linux-stable.git/drivers/s390/block, branch v5.3

s390/dasd: fix endless loop after read unit address configuration

2019-08-02T02:46:14+00:00

After getting a storage server event that causes the DASD device driver
to update its unit address configuration during a device shutdown there is
the possibility of an endless loop in the device driver.

In the system log there will be ongoing DASD error messages with RC: -19.

The reason is that the loop starting the ruac request only terminates when
the retry counter is decreased to 0. But in the sleep_on function there are
early exit paths that do not decrease the retry counter.

Prevent an endless loop by handling those cases separately.

Remove the unnecessary do..while loop since the sleep_on function takes
care of retries by itself.

Fixes: 8e09f21574ea ("[S390] dasd: add hyper PAV support to DASD device driver, part 1")
Cc: stable@vger.kernel.org # 2.6.25+
Signed-off-by: Stefan Haberland 
Reviewed-by: Jan Hoeppner 
Signed-off-by: Jens Axboe

Merge tag 'libnvdimm-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm

2019-07-18T17:52:08+00:00

Pull libnvdimm updates from Dan Williams:
 "Primarily just the virtio_pmem driver:

   - virtio_pmem

     The new virtio_pmem facility introduces a paravirtualized
     persistent memory device that allows a guest VM to use DAX
     mechanisms to access a host-file with host-page-cache. It arranges
     for MAP_SYNC to be disabled and instead triggers a host fsync()
     when a 'write-cache flush' command is sent to the virtual disk
     device.

   - Miscellaneous small fixups"

* tag 'libnvdimm-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  virtio_pmem: fix sparse warning
  xfs: disable map_sync for async flush
  ext4: disable map_sync for async flush
  dax: check synchronous mapping is supported
  dm: enable synchronous dax
  libnvdimm: add dax_dev sync flag
  virtio-pmem: Add virtio pmem driver
  libnvdimm: nd_region flush callback support
  libnvdimm, namespace: Drop uuid_t implementation detail

s390/dasd: Handle out-of-space constraint

2019-07-11T18:39:54+00:00

The storage server issues three different types of out-of-space messages
whenever the Extent Pool or Extent Repository space runs short. When a
configured warning watermark is reached, the physical space is
completeley exhausted, or the capacity constraints have been relieved, a
message is received.

A log entry for the sysadmin to react to is generated in any case. In
case the physical space is completely exhausted, sense data that reads
"no space left on device" is received. In this case, currently running
I/O will be blocked until space has either been released or added to the
extent pool, and a relieve message was received via an attention
interrupt.

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik

s390/dasd: Add discard support for ESE volumes

2019-07-11T18:39:54+00:00

ESE (Extent Space Efficient) volumes are thin-provisioned and therefore
space is only occupied with real data. In order to make previously used
space available for re-allocation again, discard support is enabled for
ESE volumes allowing the DASD driver to release said space.

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik

s390/dasd: Use ALIGN_DOWN macro

2019-07-11T18:39:54+00:00

There is now an ALIGN_DOWN macro available. Let's rather use kernel
provided macros that do the things we want.

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik

s390/dasd: Make dasd_setup_queue() a discipline function

2019-07-11T18:39:54+00:00

ECKD, FBA, and the DIAG discipline use slightly different block layer
settings. In preparation of even more diverse queue settings, make
dasd_setup_queue() a discipline function.

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik

s390/dasd: Add new ioctl to release space

2019-07-11T18:39:54+00:00

Userspace tools might have the need to release space for Extent Space
Efficient (ESE) volumes when working with such a device.

Provide the necessarry interface for such a task by implementing a new
ioctl BIODASDRAS. The ioctl uses the format_data_t data structure for
data input:

typedef struct format_data_t {
        unsigned int start_unit;        /* from track */
        unsigned int stop_unit;         /* to track */
        unsigned int blksize;           /* sectorsize */
        unsigned int intensity;
} format_data_t;

If the intensity is set to 0x40, start_unit and stop_unit are ignored
and space for the entire volume is released. Otherwise, if intensity is
set to 0, the respective range is released (if possible).

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik

s390/dasd: Add dasd_sleep_on_queue_interruptible()

2019-07-11T18:39:54+00:00

There is dasd_sleep_on() and dasd_sleep_on_interruptible() to start CCW
requests uninterruptible and interruptible. However, there is only
dasd_sleep_on_queue() to start requests from CCW queues uninterruptible.

Add dasd_sleep_on_queue_interruptible() to provide a way to start
requests from CCW queues interruptible. _dasd_sleep_on_queue() already
provides this functionality.

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik

s390/dasd: Add dynamic formatting support for ESE volumes

2019-07-11T18:39:53+00:00

A dynamic formatting is issued whenever a write request returns with
either a No Record Found error (Command Mode), Incorrect Length error
(Transport Mode), or File Protected error (Transport Mode). All three
cases mean that the tracks in question haven't been initialized in a
desired format yet.

The part of the volume that was tried to be written on is then formatted
and the original request is re-queued.

As the formatting will happen during normal I/O operations, it is quite
likely that there won't be any memory available to build the respective
request. Another two pages of memory are allocated per volume
specifically for the dynamic formatting.

The dasd_eckd_build_format() function is extended to make sure that the
original startdev is reused. Also, all formatting and format check
functions use the new memory pool exclusively now to reduce complexity.

Read operations will always return zero data when unformatted areas are
read.

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik

s390/dasd: Recognise data for ESE volumes

2019-07-11T18:39:53+00:00

In order to work with Extent Space Efficient (ESE) volumes, certain
viable information about those volumes and the corresponding extent
pool (such as extent size, configured space, allocated space, etc.) can
be provided.

Use the CCW commands Volume Storage Query and Logical Configuration
Query to receive detailed information about ESE volumes and the extent
pool respectively. These information are made accessible via internal
functions for subsequent users, and via sysfs attributes for userpsace
usage.

The new sysfs attributes reside in separate directories called capacity
and extent_pool.

attributes:
ese:
    0/1 depending on whether the volume is an ESE volume

Capacity related attributes:
space_allocated:
    Space currently allocated by the volume (in cyl)
space_configured:
    Remaining space in the extent pool (in cyl)
logical_capacity:
    The entire addressable space for this volume (in cyl)

Extent Pool related attributes:
pool_id:
    ID of the extent pool the volume in question resides in
pool_oos:
    Extent pool is out-of-space
extent_size:
    Size of a single extent in this pool
cap_at_warnlevel
    Extent pool capacity at warn level
warn_threshold:
    Threshold at which percentage of remaining extent pool space a
    warning message is issued

Signed-off-by: Jan Höppner 
Reviewed-by: Stefan Haberland 
Signed-off-by: Vasily Gorbik