linux/mdadm
1
0
Fork 0
Code Issues Activity

Merging upstream version 4.3+20240412.

Browse source 
Signed-off-by: Daniel Baumann <daniel@debian.org>
This commit is contained in:
Daniel Baumann 2025-02-14 06:34:22 +01:00
parent 8e1e76d6e6
commit 9a795ca236
Signed by: daniel
GPG key ID: FBB4F0E80A80222F
67 changed files with 2247 additions and 2747 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

98
ANNOUNCE-3.0
View file

@ -1,98 +0,0 @@
Subject: ANNOUNCE: mdadm 3.0 - A tool for managing Soft RAID under Linux
I am pleased to (finally) announce the availability of
mdadm version 3.0
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This is a major new version and as such should be treated with some
caution. However it has seen substantial testing and is considerred
to be ready for wide use.
The significant change which justifies the new major version number is
that mdadm can now handle metadata updates entirely in userspace.
This allows mdadm to support metadata formats that the kernel knows
nothing about.
Currently two such metadata formats are supported:
- DDF - The SNIA standard format
- Intel Matrix - The metadata used by recent Intel ICH controlers.
Also the approach to device names has changed significantly.
If udev is installed on the system, mdadm will not create any devices
in /dev. Rather it allows udev to manage those devices. For this to work
as expected, the included udev rules file should be installed.
If udev is not installed, mdadm will still create devices and symlinks
as required, and will also remove them when the array is stopped.
mdadm now requires all devices which do not have a standard name (mdX
or md_dX) to live in the directory /dev/md/. Names in this directory
will always be created as symlinks back to the standard name in /dev.
The man pages contain some information about the new externally managed
metadata. However see below for a more condensed overview.
Externally managed metadata introduces the concept of a 'container'.
A container is a collection of (normally) physical devices which have
a common set of metadata. A container is assembled as an md array, but
is left 'inactive'.
A container can contain one or more data arrays. These are composed from
slices (partitions?) of various devices in the container.
For example, a 5 devices DDF set can container a RAID1 using the first
half of two devices, a RAID0 using the first half of the remain 3 devices,
and a RAID5 over thte second half of all 5 devices.
A container can be created with
mdadm --create /dev/md0 -e ddf -n5 /dev/sd[abcde]
or "-e imsm" to use the Intel Matrix Storage Manager.
An array can be created within a container either by giving the
container name and the only member:
mdadm -C /dev/md1 --level raid1 -n 2 /dev/md0
or by listing the component devices
mdadm -C /dev/md2 --level raid0 -n 3 /dev/sd[cde]
To assemble a container, it is easiest just to pass each device in turn to
mdadm -I
for i in /dev/sd[abcde]
do mdadm -I $i
done
This will assemble the container and the components.
Alternately the container can be assembled explicitly
mdadm -A /dev/md0 /dev/sd[abcde]
Then the components can all be assembled with
mdadm -I /dev/md0
For each container, mdadm will start a program called "mdmon" which will
monitor the array and effect any metadata updates needed. The array is
initially assembled readonly. It is up to "mdmon" to mark the metadata
as 'dirty' and which the array to 'read-write'.
The version 0.90 and 1.x metadata formats supported by previous
versions for mdadm are still supported and the kernel still performs
the same updates it use to. The new 'mdmon' approach is only used for
newly introduced metadata types.
NeilBrown 2nd June 2009

22
ANNOUNCE-3.0.1
View file

@ -1,22 +0,0 @@
Subject: ANNOUNCE: mdadm 3.0.1 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.0.1
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This contains only minor bug fixes over 3.0. If you are using
3.0, you could consider upgrading.
The brief change log is:
- Fix various segfaults
- Fixed for --examine with containers
- Lots of other little fixes.
NeilBrown 25th September 2009

21
ANNOUNCE-3.0.2
View file

@ -1,21 +0,0 @@
Subject: ANNOUNCE: mdadm 3.0.2 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.0.2
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This just contains one bugfix over 3.0.1 - I was obviously a bit hasty
in releasing that one.
The brief change log is:
- Fix crash when hosthost is not set, as often happens in
early boot.
NeilBrown 25th September 2009

29
ANNOUNCE-3.0.3
View file

@ -1,29 +0,0 @@
Subject: ANNOUNCE: mdadm 3.0.3 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.0.3
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This contains a collection of bug fixes and minor enhancements over
3.0.1.
The brief change log is:
- Improvements for creating arrays giving just a name, like 'foo',
rather than the full '/dev/md/foo'.
- Improvements for assembling member arrays of containers.
- Improvements to test suite
- Add option to change increment for RebuildNN messages reported
by "mdadm --monitor"
- Improvements to mdmon 'hand-over' from initrd to final root.
- Handle merging of devices that have left an IMSM array and are
being re-incorporated.
- Add missing space in "--detail --brief" output.
NeilBrown 22nd October 2009

33
ANNOUNCE-3.1
View file

@ -1,33 +0,0 @@
Subject: ANNOUNCE: mdadm 3.1 - A tool for managing Soft RAID under Linux
Hot on the heals of 3.0.3 I am pleased to announce the availability of
mdadm version 3.1
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
It contains significant feature enhancements over 3.0.x
The brief change log is:
- Support --grow to change the layout of RAID4/5/6
- Support --grow to change the chunksize of raid 4/5/6
- Support --grow to change level from RAID1 -> RAID5 -> RAID6 and
back.
- Support --grow to reduce the number of devices in RAID4/5/6.
- Support restart of these grow options which assembling an array
which is partially grown.
- Assorted tests of this code, and of different RAID6 layouts.
Note that a 2.6.31 or later is needed to have access to these.
Reducing devices in a RAID4/5/6 requires 2.6.32.
Changing RAID5 to RAID1 requires 2.6.33.
You should only upgrade if you need to use, or which to test, these
features.
NeilBrown 22nd October 2009

39
ANNOUNCE-3.1.1
View file

@ -1,39 +0,0 @@
Subject: ANNOUNCE: mdadm 3.1.1 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.1.1
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This is a bugfix release over 3.1, which was withdrawn due to serious
bugs. So it might be best to ignore 3.1 and say that this is a significant
feature release over 3.0.x
Significant changes are:
- RAID level conversion between RAID1, RAID5, and RAID6 are
possible were the kernel supports it (2.6.32 at least)
- online chunksize and layout changing for RAID5 and RAID6
where the kernel supports it.
- reduce the number of devices in a RAID4/5/6 array.
- The default metadata is not v1.1. This metadata is stored at the
start of the device so is safer in many ways but could interfere with
boot loaded. The old default (0.90) is still available and fully
supported.
- The default chunksize is now 512K rather than 64K. This seems more
appropriate for modern devices.
- The default bitmap chunksize for internal bitmaps is now at least
64Meg as fine grained bitmaps tend to impact performance more for
little extra gain.
This release is believed to be stable and you should feel free to
upgrade to 3.1.1.
NeilBrown 19th November 2009

46
ANNOUNCE-3.1.2
View file

@ -1,46 +0,0 @@
Subject: ANNOUNCE: mdadm 3.1.2 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.1.2
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This is a bugfix/stability release over 3.1.1.
Significant changes are:
- The default metadata has change again (sorry about that).
It is now v1.2 and will hopefully stay that way. It turned
out there with boot-block issues with v1.1 which make it
unsuitable for a default, though in many cases it is still
suitable to use.
- Stopping a container is not permitted when members are still
active
- Add 'homehost' to the valid words for the "AUTO" config file
line. When followed by "-all", this causes mdadm to
auto-assemble any array belonging to this host, but not
auto-assemble anything else.
- Fix some bugs with "--grow --chunksize=" for changing chunksize.
- VAR_RUN can be easily changed at compile time just like ALT_RUN.
This gives distros more flexability in how to manage the
pid and sock files that mdmon needs.
- Various mdmon fixes
- Alway make bitmap 4K-aligned if at all possible.
- If mdadm.conf lists arrays which have inter-dependencies,
the previously had to be listed in the "right" order. Now
any order should work.
- Fix --force assembly of v1.x arrays which are in the process
of recovering.
- Add section on 'scrubbing' to 'md' man page.
- Various command-line-option parsing improvements.
- ... and lots of other bug fixes.
This release is believed to be stable and you should feel free to
upgrade to 3.1.2
NeilBrown 10th March 2010

46
ANNOUNCE-3.1.3
View file

@ -1,46 +0,0 @@
Subject: ANNOUNCE: mdadm 3.1.3 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.1.3
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This is a bugfix/stability release over 3.1.2
Significant changes are:
- mapfile now lives in a fixed location which default to
/dev/.mdadm/map but can be changed at compile time. This
location is choses and most distros provide it during early
boot and preserve it through. As long a /dev exists and is
writable, /dev/.mdadm will be created.
Other files file communication with mdmon live here too.
This fixes a bug reported by Debian and Gentoo users where
udev would spin in early-boot.
- IMSM and DDF metadata will not be recognised on partitions
as they should only be used on whole-disks.
- Various overflows causes by 2G drives have been addressed.
- A subarray of an IMSM contain can now be killed with
--kill-subarray. Also subarrays can be renamed with
--update-subarray
- -If (or --incremental --fail) can be used from udev to
fail and remove from all arrays a device which has been
unplugged from the system. i.e. hot-unplug-support.
- "mdadm /dev/mdX --re-add missing" will look for any device
that looks like it should be a member of /dev/mdX but isn't
and will automatically --re-add it
- Now compile with -Wextra to get extra warnings.
- Lots of minor bug fixes, documentation improvements, etcc
This release is believed to be stable and you should feel free to
upgrade to 3.1.3
It is expected that the next release will be 3.2 with a number of new
features. 3.1.4 will only happen if important bugs show up before 3.2
is stable.
NeilBrown 6th August 2010

37
ANNOUNCE-3.1.4
View file

@ -1,37 +0,0 @@
Subject: ANNOUNCE: mdadm 3.1.4 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.1.4
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This is a bugfix/stability release over 3.1.3.
3.1.3 had a couple of embarrasing regressions and a couple of other
issues surfaces which had easy fixes so I decided to make a 3.1.4
release after all.
Two fixes related to configs that aren't using udev:
- Don't remove md devices which 'standard' names on --stop
- Allow dev_open to work on read-only /dev
And fixed regressions:
- Allow --incremental to add spares to an array
- Accept --no-degraded as a deprecated option rather than
throwing an error
- Return correct success status when --incrmental assembling
a container which does not yet have enough devices.
- Don't link mdadm with pthreads, only mdmon needs it.
- Fix compiler warning due to bad use of snprintf
- Fix spare migration
This release is believed to be stable and you should feel free to
upgrade to 3.1.4
It is expected that the next release will be 3.2 with a number of new
features.
NeilBrown 31st August 2010

42
ANNOUNCE-3.1.5
View file

@ -1,42 +0,0 @@
Subject: ANNOUNCE: mdadm 3.1.5 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.1.5
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git?p=mdadm
This is a bugfix/stability release over 3.1.4. It contains all the
important bugfixes found while working on 3.2 and 3.2.1. It will be
the last 3.1.x release - 3.2.1 is expected to be released in a few days.
Changes include:
- Fixes for v1.x metadata on big-endian machines.
- man page improvements
- Improve '--detail --export' when run on partitions of an md array.
- Fix regression with removing 'failed' or 'detached' devices.
- Fixes for "--assemble --force" in various unusual cases.
- Allow '-Y' to mean --export. This was documented but not implemented.
- Various fixed for handling 'ddf' metadata. This is now more reliable
but could benefit from more interoperability testing.
- Correctly list subarrays of a container in "--detail" output.
- Improve checks on whether the requested number of devices is supported
by the metadata - both for --create and --grow.
- Don't remove partitions from a device that is being included in an
array until we are fully committed to including it.
- Allow "--assemble --update=no-bitmap" so an array with a corrupt
bitmap can still be assembled.
- Don't allow --add to succeed if it looks like a "--re-add" is probably
wanted, but cannot succeed. This avoids inadvertently turning
devices into spares when an array is failed.
This release is believed to be stable and you should feel free to
upgrade to 3.1.5
NeilBrown 23rd March 2011

77
ANNOUNCE-3.2
View file

@ -1,77 +0,0 @@
Subject: ANNOUNCE: mdadm 3.2 - A tool for managing Soft RAID under Linux (DEVEL ONLY)
I am pleased to announce the availability of
mdadm version 3.2
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm devel-3.2
http://neil.brown.name/git?p=mdadm
This is a "Developers only" release. Please don't consider using it
or making it available to others without reading the following.
By far the most significant change in this release related to the
management of reshaping arrays. This code has been substantially
re-written so that it can work with 'externally managed metadata' -
Intel's IMSM in particular. We now support level migration and
OnLine Capacity Expansion on these arrays.
However, while the code largely works it has not been tested
exhaustively so there are likely to be problems. As the reshape code
for native metadata arrays was changed as part of this rewrite these
problems could also result in regressions for reshape of native
metadata.
It is partly to encourage greater testing that this release is being
made. Any reports of problem - particular reproducible recipes for
triggering the problems - will be gratefully received.
It is hopped that a "3.2.1" release will be available in early March
which will be a bugfix release over this and can be considered
suitable for general use.
Other changes of note:
- Policy framework.
Various policy statements can be made in the mdadm.conf to guide
the behaviour of mdadm, particular with regards to how new devices
are treated by "mdadm -I".
Depending on the 'action' associated with a device (identified by
its 'path') such need devices can be automatically re-added to and
existing array that they previously fell out off, or automatically
added as a spare if they appear to contain no data.
- mdadm now has a limited understanding of partition tables. This
allows the policy framework to make decisions about partitioned
devices as well.
- --incremental --remove can be told what --path the device was on,
and this info will be recorded so that another device appearing at
the same physical location can be preferentially added to the same
array (provides the spare-same-slot action policy applied to the
path).
- A new flags "--invalid-backup" flag is available in --assemble
mode. This can be used to re-assemble an array which was stopping
in the middle of a reshape, and for which the 'backup file' is no
longer available or is corrupted. The array may have some
corruption in it at the point where reshape was up to, but at least
the rest of the array will become available.
- Various internal restructuring - more is needed.
Any feed back and bug reports are always welcomed at:
linux-raid@vger.kernel.org
And please: don't use this in production - particularly not the
--grow functionality.
NeilBrown 1st February 2011

75
ANNOUNCE-3.2.1
View file

@ -1,75 +0,0 @@
I am pleased to announce the availability of
mdadm version 3.2.1
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git/mdadm
Many of the changes in this release are of internal interest only,
restructuring and refactoring code and so forth.
Most of the bugs found and fixed during development for 3.2.1 have been
back-ported for the recently-release 3.1.5 so this release primarily
provides a few new features over 3.1.5.
They include:
- policy framework
Policy can be expressed for moving spare devices between arrays, and
for how to handle hot-plugged devices. This policy can be different
for devices plugged in to different controllers etc.
This, for example, allows a configuration where when a device is plugged
in it is immediately included in an md array as a hot spare and
possibly starts recovery immediately if an array is degraded.
- some understanding of mbr and gpt paritition tables
This is primarly to support the new hot-plug support. If a
device is plugged in and policy suggests it should have a partition table,
the partition table will be copied from a suitably similar device, and
then the partitions will hot-plug and can then be added to md arrays.
- "--incremental --remove" can remember where a device was removed from
so if a device gets plugged back in the same place, special policy applies
to it, allowing it to be included in an array even if a general hotplug
will not be included.
- enhanced reshape options, including growing a RAID0 by converting to RAID4,
restriping, and converting back. Also convertions between RAID0 and
RAID10 and between RAID1 and RAID10 are possible (with a suitably recent
kernel).
- spare migration for IMSM arrays.
Spare migration can now work across 'containers' using non-native metadata
and specifically Intel's IMSM arrays support spare migrations.
- OLCE and level migration for Intel IMSM arrays.
OnLine Capacity Expansion and level migration (e.g. RAID0 -> RAID5) is
supported for Intel Matrix Storage Manager arrays.
This support is currently 'experimental' for technical reasons. It can
be enabled with "export MDADM_EXPERIMENTAL=1"
- avoid including wayward devices
If you split a RAID1, mount the two halves as two separate degraded RAID1s,
and then later bring the two back together, it is possible that the md
metadata won't properly show that one must over-ride the other.
mdadm now does extra checking to detect this possibilty and avoid
potentially corrupting data.
- remove any possible confusion between similar options.
e.g. --brief and --bitmap were mapped to 'b' and mdadm wouldn't
notice if one was used where the other was expected.
- allow K,M,G suffixes on chunk sizes
While mdadm-3.2.1 is considered to be reasonably stable, you should
only use it if you want to try out the new features, or if you
generally like to be on the bleeding edge. If the new features are not
important to you, then 3.1.5 is probably the appropriate version to be using
until 3.2.2 comes out.
NeilBrown 28th March 2011

36
ANNOUNCE-3.2.2
View file

@ -1,36 +0,0 @@
Subject: ANNOUNCE: mdadm 3.2.2 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.2.2
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git/mdadm
This release is largely a stablising release for the 3.2 series.
Many of the changes just fix bugs introduces in 3.2 or 3.2.1.
There are some new features. They are:
- reshaping IMSM (Intel metadata) arrays is no longer 'experimental',
it should work properly and be largely compatible with IMSM drivers in
other platforms.
- --assume-clean can be used with --grow --size to avoid resyncing the
new part of the array. This is only support with very new kernels.
- RAID0 arrays can have chunksize which is not a power of 2. This has been
supported in the kernel for a while but is only now supprted by
mdadm.
- A new tool 'raid6check' is available which can check a RAID6 array,
or part of it, and report which device is most inconsistent with the
others if any stripe is inconsistent. This is still under development
and does not have a man page yet. If anyone tries it out and has any
questions or experience to report, they would be most welcome on
linux-raid@vger.kernel.org.
Future releases in the 3.2 series will only be made if bugfixes are needed.
The next release to add features is expected to be 3.3.
NeilBrown 17th June 2011

24
ANNOUNCE-3.2.3
View file

@ -1,24 +0,0 @@
Subject: ANNOUNCE: mdadm 3.2.3 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.2.3
It is available at the usual places:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://neil.brown.name/mdadm
http://neil.brown.name/git/mdadm
This release is largely a bugfix release for the 3.2 series with many
minor fixes with little or no impact.
The largest single area of change is support for reshape of Intel
IMSM arrays (OnLine Capacity Explansion and Level Migtration).
Among other fixes, this now has a better chance of surviving if a
device fails during reshape.
Upgrading is recommended - particularly if you use mdadm for IMSM
arrays - but not essential.
NeilBrown 23rd December 2011

144
ANNOUNCE-3.2.4
View file

@ -1,144 +0,0 @@
Subject: ANNOUNCE: mdadm 3.2.4 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.2.4
It is available at the usual places, now including github:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://neil.brown.name/git/mdadm
This release is largely a bugfix release for the 3.2 series with many
minor fixes with little or no impact.
"--oneline" log of changes is below. Some notable ones are:
- --offroot argument to improve interactions between mdmon and initrd
- --prefer argument to select which /dev names to display in some
circumstances.
- relax restructions on when "--add" will be allowed
- Fix bug with adding write-intent-bitmap to active array
- Now defaults to "/run/mdadm" for storing run-time files.
Upgrading is encouraged.
The next mdadm release is expected to be 3.3 with a number of new
features.
NeilBrown 9th May 2012
77b3ac8 monitor: make return from read_and_act more symbolic.
68226a8 monitor: ensure we retry soon when 'remove' fails.
8453f8d fix: Monitor sometimes crashes
90fa1a2 Work around gcc-4.7's strict aliasing checks
0c4304c fix: container creation with --incremental used.
5d1c7cd FIX: External metadata sometimes is not updated
3c20f98 FIX: mdmon check in reshape_container() can cause a problem
59ab9f5 FIX: Typo error in fprint command
9587c37 imsm: load_super_imsm_all function refactoring
ec50f7b imsm: load_imsm_super_all supports loading metadata from the device list
ca9de18 imsm: validate the number of imsm volumes per controller
30602f5 imsm: display fd in error trace when when store_imsm_mpb failes
eb155f6 mdmon: Use getopt_long() to parse command line options
08ca2ad Add --offroot argument to mdadm
da82751 Add --offroot argument to mdmon
a0963a8 Spawn mdmon with --offroot if mdadm was launched with --offroot
f878b24 imsm: fix, the second array need to have the whole available space on devices
d597705 getinfo_super1: Use MaxSector in place of sb->size
6ef8905 super1: make aread/awrite always use an aligned buffer.
de5a472 Remove avail_disks arg from 'enough'.
da8fe5a Assemble: fix --force assemble during reshape.
b10c663 config: fix handing of 'homehost' in AUTO line.
92d49ec FIX: NULL pointer to strdup() can be passed
d2bde6d imsm: FIX: No new missing disks are allowed during general migration
111e9fd FIX: Array is not run when expansion disks are added
bf5cf7c imsm: FIX: imsm_get_allowed_degradation() doesn't count degradation for raid1
50927b1 Fix: Sometimes mdmon throws core dump during reshape
78340e2 Flush mdmon before next reshape step during container operation
e174219 imsm: FIX: Chunk size migration problem
f93346e FIX: use md position to reshape restart
6a75c8c imsm: FIX: use md position to reshape restart
51d83f5 imsm: FIX: Clear migration record when migration switches to next volume.
e1dd332 FIX: restart reshape when reshape process is stopped just between 2 reshapes
1ca90aa FIX: Do not try to (continue) reshape using inactive array
9f1b0f0 config: conf_match should ignore devname when not set.
d669228 Use posix_memalign() for memory used to write bitmaps
178950e FIX: Changes in '0' case for reshape position verification
9200d41 avoid double-free upon "old buggy kernel" sysfs_read failure
4011421 Print error message if failing to write super for 1.x metadata
0011874 Use MDMON_DIR for pid files created in Monitor.c
56d1885 Assemble: don't use O_EXCL until we have checked device content.
b720636 Assemble: support assembling of a RAID0 being reshaped.
c69ffac Manage: allow --re-add to failed array.
52f07f5 Reset bad flag on map update
911cead super1: support superblocks up to 4K.
ad6db3c Create: reduce the verbosity of 'default_layout'.
b2bfdfa super1.c don't keep recalculating bitmap pointer
4122675 Define and use SUPER1_SIZE for allocations
1afa930 init_super1() memset full buffer allocated for superblock
2de0b8a match_metadata_desc1(): Use calloc instead of malloc+memset
3c0bcd4 Use 4K buffer alignment for superblock allocations
308340a Use struct align_fd to cache fd's block size for aligned reads/writes
65ed615 match_metadata_desc0(): Use calloc instead of malloc+memset
de89706 Generalize ROUND_UP() macro and introduce matching ROUND_UP_PTR()
0a2f189 super1.c: use ROUND_UP/ROUND_UP_PTR
654a381 super-intel.c: Use ROUND_UP() instead of manually coding it
42d5dfd __write_init_super_ddf(): Use posix_memalign() instead of static aligned buffer
d4633e0 Examine: fix array size calculation for RAID10.
e62b778 Assemble: improve verbose logging when including old devices.
0073a6e Remove possible crash during RAID6 -> RAID5 reshape.
69fe207 Incremental: fix adding devices with --incremental
bcbb311 Manage: replace 'return 1' with 'goto abort'.
9f58469 Manage: freeze recovery while adding multiple devices.
ae6c05a Create: round off size for RAID1 arrays.
5ca3a90 Grow: print useful error when converting RAID1->RAID5 will fail.
c07d640 Fix tests/05r1-re-add-nosupper
2d762ad Fix the new ROUND_UP macro.
fd324b0 sysfs: fixed sysfs_freeze_array array to work properly with Manage_subdevs.
5551b11 imsm: avoid overflows for disks over 1TB
97f81ee clear hi bits if not used after loading metadata from disk
e03640b simplify calculating array_blocks
29cd082 show 2TB volumes/disks support in --detail-platform
2cc699a check volume size in validate_geometry_imsm_orom
9126b9a check that no disk over 2TB is used to create container when no support
027c374 imsm: set 2tb disk attribute for spare
3556c2f Fix typo: wan -> want
15632a9 parse_size: distinguish between 0 and error.
fbdef49 Bitmap_offset is a signed number
508a7f1 super1: leave more space in front of data by default.
40110b9 Fix two typos in fprintf messages
342460c mdadm man page: fix typo
0e7f69a imsm: display maximum volumes per controller and array
36fd8cc imsm: FIX: Update function imsm_num_data_members() for Raid1/10
7abc987 imsm: FIX: Add volume size expand support to imsm_analyze_change()
f3871fd imsm: Add new metadata update for volume size expansion
54397ed imsm: Execute size change for external metatdata
016e00f FIX: Support metadata changes rollback
fbf3d20 imsm: FIX: Support metadata changes rollback
44f6f18 FIX: Extend size of raid0 array
7e7e9a4 FIX: Respect metadata size limitations
65a9798 FIX: Detect error and rollback metadata
13bcac9 imsm: Add function imsm_get_free_size()
b130333 imsm: Support setting max size for size change operation
c41e00b imsm: FIX: Component size alignment check
58d26a2 FIX: Size change is possible as standalone change only
4aecb54 FIX: Assembled second array is in read only state during reshape
ae2416e FIX: resolve make everything compilation error
480f356 Raid limit of 1024 when scanning for devices.
c2ecf5f Add --prefer option for --detail and --monitor
0a99975 Relax restrictions on when --add is permitted.
7ce0570 imsm: fix: rebuild does not continue after reboot
b51702b fix: correct extending size of raid0 array
34a1395 Fix sign extension of bitmap_offset in super1.c
012a864 Introduce sysfs_set_num_signed() and use it to set bitmap/offset
5d7b407 imsm: fix: thunderdome may drop 2tb attribute
5ffdc2d Update test for "is udev active".
96fd06e Adjust to new standard of /run
974e039 test: don't worry too much about array size.
b0a658f Grow: failing the set the per-device size is not an error.
36614e9 super-intel.c: Don't try to close negative fd
562aa10 super-intel.c: Fix resource leak from opendir()

31
ANNOUNCE-3.2.5
View file

@ -1,31 +0,0 @@
Subject: ANNOUNCE: mdadm 3.2.5 - A tool for managing Soft RAID under Linux
I am somewhat disappointed to have to announce the availability of
mdadm version 3.2.5
It is available at the usual places, now including github:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://neil.brown.name/git/mdadm
This release primarily fixes a serious regression in 3.2.4.
This regression does *not* cause any risk to data. It simply
means that adding a device with "--add" would sometime fail
when it should not.
The fix also includes a couple of minor fixes such as making
the "--layout=preserve" option to "--grow" work again.
A reminder that the default location for runtime files is now
"/run/mdadm". If you compile this for a distro that does not
have "/run", you will need to compile with an alternate setting for
MAP_DIR. e.g.
make MAP_DIR=/var/run/mdadm
or
make MAP_DIR=/dev/.mdadm
NeilBrown 18th May 2012

57
ANNOUNCE-3.2.6
View file

@ -1,57 +0,0 @@
Subject: ANNOUNCE: mdadm 3.2.6 - A tool for managing Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.2.6
It is available at the usual places, now including github:
countrycode=xx.
http://www.${countrycode}kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://neil.brown.name/git/mdadm
This is a stablity release which adds a number of bugfixs to 3.2.5.
There are no real stand-out fixes, just lots of little bits and pieces.
Below is the "git log --oneline --reverse" list of changes since
3.2.5.
NeilBrown 25th October 2012
b7e05d2 udev-rules: prevent systemd from mount devices before they are ready.
0d478e2 mdadm: Fix Segmentation fault.
42f0ca1 imsm: fix: correct checking volume's degradation
fcf2195 Monitor: fix inconsistencies in values for ->percent
5f862fb Monitor: Report NewArray when an array the disappeared, reappears.
6f51b1c Monitor: fix reporting for Fail vs FailSpare etc.
68ad53b mdmon: fix arg parsing.
517f135 Assemble: don't leak memory with fdlist.
090900c udev-rules: prevent systemd from mount devices before they are ready.
446e000 sha1.h: remove ansidecl.h header inclusion
ec894f5 Manage: zero metadata before adding to 'external' array.
3a84db5 ddf: allow a non-spare to be used to recovery a missing device.
c5d61ca ddf: hack to fix container recognition.
23084aa mdmon: fix arg processing for -a
c4e96a3 mdmon: allow --takeover when original was started with --offroot
80841df find_free_devnum: avoid auto-using names in /etc/mdadm.conf
c5c56d6 mapfile: fix mapfile rebuild for containers
aec89f6 fix segfaults in Detail()
2117ad1 Fix 'enough' function for RAID10.
0bc300d Use --offroot flag when assembling md arrays via --incrmental
ac78f24 Grow: make warning about old metadata more explicit.
14026ab Replace sha1.h with slightly older version.
6f6809f Add zlib license to crc32.c
5267ba0 Handles spaces in array names better.
c51f288 imsm: allow --assume-clean to work.
acf7076 Grow: allow --grow --continue to work for native metadata.
335d2a6 Grow: fix a couple of typos with --assume-clean usage
9ff1427 Fix open_container
3713633 mdadm: super0: do not override uuid with homehost
31bff58 Trivial bugfix and spelling fixes.
e1e539f Detail: don't report a faulty device as 'spare' or 'rebuilding'.
22a6461 super0: allow creation of array on 2TB+ devices.
a5d47a2 Create new md devices consistently
eb48676 Monitor: don't complain about non-monitorable arrays in mdadm.conf
ecdf2d7 Query: don't be confused by partition tables.
f7b75c1 Query: allow member of non-0.90 arrays to be better reported.

63
ANNOUNCE-3.3
View file

@ -1,63 +0,0 @@
Subject: ANNOUNCE: mdadm 3.3 - A tools for managing md Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.3
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://git.neil.brown.name/git/mdadm
This is a major new release so don't be too surprised if there are a
few issues. If I hear about them they will be fixed in 3.3.1.
git log reports nearly 500 changes since 3.2.6 so I won't list them
all.
Some highlights are:
- Some array reshapes can proceed without needing backup file.
This is done by changing the 'data_offset' so we never need to write
any data back over where it was before. If there is no "head space"
or "tail space" to allow data_offset to change, the old mechanism
with a backup file can still be used.
- RAID10 arrays can be reshaped to change the number of devices,
change the chunk size, or change the layout between 'near'
and 'offset'.
This will always change data_offset, and will fail if there is no
room for data_offset to be moved.
- "--assemble --update=metadata" can convert a 0.90 array to a 1.0 array.
- bad-block-logs are supported (but not heavily tested yet)
- "--assemble --update=revert-reshape" can be used to undo a reshape
that has just been started but isn't really wanted. This is very
new and while it passes basic tests it cannot be guaranteed.
- improved locking between --incremental and --assemble
- uses systemd to run "mdmon" if systemd is configured to do that.
- kernel names of md devices can be non-numeric. e.g. "md_home" rather than
"md0". This will probably confuse lots of other tools, so you need to
echo CREATE names=yes >> /etc/mdadm.conf
or the feature will not be used. (you also need a reasonably new kernel).
- "--stop" can be given a kernel name instead of a device name. i.e
mdadm --stop md4
will work even if /dev/md4 doesn't exist.
- "--detail --export" has some information about the devices in the array
- --dump and --restore can be used to backup and restore the metadata on an
array.
- Hot-replace is supported with
mdadm /dev/mdX --replace /dev/foo
and
mdadm /dev/mdX --replace /dev/foo --with /dev/bar
- Config file can be a directory in which case all "*.conf" files are
read in lexical order.
Default is to read /etc/mdadm.conf and then /etc/mdadm.conf.d
Thus
echo CREATE name=yes > /etc/mdadm.conf.d/names.conf
will also enable the use of named md devices.
- Lots of improvements to DDF support including adding support for
RAID10 (thanks Martin Wilck).
and lots of bugfixes and other little changes.
NeilBrown 3rd September 2013

23
ANNOUNCE-3.3.1
View file

@ -1,23 +0,0 @@
Subject: ANNOUNCE: mdadm 3.3.1 - A tool for managing md Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.3.1
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://git.neil.brown.name/git/mdadm.git
The main changes are:
- lots of work on "DDF" support. Hopefully it will be more stable
now. Bug reports are always welcome.
- improved interactions with 'systemd'. Where possible, background
tasks are run from systemd (if it is present) rather then forking
disassociationg from the session. This is important because udev
doesn't really let you disassociate.
though there are a number of other little bug fixes too.
NeilBrown 5th June 2014

16
ANNOUNCE-3.3.2
View file

@ -1,16 +0,0 @@
Subject: ANNOUNCE: mdadm 3.3.2 - A tool for managing md Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.3.2
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://git.neil.brown.name/git/mdadm.git
Changes since 3.3.1 are mostly little bugfixes and some man-page
updates.
NeilBrown 21st August 2014

18
ANNOUNCE-3.3.3
View file

@ -1,18 +0,0 @@
Subject: ANNOUNCE: mdadm 3.3.3 - A tool for managing md Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.3.3
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://git.neil.brown.name/git/mdadm.git
The 100 changes since 3.3.3 are mostly little bugfixes and some improvements
to the selftests.
raid6check now handle all RAID6 layouts including DDF correctly.
See git log for the rest.
NeilBrown 24th July 2015

37
ANNOUNCE-3.3.4
View file

@ -1,37 +0,0 @@
Subject: ANNOUNCE: mdadm 3.3.4 - A tool for managing md Soft RAID under Linux
I am somewhat disappointed to have to announce the availability of
mdadm version 3.3.4
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://git.neil.brown.name/git/mdadm.git
In mdadm-3.3 a change was made to how IMSM (Intel Matrix Storage
Manager) metadata was handled. Previously an IMSM array would only
be assembled if it was attached to an IMSM controller.
In 3.3 this was relaxed as there are circumstances where the
controller is not properly detected. Unfortunately this has negative
consequences which have only just come to light.
If you have an IMSM RAID1 configured and then disable RAID in the
BIOS, the metadata will remain on the devices. If you then install
some other OS on one device and then install Linux on the other, Linux
might eventually start noticing the IMSM metadata (depending a bit on whether
mdadm is included in the initramfs) and might start up the RAID1. This could
copy one device over the other, thus trashing one of the installations.
Not good.
So with this release IMSM arrays will only be assembled if attached to
an IMSM controller, or if "--force" is given to --assemble, or if the
environment variable IMSM_NO_PLATFORM is set (used primarily for
testing).
I strongly recommend upgrading to 3.3.4 if you are using 3.3 or later.
NeilBrown 3rd August 2015.

24
ANNOUNCE-3.4
View file

@ -1,24 +0,0 @@
Subject: ANNOUNCE: mdadm 3.4 - A tool for managing md Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 3.4
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://github.com/neilbrown/mdadm
git://neil.brown.name/mdadm
http://git.neil.brown.name/git/mdadm
The new second-level version number reflects significant new
functionality, particular support for journalled RAID5/6 and clustered
RAID1. This new support is probably still buggy. Please report bugs.
There are also a number of fixes for Intel's IMSM metadata support,
and an assortment of minor bug fixes.
I plan for this to be the last release of mdadm that I provide as I am
retiring from MD and mdadm maintenance. Jes Sorensen has volunteered
to oversee mdadm for the next while. Thanks Jes!
NeilBrown 28th January 2016

22
ANNOUNCE-4.0
View file

@ -1,22 +0,0 @@
Subject: ANNOUNCE: mdadm 4.0 - A tool for managing md Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 4.0
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://git.kernel.org/pub/scm/utils/mdadm/mdadm.git
http://git.kernel.org/cgit/utils/mdadm/
The update in major version number primarily indicates this is a
release by it's new maintainer. In addition it contains a large number
of fixes in particular for IMSM RAID and clustered RAID support. In
addition this release includes support for IMSM 4k sector drives,
failfast and better documentation for journaled RAID.
This is my first release of mdadm. Please thank Neil Brown for his
previous work as maintainer and blame me for all the bugs I caused
since taking over.
Jes Sorensen, 2017-01-09

16
ANNOUNCE-4.1
View file

@ -1,16 +0,0 @@
Subject: ANNOUNCE: mdadm 4.1 - A tool for managing md Soft RAID under Linux
I am pleased to announce the availability of
mdadm version 4.1
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://git.kernel.org/pub/scm/utils/mdadm/mdadm.git
http://git.kernel.org/cgit/utils/mdadm/
The update constitutes more than one year of enhancements and bug fixes
including for IMSM RAID, Partial Parity Log, clustered RAID support,
improved testing, and gcc-8 support.
Jes Sorensen, 2018-10-01

19
ANNOUNCE-4.2
View file

@ -1,19 +0,0 @@
Subject: ANNOUNCE: mdadm 4.2 - A tool for managing md Soft RAID under Linux
I am pleased to finally announce the availability of mdadm-4.2.
get 4.2 out the door soon.
It is available at the usual places:
http://www.kernel.org/pub/linux/utils/raid/mdadm/
and via git at
git://git.kernel.org/pub/scm/utils/mdadm/mdadm.git
http://git.kernel.org/cgit/utils/mdadm/
The release includes more than two years of development and bugfixes,
so it is difficult to remember everything. Highlights include
enhancements and bug fixes including for IMSM RAID, Partial Parity
Log, clustered RAID support, improved testing, and gcc-9 support.
Thank you everyone who contributed to this release!
Jes Sorensen, 2021-12-30

2
Assemble.c
View file

@ -1988,7 +1988,7 @@ int assemble_container_content(struct supertype *st, int mdfd,
* and ignoring special character on the first place.
*/
if (strcmp(sra->text_version + 1, content->text_version + 1) != 0) {
if (sysfs_set_array(content, 9003) != 0) {
if (sysfs_set_array(content) != 0) {
sysfs_free(sra);
return 1;
}

6
Build.c
View file

@ -156,12 +156,6 @@ int Build(struct mddev_ident *ident, struct mddev_dev *devlist, struct shape *s,
bitmap_fd = open(s->bitmap_file, O_RDWR);
if (bitmap_fd < 0) {
int major = BITMAP_MAJOR_HI;
#if 0
if (s->bitmap_chunk == UnSet) {
pr_err("%s cannot be opened.\n", s->bitmap_file);
goto abort;
}
#endif
bitmapsize = s->size >> 9; /* FIXME wrong for RAID10 */
if (CreateBitmap(s->bitmap_file, 1, NULL,
s->bitmap_chunk, c->delay,

368
CHANGELOG.md Normal file
View file

@ -0,0 +1,368 @@
# Release [mdadm-4.3](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-4.3)
Features:
- **IMSM_NO_PLATFORM** boot parameter support from Neil Brown.
- **--write-zeros** option support by Logan Gunthorpe.
- **IMSM** monetization by VMD register from Mateusz Grzonka.
- RST SATA under VMD support from Kevin Friedberg.
- Strong name rules from Mariusz Tkaczyk.
Fixes:
- Unify failed raid behavior from Coly Li.
- Rework of **--update** options from Mateusz Kusiak.
- **mdmon-initrd** service from Neil Brown.
- **IMSM** expand functionality rework from Mariusz Tkaczyk.
- Mdmonitor improvements from Mateusz Grzonka.
- Failed state verification from Mateusz Kusiak and Kinga Tanska.
# Release [mdadm-4.2](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-4.2)
The release includes more than two years of development and bugfixes, so it is difficult to
remember everything. Highlights include enhancements and bug fixes including for **IMSM** RAID,
Partial Parity Log, clustered RAID support, improved testing, and gcc-9 support.
# Release [mdadm-4.1](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-4.1)
The update constitutes more than one year of enhancements and bug fixes including for **IMSM**
RAID, Partial Parity Log, clustered RAID support, improved testing, and gcc-8 support.
# Release [mdadm-4.0](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-4.0)
The update in major version number primarily indicates this is a release by it's new maintainer.
In addition it contains a large number of fixes in particular for IMSM RAID and clustered RAID
support. In addition, this release includes support for IMSM 4k sector drives, failfast and better
documentation for journaled RAID.
This is my first release of mdadm. Please thank Neil Brown for his previous work as maintainer and
blame me for all the bugs I caused since taking over.
# Release [mdadm-3.4](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.4)
- Support for journalled RAID5/6 and clustered RAID1. This new support is probably still buggy.
Please report bugs.
- There are also a number of fixes for **IMSM** support and an assortment of minor bug fixes.
- I plan for this to be the last release of mdadm that I provide as I am retiring from MD and mdadm
maintenance. Jes Sorensen has volunteered to oversee mdadm for the next while. Thanks Jes!
# Release [mdadm-3.3.4](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.3.4)
**I strongly recommend upgrading to 3.3.4 if you are using 3.3 or later with IMSM.**
- **IMSM** metadata assemble fixes.
In mdadm-3.3 a change was made to how **IMSM** metadata was handled. Previously an **IMSM** array
would only be assembled if it was attached to an **IMSM** controller. In 3.3 this was relaxed as
there are circumstances where the controller is not properly detected. Unfortunately, this has
negative consequences which have only just come to light.
If you have an IMSM RAID1 configured and then disable RAID in the BIOS, the metadata will remain
on the devices. If you then install some other OS on one device and then install Linux on the
other, Linux might eventually start noticing the IMSM metadata (depending a bit on whether
mdadm is included in the initramfs) and might start up the RAID1. This could copy one device over
the other, thus trashing one of the installations.
So, with this release IMSM arrays will only be assembled if attached to an **IMSM** controller,
or if **--force** is given to **--assemble**, or if the environment variable
**IMSM_NO_PLATFORM=1** is set (used primarily for testing).
# Release [mdadm-3.3.3](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.3.3)
- The 100 changes since 3.3.3 are mostly little bugfixes and some improvements to the self-tests.
- raid6check now handle all RAID6 layouts including **DDF** correctly. See git log for the rest.
# Release [mdadm-3.3.2](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.3.2)
- Little bugfixes and some man-page updates.
# Release [mdadm-3.3.1](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.3.1)
- lots of work on **DDF** support.
- Improved interactions with **systemd**. Where possible, background tasks are run from systemd
rather than forking.
- Number of other little bug fixes too.
# Release [mdadm-3.3](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.3)
- Some array reshapes can proceed without needing backup file. This is done by changing the
data_offset* so we never need to write any data back over where it was before. If there is no
'head space' or 'tail space' to allow *data_offset* to change, the old mechanism with a backup
file can still be used.
- RAID10 arrays can be reshaped to change the number of devices, change the chunk size, or change
the layout between *near* and *offset*.
This will always change *data_offset*, and will fail if there is no room for *data_offset* to be
moved.
- **--assemble --update=metadata** can convert a **0.90** array to a **1.0** array.
- **bad-block-logs** are supported (but not heavily tested yet).
- **--assemble --update=revert-reshape** can be used to undo a reshape that has just been started
but isn't really wanted. This is very new and while it passes basic tests it cannot be
guaranteed.
- improved locking between **--incremental** and **--assemble**.
- uses systemd to run **mdmon** if systemd is configured to do that.
- kernel names of md devices can be non-numeric. e.g. "md_home" rather than
"md0". This will probably confuse lots of other tools, so you need to
**echo CREATE names=yes >> /etc/mdadm.conf** or the feature will not be used (you also need a
reasonably new kernel).
- **--stop** can be given a kernel name instead of a device name. i.e. **mdadm --stop md4** will
work even if /dev/md4 doesn't exist.
- **--detail --export** has some information about the devices in the array.
- **--dump** and **--restore** can be used to backup and restore the metadata on an array.
- Hot-replace is supported with **mdadm /dev/mdX --replace /dev/foo** and
**mdadm /dev/mdX --replace /dev/foo --with /dev/bar**.
- Config file can be a directory in which case all "*.conf" files are read in lexical order.
Default is to read **/etc/mdadm.conf** and then **/etc/mdadm.conf.d**. Thus
**echo CREATE name=yes > /etc/mdadm.conf.d/names.conf** will also enable the use of named md
devices.
- Lots of improvements to **DDF** support including adding support for RAID10 (thanks Martin Wilck).
# Release [mdadm-3.2.6](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.2.6)
- There are no real stand-out fixes, just lots of little bits and pieces.
# Release [mdadm-3.2.5](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.2.5)
- This release primarily fixes a serious regression in 3.2.4. This regression does *not* cause
any risk to data. It simply means that adding a device with **--add** would sometime fail
when it should not.
- The fix also includes a couple of minor fixes such as making the **--layout=preserve** option to
**--grow** work again.
# Release [mdadm-3.2.4](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.2.4)
- **--offroot** argument to improve interactions between mdmon and initrd.
- **--prefer** argument to select which */dev* names to display in some circumstances.
- relax restrictions on when **--add** will be allowed.
- Fix bug with adding write-intent-bitmap to active array.
- Now defaults to */run/mdadm* for storing run-time files.
# Release [mdadm-3.2.3](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.2.3)
- The largest single area of change is support for reshape of Intel IMSM arrays (OnLine Capacity
Expansion and Level Migration).
- Among other fixes, this now has a better chance of surviving if a device fails during reshape.
# Release [mdadm-3.2.2](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.2.2)
- reshaping IMSM (Intel metadata) arrays is no longer 'experimental', it should work properly and be
largely compatible with IMSM drivers in other platforms.
- **--assume-clean** can be used with **--grow --size** to avoid resyncing the new part of the
array. This is only support with very new kernels.
- RAID0 arrays can have chunksize which is not a power of 2. This has been supported in the kernel
for a while but is only now supported by mdadm.
- A new tool **raid6check** is available, which can check a RAID6 array, or part of it and report
which device is most inconsistent with the others if any stripe is inconsistent. This is still
under development and does not have a man page yet. If anyone tries it out and has any questions
or experience to report, they would be most welcome on linux-raid@vger.kernel.org.
# Release [mdadm-3.2.1](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.2.1)
- Policy framework
Policy can be expressed for moving spare devices between arrays, and for how to handle hot-plugged
devices. This policy can be different for devices plugged in to different controllers etc. This,
for example, allows a configuration where when a device is plugged in it is immediately included
in an md array as a hot spare and possibly starts recovery immediately if an array is degraded.
- Some understanding of mbr and gpt paritition tables. This is primarily to support the new
hot-plug support. If a device is plugged in and policy suggests it should have a partition table,
the partition table will be copied from a suitably similar device, and then the partitions will
hot-plug and can then be added to md arrays.
- **--incremental --remove** can remember where a device was removed from so if a device gets
plugged back in the same place, special policy applies to it, allowing it to be included in an
array even if a general hotplug will not be included.
- Enhanced reshape options, including growing a RAID0 by converting to RAID4, restriping, and
converting back. Also convertions between RAID0 and RAID10 and between RAID1 and RAID10 are
possible (with a suitably recent kernel).
- Spare migration for IMSM arrays. Spare migration can now work across 'containers' using
non-native metadata and specifically Intel's IMSM arrays support spare migrations.
- OLCE and level migration for Intel IMSM arrays. OnLine Capacity Expansion and level migration
(e.g. RAID0 -> RAID5) is supported for Intel Matrix Storage Manager arrays. This support is
currently *experimental* for technical reasons. It can be enabled with
**export MDADM_EXPERIMENTAL=1**.
- avoid including wayward devices.
If you split a RAID1, mount the two halves as two separate degraded RAID1s, and then later bring
the two back together, it is possible that the md metadata won't properly show that one must
over-ride the other. Mdadm now does extra checking to detect this possibility and avoid
potentially corrupting data.
- Remove any possible confusion between similar options. e.g. **--brief** and **--bitmap** were
mapped to 'b' and mdadm wouldn't notice if one was used where the other was expected.
- Allow K,M,G suffixes on chunk sizes.
# Release [mdadm-3.2](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.2)
- By far the most significant change in this release related to the management of reshaping arrays.
This code has been substantially re-written so that it can work with **externally managed
metadata** -Intel's IMSM in particular. We now support level migration and OnLine Capacity
Expansion on these arrays.
- Various policy statements can be made in the *mdadm.conf* to guide the behavior of mdadm,
particular with regards to how new devices are treated by **--incremental**. Depending on the
*action* associated with a device (identified by its *path*) such need devices can be
automatically re-added to and existing array that they previously fell out off, or automatically
added as a spare if they appear to contain no data.
- mdadm now has a limited understanding of partition tables. This allows the policy framework to
make decisions about partitioned devices as well.
- **--incremental --remove** can be told what **--path** the device was on, and this info will be
recorded so that another device appearing at the same physical location can be preferentially
added to the same array (provides the spare-same-slot action policy applied to the path).
- A new flags **--invalid-backup** flag is available in **--assemble** mode. This can be used to
re-assemble an array which was stopping in the middle of a reshape, and for which the
*backup file* is no longer available or is corrupted. The array may have some corruption in it
at the point where reshape was up to, but at least the rest of the array will become available.
- Policy framework.
- Various internal restructuring - more is needed.
# Release [mdadm-3.1.5](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.1.5)
- Fixes for **v1.x** metadata on big-endian machines.
- man page improvements.
- Improve **--detail --export** when run on partitions of an md array.
- Fix regression with removing *failed* or *detached* devices.
- Fixes for **--assemble --force** in various unusual cases.
- Allow **-Y** to mean **--export**. This was documented but not implemented.
- Various fixes for handling **ddf** metadata. This is now more reliable but could benefit from
more interoperability testing.
- Correctly list subarrays of a container in **--detail** output.
- Improve checks on whether the requested number of devices is supported by the metadata, both for
**--create** and **--grow**.
- Don't remove partitions from a device that is being included in an array until we are fully
committed to including it.
- Allow **--assemble --update=no-bitmap** so an array with a corrupt bitmap can still be assembled.
- Don't allow **--add** to succeed if it looks like a **--re-add** is probably wanted, but cannot
succeed. This avoids inadvertently turning devices into spares when an array is failed.
# Release [mdadm-3.1.4](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.1.4)
Two fixes related to configs that aren't using udev:
- Don't remove md devices which 'standard' names on **--stop**.
- Allow dev_open to work on read-only */dev*.
And fixed regressions:
- Allow **--incremental** to add spares to an array.
- Accept **--no-degraded** as a deprecated option rather than throwing an error.
- Return correct success status when **--incremental** assembling a container which does not yet
have enough devices.
- Don't link mdadm with pthreads, only mdmon needs it.
- Fix compiler warning due to bad use of snprintf.
# Release [mdadm-3.1.3](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.1.3)
- mapfile now lives in a fixed location which default to */dev/.mdadm/map*, but can be changed at
compile time. This location is chosen and most distros provide it during early boot and preserve
it through. As long a */dev* exists and is writable, */dev/.mdadm* will be created. Other files
communication with mdmon live here too. This fixes a bug reported by Debian and Gentoo users where
udev would spin in early-boot.
- IMSM and DDF metadata will not be recognized on partitions as they should only be used on
whole-disks.
- Various overflows causes by 2G drives have been addressed.
- A subarray of an IMSM contain can now be killed with **--kill-subarray**. Also, subarrays can be
renamed with **--update-subarray --update=name**.
- **-If** (or **--incremental --fail**) can be used from udev to fail and remove from all arrays
a device which has been unplugged from the system i.e. hot-unplug-support.
- **/dev/mdX --re-add missing** will look for any device that looks like it should be a member of
*/dev/mdX* but isn't and will automatically **--re-add** it.
- Now compile with *-Wextra* to get extra warnings.
- Lots of minor bug fixes, documentation improvements, etc.
# Release [mdadm-3.1.2](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.1.2)
- The default metadata has change again (sorry about that). It is now **v1.2** and will hopefully
stay that way. It turned out there with boot-block issues with **v1.1** which make it unsuitable
for a default, though in many cases it is still suitable to use.
- Add *homehost* to the valid words for the **AUTO** config file line. When followed by *-all*,
this causes mdadm to auto-assemble any array belonging to this host, but not auto-assemble
anything else.
- VAR_RUN can be easily changed at compile time just like ALT_RUN. This gives distros more
flexibility in how to manage the pid and sock files that mdmon needs.
- If mdadm.conf lists arrays which have inter-dependencies, the previously had to be listed in the
"right" order. Now, any order should work.
- Fix some bugs with **--grow --chunksize=**.
- Stopping a container is not permitted when members are still active.
- Various mdmon fixes.
- Alway make bitmap 4K-aligned if at all possible.
- Fix **--force** assembly of **v1.x** arrays which are in the process of recovering.
- Add section on 'scrubbing' to 'md' man page.
- Various command-line-option parsing improvements.
- ... and lots of other bug fixes.
# Release [mdadm-3.1.1](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.1.1)
- Multiple fixes for new **--grow** levels including fixes for serious data corruption
problems.
- Change default metadata to **v1.1**.
- Change default chunk size to 512K.
- Change default bitmap chunk size to 64MB.
- When **--re-add** is used, don't fall back to **--add** as this can destroy data.
# Release [mdadm-3.1](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.1)
- Support **--grow** to change the layout of RAID 4/5/6.
- Support **--grow** to change the chunk size of RAID 4/5/6.
- Support **--grow** to change level from RAID1 -> RAID5 -> RAID6 and back.
- Support **--grow** to reduce the number of devices in RAID 4/5/6.
- Support restart of these grow options which assembling an array which is partially grown.
- Assorted tests of this code, and of different RAID6 layouts.
# Release [mdadm-3.0.3](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.0.3)
- Improvements for creating arrays giving just a name, like *foo*, rather than the full
*/dev/md/foo*.
- Improvements for assembling member arrays of containers.
- Improvements to test suite.
- Add option to change increment for *RebuildNN* messages reported by **--monitor**.
- Improvements to **mdmon** hand-over from initrd to final root.
- Handle merging of devices that have left an IMSM array and are being re-incorporated.
- Add missing space in **--detail --brief** output.
# Release [mdadm-3.0.2](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.0.2)
- Fix crash when **homehost** is not set, as often happens in early boot.
# Release [mdadm-3.0.1](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.0.1)
- Fix various segfaults.
- Fixed for **--examine** with containers.
- Lots of other little fixes.
# Release [mdadm-3.0](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/log/?h=mdadm-3.0)
- Support for **externally managed metadata**, specifically DDF and IMSM.
- Depend on udev to create entries in */dev*, rather than creating them ourselves.
- Remove **--auto-update-home-hosts**.
- New config file line **auto**.
- New *ignore* and *any* options for **homehost**.
- Numerous bug fixes and minor enhancements.

306
ChangeLog
View file

@ -1,306 +0,0 @@
Please see git logs for detailed change log.
This file just contains highlight.
Changes Prior to release 3.3
- Some array reshapes can proceed without needing backup file.
This is done by changing the 'data_offset' so we never need to write
any data back over where it was before. If there is no "head space"
or "tail space" to allow data_offset to change, the old mechanism
with a backup file can still be used.
- RAID10 arrays can be reshaped to change the number of devices,
change the chunk size, or change the layout between 'near'
and 'offset'.
This will always change data_offset, and will fail if there is no
room for data_offset to be moved.
- "--assemble --update=metadata" can convert a 0.90 array to a 1.0 array.
- bad-block-logs are supported (but not heavily tested yet)
- "--assemble --update=revert-reshape" can be used to undo a reshape
that has just been started but isn't really wanted. This is very
new and while it passes basic tests it cannot be guaranteed.
- improved locking between --incremental and --assemble
- uses systemd to run "mdmon" if systemd is configured to do that.
- kernel names of md devices can be non-numeric. e.g. "md_home" rather than
"md0". This will probably confuse lots of other tools, so you need to
echo CREATE names=yes >> /etc/mdadm.conf
or the feature will not be used. (you also need a reasonably new kernel).
- "--stop" can be given a kernel name instead of a device name. i.e
mdadm --stop md4
will work even if /dev/md4 doesn't exist.
- "--detail --export" has some information about the devices in the array
- --dump and --restore can be used to backup and restore the metadata on an
array.
- Hot-replace is supported with
mdadm /dev/mdX --replace /dev/foo
and
mdadm /dev/mdX --replace /dev/foo --with /dev/bar
- Config file can be a directory in which case all "*.conf" files are
read in lexical order.
Default is to read /etc/mdadm.conf and then /etc/mdadm.conf.d
Thus
echo CREATE name=yes > /etc/mdadm.conf.d/names.conf
will also enable the use of named md devices.
- Lots of improvements to DDF support including adding support for
RAID10 (thanks Martin Wilck).
Changes Prior to release 3.2.6
- There are no real stand-out fixes, just lots of little bits and pieces.
Changes Prior to release 3.2.5
- This release primarily fixes a serious regression in 3.2.4.
This regression does *not* cause any risk to data. It simply
means that adding a device with "--add" would sometime fail
when it should not.
- The fix also includes a couple of minor fixes such as making
the "--layout=preserve" option to "--grow" work again.
Changes Prior to release 3.2.4
"--oneline" log of changes is below. Some notable ones are:
- --offroot argument to improve interactions between mdmon and initrd
- --prefer argument to select which /dev names to display in some
circumstances.
- relax restructions on when "--add" will be allowed
- Fix bug with adding write-intent-bitmap to active array
- Now defaults to "/run/mdadm" for storing run-time files.
Changes Prior to release 3.2.3
- The largest single area of change is support for reshape of Intel
IMSM arrays (OnLine Capacity Explansion and Level Migration).
- Among other fixes, this now has a better chance of surviving if a
device fails during reshape.
Changes Prior to release 3.2.2
- reshaping IMSM (Intel metadata) arrays is no longer 'experimental',
it should work properly and be largely compatible with IMSM drivers in
other platforms.
- --assume-clean can be used with --grow --size to avoid resyncing the
new part of the array. This is only support with very new kernels.
- RAID0 arrays can have chunksize which is not a power of 2. This has been
supported in the kernel for a while but is only now supprted by
mdadm.
- A new tool 'raid6check' is available which can check a RAID6 array,
or part of it, and report which device is most inconsistent with the
others if any stripe is inconsistent. This is still under development
and does not have a man page yet. If anyone tries it out and has any
questions or experience to report, they would be most welcome on
linux-raid@vger.kernel.org.
Changes Prior to release 3.2.1
- policy framework
Policy can be expressed for moving spare devices between arrays, and
for how to handle hot-plugged devices. This policy can be different
for devices plugged in to different controllers etc.
This, for example, allows a configuration where when a device is plugged
in it is immediately included in an md array as a hot spare and
possibly starts recovery immediately if an array is degraded.
- some understanding of mbr and gpt paritition tables
This is primarly to support the new hot-plug support. If a
device is plugged in and policy suggests it should have a partition table,
the partition table will be copied from a suitably similar device, and
then the partitions will hot-plug and can then be added to md arrays.
- "--incremental --remove" can remember where a device was removed from
so if a device gets plugged back in the same place, special policy applies
to it, allowing it to be included in an array even if a general hotplug
will not be included.
- enhanced reshape options, including growing a RAID0 by converting to RAID4,
restriping, and converting back. Also convertions between RAID0 and
RAID10 and between RAID1 and RAID10 are possible (with a suitably recent
kernel).
- spare migration for IMSM arrays.
Spare migration can now work across 'containers' using non-native metadata
and specifically Intel's IMSM arrays support spare migrations.
- OLCE and level migration for Intel IMSM arrays.
OnLine Capacity Expansion and level migration (e.g. RAID0 -> RAID5) is
supported for Intel Matrix Storage Manager arrays.
This support is currently 'experimental' for technical reasons. It can
be enabled with "export MDADM_EXPERIMENTAL=1"
- avoid including wayward devices
If you split a RAID1, mount the two halves as two separate degraded RAID1s,
and then later bring the two back together, it is possible that the md
metadata won't properly show that one must over-ride the other.
mdadm now does extra checking to detect this possibilty and avoid
potentially corrupting data.
- remove any possible confusion between similar options.
e.g. --brief and --bitmap were mapped to 'b' and mdadm wouldn't
notice if one was used where the other was expected.
- allow K,M,G suffixes on chunk sizes
Changes Prior to release 3.2
- By far the most significant change in this release related to the
management of reshaping arrays. This code has been substantially
re-written so that it can work with 'externally managed metadata' -
Intel's IMSM in particular. We now support level migration and
OnLine Capacity Expansion on these arrays.
- Policy framework.
Various policy statements can be made in the mdadm.conf to guide
the behaviour of mdadm, particular with regards to how new devices
are treated by "mdadm -I".
Depending on the 'action' associated with a device (identified by
its 'path') such need devices can be automatically re-added to and
existing array that they previously fell out off, or automatically
added as a spare if they appear to contain no data.
- mdadm now has a limited understanding of partition tables. This
allows the policy framework to make decisions about partitioned
devices as well.
- --incremental --remove can be told what --path the device was on,
and this info will be recorded so that another device appearing at
the same physical location can be preferentially added to the same
array (provides the spare-same-slot action policy applied to the
path).
- A new flags "--invalid-backup" flag is available in --assemble
mode. This can be used to re-assemble an array which was stopping
in the middle of a reshape, and for which the 'backup file' is no
longer available or is corrupted. The array may have some
corruption in it at the point where reshape was up to, but at least
the rest of the array will become available.
- Various internal restructuring - more is needed.
Changes Prior to release 3.1.5
- Fixes for v1.x metadata on big-endian machines.
- man page improvements
- Improve '--detail --export' when run on partitions of an md array.
- Fix regression with removing 'failed' or 'detached' devices.
- Fixes for "--assemble --force" in various unusual cases.
- Allow '-Y' to mean --export. This was documented but not implemented.
- Various fixed for handling 'ddf' metadata. This is now more reliable
but could benefit from more interoperability testing.
- Correctly list subarrays of a container in "--detail" output.
- Improve checks on whether the requested number of devices is supported
by the metadata - both for --create and --grow.
- Don't remove partitions from a device that is being included in an
array until we are fully committed to including it.
- Allow "--assemble --update=no-bitmap" so an array with a corrupt
bitmap can still be assembled.
- Don't allow --add to succeed if it looks like a "--re-add" is probably
wanted, but cannot succeed. This avoids inadvertently turning
devices into spares when an array is failed.
Changes Prior to release 3.1.4
Two fixes related to configs that aren't using udev:
- Don't remove md devices which 'standard' names on --stop
- Allow dev_open to work on read-only /dev
And fixed regressions:
- Allow --incremental to add spares to an array
- Accept --no-degraded as a deprecated option rather than
throwing an error
- Return correct success status when --incrmental assembling
a container which does not yet have enough devices.
- Don't link mdadm with pthreads, only mdmon needs it.
- Fix compiler warning due to bad use of snprintf
Changes Prior to release 3.1.3
- mapfile now lives in a fixed location which default to
/dev/.mdadm/map but can be changed at compile time. This
location is choses and most distros provide it during early
boot and preserve it through. As long a /dev exists and is
writable, /dev/.mdadm will be created.
Other files file communication with mdmon live here too.
This fixes a bug reported by Debian and Gentoo users where
udev would spin in early-boot.
- IMSM and DDF metadata will not be recognised on partitions
as they should only be used on whole-disks.
- Various overflows causes by 2G drives have been addressed.
- A subarray of an IMSM contain can now be killed with
--kill-subarray. Also subarrays can be renamed with
--update-subarray
- -If (or --incremental --fail) can be used from udev to
fail and remove from all arrays a device which has been
unplugged from the system. i.e. hot-unplug-support.
- "mdadm /dev/mdX --re-add missing" will look for any device
that looks like it should be a member of /dev/mdX but isn't
and will automatically --re-add it
- Now compile with -Wextra to get extra warnings.
- Lots of minor bug fixes, documentation improvements, etcc
Changes Prior to release 3.1.2
- The default metadata has change again (sorry about that).
It is now v1.2 and will hopefully stay that way. It turned
out there with boot-block issues with v1.1 which make it
unsuitable for a default, though in many cases it is still
suitable to use.
- Stopping a container is not permitted when members are still
active
- Add 'homehost' to the valid words for the "AUTO" config file
line. When followed by "-all", this causes mdadm to
auto-assemble any array belonging to this host, but not
auto-assemble anything else.
- Fix some bugs with "--grow --chunksize=" for changing chunksize.
- VAR_RUN can be easily changed at compile time just like ALT_RUN.
This gives distros more flexability in how to manage the
pid and sock files that mdmon needs.
- Various mdmon fixes
- Alway make bitmap 4K-aligned if at all possible.
- If mdadm.conf lists arrays which have inter-dependencies,
the previously had to be listed in the "right" order. Now
any order should work.
- Fix --force assembly of v1.x arrays which are in the process
of recovering.
- Add section on 'scrubbing' to 'md' man page.
- Various command-line-option parsing improvements.
- ... and lots of other bug fixes.
Changes Prior to release 3.1.1
- Multiple fixes for new --grow levels including fixes for
serious data corruption problems.
- Change default metadata to v1.1
- Change default chunk size to 512K
- Change default bitmap chunk size to 64Meg
- When --re-add is used, don't fall back to
--add if --re-add fails as this can destroy data.
Changes Prior to release 3.1
- Support --grow to change the layout of RAID4/5/6
- Support --grow to change the chunksize of raid 4/5/6
- Support --grow to change level from RAID1 -> RAID5 -> RAID6 and
back.
- Support --grow to reduce the number of devices in RAID4/5/6.
- Support restart of these grow options which assembling an array
which is partially grown.
- Assorted tests of this code, and of different RAID6 layouts.
Changes Prior to release 3.0.3
- Improvements for creating arrays giving just a name, like 'foo',
rather than the full '/dev/md/foo'.
- Improvements for assembling member arrays of containers.
- Improvements to test suite
- Add option to change increment for RebuildNN messages reported
by "mdadm --monitor"
- Improvements to mdmon 'hand-over' from initrd to final root.
- Handle merging of devices that have left an IMSM array and are
being re-incorporated.
- Add missing space in "--detail --brief" output.
Changes Prior to release 3.0.2
- Fix crash when hosthost is not set, as often happens in
early boot.
Changes Prior to release 3.0.1
- Fix various segfaults
- Fixed for --examine with containers
- Lots of other little fixes.
Changes Prior to release 3.0
- Support for externally managed metadata, specifically DDF and IMSM.
- Depend on udev to create entries in /dev, rather than creating them
ourselves.
- remove --auto-update-home-hosts
- new config file line "auto"
- new "<ignore>" and "any" options for "homehost"
- numerous bug fixes and minor enhancements.

58
Create.c
View file

@ -32,6 +32,10 @@
#include <sys/signalfd.h>
#include <sys/wait.h>
#ifndef FALLOC_FL_ZERO_RANGE
#define FALLOC_FL_ZERO_RANGE 16
#endif
static int round_size_and_verify(unsigned long long *size, int chunk)
{
if (*size == 0)
@ -279,8 +283,10 @@ static int add_disk_to_super(int mdfd, struct shape *s, struct context *c,
dv->devname);
return 1;
}
if (!fstat_is_blkdev(fd, dv->devname, &rdev))
if (!fstat_is_blkdev(fd, dv->devname, &rdev)) {
close(fd);
return 1;
}
info->disk.major = major(rdev);
info->disk.minor = minor(rdev);
}
@ -289,6 +295,7 @@ static int add_disk_to_super(int mdfd, struct shape *s, struct context *c,
if (st->ss->add_to_super(st, &info->disk, fd, dv->devname,
dv->data_offset)) {
ioctl(mdfd, STOP_ARRAY, NULL);
close(fd);
return 1;
}
st->ss->getinfo_super(st, info, NULL);
@ -297,6 +304,7 @@ static int add_disk_to_super(int mdfd, struct shape *s, struct context *c,
*zero_pid = write_zeroes_fork(fd, s, st, dv);
if (*zero_pid <= 0) {
ioctl(mdfd, STOP_ARRAY, NULL);
close(fd);
return 1;
}
}
@ -493,6 +501,7 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
*/
int mdfd;
unsigned long long minsize = 0, maxsize = 0;
dev_policy_t *custom_pols = NULL;
char *mindisc = NULL;
char *maxdisc = NULL;
char *name = ident->name;
@ -584,6 +593,9 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
first_missing = subdevs * 2;
second_missing = subdevs * 2;
insert_point = subdevs * 2;
if (mddev_test_and_add_drive_policies(st, &custom_pols, fd, 1))
exit(1);
}
}
if (fd >= 0)
@ -735,7 +747,7 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
close(dfd);
exit(2);
}
close(dfd);
info.array.working_disks++;
if (dnum < s->raiddisks && dv->disposition != 'j')
info.array.active_disks++;
@ -808,6 +820,11 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
}
}
if (drive_test_and_add_policies(st, &custom_pols, dfd, 1))
exit(1);
close(dfd);
if (dv->disposition == 'j')
goto skip_size_check; /* skip write journal for size check */
@ -882,6 +899,7 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
close(fd);
}
}
if (missing_disks == dnum && !have_container) {
pr_err("Subdevs can't be all missing\n");
return 1;
@ -1136,26 +1154,30 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
goto abort_locked;
}
if (did_default && c->verbose >= 0) {
if (did_default) {
if (is_subarray(info.text_version)) {
char devnm[32];
char *ep;
char devnm[MD_NAME_MAX];
struct mdinfo *mdi;
strncpy(devnm, info.text_version+1, 32);
devnm[31] = 0;
ep = strchr(devnm, '/');
if (ep)
*ep = 0;
sysfs_get_container_devnm(&info, devnm);
mdi = sysfs_read(-1, devnm, GET_VERSION);
mdi = sysfs_read(-1, devnm, GET_VERSION | GET_DEVS);
if (!mdi) {
pr_err("Cannot open sysfs for container %s\n", devnm);
goto abort_locked;
}
if (sysfs_test_and_add_drive_policies(st, &custom_pols, mdi, 1))
goto abort_locked;
if (c->verbose >= 0)
pr_info("Creating array inside %s container /dev/%s\n",
mdi->text_version, devnm);
pr_info("Creating array inside %s container %s\n",
mdi?mdi->text_version:"managed", devnm);
sysfs_free(mdi);
} else
pr_info("Defaulting to version %s metadata\n",
info.text_version);
} else if (c->verbose >= 0) {
pr_info("Defaulting to version %s metadata\n", info.text_version);
}
}
map_update(&map, fd2devnm(mdfd), info.text_version,
@ -1325,6 +1347,8 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
udev_unblock();
close(mdfd);
sysfs_uevent(&info, "change");
dev_policy_free(custom_pols);
return 0;
abort:
@ -1336,5 +1360,7 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
if (mdfd >= 0)
close(mdfd);
dev_policy_free(custom_pols);
return 1;
}

57
Detail.c
View file

@ -49,6 +49,30 @@ static int add_device(const char *dev, char ***p_devices,
return n_devices + 1;
}
/**
* detail_fname_from_uuid() - generate uuid string with special super1 handling.
* @mp: map entry to parse.
* @buf: buf to write.
*
* Hack to workaround an issue with super1 superblocks. It swapuuid set in order for assembly
* to work, but can't have it set if we want this printout to match all the other uuid printouts
* in super1.c, so we force swapuuid to 1 to make our printout match the rest of super1.
*
* Always convert uuid if host is big endian.
*/
char *detail_fname_from_uuid(struct map_ent *mp, char *buf)
{
#if __BYTE_ORDER == BIG_ENDIAN
bool swap = true;
#else
bool swap = false;
#endif
if (strncmp(mp->metadata, "1.", 2) == 0)
swap = true;
return __fname_from_uuid(mp->uuid, swap, buf, ':');
}
int Detail(char *dev, struct context *c)
{
/*
@ -226,6 +250,9 @@ int Detail(char *dev, struct context *c)
str = map_num(pers, array.level);
if (c->export) {
char nbuf[64];
struct map_ent *mp = NULL, *map = NULL;
if (array.raid_disks) {
if (str)
printf("MD_LEVEL=%s\n", str);
@ -247,32 +274,22 @@ int Detail(char *dev, struct context *c)
array.minor_version);
}
if (st && st->sb && info) {
char nbuf[64];
struct map_ent *mp, *map = NULL;
fname_from_uuid(st, info, nbuf, ':');
printf("MD_UUID=%s\n", nbuf + 5);
if (info)
mp = map_by_uuid(&map, info->uuid);
if (!mp)
mp = map_by_devnm(&map, fd2devnm(fd));
if (mp && mp->path && strncmp(mp->path, DEV_MD_DIR, DEV_MD_DIR_LEN) == 0)
if (mp) {
detail_fname_from_uuid(mp, nbuf);
printf("MD_UUID=%s\n", nbuf + 5);
if (mp->path && strncmp(mp->path, DEV_MD_DIR, DEV_MD_DIR_LEN) == 0)
printf("MD_DEVNAME=%s\n", mp->path + DEV_MD_DIR_LEN);
}
map_free(map);
if (st && st->sb) {
if (st->ss->export_detail_super)
st->ss->export_detail_super(st);
map_free(map);
} else {
struct map_ent *mp, *map = NULL;
char nbuf[64];
mp = map_by_devnm(&map, fd2devnm(fd));
if (mp) {
__fname_from_uuid(mp->uuid, 0, nbuf, ':');
printf("MD_UUID=%s\n", nbuf+5);
}
if (mp && mp->path && strncmp(mp->path, DEV_MD_DIR, DEV_MD_DIR_LEN) == 0)
printf("MD_DEVNAME=%s\n", mp->path + DEV_MD_DIR_LEN);
map_free(map);
}
if (!c->no_devices && sra) {
struct mdinfo *mdi;

32
Grow.c
View file

@ -2085,9 +2085,10 @@ int Grow_reshape(char *devname, int fd,
if (!mdmon_running(st->container_devnm))
start_mdmon(st->container_devnm);
ping_monitor(container);
if (mdmon_running(st->container_devnm) &&
st->update_tail == NULL)
st->update_tail = &st->updates;
if (mdmon_running(st->container_devnm) == false) {
pr_err("No mdmon found. Grow cannot continue.\n");
goto release;
}
}
if (s->size == MAX_SIZE)
@ -2097,11 +2098,7 @@ int Grow_reshape(char *devname, int fd,
/* got truncated to 32bit, write to
* component_size instead
*/
if (sra)
rv = sysfs_set_num(sra, NULL,
"component_size", s->size);
else
rv = -1;
rv = sysfs_set_num(sra, NULL, "component_size", s->size);
} else {
rv = md_set_array_info(fd, &array);
@ -3048,6 +3045,8 @@ static int reshape_array(char *container, int fd, char *devname,
dprintf("Cannot get array information.\n");
goto release;
}
if (st->update_tail == NULL)
st->update_tail = &st->updates;
if (array.level == 0 && info->component_size == 0) {
get_dev_size(fd, NULL, &array_size);
info->component_size = array_size / array.raid_disks;
@ -4414,19 +4413,8 @@ static void validate(int afd, int bfd, unsigned long long offset)
lseek64(afd, __le64_to_cpu(bsb2.arraystart)*512, 0);
if ((unsigned long long)read(afd, abuf, len) != len)
fail("read first from array failed");
if (memcmp(bbuf, abuf, len) != 0) {
#if 0
int i;
printf("offset=%llu len=%llu\n",
(unsigned long long)__le64_to_cpu(bsb2.arraystart)*512, len);
for (i=0; i<len; i++)
if (bbuf[i] != abuf[i]) {
printf("first diff byte %d\n", i);
break;
}
#endif
if (memcmp(bbuf, abuf, len) != 0)
fail("data1 compare failed");
}
}
if (bsb2.length2) {
unsigned long long len = __le64_to_cpu(bsb2.length2)*512;
@ -5152,9 +5140,7 @@ int Grow_continue_command(char *devname, int fd,
start_mdmon(container);
ping_monitor(container);
if (mdmon_running(container))
st->update_tail = &st->updates;
else {
if (mdmon_running(container) == false) {
pr_err("No mdmon found. Grow cannot continue.\n");
ret_val = 1;
goto Grow_continue_command_exit;

78
Incremental.c
View file

@ -833,6 +833,54 @@ container_members_max_degradation(struct map_ent *map, struct map_ent *me)
return max_degraded;
}
/**
* incremental_external_test_spare_criteria() - helper to test spare criteria.
* @st: supertype, must be not NULL, it is duplicated here.
* @container_devnm: devnm of the container.
* @disk_fd: file descriptor of device to tested.
* @verbose: verbose flag.
*
* The function is used on new drive verification path to check if it can be added to external
* container. To test spare criteria, metadata must be loaded. It duplicates super to not mess in
* original one.
* Function is executed if superblock supports get_spare_criteria(), otherwise success is returned.
*/
mdadm_status_t incremental_external_test_spare_criteria(struct supertype *st, char *container_devnm,
int disk_fd, int verbose)
{
mdadm_status_t rv = MDADM_STATUS_ERROR;
char container_devname[PATH_MAX];
struct spare_criteria sc = {0};
struct supertype *dup;
if (!st->ss->get_spare_criteria)
return MDADM_STATUS_SUCCESS;
dup = dup_super(st);
snprintf(container_devname, PATH_MAX, "/dev/%s", container_devnm);
if (dup->ss->get_spare_criteria(dup, container_devname, &sc) != 0) {
if (verbose > 1)
pr_err("Failed to get spare criteria for %s\n", container_devname);
goto out;
}
if (!disk_fd_matches_criteria(dup, disk_fd, &sc)) {
if (verbose > 1)
pr_err("Disk does not match spare criteria for %s\n", container_devname);
goto out;
}
rv = MDADM_STATUS_SUCCESS;
out:
dev_policy_free(sc.pols);
dup->ss->free_super(dup);
free(dup);
return rv;
}
static int array_try_spare(char *devname, int *dfdp, struct dev_policy *pol,
struct map_ent *target, int bare,
struct supertype *st, int verbose)
@ -873,8 +921,7 @@ static int array_try_spare(char *devname, int *dfdp, struct dev_policy *pol,
struct supertype *st2;
struct domainlist *dl = NULL;
struct mdinfo *sra;
unsigned long long devsize, freesize = 0;
struct spare_criteria sc = {0, 0};
unsigned long long freesize = 0;
if (is_subarray(mp->metadata))
continue;
@ -925,34 +972,19 @@ static int array_try_spare(char *devname, int *dfdp, struct dev_policy *pol,
if (sra->array.failed_disks == -1)
sra->array.failed_disks = container_members_max_degradation(map, mp);
get_dev_size(dfd, NULL, &devsize);
if (sra->component_size == 0) {
/* true for containers, here we must read superblock
* to obtain minimum spare size */
struct supertype *st3 = dup_super(st2);
int mdfd = open_dev(mp->devnm);
if (mdfd < 0) {
free(st3);
/* true for containers */
if (incremental_external_test_spare_criteria(st2, mp->devnm, dfd, verbose))
goto next;
}
if (st3->ss->load_container &&
!st3->ss->load_container(st3, mdfd, mp->path)) {
if (st3->ss->get_spare_criteria)
st3->ss->get_spare_criteria(st3, &sc);
st3->ss->free_super(st3);
}
free(st3);
close(mdfd);
}
if ((sra->component_size > 0 &&
st2->ss->validate_geometry(st2, sra->array.level, sra->array.layout,
if (sra->component_size > 0 &&
st2->ss->validate_geometry(st2, sra->array.level, sra->array.layout,
sra->array.raid_disks, &sra->array.chunk_size,
sra->component_size,
sra->devs ? sra->devs->data_offset : INVALID_SECTORS,
devname, &freesize, sra->consistency_policy,
0) &&
freesize < sra->component_size) ||
(sra->component_size == 0 && devsize < sc.min_size)) {
0) && freesize < sra->component_size) {
if (verbose > 1)
pr_err("not adding %s to %s as it is too small\n",
devname, mp->path);

44
MAINTAINERS.md Normal file
View file

@ -0,0 +1,44 @@
# Maintainer tools
Useful tools used in daily routines:
- [checkpatch](https://docs.kernel.org/dev-tools/checkpatch.html)
- [kup](https://korg.docs.kernel.org/kup.html)
- [Auto-publishing](https://korg.docs.kernel.org/kup.html#auto-publishing-with-git-archive-signer)
- [b4](https://b4.docs.kernel.org/en/latest/)
# Checklist before applying patch
We don't have CI testing yet, so all those steps must be performed manually:
- Style check with [checkpatch](https://docs.kernel.org/dev-tools/checkpatch.html):
This is the current code style follows. We are not strict to all rules. It must be run
by **checkpatch --no-tree**, see README.md.
- [Commit style](https://www.kernel.org/doc/html/v4.10/process/submitting-patches.html):
It doesn't need to be followed as strictly as is in kernel but changes should be logically
separated. Submitter should care at least to mention "It is used in next patches" if unused
externs/files are added in patch. We love: *Reported-by:*, *Suggested-by:*, *Fixes:* tags.
- Compilation, ideally on various gcc versions.
- Mdadm test suite execution.
- Consider requesting new tests from submitter, especially for new functionalities.
- Ensure that maintainer *sign-off* is added, before pushing.
# Making a release
Assuming that maintainer is certain that release is safe, following steps must be done:
- Update versions strings in release commit, please refer to previous releases for examples.
- Create GPG signed tag and push it to repo. Use same format as was used previously, prefixed by
**mdadm-**, e.g. **mdadm-3.1.2**, **mdadm-4.1**.
- [Auto-publishing](https://korg.docs.kernel.org/kup.html#auto-publishing-with-git-archive-signer):
Adopt script to our release tag model. When ready, push signed note to repository. If it is done
correctly, then *(sig)* is added to the package automatically generated by kernel.org automation.
There is no need to upload archive manually.
- Update CHANGELOG.md.
- Write "ANNOUNCE" mail to linux-raid@kernel.org to notify community.

4
Makefile
View file

@ -170,7 +170,7 @@ OBJS = mdadm.o config.o policy.o mdstat.o ReadMe.o uuid.o util.o maps.o lib.o u
mdopen.o super0.o super1.o super-ddf.o super-intel.o bitmap.o \
super-mbr.o super-gpt.o \
restripe.o sysfs.o sha1.o mapfile.o crc32.o sg_io.o msg.o xmalloc.o \
platform-intel.o probe_roms.o crc32c.o
platform-intel.o probe_roms.o crc32c.o drive_encryption.o
CHECK_OBJS = restripe.o uuid.o sysfs.o maps.o lib.o xmalloc.o dlink.o
@ -183,7 +183,7 @@ MON_OBJS = mdmon.o monitor.o managemon.o uuid.o util.o maps.o mdstat.o sysfs.o c
Kill.o sg_io.o dlink.o ReadMe.o super-intel.o \
super-mbr.o super-gpt.o \
super-ddf.o sha1.o crc32.o msg.o bitmap.o xmalloc.o \
platform-intel.o probe_roms.o crc32c.o
platform-intel.o probe_roms.o crc32c.o drive_encryption.o
MON_SRCS = $(patsubst %.o,%.c,$(MON_OBJS))

195
Manage.c
View file

@ -178,7 +178,7 @@ int Manage_stop(char *devname, int fd, int verbose, int will_retry)
struct map_ent *map = NULL;
struct mdinfo *mdi;
char devnm[32];
char container[32];
char container[MD_NAME_MAX] = {0};
int err;
int count;
char buf[SYSFS_MAX_BUF_SIZE];
@ -192,15 +192,9 @@ int Manage_stop(char *devname, int fd, int verbose, int will_retry)
* to stop is probably a bad idea.
*/
mdi = sysfs_read(fd, NULL, GET_LEVEL|GET_COMPONENT|GET_VERSION);
if (mdi && is_subarray(mdi->text_version)) {
char *sl;
strncpy(container, mdi->text_version+1, sizeof(container));
container[sizeof(container)-1] = 0;
sl = strchr(container, '/');
if (sl)
*sl = 0;
} else
container[0] = 0;
if (mdi && is_subarray(mdi->text_version))
sysfs_get_container_devnm(mdi, container);
close(fd);
count = 5;
while (((fd = ((devname[0] == '/')
@ -695,6 +689,99 @@ skip_re_add:
return 0;
}
/**
* manage_add_external() - Add disk to external container.
* @st: external supertype pointer, must not be NULL, superblock is released here.
* @fd: container file descriptor, must not have O_EXCL mode.
* @disk_fd: device to add file descriptor.
* @disk_name: name of the device to add.
* @disc: disk info.
*
* Superblock is released here because any open fd with O_EXCL will block sysfs_add_disk().
*/
mdadm_status_t manage_add_external(struct supertype *st, int fd, char *disk_name,
mdu_disk_info_t *disc)
{
mdadm_status_t rv = MDADM_STATUS_ERROR;
char container_devpath[MD_NAME_MAX];
struct dev_policy *pols = NULL;
struct mdinfo new_mdi;
struct mdinfo *sra = NULL;
int container_fd;
int disk_fd = -1;
snprintf(container_devpath, MD_NAME_MAX, "%s", fd2devnm(fd));
container_fd = open_dev_excl(container_devpath);
if (!is_fd_valid(container_fd)) {
pr_err("Failed to get exclusive access to container %s\n", container_devpath);
return MDADM_STATUS_ERROR;
}
/* Check if metadata handler is able to accept the drive */
if (!st->ss->validate_geometry(st, LEVEL_CONTAINER, 0, 1, NULL, 0, 0, disk_name, NULL,
0, 1))
goto out;
if (mddev_test_and_add_drive_policies(st, &pols, container_fd, 1))
goto out;
Kill(disk_name, NULL, 0, -1, 0);
disk_fd = dev_open(disk_name, O_RDWR | O_EXCL | O_DIRECT);
if (!is_fd_valid(disk_fd)) {
pr_err("Failed to exclusively open %s\n", disk_name);
goto out;
}
if (drive_test_and_add_policies(st, &pols, disk_fd, 1))
goto out;
if (st->ss->add_to_super(st, disc, disk_fd, disk_name, INVALID_SECTORS))
goto out;
if (!mdmon_running(st->container_devnm))
st->ss->sync_metadata(st);
sra = sysfs_read(container_fd, NULL, 0);
if (!sra) {
pr_err("Failed to read sysfs for %s\n", disk_name);
goto out;
}
sra->array.level = LEVEL_CONTAINER;
/* Need to set data_offset and component_size */
st->ss->getinfo_super(st, &new_mdi, NULL);
new_mdi.disk.major = disc->major;
new_mdi.disk.minor = disc->minor;
new_mdi.recovery_start = 0;
st->ss->free_super(st);
if (sysfs_add_disk(sra, &new_mdi, 0) != 0) {
pr_err("Failed to add %s to container %s\n", disk_name, container_devpath);
goto out;
}
ping_monitor(container_devpath);
rv = MDADM_STATUS_SUCCESS;
out:
close(container_fd);
dev_policy_free(pols);
if (sra)
sysfs_free(sra);
if (rv != MDADM_STATUS_SUCCESS && is_fd_valid(disk_fd))
/* Metadata handler records this descriptor, so release it only on failure. */
close(disk_fd);
if (st->sb)
st->ss->free_super(st);
return rv;
}
int Manage_add(int fd, int tfd, struct mddev_dev *dv,
struct supertype *tst, mdu_array_info_t *array,
int force, int verbose, char *devname,
@ -794,25 +881,23 @@ int Manage_add(int fd, int tfd, struct mddev_dev *dv,
* simply re-add it.
*/
if (array->not_persistent == 0) {
if (array->not_persistent == 0 && dv->disposition != 'S') {
int rv = 0;
dev_st = dup_super(tst);
dev_st->ss->load_super(dev_st, tfd, NULL);
if (dev_st->sb && dv->disposition != 'S') {
int rv;
rv = attempt_re_add(fd, tfd, dv, dev_st, tst,
rdev, update, devname,
verbose, array);
if (dev_st->sb) {
rv = attempt_re_add(fd, tfd, dv, dev_st, tst, rdev, update,
devname, verbose, array);
dev_st->ss->free_super(dev_st);
if (rv) {
free(dev_st);
return rv;
}
}
if (dev_st) {
dev_st->ss->free_super(dev_st);
free(dev_st);
}
free(dev_st);
if (rv)
return rv;
}
if (dv->disposition == 'M') {
if (verbose > 0)
@ -968,68 +1053,8 @@ int Manage_add(int fd, int tfd, struct mddev_dev *dv,
if (dv->failfast == FlagSet)
disc.state |= (1 << MD_DISK_FAILFAST);
if (tst->ss->external) {
/* add a disk
* to an external metadata container */
struct mdinfo new_mdi;
struct mdinfo *sra;
int container_fd;
char devnm[32];
int dfd;
strcpy(devnm, fd2devnm(fd));
container_fd = open_dev_excl(devnm);
if (container_fd < 0) {
pr_err("add failed for %s: could not get exclusive access to container\n",
dv->devname);
tst->ss->free_super(tst);
if (manage_add_external(tst, fd, dv->devname, &disc) != MDADM_STATUS_SUCCESS)
goto unlock;
}
/* Check if metadata handler is able to accept the drive */
if (!tst->ss->validate_geometry(tst, LEVEL_CONTAINER, 0, 1, NULL,
0, 0, dv->devname, NULL, 0, 1)) {
close(container_fd);
goto unlock;
}
Kill(dv->devname, NULL, 0, -1, 0);
dfd = dev_open(dv->devname, O_RDWR | O_EXCL|O_DIRECT);
if (tst->ss->add_to_super(tst, &disc, dfd,
dv->devname, INVALID_SECTORS)) {
close(dfd);
close(container_fd);
goto unlock;
}
if (!mdmon_running(tst->container_devnm))
tst->ss->sync_metadata(tst);
sra = sysfs_read(container_fd, NULL, 0);
if (!sra) {
pr_err("add failed for %s: sysfs_read failed\n",
dv->devname);
close(container_fd);
tst->ss->free_super(tst);
goto unlock;
}
sra->array.level = LEVEL_CONTAINER;
/* Need to set data_offset and component_size */
tst->ss->getinfo_super(tst, &new_mdi, NULL);
new_mdi.disk.major = disc.major;
new_mdi.disk.minor = disc.minor;
new_mdi.recovery_start = 0;
/* Make sure fds are closed as they are O_EXCL which
* would block add_disk */
tst->ss->free_super(tst);
if (sysfs_add_disk(sra, &new_mdi, 0) != 0) {
pr_err("add new device to external metadata failed for %s\n", dv->devname);
close(container_fd);
sysfs_free(sra);
goto unlock;
}
ping_monitor(devnm);
sysfs_free(sra);
close(container_fd);
} else {
tst->ss->free_super(tst);
if (ioctl(fd, ADD_NEW_DISK, &disc)) {

66
Monitor.c
View file

@ -451,17 +451,19 @@ static int check_one_sharer(int scan)
return 2;
}
if (access(AUTOREBUILD_PID_PATH, F_OK) != 0)
return 0;
fp = fopen(AUTOREBUILD_PID_PATH, "r");
if (!fp) {
/* PID file does not exist */
if (errno == ENOENT)
return 0;
if (!is_file(AUTOREBUILD_PID_PATH)) {
pr_err("%s is not a regular file.\n", AUTOREBUILD_PID_PATH);
pr_err("Cannot open %s file.\n", AUTOREBUILD_PID_PATH);
return 2;
}
fp = fopen(AUTOREBUILD_PID_PATH, "r");
if (!fp) {
pr_err("Cannot open %s file.\n", AUTOREBUILD_PID_PATH);
if (!is_file(AUTOREBUILD_PID_PATH)) {
pr_err("%s is not a regular file.\n", AUTOREBUILD_PID_PATH);
fclose(fp);
return 2;
}
@ -1006,34 +1008,6 @@ static int add_new_arrays(struct mdstat_ent *mdstat, struct state **statelist)
return new_found;
}
static int get_required_spare_criteria(struct state *st,
struct spare_criteria *sc)
{
int fd;
if (!st->metadata || !st->metadata->ss->get_spare_criteria) {
sc->min_size = 0;
sc->sector_size = 0;
return 0;
}
fd = open(st->devname, O_RDONLY);
if (fd < 0)
return 1;
if (st->metadata->ss->external)
st->metadata->ss->load_container(st->metadata, fd, st->devname);
else
st->metadata->ss->load_super(st->metadata, fd, st->devname);
close(fd);
if (!st->metadata->sb)
return 1;
st->metadata->ss->get_spare_criteria(st->metadata, sc);
st->metadata->ss->free_super(st->metadata);
return 0;
}
static int check_donor(struct state *from, struct state *to)
{
struct state *sub;
@ -1068,22 +1042,12 @@ static dev_t choose_spare(struct state *from, struct state *to,
for (d = from->raid; !dev && d < MAX_DISKS; d++) {
if (from->devid[d] > 0 && from->devstate[d] == 0) {
struct dev_policy *pol;
unsigned long long dev_size;
unsigned int dev_sector_size;
if (to->metadata->ss->external &&
test_partition_from_id(from->devid[d]))
continue;
if (sc->min_size &&
dev_size_from_id(from->devid[d], &dev_size) &&
dev_size < sc->min_size)
continue;
if (sc->sector_size &&
dev_sector_size_from_id(from->devid[d],
&dev_sector_size) &&
sc->sector_size != dev_sector_size)
if (devid_matches_criteria(to->metadata, from->devid[d], sc) == false)
continue;
pol = devid_policy(from->devid[d]);
@ -1168,12 +1132,12 @@ static void try_spare_migration(struct state *statelist)
{
struct state *from;
struct state *st;
struct spare_criteria sc;
link_containers_with_subarrays(statelist);
for (st = statelist; st; st = st->next)
if (st->active < st->raid && st->spare == 0 && !st->err) {
struct domainlist *domlist = NULL;
struct spare_criteria sc = {0};
int d;
struct state *to = st;
@ -1186,8 +1150,11 @@ static void try_spare_migration(struct state *statelist)
/* member of a container */
to = to->parent;
if (get_required_spare_criteria(to, &sc))
continue;
if (to->metadata->ss->get_spare_criteria)
if (to->metadata->ss->get_spare_criteria(to->metadata, to->devname,
&sc))
continue;
if (to->metadata->ss->external) {
/* We must make sure there is
* no suitable spare in container already.
@ -1228,6 +1195,7 @@ static void try_spare_migration(struct state *statelist)
}
}
domain_free(domlist);
dev_policy_free(sc.pols);
}
}

122
README.initramfs
View file

@ -1,122 +0,0 @@
Assembling md arrays at boot time.
---------------------------------
December 2005
These notes apply to 2.6 kernels only and, in some cases,
to 2.6.15 or later.
Md arrays can be assembled at boot time using the 'autodetect' functionality
which is triggered by storing components of an array in partitions of type
'fd' - Linux Raid Autodetect.
They can also be assembled by specifying the component devices in a
kernel parameter such as
md=0,/dev/sda,/dev/sdb
In this case, /dev/md0 will be assembled (because of the 0) from the listed
devices.
These mechanisms, while useful, do not provide complete functionality
and are unlikely to be extended. The preferred way to assemble md
arrays at boot time is using 'mdadm'. To assemble an array which
contains the root filesystem, mdadm needs to be run before that
filesystem is mounted, and so needs to be run from an initial-ram-fs.
It is how this can work that is the primary focus of this document.
It should be noted up front that only the array containing the root
filesystem should be assembled from the initramfs. Any other arrays
should be assembled under the control of files on the main filesystem
as this enhanced flexibility and maintainability.
A minimal initramfs for assembling md arrays can be created using 3
files and one directory. These are:
/bin Directory
/bin/mdadm statically linked mdadm binary
/bin/busybox statically linked busybox binary
/bin/sh hard link to /bin/busybox
/init a shell script which call mdadm appropriately.
An example init script is:
==============================================
#!/bin/sh
echo 'Auto-assembling boot md array'
mkdir /proc
mount -t proc proc /proc
if [ -n "$rootuuid" ]
then arg=--uuid=$rootuuid
elif [ -n "$mdminor" ]
then arg=--super-minor=$mdminor
else arg=--super-minor=0
fi
echo "Using $arg"
mdadm -Acpartitions $arg --auto=part /dev/mda
cd /
mount /dev/mda1 /root || mount /dev/mda /root
umount /proc
cd /root
exec chroot . /sbin/init < /dev/console > /dev/console 2>&1
=============================================
This could certainly be extended, or merged into a larger init script.
Though tested and in production use, it is not presented here as
"The Right Way" to do it, but as a useful example.
Some key points are:
/proc needs to be mounted so that /proc/partitions can be accessed
by mdadm, and so that /proc/filesystems can be accessed by mount.
The uuid of the array can be passed in as a kernel parameter
(rootuuid). As the kernel doesn't use this value, it is made available
in the environment for /init
If no uuid is given, we default to md0, (--super-minor=0) which is a
commonly used to store the root filesystem. This may not work in
all situations.
We assemble the array as a partitionable array (/dev/mda) even if we
end up using the whole array. There is no cost in using the partitionable
interface, and in this context it is simpler.
We try mounting both /dev/mda1 and /dev/mda as they are the most like
part of the array to contain the root filesystem.
The --auto flag is given to mdadm so that it will create /dev/md*
files automatically. This is needed as /dev will not contain
and md files, and udev will not create them (as udev only created device
files after the device exists, and mdadm need the device file to create
the device). Note that the created md files may not exist in /dev
of the mounted root filesystem. This needs to be deal with separately
from mdadm - possibly using udev.
We do not need to create device files for the components which will
be assembled into /dev/mda. mdadm finds the major/minor numbers from
/proc/partitions and creates a temporary /dev file if one doesn't already
exist.
The script "mkinitramfs" which is included with the mdadm distribution
can be used to create a minimal initramfs. It creates a file called
'init.cpio.gz' which can be specified as an 'initrd' to lilo or grub
(or whatever boot loader is being used).
Resume from an md array
-----------------------
If you want to make use of the suspend-to-disk/resume functionality in Linux,
and want to have swap on an md array, you will need to assemble the array
before resume is possible.
However, because the array is active in the resumed image, you do not want
anything written to any drives during the resume process, such as superblock
updates or array resync.
This can be achieved in 2.6.15-rc1 and later kernels using the
'start_readonly' module parameter.
Simply include the command
echo 1 > /sys/module/md_mod/parameters/start_ro
before assembling the array with 'mdadm'.
You can then echo
9:0
or whatever is appropriate to /sys/power/resume to trigger the resume.

83
README.md Normal file
View file

@ -0,0 +1,83 @@
**mdadm** is a utility used to create and manage **software RAID** devices implemented through
**Multiple devices driver (MD)** in kernel. It supports following RAID metadata formats:
* [Linux native RAID](https://raid.wiki.kernel.org/index.php/RAID_superblock_formats):
Known as **native** or **native RAID**. First and default metadata format. Metadata management
is implemented in **MD driver**.
* Matrix Storage Manager Support (no reference, metadata format documentation is proprietary).
Known as **IMSM**. Metadata format developed and maintained by **Intel®** as a part of **VROC**
solution. There are some functional differences between **native** and **imsm**. The most
important difference is that the metadata is managed from userspace.
**CAUTION:** **imsm** is compatible with **Intel RST**, however it is not officially supported.
You are using it on your own risk.
* [Common RAID DDF Specification Revision](https://www.snia.org/sites/default/files/SNIA_DDF_Technical_Position_v2.0.pdf)
**IMPORTANT:** DDF is in **maintenance only** mode. There is no active development around it.
Please do not use it in new solutions.
# How to Contribute
**mdadm** is hosted on [kernel.org](https://kernel.org/). You can access repository
[here](https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git).
It is maintained similarly to kernel, using *mailing list*. Patches must be send through email.
Please familiarize with general kernel
[submitting patches](https://www.kernel.org/doc/html/v4.17/process/submitting-patches.html)
documentation. Formatting, tags and commit message guidelines applies to **mdadm**.
## Sending patches step-by-step
To maximize change of patches being taken, follow this instruction when submitting:
1. Create possibly logically separated commits and generate patches:
Use ``git format-patch --cover-letter --signoff -v <nr>`` to create patches:
* ``--cover-letter`` can be skipped if it is only one patch;
* ``--signoff`` adds sign-off tag;
* ``-v <nr>`` indicates review revision number, sender should increment it before resending.
2. Check style of every patch with kernel
[checkpatch](https://docs.kernel.org/dev-tools/checkpatch.html) script:
It is important to keep same coding style that is why in **mdadm**
[kernel coding style](https://www.kernel.org/doc/html/v4.10/process/coding-style.html)
is preferred. ``checkpath --no-tree <patch_file>`` can be used to verify patches.
Following checkpatch issues can be ignored:
- New typedefs.
- comparing with *True/False*.
- kernel *MAINTAINERS* file warning.
- *extern* keyword in headers.
3. Send patches using ``git send-mail --to=linux-raid@vger.kernel.org <cover-letter> <patch1> <patch2> (...)``
# Maintainers
It is good practice to add **mdadm maintainers** to recipients for patches:
- Jes Sorensen <jes@trained-monkey.org>;
- Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>;
Adding **MD maintainers** could be reasonable, especially if patches may affect MD driver:
- Song Liu <song@kernel.org>;
- Yu Kuai <yukuai3@huawei.com>;
# Reviewers
**mdadm** utility is not part of kernel tree, so there is no certificated *Reviewers* list. Everyone
can comment on mailing list, last decision (and merging) belongs to maintainers.
# Minimal supported kernel version
We do not support kernel versions below **v3.10**. Please be aware that maintainers may remove
workarounds and fixes for legacy issues.
# License
It is released under the terms of the **GNU General Public License version 2** as published
by the **Free Software Foundation**.

213
TODO
View file

@ -1,213 +0,0 @@
- add 'name' field to metadata type and use it.
- use validate_geometry more
- metadata should be able to check/reject bitmap stuff.
DDF:
Three new metadata types:
ddf - used only to create a container.
ddf-bvd - used to create an array in a container
ddf-svd - used to create a secondary array from bvds.
Usage:
mdadm -C /dev/ddf1 /dev/sd[abcdef]
mdadm -C /dev/md1 -e ddf /dev/sd[a-f]
mdadm -C /dev/md1 -l container /dev/sd[a-f]
Each of these create a new ddf container using all those
devices. The name 'ddf*' signals that ddf metadata should be used.
'-e ddf' only supports one level - 'container'. 'container' is only
supported by ddf.
mdadm -C /dev/md1 -l0 -n4 /dev/ddf1 # or maybe not ???
mdadm -C /dev/md1 -l1 -n2 /dev/sda /dev/sdb
If exactly one device is given, and it is a container, we select
devices from that container.
If devices are given that are already in use, they must be in use by
a container, and the array is created in the container.
If devices given are bvds, we slip under the hood to make
the svd arrays.
mdadm -A /dev/ddf ......
base drives make a container. Anything in that container is started
auto-read-only.
if /dev/ddf is already assembled, we assemble bvds and svds inside it.
2005-dec-20
Want an incremental assembly mode to work nicely with udev.
Core usage would be something like
mdadm --incr-assemble /dev/newdevice
This would
- examine the device to determine uuid etc.
- look for a match in /etc/mdadm.conf, abort if not found
- find that device and collect current contents
- perform an 'assemble' analysis to make sure we have the best set of devices.
- remove or add devices as appropriate
- possibly start the array if it was complete
Other usages could involve
- specify which array to auto-add to.
This requires an existing array for uuid matching... is there any point?
-
2004-june-02
* Don't print 'errors' flag, it is meaningless. DONE
* Handle new superblock format
* create device file on demand, particularly partitionable devices. DONE
BUT figure a way to create the partition devices.
auto=partN
* Use Event: interface to listen for events. DONE, untested
* Make sure mdadm -As can assemble multi-level RAIDs ok.
* --build to build raid1 or multipath arrays
clean or not ???
----------------------------------------------------------------------------
* mdadm --monitor to monitor failed multipath paths and re-instate them.
* Maybe make "--help" fit in 80x24 and have a --long-help with more info. DONE
* maybe "missing" instead of <bold>missing</> in doco DONE
* possibly wait for resync to start, or even finish while assembling.- NO
* -Db should have a devices= entry if possible. - DONE
* when assembling multipath arrays, ignore any error indicators. - DONE
* rationalise --monitor usage:
mdadm --monitor
doesn't do as expected. DONE
* --assemble could have a --update option. - DONE
following word can be:
sparc2.2
super-minor
* mdadm /dev/md11, where md11 is raid0 can segfault, particularly when looking in the
[UU_UUU] string ... which doesn't exist !
It should be more sensible. DONE
Example:
from Raimund Sacherer <raimund.sacherer@ngit.at>
mke2fs -m0 -q /dev/ram1 300
mount -n -t ext2 /dev/ram1 /tmp
echo DEVICE /dev/[sh]* >> /tmp/mdadm.conf
mdadm -Esb /dev/[sh]* 2>/dev/null >> /tmp/mdadm.conf
mdadm -ARsc /tmp/mdadm.conf
umount /tmp
?? Allow -S /dev/md? - current complains subsequent not a/d/r - DONE
* new "Query" mode to subsume --detail and --examine.
--query or -Q, takes a device and tells if it is an MD device,
and also tells in a raid superblock is found.
DONE
* write mdstat.c to parse /proc/mdstat file
Build list of arrays: name, rebuild-percent
DONE
* parse /proc/partitions and map major/minor into /dev/* names,
and use that for default DEVICE list ????
* --detail --scan to read /proc/mdstat, and then iterate over these,
but assume --brief. --verbose can override
check each subdevice to see if it is in conf_get_devs.
Warn if not.
DONE, but don't warn yet...
* Support multipath ... maybe...
maybe DONE
* --follow to syslog
* --follow to move spares around DONE
* --follow to notice other events: DONE
rebuild started
spare activated
spare removed
spare added
------------------------------------
- --examine --scan scans all drives and build an mdadm.conf file DONE
- check superblock checksum in examine DONE
- report "chunk" or "rounding" depending on raid level DONE
- report "linear" instead of "-1" for raid level DONE
- decode ayout depending on raid level DONE
- --verbose and --force flags. DONE
- set md_minor, *_disks for Create - DONE
- for create raid5, how to choose between
all working, but not insync
one missing, one spare, insync DONE (--force)
- and for raid1 - some failed drives... (missing)
- when RUN_ARRAY, make sure *_disks counts are right
- get --detail to extract extra stuff from superblock,
like uuid DONE
- --detail --brief to give a config file line DONE
- parse config file. DONE
- test...
- when --assemble --scan, if an underlying device is an md device,
then try to assemble that device first.
- mdadm -S /dev/md0 /dev/md1 gives internal error FIXED
- mdadm --detail --scan print summary of what it can find? DONE
---------
Assemble doesn't add spares. - DONE
Create to allow "missing" name for devices.
Create to accept "--force" for do exactly what is requested
- get Assemble to upgrade devices if force flag.
ARRAY lines in config file to have super_minor=n
ARRAY lines in config file to have device=pattern, and only accept
those devices
If UUID given, insist on that
If not, but super_minor given, require all found with that minor
to have same uuid
If only device given, all valid supers on those devices must have
same uuid
allow /dev/mdX as first argument before any options
Possible --dry-run option for create and assemble--force
Assemble to check that all devices mentioned in superblock
are present.
New mode: --Monitor (or --Follow)
Periodically check status of all arrays (listed in config file).
Log every event and apparent cause - or differences
Email and alert - or run a program - for important events
Move spares around if necessary.
An Array line can have a spare-group= field that indicates that
the array shares spares with other arrays with the same
spare-group name.
If an array has a failed and no spares, then check all other
arrays in the spare group. If one has no failures and a spare,
then consider that spare.
Choose the smallest considered spare that is large enough.
If there is one, then hot-remove it from it's home, and
hot-add it to the array in question.
--mail-to address
--alert-handler program
Will also extract information from /proc/mdstat if present,
and consider 20% marks in rebuild as events.
Events are:
drive fails - causes mail to be sent
rebuild started
spare activated
spare removed
spare added

25
config.c
View file

@ -81,7 +81,7 @@ char DefaultAltConfDir[] = CONFFILE2 ".d";
enum linetype { Devices, Array, Mailaddr, Mailfrom, Program, CreateDev,
Homehost, HomeCluster, AutoMode, Policy, PartPolicy, Sysfs,
MonitorDelay, LTEnd };
MonitorDelay, EncryptionNoVerify, LTEnd };
char *keywords[] = {
[Devices] = "devices",
[Array] = "array",
@ -96,6 +96,7 @@ char *keywords[] = {
[PartPolicy]="part-policy",
[Sysfs] = "sysfs",
[MonitorDelay] = "monitordelay",
[EncryptionNoVerify] = "ENCRYPTION_NO_VERIFY",
[LTEnd] = NULL
};
@ -729,6 +730,19 @@ void monitordelayline(char *line)
}
}
static bool sata_opal_encryption_no_verify;
void encryption_no_verify_line(char *line)
{
char *word;
for (word = dl_next(line); word != line; word = dl_next(word)) {
if (strcasecmp(word, "sata_opal") == 0)
sata_opal_encryption_no_verify = true;
else
pr_err("unrecognised word on ENCRYPTION_NO_VERIFY line: %s\n", word);
}
}
char auto_yes[] = "yes";
char auto_no[] = "no";
char auto_homehost[] = "homehost";
@ -913,6 +927,9 @@ void conf_file(FILE *f)
case MonitorDelay:
monitordelayline(line);
break;
case EncryptionNoVerify:
encryption_no_verify_line(line);
break;
default:
pr_err("Unknown keyword %s\n", line);
}
@ -1075,6 +1092,12 @@ int conf_get_monitor_delay(void)
return monitor_delay;
}
bool conf_get_sata_opal_encryption_no_verify(void)
{
load_conffile();
return sata_opal_encryption_no_verify;
}
struct createinfo *conf_get_create_info(void)
{
load_conffile();

0
external-reshape-design.txt → documentation/external-reshape-design.txt
View file

0
mdadm.conf-example → documentation/mdadm.conf-example
View file

0
mdmon-design.txt → documentation/mdmon-design.txt
View file

724
drive_encryption.c Normal file
View file

@ -0,0 +1,724 @@
// SPDX-License-Identifier: GPL-2.0-only
/*
* Read encryption information for Opal and ATA devices.
*
* Copyright (C) 2024 Intel Corporation
* Author: Blazej Kucman <blazej.kucman@intel.com>
*/
#include "mdadm.h"
#include <asm/types.h>
#include <linux/nvme_ioctl.h>
#include <scsi/sg.h>
#include <scsi/scsi.h>
#include "drive_encryption.h"
#define DEFAULT_SECTOR_SIZE (512)
/*
* Opal defines
* TCG Storage Opal SSC 2.01 chapter 3.3.3
* NVM ExpressTM Revision 1.4c, chapter 5
*/
#define TCG_SECP_01 (0x01)
#define TCG_SECP_00 (0x00)
#define OPAL_DISCOVERY_COMID (0x0001)
#define OPAL_LOCKING_FEATURE (0x0002)
#define OPAL_IO_BUFFER_LEN 2048
#define OPAL_DISCOVERY_FEATURE_HEADER_LEN (4)
/*
* NVMe defines
* NVM ExpressTM Revision 1.4c, chapter 5
*/
#define NVME_SECURITY_RECV (0x82)
#define NVME_IDENTIFY (0x06)
#define NVME_IDENTIFY_RESPONSE_LEN 4096
#define NVME_OACS_BYTE_POSITION (256)
#define NVME_IDENTIFY_CONTROLLER_DATA (1)
/*
* ATA defines
* ATA/ATAPI Command Set ATA8-ACS
* SCSI / ATA Translation - 3 (SAT-3)
* SCSI Primary Commands - 4 (SPC-4)
* AT Attachment-8 - ATA Serial Transport (ATA8-AST)
* ATA Command Pass-Through
*/
#define ATA_IDENTIFY (0xec)
#define ATA_TRUSTED_RECEIVE (0x5c)
#define ATA_SECURITY_WORD_POSITION (128)
#define HDIO_DRIVE_CMD (0x031f)
#define ATA_TRUSTED_COMPUTING_POS (48)
#define ATA_PASS_THROUGH_12 (0xa1)
#define ATA_IDENTIFY_RESPONSE_LEN (512)
#define ATA_PIO_DATA_IN (4)
#define SG_CHECK_CONDITION (0x02)
#define ATA_STATUS_RETURN_DESCRIPTOR (0x09)
#define ATA_PT_INFORMATION_AVAILABLE_ASCQ (0x1d)
#define ATA_PT_INFORMATION_AVAILABLE_ASC (0x00)
#define ATA_INQUIRY_LENGTH (0x0c)
#define SG_INTERFACE_ID 'S'
#define SG_IO_TIMEOUT (60000)
#define SG_SENSE_SIZE (32)
#define SENSE_DATA_CURRENT_FIXED (0x70)
#define SENSE_DATA_CURRENT_DESC (0x72)
#define SENSE_CURRENT_RES_DESC_POS (8)
#define SG_DRIVER_SENSE (0x08)
typedef enum drive_feature_support_status {
/* Drive feature is supported. */
DRIVE_FEAT_SUP_ST = 0,
/* Drive feature is not supported. */
DRIVE_FEAT_NOT_SUP_ST,
/* Drive feature support check failed. */
DRIVE_FEAT_CHECK_FAILED_ST
} drive_feat_sup_st;
/* TCG Storage Opal SSC 2.01 chapter 3.1.1.3 */
typedef struct opal_locking_feature {
/* feature header */
__u16 feature_code;
__u8 reserved : 4;
__u8 version : 4;
__u8 description_length;
/* feature description */
__u8 locking_supported : 1;
__u8 locking_enabled : 1;
__u8 locked : 1;
__u8 media_encryption : 1;
__u8 mbr_enabled : 1;
__u8 mbr_done : 1;
__u8 mbr_shadowing_not_supported : 1;
__u8 hw_reset_for_dor_supported : 1;
__u8 reserved1[11];
} __attribute__((__packed__)) opal_locking_feature_t;
/* TCG Storage Opal SSC 2.01 chapter 3.1.1.1 */
typedef struct opal_level0_header {
__u32 length;
__u32 version;
__u64 reserved;
__u8 vendor_specific[32];
} opal_level0_header_t;
/**
* NVM ExpressTM Revision 1.4c, Figure 249
* Structure specifies only OACS filed, which is needed in the current use case.
*/
typedef struct nvme_identify_ctrl {
__u8 reserved[255];
__u16 oacs;
__u8 reserved2[3839];
} nvme_identify_ctrl_t;
/* SCSI Primary Commands - 4 (SPC-4), Table 512 */
typedef struct supported_security_protocols {
__u8 reserved[6];
__u16 list_length;
__u8 list[504];
} supported_security_protocols_t;
/* ATA/ATAPI Command Set - 3 (ACS-3), Table 45 */
typedef struct ata_security_status {
__u16 security_supported : 1;
__u16 security_enabled : 1;
__u16 security_locked : 1;
__u16 security_frozen : 1;
__u16 security_count_expired : 1;
__u16 enhanced_security_erase_supported : 1;
__u16 reserved1 : 2;
__u16 security_level : 1;
__u16 reserved2 : 7;
} __attribute__((__packed__)) ata_security_status_t;
/* ATA/ATAPI Command Set - 3 (ACS-3), Table 45 */
typedef struct ata_trusted_computing {
__u16 tc_feature :1;
__u16 reserved : 13;
__u16 var1 : 1;
__u16 var2 : 1;
} __attribute__((__packed__)) ata_trusted_computing_t;
mapping_t encryption_ability_map[] = {
{ "None", ENC_ABILITY_NONE },
{ "Other", ENC_ABILITY_OTHER },
{ "SED", ENC_ABILITY_SED },
{ NULL, UnSet }
};
mapping_t encryption_status_map[] = {
{ "Unencrypted", ENC_STATUS_UNENCRYPTED },
{ "Locked", ENC_STATUS_LOCKED },
{ "Unlocked", ENC_STATUS_UNLOCKED },
{ NULL, UnSet }
};
/**
* get_encryption_ability_string() - get encryption ability name string.
* @ability: encryption ability enum.
*
* Return: encryption ability string.
*/
const char *get_encryption_ability_string(enum encryption_ability ability)
{
return map_num_s(encryption_ability_map, ability);
}
/**
* get_encryption_status_string() - get encryption status name string.
* @ability: encryption status enum.
*
* Return: encryption status string.
*/
const char *get_encryption_status_string(enum encryption_status status)
{
return map_num_s(encryption_status_map, status);
}
/**
* get_opal_locking_feature_description() - get opal locking feature description.
* @response: response from Opal Discovery Level 0.
*
* Based on the documentation TCG Storage Opal SSC 2.01 chapter 3.1.1,
* a Locking feature is searched for in Opal Level 0 Discovery response.
*
* Return: if locking feature is found, pointer to struct %opal_locking_feature_t, NULL otherwise.
*/
static opal_locking_feature_t *get_opal_locking_feature_description(__u8 *response)
{
opal_level0_header_t *response_header = (opal_level0_header_t *)response;
int features_length = __be32_to_cpu(response_header->length);
int current_position = sizeof(*response_header);
while (current_position < features_length) {
opal_locking_feature_t *feature;
feature = (opal_locking_feature_t *)(response + current_position);
if (__be16_to_cpu(feature->feature_code) == OPAL_LOCKING_FEATURE)
return feature;
current_position += feature->description_length + OPAL_DISCOVERY_FEATURE_HEADER_LEN;
}
return NULL;
}
/**
* nvme_security_recv_ioctl() - nvme security receive ioctl.
* @disk_fd: a disk file descriptor.
* @sec_protocol: security protocol.
* @comm_id: command id.
* @response_buffer: response buffer to fill out.
* @buf_size: response buffer size.
* @verbose: verbose flag.
*
* Based on the documentations TCG Storage Opal SSC 2.01 chapter 3.3.3 and
* NVM ExpressTM Revision 1.4c, chapter 5.25,
* read security receive command via ioctl().
* On success, @response_buffer is completed.
*
* Return: %MDADM_STATUS_SUCCESS on success, %MDADM_STATUS_ERROR otherwise.
*/
static mdadm_status_t
nvme_security_recv_ioctl(int disk_fd, __u8 sec_protocol, __u16 comm_id, void *response_buffer,
size_t buf_size, const int verbose)
{
struct nvme_admin_cmd nvme_cmd = {0};
int status;
nvme_cmd.opcode = NVME_SECURITY_RECV;
nvme_cmd.cdw10 = sec_protocol << 24 | comm_id << 8;
nvme_cmd.cdw11 = buf_size;
nvme_cmd.data_len = buf_size;
nvme_cmd.addr = (__u64)response_buffer;
status = ioctl(disk_fd, NVME_IOCTL_ADMIN_CMD, &nvme_cmd);
if (status != 0) {
pr_vrb("Failed to read NVMe security receive ioctl() for device /dev/%s, status: %d\n",
fd2kname(disk_fd), status);
return MDADM_STATUS_ERROR;
}
return MDADM_STATUS_SUCCESS;
}
/**
* nvme_identify_ioctl() - NVMe identify ioctl.
* @disk_fd: a disk file descriptor.
* @response_buffer: response buffer to fill out.
* @buf_size: response buffer size.
* @verbose: verbose flag.
*
* Based on the documentations TCG Storage Opal SSC 2.01 chapter 3.3.3 and
* NVM ExpressTM Revision 1.4c, chapter 5.25,
* read NVMe identify via ioctl().
* On success, @response_buffer will be completed.
*
* Return: %MDADM_STATUS_SUCCESS on success, %MDADM_STATUS_ERROR otherwise.
*/
static mdadm_status_t
nvme_identify_ioctl(int disk_fd, void *response_buffer, size_t buf_size, const int verbose)
{
struct nvme_admin_cmd nvme_cmd = {0};
int status;
nvme_cmd.opcode = NVME_IDENTIFY;
nvme_cmd.cdw10 = NVME_IDENTIFY_CONTROLLER_DATA;
nvme_cmd.data_len = buf_size;
nvme_cmd.addr = (__u64)response_buffer;
status = ioctl(disk_fd, NVME_IOCTL_ADMIN_CMD, &nvme_cmd);
if (status != 0) {
pr_vrb("Failed to read NVMe identify ioctl() for device /dev/%s, status: %d\n",
fd2kname(disk_fd), status);
return MDADM_STATUS_ERROR;
}
return MDADM_STATUS_SUCCESS;
}
/**
* is_sec_prot_01h_supported() - check if security protocol 01h supported.
* @security_protocols: struct with response from disk (NVMe, SATA) describing supported
* security protocols.
*
* Return: true if TCG_SECP_01 found, false otherwise.
*/
static bool is_sec_prot_01h_supported(supported_security_protocols_t *security_protocols)
{
int list_length = be16toh(security_protocols->list_length);
int index;
for (index = 0 ; index < list_length; index++) {
if (security_protocols->list[index] == TCG_SECP_01)
return true;
}
return false;
}
/**
* is_sec_prot_01h_supported_nvme() - check if security protocol 01h supported for given NVMe disk.
* @disk_fd: a disk file descriptor.
* @verbose: verbose flag.
*
* Return: %DRIVE_FEAT_SUP_ST if TCG_SECP_01 supported, %DRIVE_FEAT_NOT_SUP_ST if not supported,
* %DRIVE_FEAT_CHECK_FAILED_ST if failed to check.
*/
static drive_feat_sup_st is_sec_prot_01h_supported_nvme(int disk_fd, const int verbose)
{
supported_security_protocols_t security_protocols = {0};
/* security_protocol: TCG_SECP_00, comm_id: not applicable */
if (nvme_security_recv_ioctl(disk_fd, TCG_SECP_00, 0x0, &security_protocols,
sizeof(security_protocols), verbose))
return DRIVE_FEAT_CHECK_FAILED_ST;
if (is_sec_prot_01h_supported(&security_protocols))
return DRIVE_FEAT_SUP_ST;
return DRIVE_FEAT_NOT_SUP_ST;
}
/**
* is_nvme_sec_send_recv_supported() - check if Security Send and Security Receive is supported.
* @disk_fd: a disk file descriptor.
* @verbose: verbose flag.
*
* Check if "Optional Admin Command Support" bit 0 is set in NVMe identify.
* Bit 0 set to 1 means controller supports the Security Send and Security Receive commands.
*
* Return: %DRIVE_FEAT_SUP_ST if security send/receive supported,
* %DRIVE_FEAT_NOT_SUP_ST if not supported, %DRIVE_FEAT_CHECK_FAILED_ST if check failed.
*/
static drive_feat_sup_st is_nvme_sec_send_recv_supported(int disk_fd, const int verbose)
{
nvme_identify_ctrl_t nvme_identify = {0};
int status = 0;
status = nvme_identify_ioctl(disk_fd, &nvme_identify, sizeof(nvme_identify), verbose);
if (status)
return DRIVE_FEAT_CHECK_FAILED_ST;
if ((__le16_to_cpu(nvme_identify.oacs) & 0x1) == 0x1)
return DRIVE_FEAT_SUP_ST;
return DRIVE_FEAT_NOT_SUP_ST;
}
/**
* get_opal_encryption_information() - get Opal encryption information.
* @buffer: buffer with Opal Level 0 Discovery response.
* @information: struct to fill out, describing encryption status of disk.
*
* If Locking feature frame is in response from Opal Level 0 discovery, &encryption_information_t
* structure is completed with status and ability otherwise the status is set to &None.
* For possible encryption statuses and abilities,
* please refer to enums &encryption_status and &encryption_ability.
*
* Return: %MDADM_STATUS_SUCCESS on success, %MDADM_STATUS_ERROR otherwise.
*/
static mdadm_status_t get_opal_encryption_information(__u8 *buffer,
encryption_information_t *information)
{
opal_locking_feature_t *opal_locking_feature =
get_opal_locking_feature_description(buffer);
if (!opal_locking_feature)
return MDADM_STATUS_ERROR;
if (opal_locking_feature->locking_supported == 1) {
information->ability = ENC_ABILITY_SED;
if (opal_locking_feature->locking_enabled == 0)
information->status = ENC_STATUS_UNENCRYPTED;
else if (opal_locking_feature->locked == 1)
information->status = ENC_STATUS_LOCKED;
else
information->status = ENC_STATUS_UNLOCKED;
} else {
information->ability = ENC_ABILITY_NONE;
information->status = ENC_STATUS_UNENCRYPTED;
}
return MDADM_STATUS_SUCCESS;
}
/**
* get_nvme_opal_encryption_information() - get NVMe Opal encryption information.
* @disk_fd: a disk file descriptor.
* @information: struct to fill out, describing encryption status of disk.
* @verbose: verbose flag.
*
* In case the disk supports Opal Level 0 discovery, &encryption_information_t structure
* is completed with status and ability based on ioctl response,
* otherwise the ability is set to %ENC_ABILITY_NONE and &status to %ENC_STATUS_UNENCRYPTED.
* As the current use case does not need the knowledge of Opal support, if there is no support,
* %MDADM_STATUS_SUCCESS will be returned, with the values described above.
* For possible encryption statuses and abilities,
* please refer to enums &encryption_status and &encryption_ability.
*
* %MDADM_STATUS_SUCCESS on success, %MDADM_STATUS_ERROR otherwise.
*/
mdadm_status_t
get_nvme_opal_encryption_information(int disk_fd, encryption_information_t *information,
const int verbose)
{
__u8 buffer[OPAL_IO_BUFFER_LEN];
int sec_send_recv_supported = 0;
int protocol_01h_supported = 0;
mdadm_status_t status;
information->ability = ENC_ABILITY_NONE;
information->status = ENC_STATUS_UNENCRYPTED;
sec_send_recv_supported = is_nvme_sec_send_recv_supported(disk_fd, verbose);
if (sec_send_recv_supported == DRIVE_FEAT_CHECK_FAILED_ST)
return MDADM_STATUS_ERROR;
/* Opal not supported */
if (sec_send_recv_supported == DRIVE_FEAT_NOT_SUP_ST)
return MDADM_STATUS_SUCCESS;
/**
* sec_send_recv_supported determine that it should be possible to read
* supported sec protocols
*/
protocol_01h_supported = is_sec_prot_01h_supported_nvme(disk_fd, verbose);
if (protocol_01h_supported == DRIVE_FEAT_CHECK_FAILED_ST)
return MDADM_STATUS_ERROR;
/* Opal not supported */
if (sec_send_recv_supported == DRIVE_FEAT_SUP_ST &&
protocol_01h_supported == DRIVE_FEAT_NOT_SUP_ST)
return MDADM_STATUS_SUCCESS;
if (nvme_security_recv_ioctl(disk_fd, TCG_SECP_01, OPAL_DISCOVERY_COMID, (void *)&buffer,
OPAL_IO_BUFFER_LEN, verbose))
return MDADM_STATUS_ERROR;
status = get_opal_encryption_information((__u8 *)&buffer, information);
if (status)
pr_vrb("Locking feature description not found in Level 0 discovery response. Device /dev/%s.\n",
fd2kname(disk_fd));
if (information->ability == ENC_ABILITY_NONE)
assert(information->status == ENC_STATUS_UNENCRYPTED);
return status;
}
/**
* ata_pass_through12_ioctl() - ata pass through12 ioctl.
* @disk_fd: a disk file descriptor.
* @ata_command: ata command.
* @sec_protocol: security protocol.
* @comm_id: additional command id.
* @response_buffer: response buffer to fill out.
* @buf_size: response buffer size.
* @verbose: verbose flag.
*
* Based on the documentations ATA Command Pass-Through, chapter 13.2.2 and
* ATA Translation - 3 (SAT-3), send read ata pass through 12 command via ioctl().
* On success, @response_buffer will be completed.
*
* Return: %MDADM_STATUS_SUCCESS on success, %MDADM_STATUS_ERROR on fail.
*/
static mdadm_status_t
ata_pass_through12_ioctl(int disk_fd, __u8 ata_command, __u8 sec_protocol, __u16 comm_id,
void *response_buffer, size_t buf_size, const int verbose)
{
__u8 cdb[ATA_INQUIRY_LENGTH] = {0};
__u8 sense[SG_SENSE_SIZE] = {0};
__u8 *sense_desc = NULL;
sg_io_hdr_t sg = {0};
/*
* ATA Command Pass-Through, chapter 13.2.2
* SCSI Primary Commands - 4 (SPC-4)
* ATA Translation - 3 (SAT-3)
*/
cdb[0] = ATA_PASS_THROUGH_12;
/* protocol, bits 1-4 */
cdb[1] = ATA_PIO_DATA_IN << 1;
/* Bytes: CK_COND=1, T_DIR = 1, BYTE_BLOCK = 1, Length in Sector Count = 2 */
cdb[2] = 0x2E;
cdb[3] = sec_protocol;
/* Sector count */
cdb[4] = buf_size / DEFAULT_SECTOR_SIZE;
cdb[6] = (comm_id) & 0xFF;
cdb[7] = (comm_id >> 8) & 0xFF;
cdb[9] = ata_command;
sg.interface_id = SG_INTERFACE_ID;
sg.cmd_len = sizeof(cdb);
sg.mx_sb_len = sizeof(sense);
sg.dxfer_direction = SG_DXFER_FROM_DEV;
sg.dxfer_len = buf_size;
sg.dxferp = response_buffer;
sg.cmdp = cdb;
sg.sbp = sense;
sg.timeout = SG_IO_TIMEOUT;
sg.usr_ptr = NULL;
if (ioctl(disk_fd, SG_IO, &sg) < 0) {
pr_vrb("Failed ata passthrough12 ioctl. Device: /dev/%s.\n", fd2kname(disk_fd));
return MDADM_STATUS_ERROR;
}
if ((sg.status && sg.status != SG_CHECK_CONDITION) || sg.host_status ||
(sg.driver_status && sg.driver_status != SG_DRIVER_SENSE)) {
pr_vrb("Failed ata passthrough12 ioctl. Device: /dev/%s.\n", fd2kname(disk_fd));
pr_vrb("SG_IO error: ATA_12 Status: %d Host Status: %d, Driver Status: %d\n",
sg.status, sg.host_status, sg.driver_status);
return MDADM_STATUS_ERROR;
}
/* verify expected sense response code */
if (!(sense[0] == SENSE_DATA_CURRENT_DESC || sense[0] == SENSE_DATA_CURRENT_FIXED)) {
pr_vrb("Failed ata passthrough12 ioctl. Device: /dev/%s.\n", fd2kname(disk_fd));
return MDADM_STATUS_ERROR;
}
sense_desc = sense + SENSE_CURRENT_RES_DESC_POS;
/* verify sense data current response with descriptor format */
if (sense[0] == SENSE_DATA_CURRENT_DESC &&
!(sense_desc[0] == ATA_STATUS_RETURN_DESCRIPTOR &&
sense_desc[1] == ATA_INQUIRY_LENGTH)) {
pr_vrb("Failed ata passthrough12 ioctl. Device: /dev/%s. Sense data ASC: %d, ASCQ: %d.\n",
fd2kname(disk_fd), sense[2], sense[3]);
return MDADM_STATUS_ERROR;
}
/* verify sense data current response with fixed format */
if (sense[0] == SENSE_DATA_CURRENT_FIXED &&
!(sense[12] == ATA_PT_INFORMATION_AVAILABLE_ASC &&
sense[13] == ATA_PT_INFORMATION_AVAILABLE_ASCQ)) {
pr_vrb("Failed ata passthrough12 ioctl. Device: /dev/%s. Sense data ASC: %d, ASCQ: %d.\n",
fd2kname(disk_fd), sense[12], sense[13]);
return MDADM_STATUS_ERROR;
}
return MDADM_STATUS_SUCCESS;
}
/**
* is_sec_prot_01h_supported_ata() - check if security protocol 01h supported for given SATA disk.
* @disk_fd: a disk file descriptor.
* @verbose: verbose flag.
*
* Return: %DRIVE_FEAT_SUP_ST if TCG_SECP_01 supported, %DRIVE_FEAT_NOT_SUP_ST if not supported,
* %DRIVE_FEAT_CHECK_FAILED_ST if failed.
*/
static drive_feat_sup_st is_sec_prot_01h_supported_ata(int disk_fd, const int verbose)
{
supported_security_protocols_t security_protocols;
mdadm_status_t result = ata_pass_through12_ioctl(disk_fd, ATA_TRUSTED_RECEIVE, TCG_SECP_00,
0x0, &security_protocols,
sizeof(security_protocols), verbose);
if (result)
return DRIVE_FEAT_CHECK_FAILED_ST;
if (is_sec_prot_01h_supported(&security_protocols))
return DRIVE_FEAT_SUP_ST;
return DRIVE_FEAT_NOT_SUP_ST;
}
/**
* is_ata_trusted_computing_supported() - check if ata trusted computing supported.
* @buffer: buffer with ATA identify response, not NULL.
*
* Return: true if trusted computing bit set, false otherwise.
*/
bool is_ata_trusted_computing_supported(__u16 *buffer)
{
/* Added due to warnings from the compiler about a possible uninitialized variable below. */
assert(buffer);
__u16 security_tc_frame = __le16_to_cpu(buffer[ATA_TRUSTED_COMPUTING_POS]);
ata_trusted_computing_t *security_tc = (ata_trusted_computing_t *)&security_tc_frame;
if (security_tc->tc_feature == 1)
return true;
return false;
}
/**
* get_ata_standard_security_status() - get ATA disk encryption information from ATA identify.
* @buffer: buffer with response from ATA identify, not NULL.
* @information: struct to fill out, describing encryption status of disk.
*
* The function based on the Security status frame from ATA identify,
* completed encryption information.
* For possible encryption statuses and abilities,
* please refer to enums &encryption_status and &encryption_ability.
*
* Return: %MDADM_STATUS_SUCCESS on success, %MDADM_STATUS_ERROR on fail.
*/
static mdadm_status_t get_ata_standard_security_status(__u16 *buffer,
struct encryption_information *information)
{
/* Added due to warnings from the compiler about a possible uninitialized variable below. */
assert(buffer);
__u16 security_status_frame = __le16_to_cpu(buffer[ATA_SECURITY_WORD_POSITION]);
ata_security_status_t *security_status = (ata_security_status_t *)&security_status_frame;
if (!security_status->security_supported) {
information->ability = ENC_ABILITY_NONE;
information->status = ENC_STATUS_UNENCRYPTED;
return MDADM_STATUS_SUCCESS;
}
information->ability = ENC_ABILITY_OTHER;
if (security_status->security_enabled == 0)
information->status = ENC_STATUS_UNENCRYPTED;
else if (security_status->security_locked == 1)
information->status = ENC_STATUS_LOCKED;
else
information->status = ENC_STATUS_UNLOCKED;
return MDADM_STATUS_SUCCESS;
}
/**
* is_ata_opal() - check if SATA disk support Opal.
* @disk_fd: a disk file descriptor.
* @buffer: buffer with ATA identify response.
* @verbose: verbose flag.
*
* Return: %DRIVE_FEAT_SUP_ST if TCG_SECP_01 supported, %DRIVE_FEAT_NOT_SUP_ST if not supported,
* %DRIVE_FEAT_CHECK_FAILED_ST if failed to check.
*/
static drive_feat_sup_st is_ata_opal(int disk_fd, __u16 *buffer_identify, const int verbose)
{
bool tc_status = is_ata_trusted_computing_supported(buffer_identify);
drive_feat_sup_st tcg_sec_prot_status;
if (!tc_status)
return DRIVE_FEAT_NOT_SUP_ST;
tcg_sec_prot_status = is_sec_prot_01h_supported_ata(disk_fd, verbose);
if (tcg_sec_prot_status == DRIVE_FEAT_CHECK_FAILED_ST) {
pr_vrb("Failed to verify if security protocol 01h supported. Device /dev/%s.\n",
fd2kname(disk_fd));
return DRIVE_FEAT_CHECK_FAILED_ST;
}
if (tc_status && tcg_sec_prot_status == DRIVE_FEAT_SUP_ST)
return DRIVE_FEAT_SUP_ST;
return DRIVE_FEAT_NOT_SUP_ST;
}
/**
* get_ata_encryption_information() - get ATA disk encryption information.
* @disk_fd: a disk file descriptor.
* @information: struct to fill out, describing encryption status of disk.
* @verbose: verbose flag.
*
* The function reads information about encryption, if the disk supports Opal,
* the information is completed based on Opal Level 0 discovery, otherwise,
* based on ATA security status frame from ATA identification response.
* For possible encryption statuses and abilities,
* please refer to enums &encryption_status and &encryption_ability.
*
* Based on the documentations ATA/ATAPI Command Set ATA8-ACS and
* AT Attachment-8 - ATA Serial Transport (ATA8-AST).
*
* Return: %MDADM_STATUS_SUCCESS on success, %MDADM_STATUS_ERROR on fail.
*/
mdadm_status_t
get_ata_encryption_information(int disk_fd, struct encryption_information *information,
const int verbose)
{
__u8 buffer_opal_level0_discovery[OPAL_IO_BUFFER_LEN] = {0};
__u16 buffer_identify[ATA_IDENTIFY_RESPONSE_LEN] = {0};
drive_feat_sup_st ata_opal_status;
mdadm_status_t status;
/* Get disk ATA identification */
status = ata_pass_through12_ioctl(disk_fd, ATA_IDENTIFY, 0x0, 0x0, buffer_identify,
sizeof(buffer_identify), verbose);
if (status == MDADM_STATUS_ERROR)
return MDADM_STATUS_ERROR;
/* Possible OPAL support, further checks require tpm_enabled.*/
if (is_ata_trusted_computing_supported(buffer_identify)) {
/* OPAL SATA encryption checking disabled. */
if (conf_get_sata_opal_encryption_no_verify())
return MDADM_STATUS_SUCCESS;
if (!sysfs_is_libata_allow_tpm_enabled(verbose)) {
pr_vrb("Detected SATA drive /dev/%s with Trusted Computing support.\n",
fd2kname(disk_fd));
pr_vrb("Cannot verify encryption state. Requires libata.tpm_enabled=1.\n");
return MDADM_STATUS_ERROR;
}
}
ata_opal_status = is_ata_opal(disk_fd, buffer_identify, verbose);
if (ata_opal_status == DRIVE_FEAT_CHECK_FAILED_ST)
return MDADM_STATUS_ERROR;
if (ata_opal_status == DRIVE_FEAT_NOT_SUP_ST)
return get_ata_standard_security_status(buffer_identify, information);
/* SATA Opal */
status = ata_pass_through12_ioctl(disk_fd, ATA_TRUSTED_RECEIVE, TCG_SECP_01,
OPAL_DISCOVERY_COMID, buffer_opal_level0_discovery,
OPAL_IO_BUFFER_LEN, verbose);
if (status != MDADM_STATUS_SUCCESS)
return MDADM_STATUS_ERROR;
return get_opal_encryption_information(buffer_opal_level0_discovery, information);
}

37
drive_encryption.h Normal file
View file

@ -0,0 +1,37 @@
/* SPDX-License-Identifier: GPL-2.0-only */
/*
* Read encryption information for Opal and ATA devices.
*
* Copyright (C) 2024 Intel Corporation
* Author: Blazej Kucman <blazej.kucman@intel.com>
*/
typedef enum encryption_status {
/* The drive is not currently encrypted. */
ENC_STATUS_UNENCRYPTED = 0,
/* The drive is encrypted and the data is not accessible. */
ENC_STATUS_LOCKED,
/* The drive is encrypted but the data is accessible in unencrypted form. */
ENC_STATUS_UNLOCKED
} encryption_status_t;
typedef enum encryption_ability {
ENC_ABILITY_NONE = 0,
ENC_ABILITY_OTHER,
/* Self encrypted drive */
ENC_ABILITY_SED
} encryption_ability_t;
typedef struct encryption_information {
encryption_ability_t ability;
encryption_status_t status;
} encryption_information_t;
mdadm_status_t
get_nvme_opal_encryption_information(int disk_fd, struct encryption_information *information,
const int verbose);
mdadm_status_t
get_ata_encryption_information(int disk_fd, struct encryption_information *information,
const int verbose);
const char *get_encryption_ability_string(enum encryption_ability ability);
const char *get_encryption_status_string(enum encryption_status status);

284
inventory
View file

@ -1,284 +0,0 @@
.gitignore
ANNOUNCE-3.0
ANNOUNCE-3.0.1
ANNOUNCE-3.0.2
ANNOUNCE-3.0.3
ANNOUNCE-3.1
ANNOUNCE-3.1.1
ANNOUNCE-3.1.2
ANNOUNCE-3.1.3
ANNOUNCE-3.1.4
ANNOUNCE-3.1.5
ANNOUNCE-3.2
ANNOUNCE-3.2.1
ANNOUNCE-3.2.2
ANNOUNCE-3.2.3
ANNOUNCE-3.2.4
ANNOUNCE-3.2.5
ANNOUNCE-3.2.6
ANNOUNCE-3.3
ANNOUNCE-3.3.1
ANNOUNCE-3.3.2
ANNOUNCE-3.3.3
ANNOUNCE-3.3.4
ANNOUNCE-3.4
ANNOUNCE-4.0
ANNOUNCE-4.1
ANNOUNCE-4.2
Assemble.c
Build.c
COPYING
ChangeLog
Create.c
Detail.c
Dump.c
Examine.c
Grow.c
INSTALL
Incremental.c
Kill.c
Makefile
Manage.c
Monitor.c
Query.c
README.initramfs
ReadMe.c
TODO
bitmap.c
bitmap.h
clustermd_tests/
clustermd_tests/00r10_Create
clustermd_tests/00r1_Create
clustermd_tests/01r10_Grow_bitmap-switch
clustermd_tests/01r10_Grow_resize
clustermd_tests/01r1_Grow_add
clustermd_tests/01r1_Grow_bitmap-switch
clustermd_tests/01r1_Grow_resize
clustermd_tests/02r10_Manage_add
clustermd_tests/02r10_Manage_add-spare
clustermd_tests/02r10_Manage_re-add
clustermd_tests/02r1_Manage_add
clustermd_tests/02r1_Manage_add-spare
clustermd_tests/02r1_Manage_re-add
clustermd_tests/03r10_switch-recovery
clustermd_tests/03r10_switch-resync
clustermd_tests/03r1_switch-recovery
clustermd_tests/03r1_switch-resync
clustermd_tests/cluster_conf
clustermd_tests/func.sh
config.c
coverity-gcc-hack.h
crc32.c
crc32.h
crc32c.c
dlink.c
dlink.h
external-reshape-design.txt
inventory
lib.c
makedist
managemon.c
mapfile.c
maps.c
md.4
md5.h
md_p.h
md_u.h
mdadm.8.in
mdadm.c
mdadm.conf-example
mdadm.conf.5
mdadm.h
mdadm.spec
mdmon-design.txt
mdmon.8
mdmon.c
mdmon.h
mdopen.c
mdstat.c
misc/
misc/mdcheck
misc/syslog-events
mkinitramfs
monitor.c
msg.c
msg.h
part.h
platform-intel.c
platform-intel.h
policy.c
probe_roms.c
probe_roms.h
pwgr.c
raid5extend.c
raid6check.8
raid6check.c
restripe.c
sg_io.c
sha1.c
sha1.h
super-ddf.c
super-gpt.c
super-intel.c
super-mbr.c
super0.c
super1.c
swap_super.c
sysfs.c
systemd/
systemd/SUSE-mdadm_env.sh
systemd/mdadm-grow-continue@.service
systemd/mdadm-last-resort@.service
systemd/mdadm-last-resort@.timer
systemd/mdadm.shutdown
systemd/mdcheck_continue.service
systemd/mdcheck_continue.timer
systemd/mdcheck_start.service
systemd/mdcheck_start.timer
systemd/mdmon@.service
systemd/mdmonitor-oneshot.service
systemd/mdmonitor-oneshot.timer
systemd/mdmonitor.service
test
tests/
tests/00linear
tests/00multipath
tests/00names
tests/00raid0
tests/00raid1
tests/00raid10
tests/00raid4
tests/00raid5
tests/00raid6
tests/00readonly
tests/01r1fail
tests/01r5fail
tests/01r5integ
tests/01raid6integ
tests/01replace
tests/02lineargrow
tests/02r1add
tests/02r1grow
tests/02r5grow
tests/02r6grow
tests/03assem-incr
tests/03r0assem
tests/03r5assem
tests/03r5assem-failed
tests/03r5assemV1
tests/04r0update
tests/04r1update
tests/04r5swap
tests/04update-metadata
tests/04update-uuid
tests/05r1-add-internalbitmap
tests/05r1-add-internalbitmap-v1a
tests/05r1-add-internalbitmap-v1b
tests/05r1-add-internalbitmap-v1c
tests/05r1-bitmapfile
tests/05r1-failfast
tests/05r1-grow-external
tests/05r1-grow-internal
tests/05r1-grow-internal-1
tests/05r1-internalbitmap
tests/05r1-internalbitmap-v1a
tests/05r1-internalbitmap-v1b
tests/05r1-internalbitmap-v1c
tests/05r1-n3-bitmapfile
tests/05r1-re-add
tests/05r1-re-add-nosuper
tests/05r1-remove-internalbitmap
tests/05r1-remove-internalbitmap-v1a
tests/05r1-remove-internalbitmap-v1b
tests/05r1-remove-internalbitmap-v1c
tests/05r5-bitmapfile
tests/05r5-internalbitmap
tests/05r6-bitmapfile
tests/05r6tor0
tests/06name
tests/06sysfs
tests/06wrmostly
tests/07autoassemble
tests/07autodetect
tests/07changelevelintr
tests/07changelevels
tests/07layouts
tests/07reshape5intr
tests/07revert-grow
tests/07revert-inplace
tests/07revert-shrink
tests/07testreshape5
tests/09imsm-assemble
tests/09imsm-create-fail-rebuild
tests/09imsm-overlap
tests/10ddf-assemble-missing
tests/10ddf-create
tests/10ddf-create-fail-rebuild
tests/10ddf-fail-create-race
tests/10ddf-fail-readd
tests/10ddf-fail-readd-readonly
tests/10ddf-fail-spare
tests/10ddf-fail-stop-readd
tests/10ddf-fail-twice
tests/10ddf-fail-two-spares
tests/10ddf-geometry
tests/10ddf-incremental-wrong-order
tests/10ddf-sudden-degraded
tests/11spare-migration
tests/12imsm-r0_2d-grow-r0_3d
tests/12imsm-r0_2d-grow-r0_4d
tests/12imsm-r0_2d-grow-r0_5d
tests/12imsm-r0_3d-grow-r0_4d
tests/12imsm-r5_3d-grow-r5_4d
tests/12imsm-r5_3d-grow-r5_5d
tests/13imsm-r0_r0_2d-grow-r0_r0_4d
tests/13imsm-r0_r0_2d-grow-r0_r0_5d
tests/13imsm-r0_r0_3d-grow-r0_r0_4d
tests/13imsm-r0_r5_3d-grow-r0_r5_4d
tests/13imsm-r0_r5_3d-grow-r0_r5_5d
tests/13imsm-r5_r0_3d-grow-r5_r0_4d
tests/13imsm-r5_r0_3d-grow-r5_r0_5d
tests/14imsm-r0_3d-r5_3d-migrate-r5_4d-r5_4d
tests/14imsm-r0_3d_no_spares-migrate-r5_3d
tests/14imsm-r0_r0_2d-takeover-r10_4d
tests/14imsm-r10_4d-grow-r10_5d
tests/14imsm-r10_r5_4d-takeover-r0_2d
tests/14imsm-r1_2d-grow-r1_3d
tests/14imsm-r1_2d-takeover-r0_2d
tests/14imsm-r5_3d-grow-r5_5d-no-spares
tests/14imsm-r5_3d-migrate-r4_3d
tests/15imsm-r0_3d_64k-migrate-r0_3d_256k
tests/15imsm-r5_3d_4k-migrate-r5_3d_256k
tests/15imsm-r5_3d_64k-migrate-r5_3d_256k
tests/15imsm-r5_6d_4k-migrate-r5_6d_256k
tests/15imsm-r5_r0_3d_64k-migrate-r5_r0_3d_256k
tests/16imsm-r0_3d-migrate-r5_4d
tests/16imsm-r0_5d-migrate-r5_6d
tests/16imsm-r5_3d-migrate-r0_3d
tests/16imsm-r5_5d-migrate-r0_5d
tests/18imsm-1d-takeover-r0_1d
tests/18imsm-1d-takeover-r1_2d
tests/18imsm-r0_2d-takeover-r10_4d
tests/18imsm-r10_4d-takeover-r0_2d
tests/18imsm-r1_2d-takeover-r0_1d
tests/19raid6auto-repair
tests/19raid6check
tests/19raid6repair
tests/19repair-does-not-destroy
tests/20raid5journal
tests/21raid5cache
tests/ToTest
tests/env-ddf-template
tests/env-imsm-template
tests/func.sh
tests/imsm-grow-template
tests/utils
udev-md-clustered-confirm-device.rules
udev-md-raid-arrays.rules
udev-md-raid-assembly.rules
udev-md-raid-creating.rules
udev-md-raid-safe-timeouts.rules
util.c
uuid.c
xmalloc.c

96
makedist
View file

@ -1,96 +0,0 @@
#!/bin/sh
# avoid silly sorting
export LANG=C
arg=$1
target=~/public_html/source/mdadm
if [ " $arg" = " test" ]
then
target=/tmp/mdadm-test
rm -rf $target
mkdir -p $target
fi
if [ -d $target ]
then :
else echo $target is not a directory
exit 2
fi
set `grep '^#define VERSION' ReadMe.c `
version=`echo $3 | sed -e 's/"//g'`
grep "^.TH MDADM 8 .. v$version" mdadm.8.in > /dev/null 2>&1 ||
{
echo mdadm.8.in does not mention version $version.
exit 1
}
grep "^.TH MDMON 8 .. v$version" mdmon.8 > /dev/null 2>&1 ||
{
echo mdmon.8 does not mention version $version.
exit 1
}
rpmv=`echo $version | tr - _`
grep "^Version: *$rpmv$" mdadm.spec > /dev/null 2>&1 ||
{
echo mdadm.spec does not mention version $version.
exit 1
}
if [ -f ANNOUNCE-$version ]
then :
else
echo ANNOUNCE-$version does not exist
exit 1
fi
if grep "^ANNOUNCE-$version\$" inventory
then :
else { cat inventory ; echo ANNOUNCE-$version ; } | sort -o inventory
fi
echo version = $version
base=mdadm-$rpmv.tar.gz
if [ " $arg" != " diff" ]
then
if [ -f $target/$base ]
then
echo $target/$base exists.
exit 1
fi
trap "rm $target/$base; exit" 1 2 3
git archive --prefix=mdadm-$rpmv/ HEAD | gzip --best > $target/$base
chmod a+r $target/$base
ls -l $target/$base
if tar tzf $target/$base | sed 's,[^/]*/,,' | sort | diff -u inventory -
then : correct files found
else echo "Extra files, or inventory is out-of-date"
rm $target/$base
exit 1
fi
rpmbuild -ta $target/$base || exit 1
find ~/rpmbuild/RPMS -name "*mdadm-$version-*" \
-exec cp {} $target/RPM \;
cp ANNOUNCE-$version $target/ANNOUNCE
cp ChangeLog $target/ChangeLog
if [ " $arg" != " test" ]
then
echo -n "Confirm signing this release? "
read a
if [ " $a" != " y" ]; then echo OK - bye. ; exit 1; fi
if zcat $target/$base | gpg -ba > $target/$base.sign && gpg -ba $target/ANNOUNCE
then
kup put $target/$base $target/$base.sign \
/pub/linux/utils/raid/mdadm/mdadm-$version.tar.gz
kup put $target/ANNOUNCE $target/ANNOUNCE.asc /pub/linux/utils/raid/mdadm/ANNOUNCE
else
echo signing failed
exit 1
fi
fi
else
if [ ! -f $target/$base ]
then
echo $target/$base does not exist.
exit 1
fi
( cd .. ; ln -s mdadm.v2 mdadm-$version ; tar chf - --exclude=.git --exclude="TAGS" --exclude='*,v' --exclude='*~' --exclude='*.o' --exclude mdadm --exclude=mdadm'.[^ch0-9]' --exclude=RCS mdadm-$version ; rm mdadm-$version ) | gzip --best > /var/tmp/mdadm-new.tgz
mkdir /var/tmp/mdadm-old ; zcat $target/$base | ( cd /var/tmp/mdadm-old ; tar xf - )
mkdir /var/tmp/mdadm-new ; zcat /var/tmp/mdadm-new.tgz | ( cd /var/tmp/mdadm-new ; tar xf - )
diff -ru /var/tmp/mdadm-old /var/tmp/mdadm-new
rm -rf /var/tmp/mdadm-old /var/tmp/mdadm-new /var/tmp/mdadm-new.tgz
fi

21
mdadm.8.in
View file

@ -3179,7 +3179,7 @@ environment. This can be useful for testing or for disaster
recovery. You should be aware that interoperability may be
compromised by setting this value.
These change can also be suppressed by adding
These change can also be suppressed by adding
.B mdadm.imsm.test=1
to the kernel command line. This makes it easy to test IMSM
code in a virtual machine that doesn't have IMSM virtual hardware.
@ -3454,6 +3454,25 @@ is any string. These names are supported by
since version 3.3 provided they are enabled in
.IR mdadm.conf .
.SH UNDERSTANDING OUTPUT
.TP
EXAMINE
.TP
.B checkpoint
Checkpoint value is reported when array is performing some action including
resync, recovery or reshape. Checkpoints allow resuming action from certain
point if it was interrupted.
Checkpoint is reported as combination of two values: current migration unit
and number of blocks per unit. By multiplying those values and dividing by
array size checkpoint progress percentage can be obtained in relation to
current progress reported in /proc/mdstat. Checkpoint is also related to (and
sometimes based on) sysfs entry sync_completed but depending on action units
may differ. Even if units are the same, it should not be expected that
checkpoint and sync_completed will be exact match nor updated simultaneously.
.SH NOTE
.I mdadm
was previously known as

16
mdadm.conf.5.in
View file

@ -636,6 +636,20 @@ If multiple
.B MINITORDELAY
lines are provided, only first non-zero value is considered.
.TP
.B ENCRYPTION_NO_VERIFY
The
.B ENCRYPTION_NO_VERIFY
disables encryption verification for devices with particular encryption support detected.
Currently, only verification of SATA OPAL encryption can be disabled.
It does not disable ATA security encryption verification.
Currently effective only for
.I IMSM
metadata.
Available parameter
.I "sata_opal".
.SH FILES
.SS {CONFFILE}
@ -744,6 +758,8 @@ SYSFS uuid=bead5eb6:31c17a27:da120ba2:7dfda40d group_thread_cnt=4
sync_speed_max=1000000
.br
MONITORDELAY 60
.br
ENCRYPTION_NO_VERIFY sata_opal
.SH SEE ALSO
.BR mdadm (8),

104
mdadm.h
View file

@ -430,8 +430,10 @@ struct createinfo {
};
struct spare_criteria {
bool criteria_set;
unsigned long long min_size;
unsigned int sector_size;
struct dev_policy *pols;
};
typedef enum mdadm_status {
@ -776,6 +778,8 @@ enum sysfs_read_flags {
#define SYSFS_MAX_BUF_SIZE 64
extern void sysfs_get_container_devnm(struct mdinfo *mdi, char *buf);
/* If fd >= 0, get the array it is open on,
* else use devnm.
*/
@ -807,7 +811,7 @@ extern int sysfs_attribute_available(struct mdinfo *sra, struct mdinfo *dev,
extern int sysfs_get_str(struct mdinfo *sra, struct mdinfo *dev,
char *name, char *val, int size);
extern int sysfs_set_safemode(struct mdinfo *sra, unsigned long ms);
extern int sysfs_set_array(struct mdinfo *info, int vers);
extern int sysfs_set_array(struct mdinfo *info);
extern int sysfs_add_disk(struct mdinfo *sra, struct mdinfo *sd, int resume);
extern int sysfs_disk_to_scsi_id(int fd, __u32 *id);
extern int sysfs_unique_holder(char *devnm, long rdev);
@ -849,6 +853,7 @@ extern int restore_stripes(int *dest, unsigned long long *offsets,
int source, unsigned long long read_offset,
unsigned long long start, unsigned long long length,
char *src_buf);
extern bool sysfs_is_libata_allow_tpm_enabled(const int verbose);
#ifndef Sendmail
#define Sendmail "/usr/lib/sendmail -t"
@ -937,6 +942,23 @@ struct reshape {
unsigned long long new_size; /* New size of array in sectors */
};
/**
* struct dev_policy - Data structure for policy management.
* @next: pointer to next dev_policy.
* @name: policy name, category.
* @metadata: the metadata type it affects.
* @value: value of the policy.
*
* The functions to manipulate dev_policy lists do not free elements, so they must be statically
* allocated. @name and @metadata can be compared by address.
*/
typedef struct dev_policy {
struct dev_policy *next;
char *name;
const char *metadata;
const char *value;
} dev_policy_t;
/* A superswitch provides entry point to a metadata handler.
*
* The superswitch primarily operates on some "metadata" that
@ -1115,10 +1137,9 @@ extern struct superswitch {
* Return spare criteria for array:
* - minimum disk size can be used in array;
* - sector size can be used in array.
* Return values: 0 - for success and -EINVAL on error.
*/
int (*get_spare_criteria)(struct supertype *st,
struct spare_criteria *sc);
mdadm_status_t (*get_spare_criteria)(struct supertype *st, char *mddev_path,
struct spare_criteria *sc);
/* Find somewhere to put a bitmap - possibly auto-size it - and
* update the metadata to record this. The array may be newly
* created, in which case data_size may be updated, or it might
@ -1166,6 +1187,25 @@ extern struct superswitch {
char *subdev, unsigned long long *freesize,
int consistency_policy, int verbose);
/**
* test_and_add_drive_policies() - test new and add custom policies from metadata handler.
* @pols: list of currently recorded policies.
* @disk_fd: file descriptor of the device to check.
* @verbose: verbose flag.
*
* Used by IMSM to verify all drives in container/array, against requirements not recored
* in superblock, like controller type for IMSM. It should check all drives even if
* they are not actually used, because mdmon or kernel are free to use any drive assigned to
* container automatically.
*
* Generating and comparison methods belong to metadata handler. It is not mandatory to be
* implemented.
*
* Return: MDADM_STATUS_SUCCESS is expected on success.
*/
mdadm_status_t (*test_and_add_drive_policies)(dev_policy_t **pols, int disk_fd,
const int verbose);
/* Return a linked list of 'mdinfo' structures for all arrays
* in the container. For non-containers, it is like
* getinfo_super with an allocated mdinfo.*/
@ -1247,21 +1287,6 @@ extern struct superswitch {
*/
struct mdinfo *(*activate_spare)(struct active_array *a,
struct metadata_update **updates);
/*
* Return statically allocated string that represents metadata specific
* controller domain of the disk. The domain is used in disk domain
* matching functions. Disks belong to the same domain if the they have
* the same domain from mdadm.conf and belong the same metadata domain.
* Returning NULL or not providing this handler means that metadata
* does not distinguish the differences between disks that belong to
* different controllers. They are in the domain specified by
* configuration file (mdadm.conf).
* In case when the metadata has the notion of domains based on disk
* it shall return NULL for disks that do not belong to the controller
* the supported domains. Such disks will form another domain and won't
* be mixed with supported ones.
*/
const char *(*get_disk_controller_domain)(const char *path);
/* for external backup area */
int (*recover_backup)(struct supertype *st, struct mdinfo *info);
@ -1368,27 +1393,8 @@ extern struct supertype *dup_super(struct supertype *st);
extern int get_dev_size(int fd, char *dname, unsigned long long *sizep);
extern int get_dev_sector_size(int fd, char *dname, unsigned int *sectsizep);
extern int must_be_container(int fd);
extern int dev_size_from_id(dev_t id, unsigned long long *size);
extern int dev_sector_size_from_id(dev_t id, unsigned int *size);
void wait_for(char *dev, int fd);
/*
* Data structures for policy management.
* Each device can have a policy structure that lists
* various name/value pairs each possibly with a metadata associated.
* The policy list is sorted by name/value/metadata
*/
struct dev_policy {
struct dev_policy *next;
char *name; /* None of these strings are allocated. They are
* all just references to strings which are known
* to exist elsewhere.
* name and metadata can be compared by address equality.
*/
const char *metadata;
const char *value;
};
extern char pol_act[], pol_domain[], pol_metadata[], pol_auto[];
/* iterate over the sublist starting at list, having the same
@ -1430,10 +1436,16 @@ extern struct dev_policy *disk_policy(struct mdinfo *disk);
extern struct dev_policy *devid_policy(int devid);
extern void dev_policy_free(struct dev_policy *p);
//extern void pol_new(struct dev_policy **pol, char *name, char *val, char *metadata);
extern void pol_add(struct dev_policy **pol, char *name, char *val, char *metadata);
extern struct dev_policy *pol_find(struct dev_policy *pol, char *name);
extern mdadm_status_t drive_test_and_add_policies(struct supertype *st, dev_policy_t **pols,
int fd, const int verbose);
extern mdadm_status_t sysfs_test_and_add_drive_policies(struct supertype *st, dev_policy_t **pols,
struct mdinfo *mdi, const int verbose);
extern mdadm_status_t mddev_test_and_add_drive_policies(struct supertype *st, dev_policy_t **pols,
int array_fd, const int verbose);
enum policy_action {
act_default,
act_include,
@ -1661,6 +1673,7 @@ extern char *conf_get_program(void);
extern char *conf_get_homehost(int *require_homehostp);
extern char *conf_get_homecluster(void);
extern int conf_get_monitor_delay(void);
extern bool conf_get_sata_opal_encryption_no_verify(void);
extern char *conf_line(FILE *file);
extern char *conf_word(FILE *file, int allow_key);
extern void print_quoted(char *str);
@ -1685,8 +1698,7 @@ extern const int uuid_zero[4];
extern int same_uuid(int a[4], int b[4], int swapuuid);
extern void copy_uuid(void *a, int b[4], int swapuuid);
extern char *__fname_from_uuid(int id[4], int swap, char *buf, char sep);
extern char *fname_from_uuid(struct supertype *st,
struct mdinfo *info, char *buf, char sep);
extern char *fname_from_uuid(struct mdinfo *info, char *buf);
extern unsigned long calc_csum(void *super, int bytes);
extern int enough(int level, int raid_disks, int layout, int clean,
char *avail);
@ -1708,6 +1720,9 @@ extern int assemble_container_content(struct supertype *st, int mdfd,
#define INCR_UNSAFE 2
#define INCR_ALREADY 4
#define INCR_YES 8
extern bool devid_matches_criteria(struct supertype *st, dev_t devid, struct spare_criteria *sc);
extern bool disk_fd_matches_criteria(struct supertype *st, int disk_fd, struct spare_criteria *sc);
extern struct mdinfo *container_choose_spares(struct supertype *st,
struct spare_criteria *criteria,
struct domainlist *domlist,
@ -1856,11 +1871,10 @@ static inline char *to_subarray(struct mdstat_ent *ent, char *container)
*/
static inline sighandler_t signal_s(int sig, sighandler_t handler)
{
struct sigaction new_act;
struct sigaction old_act;
struct sigaction new_act = {0};
struct sigaction old_act = {0};
new_act.sa_handler = handler;
new_act.sa_flags = 0;
if (sigaction(sig, &new_act, &old_act) == 0)
return old_act.sa_handler;
@ -1899,6 +1913,8 @@ static inline int xasprintf(char **strp, const char *fmt, ...) {
#define pr_info(fmt, args...) printf("%s: "fmt, Name, ##args)
#define pr_vrb(fmt, arg...) ((void)(verbose && pr_err(fmt, ##arg)))
void *xmalloc(size_t len);
void *xrealloc(void *ptr, size_t len);
void *xcalloc(size_t num, size_t size);

47
mdadm.spec
View file

@ -1,47 +0,0 @@
Summary: mdadm is used for controlling Linux md devices (aka RAID arrays)
Name: mdadm
Version: 4.3
Release: 1
Source: https://www.kernel.org/pub/linux/utils/raid/mdadm/mdadm-%{version}.tar.gz
URL: https://neil.brown.name/blog/mdadm
License: GPL
Group: Utilities/System
BuildRoot: %{_tmppath}/%{name}-root
Obsoletes: mdctl
%description
mdadm is a program that can be used to create, manage, and monitor
Linux MD (Software RAID) devices.
%prep
%setup -q
# we want to install in /sbin, not /usr/sbin...
%define _exec_prefix %{nil}
%build
# This is a debatable issue. The author of this RPM spec file feels that
# people who install RPMs (especially given that the default RPM options
# will strip the binary) are not going to be running gdb against the
# program.
make CXFLAGS="$RPM_OPT_FLAGS" SYSCONFDIR="%{_sysconfdir}"
%install
make DESTDIR=$RPM_BUILD_ROOT MANDIR=%{_mandir} BINDIR=%{_sbindir} install
install -D -m644 mdadm.conf-example $RPM_BUILD_ROOT/%{_sysconfdir}/mdadm.conf
%clean
rm -rf $RPM_BUILD_ROOT
%files
%defattr(-,root,root)
%doc TODO ChangeLog mdadm.conf-example COPYING
%{_sbindir}/mdadm
%{_sbindir}/mdmon
/usr/lib/udev/rules.d/01-md-raid-creating.rules
/usr/lib/udev/rules.d/63-md-raid-arrays.rules
/usr/lib/udev/rules.d/64-md-raid-assembly.rules
/usr/lib/udev/rules.d/69-md-clustered-confirm-device.rules
%config(noreplace,missingok)/%{_sysconfdir}/mdadm.conf
%{_mandir}/man*/md*
%changelog

21
mdmon.c
View file

@ -302,12 +302,12 @@ static int mdmon(char *devnm, int must_fork, int takeover);
int main(int argc, char *argv[])
{
char *container_name = NULL;
char *devnm = NULL;
int status = 0;
int opt;
int all = 0;
int takeover = 0;
int dofork = 1;
int mdfd = -1;
bool help = false;
static struct option options[] = {
{"all", 0, NULL, 'a'},
@ -410,19 +410,20 @@ int main(int argc, char *argv[])
free_mdstat(mdstat);
return status;
} else {
int mdfd = open_mddev(container_name, 0);
devnm = fd2devnm(mdfd);
}
mdfd = open_mddev(container_name, 0);
if (is_fd_valid(mdfd)) {
char *devnm = fd2devnm(mdfd);
close(mdfd);
if (devnm)
return mdmon(devnm, dofork && do_fork(), takeover);
}
if (!devnm) {
pr_err("%s is not a valid md device name\n",
container_name);
return 1;
}
return mdmon(devnm, dofork && do_fork(), takeover);
pr_err("%s is not a valid md device name\n", container_name);
return 1;
}
static int mdmon(char *devnm, int must_fork, int takeover)

55
mkinitramfs
View file

@ -1,55 +0,0 @@
#!/bin/sh
# make sure we are being run in the right directory...
if [ -f mkinitramfs ]
then :
else
echo >&2 mkinitramfs must be run from the mdadm source directory.
exit 1
fi
if [ -f /bin/busybox ]
then : good, it exists
case `file /bin/busybox` in
*statically* ) : good ;;
* ) echo >&2 mkinitramfs: /bin/busybox is not statically linked: cannot proceed.
exit 1
esac
else
echo >&2 "mkinitramfs: /bin/busybox doesn't exist - please install it statically linked."
exit 1
fi
rm -rf initramfs
mkdir initramfs
mkdir initramfs/bin
make mdadm.static
cp mdadm.static initramfs/bin/mdadm
cp /bin/busybox initramfs/bin/busybox
ln initramfs/bin/busybox initramfs/bin/sh
cat <<- END > initramfs/init
#!/bin/sh
echo 'Auto-assembling boot md array'
mkdir /proc
mount -t proc proc /proc
if [ -n "$rootuuid" ]
then arg=--uuid=$rootuuid
elif [ -n "$mdminor" ]
then arg=--super-minor=$mdminor
else arg=--super-minor=0
fi
echo "Using $arg"
mdadm -Acpartitions $arg --auto=part /dev/mda
cd /
mount /dev/mda1 /root || mount /dev/mda /root
umount /proc
cd /root
exec chroot . /sbin/init < /dev/console > /dev/console 2>&1
END
chmod +x initramfs/init
(cd initramfs
find init bin | cpio -o -H newc | gzip --best
) > init.cpio.gz
rm -rf initramfs
ls -l init.cpio.gz

65
monitor.c
View file

@ -412,6 +412,7 @@ static int read_and_act(struct active_array *a, fd_set *fds)
int ret = 0;
int count = 0;
struct timeval tv;
bool write_checkpoint = false;
a->next_state = bad_word;
a->next_action = bad_action;
@ -564,52 +565,40 @@ static int read_and_act(struct active_array *a, fd_set *fds)
}
}
/* Check for recovery checkpoint notifications. We need to be a
* minimum distance away from the last checkpoint to prevent
* over checkpointing. Note reshape checkpointing is handled
* in the second branch.
*/
if (sync_completed > a->last_checkpoint &&
sync_completed - a->last_checkpoint > a->info.component_size >> 4 &&
a->curr_action > reshape) {
/* A (non-reshape) sync_action has reached a checkpoint.
* Record the updated position in the metadata
*/
a->last_checkpoint = sync_completed;
a->container->ss->set_array_state(a, a->curr_state <= clean);
} else if ((a->curr_action == idle && a->prev_action == reshape) ||
(a->curr_action == reshape &&
sync_completed > a->last_checkpoint)) {
/* Reshape has progressed or completed so we need to
* update the array state - and possibly the array size
*/
/* Update reshape checkpoint, depending if it finished or progressed */
if (a->curr_action == idle && a->prev_action == reshape) {
char buf[SYSFS_MAX_BUF_SIZE];
if (sync_completed != 0)
a->last_checkpoint = sync_completed;
/* We might need to update last_checkpoint depending on
* the reason that reshape finished.
* if array reshape is really finished:
* set check point to the end, this allows
* set_array_state() to finalize reshape in metadata
* if reshape if broken: do not set checkpoint to the end
* this allows for reshape restart from checkpoint
/*
* If reshape really finished, set checkpoint to the end to finalize it.
* Do not set checkpoint if reshape is broken.
* Reshape will restart from last checkpoint.
*/
if ((a->curr_action != reshape) &&
(a->prev_action == reshape)) {
char buf[SYSFS_MAX_BUF_SIZE];
if ((sysfs_get_str(&a->info, NULL,
"reshape_position",
buf,
sizeof(buf)) >= 0) &&
str_is_none(buf) == true)
if (sysfs_get_str(&a->info, NULL, "reshape_position", buf, sizeof(buf)) >= 0)
if (str_is_none(buf) == true)
a->last_checkpoint = a->info.component_size;
}
a->container->ss->set_array_state(a, a->curr_state <= clean);
a->last_checkpoint = sync_completed;
write_checkpoint = true;
}
if (sync_completed > a->last_checkpoint)
if (a->curr_action >= reshape && sync_completed > a->last_checkpoint) {
/* Update checkpoint if neither reshape nor idle action */
a->last_checkpoint = sync_completed;
write_checkpoint = true;
}
/* Save checkpoint */
if (write_checkpoint) {
a->container->ss->set_array_state(a, a->curr_state <= clean);
if (a->curr_action <= reshape)
a->last_checkpoint = sync_completed;
}
if (sync_completed >= a->info.component_size)
a->last_checkpoint = 0;

1
platform-intel.h
View file

@ -262,7 +262,6 @@ int disk_attached_to_hba(int fd, const char *hba_path);
int devt_attached_to_hba(dev_t dev, const char *hba_path);
char *devt_to_devpath(dev_t dev, int dev_level, char *buf);
int path_attached_to_hba(const char *disk_path, const char *hba_path);
const char *get_sys_dev_type(enum sys_dev_type);
const struct orom_entry *get_orom_entry_by_device_id(__u16 dev_id);
const struct imsm_orom *get_orom_by_device_id(__u16 device_id);
struct sys_dev *device_by_id(__u16 device_id);

110
policy.c
View file

@ -365,7 +365,6 @@ struct dev_policy *path_policy(char **paths, char *type)
{
struct pol_rule *rules;
struct dev_policy *pol = NULL;
int i;
rules = config_rules;
@ -380,23 +379,104 @@ struct dev_policy *path_policy(char **paths, char *type)
rules = rules->next;
}
/* Now add any metadata-specific internal knowledge
* about this path
*/
for (i=0; paths && paths[0] && superlist[i]; i++)
if (superlist[i]->get_disk_controller_domain) {
const char *d =
superlist[i]->get_disk_controller_domain(
paths[0]);
if (d)
pol_new(&pol, pol_domain, d, superlist[i]->name);
}
pol_sort(&pol);
pol_dedup(pol);
return pol;
}
/**
* drive_test_and_add_policies() - get policies for drive and add them to pols.
* @st: supertype.
* @pols: pointer to pointer of first list entry, cannot be NULL, may point to NULL.
* @fd: device descriptor.
* @verbose: verbose flag.
*
* If supertype doesn't support this functionality return success. Use metadata handler to get
* policies.
*/
mdadm_status_t drive_test_and_add_policies(struct supertype *st, dev_policy_t **pols, int fd,
const int verbose)
{
if (!st->ss->test_and_add_drive_policies)
return MDADM_STATUS_SUCCESS;
if (st->ss->test_and_add_drive_policies(pols, fd, verbose) == MDADM_STATUS_SUCCESS) {
/* After successful call list cannot be empty */
assert(*pols);
return MDADM_STATUS_SUCCESS;
}
return MDADM_STATUS_ERROR;
}
/**
* sysfs_test_and_add_policies() - get policies for mddev and add them to pols.
* @st: supertype.
* @pols: pointer to pointer of first list entry, cannot be NULL, may point to NULL.
* @mdi: mdinfo describes the MD array, must have GET_DISKS option.
* @verbose: verbose flag.
*
* If supertype doesn't support this functionality return success. To get policies, all disks
* connected to mddev are analyzed.
*/
mdadm_status_t sysfs_test_and_add_drive_policies(struct supertype *st, dev_policy_t **pols,
struct mdinfo *mdi, const int verbose)
{
struct mdinfo *sd;
if (!st->ss->test_and_add_drive_policies)
return MDADM_STATUS_SUCCESS;
for (sd = mdi->devs; sd; sd = sd->next) {
char *devpath = map_dev(sd->disk.major, sd->disk.minor, 0);
int fd = dev_open(devpath, O_RDONLY);
int rv;
if (!is_fd_valid(fd)) {
pr_err("Cannot open fd for %s\n", devpath);
return MDADM_STATUS_ERROR;
}
rv = drive_test_and_add_policies(st, pols, fd, verbose);
close(fd);
if (rv)
return MDADM_STATUS_ERROR;
}
return MDADM_STATUS_SUCCESS;
}
/**
* mddev_test_and_add_policies() - get policies for mddev and add them to pols.
* @st: supertype.
* @pols: pointer to pointer of first list entry, cannot be NULL, may point to NULL.
* @array_fd: MD device descriptor.
* @verbose: verbose flag.
*
* If supertype doesn't support this functionality return success. Use fd to extract disks.
*/
mdadm_status_t mddev_test_and_add_drive_policies(struct supertype *st, dev_policy_t **pols,
int array_fd, const int verbose)
{
struct mdinfo *sra;
int ret;
if (!st->ss->test_and_add_drive_policies)
return MDADM_STATUS_SUCCESS;
sra = sysfs_read(array_fd, NULL, GET_DEVS);
if (!sra) {
pr_err("Cannot load sysfs for %s\n", fd2devnm(array_fd));
return MDADM_STATUS_ERROR;
}
ret = sysfs_test_and_add_drive_policies(st, pols, sra, verbose);
sysfs_free(sra);
return ret;
}
void pol_add(struct dev_policy **pol,
char *name, char *val,
char *metadata)
@ -679,7 +759,6 @@ int domain_test(struct domainlist *dom, struct dev_policy *pol,
* 1: has domains, all match
*/
int found_any = -1;
int has_one_domain = 1;
struct dev_policy *p;
pol = pol_find(pol, pol_domain);
@ -689,9 +768,6 @@ int domain_test(struct domainlist *dom, struct dev_policy *pol,
dom = dom->next;
if (!dom || strcmp(dom->dom, p->value) != 0)
return 0;
if (has_one_domain && metadata && strcmp(metadata, "imsm") == 0)
found_any = -1;
has_one_domain = 0;
}
return found_any;
}

11
super-ddf.c
View file

@ -1617,7 +1617,7 @@ static void brief_examine_super_ddf(struct supertype *st, int verbose)
struct mdinfo info;
char nbuf[64];
getinfo_super_ddf(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
printf("ARRAY metadata=ddf UUID=%s\n", nbuf + 5);
}
@ -1632,7 +1632,7 @@ static void brief_examine_subarrays_ddf(struct supertype *st, int verbose)
unsigned int i;
char nbuf[64];
getinfo_super_ddf(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
for (i = 0; i < be16_to_cpu(ddf->virt->max_vdes); i++) {
struct virtual_entry *ve = &ddf->virt->entries[i];
@ -1645,7 +1645,7 @@ static void brief_examine_subarrays_ddf(struct supertype *st, int verbose)
ddf->currentconf =&vcl;
vcl.vcnum = i;
uuid_from_super_ddf(st, info.uuid);
fname_from_uuid(st, &info, nbuf1, ':');
fname_from_uuid(&info, nbuf1);
_ddf_array_name(namebuf, ddf, i);
printf("ARRAY%s%s container=%s member=%d UUID=%s\n",
namebuf[0] == '\0' ? "" : " " DEV_MD_DIR, namebuf,
@ -1658,7 +1658,7 @@ static void export_examine_super_ddf(struct supertype *st)
struct mdinfo info;
char nbuf[64];
getinfo_super_ddf(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
printf("MD_METADATA=ddf\n");
printf("MD_LEVEL=container\n");
printf("MD_UUID=%s\n", nbuf+5);
@ -1798,7 +1798,7 @@ static void brief_detail_super_ddf(struct supertype *st, char *subarray)
return;
else
uuid_of_ddf_subarray(ddf, vcnum, info.uuid);
fname_from_uuid(st, &info, nbuf,':');
fname_from_uuid(&info, nbuf);
printf(" UUID=%s", nbuf + 5);
}
@ -5162,6 +5162,7 @@ struct superswitch super_ddf = {
.default_geometry = default_geometry_ddf,
.external = 1,
.swapuuid = 0,
/* for mdmon */
.open_new = ddf_open_new,

397
super-intel.c
View file

@ -27,6 +27,7 @@
#include <scsi/sg.h>
#include <ctype.h>
#include <dirent.h>
#include "drive_encryption.h"
/* MPB == Metadata Parameter Block */
#define MPB_SIGNATURE "Intel Raid ISM Cfg Sig. "
@ -393,8 +394,6 @@ struct md_list {
struct md_list *next;
};
#define pr_vrb(fmt, arg...) (void) (verbose && pr_err(fmt, ##arg))
static __u8 migr_type(struct imsm_dev *dev)
{
if (dev->vol.migr_type == MIGR_VERIFY &&
@ -1736,48 +1735,6 @@ static __u32 imsm_min_reserved_sectors(struct intel_super *super)
return (remainder < rv) ? remainder : rv;
}
/*
* Return minimum size of a spare and sector size
* that can be used in this array
*/
int get_spare_criteria_imsm(struct supertype *st, struct spare_criteria *c)
{
struct intel_super *super = st->sb;
struct dl *dl;
struct extent *e;
int i;
unsigned long long size = 0;
c->min_size = 0;
c->sector_size = 0;
if (!super)
return -EINVAL;
/* find first active disk in array */
dl = super->disks;
while (dl && (is_failed(&dl->disk) || dl->index == -1))
dl = dl->next;
if (!dl)
return -EINVAL;
/* find last lba used by subarrays */
e = get_extents(super, dl, 0);
if (!e)
return -EINVAL;
for (i = 0; e[i].size; i++)
continue;
if (i > 0)
size = e[i-1].start + e[i-1].size;
free(e);
/* add the amount of space needed for metadata */
size += imsm_min_reserved_sectors(super);
c->min_size = size * 512;
c->sector_size = super->sector_size;
return 0;
}
static bool is_gen_migration(struct imsm_dev *dev);
#define IMSM_4K_DIV 8
@ -2259,7 +2216,7 @@ static void examine_super_imsm(struct supertype *st, char *homehost)
else
printf("not supported\n");
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
printf(" UUID : %s\n", nbuf + 5);
sum = __le32_to_cpu(mpb->check_sum);
printf(" Checksum : %08x %s\n", sum,
@ -2284,7 +2241,7 @@ static void examine_super_imsm(struct supertype *st, char *homehost)
super->current_vol = i;
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
print_imsm_dev(super, dev, nbuf + 5, super->disks->index);
}
for (i = 0; i < mpb->num_disks; i++) {
@ -2309,7 +2266,7 @@ static void brief_examine_super_imsm(struct supertype *st, int verbose)
char nbuf[64];
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
printf("ARRAY metadata=imsm UUID=%s\n", nbuf + 5);
}
@ -2326,13 +2283,13 @@ static void brief_examine_subarrays_imsm(struct supertype *st, int verbose)
return;
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
for (i = 0; i < super->anchor->num_raid_devs; i++) {
struct imsm_dev *dev = get_imsm_dev(super, i);
super->current_vol = i;
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf1, ':');
fname_from_uuid(&info, nbuf1);
printf("ARRAY " DEV_MD_DIR "%.16s container=%s member=%d UUID=%s\n",
dev->volume, nbuf + 5, i, nbuf1 + 5);
}
@ -2346,7 +2303,7 @@ static void export_examine_super_imsm(struct supertype *st)
char nbuf[64];
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
printf("MD_METADATA=imsm\n");
printf("MD_LEVEL=container\n");
printf("MD_UUID=%s\n", nbuf+5);
@ -2366,7 +2323,7 @@ static void detail_super_imsm(struct supertype *st, char *homehost,
super->current_vol = strtoul(subarray, NULL, 10);
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
printf("\n UUID : %s\n", nbuf + 5);
super->current_vol = temp_vol;
@ -2383,7 +2340,7 @@ static void brief_detail_super_imsm(struct supertype *st, char *subarray)
super->current_vol = strtoul(subarray, NULL, 10);
getinfo_super_imsm(st, &info, NULL);
fname_from_uuid(st, &info, nbuf, ':');
fname_from_uuid(&info, nbuf);
printf(" UUID=%s", nbuf + 5);
super->current_vol = temp_vol;
@ -2393,12 +2350,41 @@ static int imsm_read_serial(int fd, char *devname, __u8 *serial,
size_t serial_buf_len);
static void fd2devname(int fd, char *name);
static int ahci_enumerate_ports(const char *hba_path, int port_count, int host_base, int verbose)
void print_encryption_information(int disk_fd, enum sys_dev_type hba_type)
{
struct encryption_information information = {0};
mdadm_status_t status = MDADM_STATUS_SUCCESS;
const char *indent = " ";
switch (hba_type) {
case SYS_DEV_VMD:
case SYS_DEV_NVME:
status = get_nvme_opal_encryption_information(disk_fd, &information, 1);
break;
case SYS_DEV_SATA:
case SYS_DEV_SATA_VMD:
status = get_ata_encryption_information(disk_fd, &information, 1);
break;
default:
return;
}
if (status) {
pr_err("Failed to get drive encryption information.\n");
return;
}
printf("%sEncryption(Ability|Status): %s|%s\n", indent,
get_encryption_ability_string(information.ability),
get_encryption_status_string(information.status));
}
static int ahci_enumerate_ports(struct sys_dev *hba, int port_count, int host_base, int verbose)
{
/* dump an unsorted list of devices attached to AHCI Intel storage
* controller, as well as non-connected ports
*/
int hba_len = strlen(hba_path) + 1;
int hba_len = strlen(hba->path) + 1;
struct dirent *ent;
DIR *dir;
char *path = NULL;
@ -2434,7 +2420,7 @@ static int ahci_enumerate_ports(const char *hba_path, int port_count, int host_b
path = devt_to_devpath(makedev(major, minor), 1, NULL);
if (!path)
continue;
if (!path_attached_to_hba(path, hba_path)) {
if (!path_attached_to_hba(path, hba->path)) {
free(path);
path = NULL;
continue;
@ -2537,6 +2523,8 @@ static int ahci_enumerate_ports(const char *hba_path, int port_count, int host_b
printf(" (%s)\n", buf);
else
printf(" ()\n");
print_encryption_information(fd, hba->type);
close(fd);
}
free(path);
@ -2601,6 +2589,8 @@ static int print_nvme_info(struct sys_dev *hba)
else
printf("()\n");
print_encryption_information(fd, hba->type);
skip:
close_fd(&fd);
}
@ -2856,7 +2846,7 @@ static int detail_platform_imsm(int verbose, int enumerate_only, char *controlle
hba->path, get_sys_dev_type(hba->type));
if (hba->type == SYS_DEV_SATA || hba->type == SYS_DEV_SATA_VMD) {
host_base = ahci_get_port_count(hba->path, &port_count);
if (ahci_enumerate_ports(hba->path, port_count, host_base, verbose)) {
if (ahci_enumerate_ports(hba, port_count, host_base, verbose)) {
if (verbose > 0)
pr_err("failed to enumerate ports on %s controller at %s.\n",
get_sys_dev_type(hba->type), hba->pci_id);
@ -4174,17 +4164,17 @@ static int imsm_read_serial(int fd, char *devname,
memset(buf, 0, sizeof(buf));
if (check_env("IMSM_DEVNAME_AS_SERIAL")) {
memset(serial, 0, serial_buf_len);
fd2devname(fd, (char *) serial);
return 0;
}
rv = nvme_get_serial(fd, buf, sizeof(buf));
if (rv)
rv = scsi_get_serial(fd, buf, sizeof(buf));
if (rv && check_env("IMSM_DEVNAME_AS_SERIAL")) {
memset(serial, 0, MAX_RAID_SERIAL_LEN);
fd2devname(fd, (char *) serial);
return 0;
}
if (rv != 0) {
if (devname)
pr_err("Failed to retrieve serial for %s\n",
@ -8771,6 +8761,9 @@ static int imsm_set_array_state(struct active_array *a, int consistent)
super->updates_pending++;
}
if (a->prev_action == idle)
goto skip_mark_checkpoint;
mark_checkpoint:
/* skip checkpointing for general migration,
* it is controlled in mdadm
@ -11259,39 +11252,247 @@ abort:
return retval;
}
static char disk_by_path[] = "/dev/disk/by-path/";
static const char *imsm_get_disk_controller_domain(const char *path)
/**
* test_and_add_drive_controller_policy_imsm() - add disk controller to policies list.
* @type: Policy type to search on list.
* @pols: List of currently recorded policies.
* @disk_fd: File descriptor of the device to check.
* @hba: The hba disk is attached, could be NULL if verification is disabled.
* @verbose: verbose flag.
*
* IMSM cares about drive physical placement. If @hba is not set, it adds unknown policy.
* If there is no controller policy on pols we are free to add first one. If there is a policy then,
* new must be the same - no controller mixing allowed.
*/
static mdadm_status_t
test_and_add_drive_controller_policy_imsm(const char * const type, dev_policy_t **pols, int disk_fd,
struct sys_dev *hba, const int verbose)
{
char disk_path[PATH_MAX];
char *drv=NULL;
struct stat st;
const char *controller_policy = get_sys_dev_type(SYS_DEV_UNKNOWN);
struct dev_policy *pol = pol_find(*pols, (char *)type);
char devname[MAX_RAID_SERIAL_LEN];
strncpy(disk_path, disk_by_path, PATH_MAX);
strncat(disk_path, path, PATH_MAX - strlen(disk_path) - 1);
if (stat(disk_path, &st) == 0) {
struct sys_dev* hba;
char *path;
if (hba)
controller_policy = get_sys_dev_type(hba->type);
path = devt_to_devpath(st.st_rdev, 1, NULL);
if (path == NULL)
return "unknown";
hba = find_disk_attached_hba(-1, path);
if (hba && hba->type == SYS_DEV_SAS)
drv = "isci";
else if (hba && (hba->type == SYS_DEV_SATA || hba->type == SYS_DEV_SATA_VMD))
drv = "ahci";
else if (hba && hba->type == SYS_DEV_VMD)
drv = "vmd";
else if (hba && hba->type == SYS_DEV_NVME)
drv = "nvme";
else
drv = "unknown";
dprintf("path: %s hba: %s attached: %s\n",
path, (hba) ? hba->path : "NULL", drv);
free(path);
if (!pol) {
pol_add(pols, (char *)type, (char *)controller_policy, "imsm");
return MDADM_STATUS_SUCCESS;
}
return drv;
if (strcmp(pol->value, controller_policy) == 0)
return MDADM_STATUS_SUCCESS;
fd2devname(disk_fd, devname);
pr_vrb("Intel(R) raid controller \"%s\" found for %s, but \"%s\" was detected earlier\n",
controller_policy, devname, pol->value);
pr_vrb("Disks under different controllers cannot be used, aborting\n");
return MDADM_STATUS_ERROR;
}
/**
* test_and_add_drive_encryption_policy_imsm() - add disk encryption to policies list.
* @type: policy type to search in the list.
* @pols: list of currently recorded policies.
* @disk_fd: file descriptor of the device to check.
* @hba: The hba to which the drive is attached, could be NULL if verification is disabled.
* @verbose: verbose flag.
*
* IMSM cares about drive encryption state. It is not allowed to mix disks with different
* encryption state within one md device.
* If there is no encryption policy on pols we are free to add first one.
* If there is a policy then, new must be the same.
*/
static mdadm_status_t
test_and_add_drive_encryption_policy_imsm(const char * const type, dev_policy_t **pols, int disk_fd,
struct sys_dev *hba, const int verbose)
{
struct dev_policy *expected_policy = pol_find(*pols, (char *)type);
struct encryption_information information = {0};
char *encryption_state = "Unknown";
int status = MDADM_STATUS_SUCCESS;
bool encryption_checked = true;
char devname[PATH_MAX];
if (!hba)
goto check_policy;
switch (hba->type) {
case SYS_DEV_NVME:
case SYS_DEV_VMD:
status = get_nvme_opal_encryption_information(disk_fd, &information, verbose);
break;
case SYS_DEV_SATA:
case SYS_DEV_SATA_VMD:
status = get_ata_encryption_information(disk_fd, &information, verbose);
break;
default:
encryption_checked = false;
}
if (status) {
fd2devname(disk_fd, devname);
pr_vrb("Failed to read encryption information of device %s\n", devname);
return MDADM_STATUS_ERROR;
}
if (encryption_checked) {
if (information.status == ENC_STATUS_LOCKED) {
fd2devname(disk_fd, devname);
pr_vrb("Device %s is in Locked state, cannot use. Aborting.\n", devname);
return MDADM_STATUS_ERROR;
}
encryption_state = (char *)get_encryption_status_string(information.status);
}
check_policy:
if (expected_policy) {
if (strcmp(expected_policy->value, encryption_state) == 0)
return MDADM_STATUS_SUCCESS;
fd2devname(disk_fd, devname);
pr_vrb("Encryption status \"%s\" detected for disk %s, but \"%s\" status was detected eariler.\n",
encryption_state, devname, expected_policy->value);
pr_vrb("Disks with different encryption status cannot be used.\n");
return MDADM_STATUS_ERROR;
}
pol_add(pols, (char *)type, encryption_state, "imsm");
return MDADM_STATUS_SUCCESS;
}
struct imsm_drive_policy {
char *type;
mdadm_status_t (*test_and_add_drive_policy)(const char * const type,
struct dev_policy **pols, int disk_fd,
struct sys_dev *hba, const int verbose);
};
struct imsm_drive_policy imsm_policies[] = {
{"controller", test_and_add_drive_controller_policy_imsm},
{"encryption", test_and_add_drive_encryption_policy_imsm}
};
mdadm_status_t test_and_add_drive_policies_imsm(struct dev_policy **pols, int disk_fd,
const int verbose)
{
struct imsm_drive_policy *imsm_pol;
struct sys_dev *hba = NULL;
char path[PATH_MAX];
mdadm_status_t ret;
unsigned int i;
/* If imsm platform verification is disabled, do not search for hba. */
if (check_no_platform() != 1) {
if (!diskfd_to_devpath(disk_fd, 1, path)) {
pr_vrb("IMSM: Failed to retrieve device path by file descriptor.\n");
return MDADM_STATUS_ERROR;
}
hba = find_disk_attached_hba(disk_fd, path);
if (!hba) {
pr_vrb("IMSM: Failed to find hba for %s\n", path);
return MDADM_STATUS_ERROR;
}
}
for (i = 0; i < ARRAY_SIZE(imsm_policies); i++) {
imsm_pol = &imsm_policies[i];
ret = imsm_pol->test_and_add_drive_policy(imsm_pol->type, pols, disk_fd, hba,
verbose);
if (ret != MDADM_STATUS_SUCCESS)
/* Inherit error code */
return ret;
}
return MDADM_STATUS_SUCCESS;
}
/**
* get_spare_criteria_imsm() - set spare criteria.
* @st: supertype.
* @mddev_path: path to md device devnode, it must be container.
* @c: spare_criteria struct to fill, not NULL.
*
* If superblock is not loaded, use mddev_path to load_container. It must be given in this case.
* Filles size and sector size accordingly to superblock.
*/
mdadm_status_t get_spare_criteria_imsm(struct supertype *st, char *mddev_path,
struct spare_criteria *c)
{
mdadm_status_t ret = MDADM_STATUS_ERROR;
bool free_superblock = false;
unsigned long long size = 0;
struct intel_super *super;
struct extent *e;
struct dl *dl;
int i;
/* If no superblock and no mddev_path, we cannot load superblock. */
assert(st->sb || mddev_path);
if (mddev_path) {
int fd = open(mddev_path, O_RDONLY);
mdadm_status_t rv;
if (!is_fd_valid(fd))
return MDADM_STATUS_ERROR;
if (!st->sb) {
if (load_container_imsm(st, fd, st->devnm)) {
close(fd);
return MDADM_STATUS_ERROR;
}
free_superblock = true;
}
rv = mddev_test_and_add_drive_policies(st, &c->pols, fd, 0);
close(fd);
if (rv != MDADM_STATUS_SUCCESS)
goto out;
}
super = st->sb;
/* find first active disk in array */
dl = super->disks;
while (dl && (is_failed(&dl->disk) || dl->index == -1))
dl = dl->next;
if (!dl)
goto out;
/* find last lba used by subarrays */
e = get_extents(super, dl, 0);
if (!e)
goto out;
for (i = 0; e[i].size; i++)
continue;
if (i > 0)
size = e[i - 1].start + e[i - 1].size;
free(e);
/* add the amount of space needed for metadata */
size += imsm_min_reserved_sectors(super);
c->min_size = size * 512;
c->sector_size = super->sector_size;
c->criteria_set = true;
ret = MDADM_STATUS_SUCCESS;
out:
if (free_superblock)
free_super_imsm(st);
if (ret != MDADM_STATUS_SUCCESS)
c->criteria_set = false;
return ret;
}
static char *imsm_find_array_devnm_by_subdev(int subdev, char *container)
@ -11422,10 +11623,15 @@ static int imsm_reshape_is_allowed_on_container(struct supertype *st,
*/
static struct mdinfo *get_spares_for_grow(struct supertype *st)
{
struct spare_criteria sc;
struct spare_criteria sc = {0};
struct mdinfo *spares;
get_spare_criteria_imsm(st, &sc);
return container_choose_spares(st, &sc, NULL, NULL, NULL, 0);
get_spare_criteria_imsm(st, NULL, &sc);
spares = container_choose_spares(st, &sc, NULL, NULL, NULL, 0);
dev_policy_free(sc.pols);
return spares;
}
/******************************************************************************
@ -12987,7 +13193,7 @@ struct superswitch super_imsm = {
.update_subarray = update_subarray_imsm,
.load_container = load_container_imsm,
.default_geometry = default_geometry_imsm,
.get_disk_controller_domain = imsm_get_disk_controller_domain,
.test_and_add_drive_policies = test_and_add_drive_policies_imsm,
.reshape_super = imsm_reshape_super,
.manage_reshape = imsm_manage_reshape,
.recover_backup = recover_backup_imsm,
@ -13020,6 +13226,7 @@ struct superswitch super_imsm = {
.validate_ppl = validate_ppl_imsm,
.external = 1,
.swapuuid = 0,
.name = "imsm",
/* for mdmon */

2
super0.c
View file

@ -1369,5 +1369,7 @@ struct superswitch super0 = {
.locate_bitmap = locate_bitmap0,
.write_bitmap = write_bitmap0,
.free_super = free_super0,
.swapuuid = 0,
.name = "0.90",
};

16
super1.c
View file

@ -575,17 +575,6 @@ static void examine_super1(struct supertype *st, char *homehost)
inconsistent = 1;
}
}
#if 0
/* This is confusing too */
faulty = 0;
for (i = 0; i < __le32_to_cpu(sb->max_dev); i++) {
int role = __le16_to_cpu(sb->dev_roles[i]);
if (role == MD_DISK_ROLE_FAULTY)
faulty++;
}
if (faulty)
printf(" %d failed", faulty);
#endif
printf(" ('A' == active, '.' == missing, 'R' == replacing)");
printf("\n");
for (d = 0; d < __le32_to_cpu(sb->max_dev); d++) {
@ -1752,7 +1741,10 @@ static int add_to_super1(struct supertype *st, mdu_disk_info_t *dk,
di->devname = devname;
di->disk = *dk;
di->data_offset = data_offset;
get_dev_size(fd, NULL, &di->dev_size);
if (is_fd_valid(fd))
get_dev_size(fd, NULL, &di->dev_size);
di->next = NULL;
*dip = di;

58
sysfs.c
View file

@ -74,6 +74,29 @@ void sysfs_free(struct mdinfo *sra)
}
}
/**
* sysfs_get_container_devnm() - extract container device name.
* @mdi: md_info describes member array, with GET_VERSION option.
* @buf: buf to fill, must be MD_NAME_MAX.
*
* External array version is in format {/,-}<container_devnm>/<array_index>
* Extract container_devnm from it and safe it in @buf.
*/
void sysfs_get_container_devnm(struct mdinfo *mdi, char *buf)
{
char *p;
assert(is_subarray(mdi->text_version));
/* Skip first special sign */
snprintf(buf, MD_NAME_MAX, "%s", mdi->text_version + 1);
/* Remove array index */
p = strchr(buf, '/');
if (p)
*p = 0;
}
int sysfs_open(char *devnm, char *devname, char *attr)
{
char fname[MAX_SYSFS_PATH_LEN];
@ -655,7 +678,7 @@ int sysfs_set_safemode(struct mdinfo *sra, unsigned long ms)
return sysfs_set_str(sra, NULL, "safe_mode_delay", delay);
}
int sysfs_set_array(struct mdinfo *info, int vers)
int sysfs_set_array(struct mdinfo *info)
{
int rv = 0;
char ver[100];
@ -679,9 +702,7 @@ int sysfs_set_array(struct mdinfo *info, int vers)
if (strlen(buf) >= 9 && buf[9] == '-')
ver[9] = '-';
if ((vers % 100) < 2 ||
sysfs_set_str(info, NULL, "metadata_version",
ver) < 0) {
if (sysfs_set_str(info, NULL, "metadata_version", ver) < 0) {
pr_err("This kernel does not support external metadata.\n");
return 1;
}
@ -1100,3 +1121,32 @@ void sysfsline(char *line)
sr->next = sysfs_rules;
sysfs_rules = sr;
}
/**
* sysfs_is_libata_allow_tpm_enabled() - check if libata allow_tmp is enabled.
* @verbose: verbose flag.
*
* Check if libata allow_tmp flag is set, this is required for SATA Opal Security commands to work.
*
* Return: true if allow_tpm enable, false otherwise.
*/
bool sysfs_is_libata_allow_tpm_enabled(const int verbose)
{
const char *path = "/sys/module/libata/parameters/allow_tpm";
const char *expected_value = "1";
int fd = open(path, O_RDONLY);
char buf[3];
if (!is_fd_valid(fd)) {
pr_vrb("Failed open file descriptor to %s. Cannot check libata allow_tpm param.\n",
path);
return false;
}
sysfs_fd_get_str(fd, buf, sizeof(buf));
close(fd);
if (strncmp(buf, expected_value, 1) == 0)
return true;
return false;
}

27
test
View file

@ -1,11 +1,12 @@
#!/bin/bash
#
# run test suite for mdadm
mdadm=$PWD/mdadm
mdadm=`which mdadm`
targetdir="/var/tmp"
logdir="$targetdir"
config=/tmp/mdadm.conf
testdir=$PWD/tests
system_speed_limit=`cat /proc/sys/dev/raid/speed_limit_max`
devlist=
savelogs=0
@ -20,9 +21,6 @@ DEVTYPE=loop
INTEGRITY=yes
LVM_VOLGROUP=mdtest
# make sure to test local mdmon, not system one
export MDADM_NO_SYSTEMCTL=1
# assume md0, md1, md2 exist in /dev
md0=/dev/md0
md1=/dev/md1
@ -41,7 +39,10 @@ ctrl_c() {
ctrl_c_error=1
}
# mdadm always adds --quiet, and we want to see any unexpected messages
restore_system_speed_limit() {
echo $system_speed_limit > /proc/sys/dev/raid/speed_limit_max
}
mdadm() {
rm -f $targetdir/stderr
case $* in
@ -63,10 +64,10 @@ mdadm() {
$mdadm --zero $args > /dev/null
}
done
$mdadm 2> $targetdir/stderr --quiet "$@" --auto=yes
$mdadm 2> $targetdir/stderr "$@" --auto=yes
;;
* )
$mdadm 2> $targetdir/stderr --quiet "$@"
$mdadm 2> $targetdir/stderr "$@"
;;
esac
rv=$?
@ -99,8 +100,6 @@ do_test() {
fi
rm -f $targetdir/stderr
# this might have been reset: restore the default.
echo 2000 > /proc/sys/dev/raid/speed_limit_max
do_clean
# source script in a subshell, so it has access to our
# namespace, but cannot change it.
@ -122,6 +121,7 @@ do_test() {
echo " (KNOWN BROKEN TEST: $_broken_msg)"
fi
fi
restore_system_speed_limit
[ "$savelogs" == "1" ] &&
mv -f $targetdir/log $logdir/$_basename.log
[ "$ctrl_c_error" == "1" ] && exit 1
@ -299,7 +299,15 @@ parse_args() {
done
}
print_warning() {
cat <<-EOF
Warning! Tests are performed on system level mdadm!
If you want to test local build, you need to install it first!
EOF
}
main() {
print_warning
do_setup
echo "Testing on linux-$(uname -r) kernel"
@ -329,6 +337,7 @@ main() {
break
fi
done
exit 0
}

1
tests/func.sh
View file

@ -213,7 +213,6 @@ do_setup() {
path1=$dev7
ulimit -c unlimited
[ -f /proc/mdstat ] || modprobe md_mod
echo 2000 > /proc/sys/dev/raid/speed_limit_max
echo 0 > /sys/module/md_mod/parameters/start_ro
}

3
udev.c
View file

@ -26,7 +26,10 @@
#include <signal.h>
#include <limits.h>
#include <syslog.h>
#ifndef NO_LIBUDEV
#include <libudev.h>
#endif
static char *unblock_path;

144
util.c
View file

@ -36,7 +36,7 @@
#include <ctype.h>
#include <dirent.h>
#include <dlfcn.h>
#include <limits.h>
/*
* following taken from linux/blkpg.h because they aren't
@ -589,19 +589,21 @@ char *__fname_from_uuid(int id[4], int swap, char *buf, char sep)
}
char *fname_from_uuid(struct supertype *st, struct mdinfo *info,
char *buf, char sep)
/**
* fname_from_uuid() - generate uuid string. Should not be used with super1.
* @info: info with uuid
* @buf: buf to fill.
*
* This routine should not be used with super1. See detail_fname_from_uuid() for details. It does
* not use superswitch swapuuid as it should be 0 but it has to do UUID conversion if host is big
* endian- left for backward compatibility.
*/
char *fname_from_uuid(struct mdinfo *info, char *buf)
{
// dirty hack to work around an issue with super1 superblocks...
// super1 superblocks need swapuuid set in order for assembly to
// work, but can't have it set if we want this printout to match
// all the other uuid printouts in super1.c, so we force swapuuid
// to 1 to make our printout match the rest of super1
#if __BYTE_ORDER == BIG_ENDIAN
return __fname_from_uuid(info->uuid, 1, buf, sep);
return __fname_from_uuid(info->uuid, true, buf, ':');
#else
return __fname_from_uuid(info->uuid, (st->ss == &super1) ? 1 :
st->ss->swapuuid, buf, sep);
return __fname_from_uuid(info->uuid, false, buf, ':');
#endif
}
@ -1266,40 +1268,6 @@ struct supertype *super_by_fd(int fd, char **subarrayp)
return st;
}
int dev_size_from_id(dev_t id, unsigned long long *size)
{
char buf[20];
int fd;
sprintf(buf, "%d:%d", major(id), minor(id));
fd = dev_open(buf, O_RDONLY);
if (fd < 0)
return 0;
if (get_dev_size(fd, NULL, size)) {
close(fd);
return 1;
}
close(fd);
return 0;
}
int dev_sector_size_from_id(dev_t id, unsigned int *size)
{
char buf[20];
int fd;
sprintf(buf, "%d:%d", major(id), minor(id));
fd = dev_open(buf, O_RDONLY);
if (fd < 0)
return 0;
if (get_dev_sector_size(fd, NULL, size)) {
close(fd);
return 1;
}
close(fd);
return 0;
}
struct supertype *dup_super(struct supertype *orig)
{
struct supertype *st;
@ -1899,8 +1867,7 @@ int set_array_info(int mdfd, struct supertype *st, struct mdinfo *info)
int rv;
if (st->ss->external)
return sysfs_set_array(info, 9003);
return sysfs_set_array(info);
memset(&inf, 0, sizeof(inf));
inf.major_version = info->array.major_version;
inf.minor_version = info->array.minor_version;
@ -2088,6 +2055,65 @@ void append_metadata_update(struct supertype *st, void *buf, int len)
unsigned int __invalid_size_argument_for_IOC = 0;
#endif
/**
* disk_fd_matches_criteria() - check if device matches spare criteria.
* @st: supertype, not NULL.
* @disk_fd: file descriptor of the disk.
* @sc: criteria to test.
*
* Return: true if disk matches criteria, false otherwise.
*/
bool disk_fd_matches_criteria(struct supertype *st, int disk_fd, struct spare_criteria *sc)
{
unsigned int dev_sector_size = 0;
unsigned long long dev_size = 0;
if (!sc->criteria_set)
return true;
if (!get_dev_size(disk_fd, NULL, &dev_size) || dev_size < sc->min_size)
return false;
if (!get_dev_sector_size(disk_fd, NULL, &dev_sector_size) ||
sc->sector_size != dev_sector_size)
return false;
if (drive_test_and_add_policies(st, &sc->pols, disk_fd, 0))
return false;
return true;
}
/**
* devid_matches_criteria() - check if device referenced by devid matches spare criteria.
* @st: supertype, not NULL.
* @devid: devid of the device to check.
* @sc: criteria to test.
*
* Return: true if disk matches criteria, false otherwise.
*/
bool devid_matches_criteria(struct supertype *st, dev_t devid, struct spare_criteria *sc)
{
char buf[NAME_MAX];
bool ret;
int fd;
if (!sc->criteria_set)
return true;
snprintf(buf, NAME_MAX, "%d:%d", major(devid), minor(devid));
fd = dev_open(buf, O_RDONLY);
if (!is_fd_valid(fd))
return false;
/* Error code inherited */
ret = disk_fd_matches_criteria(st, fd, sc);
close(fd);
return ret;
}
/* Pick all spares matching given criteria from a container
* if min_size == 0 do not check size
* if domlist == NULL do not check domains
@ -2111,28 +2137,13 @@ struct mdinfo *container_choose_spares(struct supertype *st,
dp = &disks->devs;
disks->array.spare_disks = 0;
while (*dp) {
int found = 0;
bool found = false;
d = *dp;
if (d->disk.state == 0) {
/* check if size is acceptable */
unsigned long long dev_size;
unsigned int dev_sector_size;
int size_valid = 0;
int sector_size_valid = 0;
dev_t dev = makedev(d->disk.major,d->disk.minor);
if (!criteria->min_size ||
(dev_size_from_id(dev, &dev_size) &&
dev_size >= criteria->min_size))
size_valid = 1;
if (!criteria->sector_size ||
(dev_sector_size_from_id(dev, &dev_sector_size) &&
criteria->sector_size == dev_sector_size))
sector_size_valid = 1;
found = size_valid && sector_size_valid;
found = devid_matches_criteria(st, dev, criteria);
/* check if domain matches */
if (found && domlist) {
@ -2141,7 +2152,8 @@ struct mdinfo *container_choose_spares(struct supertype *st,
pol_add(&pol, pol_domain,
spare_group, NULL);
if (domain_test(domlist, pol, metadata) != 1)
found = 0;
found = false;
dev_policy_free(pol);
}
}