2018-10-20 - Jan Engelhardt <email@example.com>
- Update to new upstream release 5.7.1
* Improved handling of non-SCSI zoned block devices.
* Change the behavior of zbc_open() to return an error when
trying to open a regular block device that was not
initialized for the emulation mode using the zbc_set_zones.
2018-03-18 - firstname.lastname@example.org
- Update to new upstream release 5.5.0
* Fixed various problems with the fake backend driver
(emulation mode), in particular when using a 4K sector disk
as backend storage.
* Properly set up direct I/O operation with the SG_IO interface
(SCSI and ATA backend drivers).
* Properly handle the maximum and optimal number of open zones
for drives with no limit and drives not reporting any limit.
* Skip open zone resource tests for drives with no limit or an
unreported limit on the maximum/optimal number of open zones.
* Add a new test to verify that a drive returns the correct
ASC/ASCQ sense error codes when writing to a full zone (HM
2017-08-08 - email@example.com
- Update to new upstream release 5.3.1
* A rewrite of libzbc introducing significant changes to the
API. Support for draft standards 4.x is dropped. Compliance
to the final ZBC/ZAC standards is improved.
* All functions now operate using 512B sector units, regardless
of the device actual logical block size. This is in line with
Linux block device management and can simplify application
implementation by unifying the addressing unit used when
devices with different logical block sizes (512B or 4096B)
are being used. Applications must however still comply with
write request alignment to the device physical block size,
which may be different from the 512B sector size unit used.
* Improved read and write functions (zbc_pread and zbc_pwrite)
to not limit the request size to the device maximum request
data transfer size (calls are internally split into multiple
2016-11-29 - firstname.lastname@example.org
- Update to new upstream release 4.3.3
* block backend driver: support latest "official" kernel
interface queued up for kernel 4.10 release.
2016-09-29 - email@example.com
- Update to new upstream release 4.3.0
* Properly handle detection of zoned block devices using sysfs
* Fixed handling of list view so that an operation on a zone does
not reset the current selection and view whenever possible.
* Improved handling of the list view using a filter model.
* Added support for direct I/Os (O_DIRECT). Note: this does not
work with SG node device files.
* Added maximum read/write command size in the device info
* Initialize maximum device read/write I/O size for each backend
* Check read/write command size against the maximum allowed for
* Fixed get capacity: the code is now common for ATA and SCSI.
This fixes problems in SCSI backend when RC_BASIS is 0 (report
zones was being executed without the max logical sectors being
set), and also adds RC_BASIS handling for ATA drives (this was
2016-07-01 - firstname.lastname@example.org
- Update to new upstream release 4.1.1
* This release implements support for ZBC and ZAC revision 4 and
above draft standards.
* Fixed the fake backend driver to pass all tests even if the
device specified is very small (i.e. it has only a few sequential
* Display SG_IO ioctl error message only when the log level is set
2016-02-13 - email@example.com
- Update to new git snapshot 2.0.0.g32
* Support for ZBC and ZAC r04
* Added block device backend driver for use with kernels providing
SMR support and exposing host-managed disks as block devices.