovery method (e820 type-12) will only ever be used to describe small quantities of persistent memory. Larger capacities will be described via the ACPI NFIT. When "allocate struct page from pmem" support is added this default policy can be overridden by assigning a legacy pmem namespace to a pfn device, however this would be only be necessary if a platform used the legacy mechanism to define a very large range. Cc: Christoph Hellwig Signed-off-by: Dan Williams

libnvdimm, pfn: 'struct page' provider infrastructure

2015-08-29T03:39:36+00:00

Implement the base infrastructure for libnvdimm PFN devices. Similar to BTT devices they take a namespace as a backing device and layer functionality on top. In this case the functionality is reserving space for an array of 'struct page' entries to be handed out through pfn_to_page(). For now this is just the basic libnvdimm-device-model for configuring the base PFN device. As the namespace claiming mechanism for PFN devices is mostly identical to BTT devices drivers/nvdimm/claim.c is created to house the common bits. Cc: Ross Zwisler Signed-off-by: Dan Williams

libnvdimm, btt: write and validate parent_uuid

2015-08-14T17:43:04+00:00

When a BTT is instantiated on a namespace it must validate the namespace uuid matches the 'parent_uuid' stored in the btt superblock. This property enforces that changing the namespace UUID invalidates all former BTT instances on that storage. For "IO namespaces" that don't have a label or UUID, the parent_uuid is set to zero, and this validation is skipped. For such cases, old BTTs have to be invalidated by forcing the namespace to raw mode, and overwriting the BTT info blocks. Based on a patch by Dan Williams Signed-off-by: Vishal Verma Signed-off-by: Dan Williams

libnvdimm: Add sysfs numa_node to NVDIMM devices

2015-06-26T15:23:38+00:00

Add support of sysfs 'numa_node' to I/O-related NVDIMM devices under /sys/bus/nd/devices, regionN, namespaceN.0, and bttN.x. An example of numa_node values on a 2-socket system with a single NVDIMM range on each socket is shown below. /sys/bus/nd/devices |-- btt0.0/numa_node:0 |-- btt1.0/numa_node:1 |-- btt1.1/numa_node:1 |-- namespace0.0/numa_node:0 |-- namespace1.0/numa_node:1 |-- region0/numa_node:0 |-- region1/numa_node:1 These numa_node files are then linked under the block class of their device names. /sys/class/block/pmem0/device/numa_node:0 /sys/class/block/pmem1s/device/numa_node:1 This enables numactl(8) to accept 'block:' and 'file:' paths of pmem and btt devices as shown in the examples below. numactl --preferred block:pmem0 --show numactl --preferred file:/dev/pmem1s --show Signed-off-by: Toshi Kani Signed-off-by: Dan Williams