the underlying media fails or becomes inaccessible. For example
when a USB flash memory card hosting a UFS filesystem is unplugged.
The strategy for handling disk I/O errors when soft updates are
enabled is to stop writing to the disk of the affected file system
but continue to accept I/O requests and report that all future
writes by the file system to that disk actually succeed. Then
initiate an asynchronous forced unmount of the affected file system.
There are two cases for disk I/O errors:
- ENXIO, which means that this disk is gone and the lower layers
of the storage stack already guarantee that no future I/O to
this disk will succeed.
- EIO (or most other errors), which means that this particular
I/O request has failed but subsequent I/O requests to this
disk might still succeed.
For ENXIO, we can just clear the error and continue, because we
know that the file system cannot affect the on-disk state after we
see this error. For EIO or other errors, we arrange for the geom_vfs
layer to reject all future I/O requests with ENXIO just like is
done when the geom_vfs is orphaned. In both cases, the file system
code can just clear the error and proceed with the forcible unmount.
This new treatment of I/O errors is needed for writes of any buffer
that is involved in a dependency. Most dependencies are described
by a structure attached to the buffer's b_dep field. But some are
created and processed as a result of the completion of the dependencies
attached to the buffer.
Clearing of some dependencies require a read. For example if there
is a dependency that requires an inode to be written, the disk block
containing that inode must be read, the updated inode copied into
place in that buffer, and the buffer then written back to disk.
Often the needed buffer is already in memory and can be used. But
if it needs to be read from the disk, the read will fail, so we
fabricate a buffer full of zeroes and pretend that the read succeeded.
This zero'ed buffer can be updated and written back to disk.
The only case where a buffer full of zeros causes the code to do
the wrong thing is when reading an inode buffer containing an inode
that still has an inode dependency in memory that will reinitialize
the effective link count (i_effnlink) based on the actual link count
(i_nlink) that we read. To handle this case we now store the i_nlink
value that we wrote in the inode dependency so that it can be
restored into the zero'ed buffer thus keeping the tracking of the
inode link count consistent.
Because applications depend on knowing when an attempt to write
their data to stable storage has failed, the fsync(2) and msync(2)
system calls need to return errors if data fails to be written to
stable storage. So these operations return ENXIO for every call
made on files in a file system where we have otherwise been ignoring
I/O errors.
Coauthered by: mckusick
Reviewed by: kib
Tested by: Peter Holm
Approved by: mckusick (mentor)
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D24088
Some crypto consumers such as GELI and KTLS for file-backed sendfile
need to store their output in a separate buffer from the input.
Currently these consumers copy the contents of the input buffer into
the output buffer and queue an in-place crypto operation on the output
buffer. Using a separate output buffer avoids this copy.
- Create a new 'struct crypto_buffer' describing a crypto buffer
containing a type and type-specific fields. crp_ilen is gone,
instead buffers that use a flat kernel buffer have a cb_buf_len
field for their length. The length of other buffer types is
inferred from the backing store (e.g. uio_resid for a uio).
Requests now have two such structures: crp_buf for the input buffer,
and crp_obuf for the output buffer.
- Consumers now use helper functions (crypto_use_*,
e.g. crypto_use_mbuf()) to configure the input buffer. If an output
buffer is not configured, the request still modifies the input
buffer in-place. A consumer uses a second set of helper functions
(crypto_use_output_*) to configure an output buffer.
- Consumers must request support for separate output buffers when
creating a crypto session via the CSP_F_SEPARATE_OUTPUT flag and are
only permitted to queue a request with a separate output buffer on
sessions with this flag set. Existing drivers already reject
sessions with unknown flags, so this permits drivers to be modified
to support this extension without requiring all drivers to change.
- Several data-related functions now have matching versions that
operate on an explicit buffer (e.g. crypto_apply_buf,
crypto_contiguous_subsegment_buf, bus_dma_load_crp_buf).
- Most of the existing data-related functions operate on the input
buffer. However crypto_copyback always writes to the output buffer
if a request uses a separate output buffer.
- For the regions in input/output buffers, the following conventions
are followed:
- AAD and IV are always present in input only and their
fields are offsets into the input buffer.
- payload is always present in both buffers. If a request uses a
separate output buffer, it must set a new crp_payload_start_output
field to the offset of the payload in the output buffer.
- digest is in the input buffer for verify operations, and in the
output buffer for compute operations. crp_digest_start is relative
to the appropriate buffer.
- Add a crypto buffer cursor abstraction. This is a more general form
of some bits in the cryptosoft driver that tried to always use uio's.
However, compared to the original code, this avoids rewalking the uio
iovec array for requests with multiple vectors. It also avoids
allocate an iovec array for mbufs and populating it by instead walking
the mbuf chain directly.
- Update the cryptosoft(4) driver to support separate output buffers
making use of the cursor abstraction.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D24545
The alias needs to be part of the provider instead of the geom to work
properly. To bind the DEV geom, we need to look at the provider's names and
aliases and create the dev entries from there. If this lives in the GEOM, then
it won't propigate down the tree properly. Remove it from geom, add it provider.
Update geli, gmountver, gnop, gpart, and guzip to use it, which handles the bulk
of the uses in FreeBSD. I think this is all the providers that create a new name
based on their parent's name.
Geom_mirror initialization occurs in spurts and the present of a
non-destroyed g_mirror softc does not always indicate that the geom has
launched (i.e., has an sc_provider).
Some gmirror(8) commands (via g_mirror_ctl) depend on a g_mirror's
sc_provider (insert and resize). For those commands, g_mirror_ctl is
modified to sleep-poll in an interruptible way until the target geom is
either launched or destroyed.
Reviewed by: markj
Tested by: markj
Sponsored by: Dell EMC Isilon
Differential Revision: https://reviews.freebsd.org/D24780
This removes support for reading and writing volumes using the
following algorithms:
- Triple DES
- Blowfish
- MD5 HMAC integrity
In addition, this commit adds an explicit whitelist of supported
algorithms to give a better error message when an invalid or
unsupported algorithm is used by an existing volume.
Reviewed by: cem
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D24343
- The linked list of cryptoini structures used in session
initialization is replaced with a new flat structure: struct
crypto_session_params. This session includes a new mode to define
how the other fields should be interpreted. Available modes
include:
- COMPRESS (for compression/decompression)
- CIPHER (for simply encryption/decryption)
- DIGEST (computing and verifying digests)
- AEAD (combined auth and encryption such as AES-GCM and AES-CCM)
- ETA (combined auth and encryption using encrypt-then-authenticate)
Additional modes could be added in the future (e.g. if we wanted to
support TLS MtE for AES-CBC in the kernel we could add a new mode
for that. TLS modes might also affect how AAD is interpreted, etc.)
The flat structure also includes the key lengths and algorithms as
before. However, code doesn't have to walk the linked list and
switch on the algorithm to determine which key is the auth key vs
encryption key. The 'csp_auth_*' fields are always used for auth
keys and settings and 'csp_cipher_*' for cipher. (Compression
algorithms are stored in csp_cipher_alg.)
- Drivers no longer register a list of supported algorithms. This
doesn't quite work when you factor in modes (e.g. a driver might
support both AES-CBC and SHA2-256-HMAC separately but not combined
for ETA). Instead, a new 'crypto_probesession' method has been
added to the kobj interface for symmteric crypto drivers. This
method returns a negative value on success (similar to how
device_probe works) and the crypto framework uses this value to pick
the "best" driver. There are three constants for hardware
(e.g. ccr), accelerated software (e.g. aesni), and plain software
(cryptosoft) that give preference in that order. One effect of this
is that if you request only hardware when creating a new session,
you will no longer get a session using accelerated software.
Another effect is that the default setting to disallow software
crypto via /dev/crypto now disables accelerated software.
Once a driver is chosen, 'crypto_newsession' is invoked as before.
- Crypto operations are now solely described by the flat 'cryptop'
structure. The linked list of descriptors has been removed.
A separate enum has been added to describe the type of data buffer
in use instead of using CRYPTO_F_* flags to make it easier to add
more types in the future if needed (e.g. wired userspace buffers for
zero-copy). It will also make it easier to re-introduce separate
input and output buffers (in-kernel TLS would benefit from this).
Try to make the flags related to IV handling less insane:
- CRYPTO_F_IV_SEPARATE means that the IV is stored in the 'crp_iv'
member of the operation structure. If this flag is not set, the
IV is stored in the data buffer at the 'crp_iv_start' offset.
- CRYPTO_F_IV_GENERATE means that a random IV should be generated
and stored into the data buffer. This cannot be used with
CRYPTO_F_IV_SEPARATE.
If a consumer wants to deal with explicit vs implicit IVs, etc. it
can always generate the IV however it needs and store partial IVs in
the buffer and the full IV/nonce in crp_iv and set
CRYPTO_F_IV_SEPARATE.
The layout of the buffer is now described via fields in cryptop.
crp_aad_start and crp_aad_length define the boundaries of any AAD.
Previously with GCM and CCM you defined an auth crd with this range,
but for ETA your auth crd had to span both the AAD and plaintext
(and they had to be adjacent).
crp_payload_start and crp_payload_length define the boundaries of
the plaintext/ciphertext. Modes that only do a single operation
(COMPRESS, CIPHER, DIGEST) should only use this region and leave the
AAD region empty.
If a digest is present (or should be generated), it's starting
location is marked by crp_digest_start.
Instead of using the CRD_F_ENCRYPT flag to determine the direction
of the operation, cryptop now includes an 'op' field defining the
operation to perform. For digests I've added a new VERIFY digest
mode which assumes a digest is present in the input and fails the
request with EBADMSG if it doesn't match the internally-computed
digest. GCM and CCM already assumed this, and the new AEAD mode
requires this for decryption. The new ETA mode now also requires
this for decryption, so IPsec and GELI no longer do their own
authentication verification. Simple DIGEST operations can also do
this, though there are no in-tree consumers.
To eventually support some refcounting to close races, the session
cookie is now passed to crypto_getop() and clients should no longer
set crp_sesssion directly.
- Assymteric crypto operation structures should be allocated via
crypto_getkreq() and freed via crypto_freekreq(). This permits the
crypto layer to track open asym requests and close races with a
driver trying to unregister while asym requests are in flight.
- crypto_copyback, crypto_copydata, crypto_apply, and
crypto_contiguous_subsegment now accept the 'crp' object as the
first parameter instead of individual members. This makes it easier
to deal with different buffer types in the future as well as
separate input and output buffers. It's also simpler for driver
writers to use.
- bus_dmamap_load_crp() loads a DMA mapping for a crypto buffer.
This understands the various types of buffers so that drivers that
use DMA do not have to be aware of different buffer types.
- Helper routines now exist to build an auth context for HMAC IPAD
and OPAD. This reduces some duplicated work among drivers.
- Key buffers are now treated as const throughout the framework and in
device drivers. However, session key buffers provided when a session
is created are expected to remain alive for the duration of the
session.
- GCM and CCM sessions now only specify a cipher algorithm and a cipher
key. The redundant auth information is not needed or used.
- For cryptosoft, split up the code a bit such that the 'process'
callback now invokes a function pointer in the session. This
function pointer is set based on the mode (in effect) though it
simplifies a few edge cases that would otherwise be in the switch in
'process'.
It does split up GCM vs CCM which I think is more readable even if there
is some duplication.
- I changed /dev/crypto to support GMAC requests using CRYPTO_AES_NIST_GMAC
as an auth algorithm and updated cryptocheck to work with it.
- Combined cipher and auth sessions via /dev/crypto now always use ETA
mode. The COP_F_CIPHER_FIRST flag is now a no-op that is ignored.
This was actually documented as being true in crypto(4) before, but
the code had not implemented this before I added the CIPHER_FIRST
flag.
- I have not yet updated /dev/crypto to be aware of explicit modes for
sessions. I will probably do that at some point in the future as well
as teach it about IV/nonce and tag lengths for AEAD so we can support
all of the NIST KAT tests for GCM and CCM.
- I've split up the exising crypto.9 manpage into several pages
of which many are written from scratch.
- I have converted all drivers and consumers in the tree and verified
that they compile, but I have not tested all of them. I have tested
the following drivers:
- cryptosoft
- aesni (AES only)
- blake2
- ccr
and the following consumers:
- cryptodev
- IPsec
- ktls_ocf
- GELI (lightly)
I have not tested the following:
- ccp
- aesni with sha
- hifn
- kgssapi_krb5
- ubsec
- padlock
- safe
- armv8_crypto (aarch64)
- glxsb (i386)
- sec (ppc)
- cesa (armv7)
- cryptocteon (mips64)
- nlmsec (mips64)
Discussed with: cem
Relnotes: yes
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D23677
GELI used to fail with EINVAL when a read request spanned a disk
sector whose contents did not match the sector's authentication tag.
The recently-added EINTEGRITY more closely matches to the error in
this case.
Reviewed by: cem, mckusick
MFC after: 2 weeks
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D24131
r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are
still not MPSAFE (or already are but aren’t properly marked).
Use it in preparation for a general review of all nodes.
This is non-functional change that adds annotations to SYSCTL_NODE and
SYSCTL_PROC nodes using one of the soon-to-be-required flags.
Mark all obvious cases as MPSAFE. All entries that haven't been marked
as MPSAFE before are by default marked as NEEDGIANT
Approved by: kib (mentor, blanket)
Commented by: kib, gallatin, melifaro
Differential Revision: https://reviews.freebsd.org/D23718
r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are
still not MPSAFE (or already are but aren’t properly marked).
Use it in preparation for a general review of all nodes.
This is non-functional change that adds annotations to SYSCTL_NODE and
SYSCTL_PROC nodes using one of the soon-to-be-required flags.
Approved by: kib (mentor, blanket)
Differential Revision: https://reviews.freebsd.org/D23637
Currently the installer will tag geliboot partitions with both BOOT and
GELIBOOT; the former allows the kernel to taste it at boot, while the latter
is what loaders keys off of.
However, it seems reasonable to assume that if a provider's been tagged with
GELIBOOT that the kernel should also take that as a hint to taste/attach at
boot. This would allow us to stop tagging GELIBOOT partitions with BOOT in
bsdinstall, but I'm not sure that there's a compelling reason to do so any
time soon.
Reviewed by: oshogbo
Differential Revision: https://reviews.freebsd.org/D23387
For the moment, supress the operation not supported messages at this level. In
the fullness of time, we will have better error tracking so we can diagnose
issues in the future.
Reviewed by: scottl@
If the bio was split into several bios going down, completion computes
bio_completed of the original bio as sum of the bio_completes of the
splits. For BIO_SETUP, bio_length means something different than the
length. it is the requested speedup amount, and is duplicated into the
splits, which is in fact reasonable, since we cannot know how the
previous activity was distributed among subordinate geoms. Obviously,
the sum of n bio_length is greater than bio_length for n > 1, which
triggers assert that bio_length >= bio_completed for e.g. geom_stripe
and geom_raid3.
Fix this by reassigning bio_completed from bio_length for completed
BIO_SPEEDED, I do not think it really mattters what we return in
bio_completed.
Reported and tested by: pho
Reviewed by: imp
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D23380
%20%20%20 is ugly and doesn't really help make human-readable devfs names.
PR: 243318
Reported by: Peter Eriksson <pen AT lysator.liu.se>
Relnotes: yes
It turns out there's a problem with using g_io to send the speedup. It leads to
a race when there's a resource shortage when a disk fails.
Instead, send BIO_SPEEDUP via struct buf. This is pretty straight forward,
except we need to transfer the bio_flags from b_ioflags for BIO_SPEEDUP commands
in g_vfs_strategy.
Reviewed by: kirk, chs
Differential Revision: https://reviews.freebsd.org/D23117
While some geom layers pass unknown commands down, not all do. For the ones that
don't, pass BIO_SPEEDUP down to the providers that constittue the geom, as
applicable. No changes to vinum or virstor because I was unsure how to add this
support, and I'm also unsure how to test these. gvinum doesn't implement
BIO_FLUSH either, so it may just be poorly maintained. gvirstor is for testing
and not supportig BIO_SPEEDUP is fine.
Reviewed by: chs
Differential Revision: https://reviews.freebsd.org/D23183
The previous behavior of leaving VI_OWEINACT vnodes on the active list without
a hold count is eliminated. Hold count is kept and inactive processing gets
explicitly deferred by setting the VI_DEFINACT flag. The syncer is then
responsible for vdrop.
Reviewed by: kib (previous version)
Tested by: pho (in a larger patch, previous version)
Differential Revision: https://reviews.freebsd.org/D23036
Combined with earlier nstart/nend removal it allows to remove several locks
from request path of GEOM and few other places. It would be cool if we had
more SMP-friendly statistics, but this helps too.
Sponsored by: iXsystems, Inc.
Those counters were abused for decade to workaround broken orphanization
process in different classes by delaying the call while there are active
requests. But from one side it did not close all the races, while from
another was quite expensive on SMP due to trashing twice per request cache
lines of consumer and provider and requiring locks. It lost its sense
after I manually went through all the GEOM classes in base and made
orphanization wait for either provider close or request completion.
Consumer counters are still used under INVARIANTS to detect premature
consumer close and detach. Provider counters are removed completely.
Sponsored by: iXsystems, Inc.
This code was not actively maintained since it was introduced 10 years ago.
It lacks support for many later GEOM features, such as direct dispatch,
unmapped I/O, stripesize/stripeoffset, resize, etc. Plus it is the only
remaining use of GEOM nstart/nend request counters, used there to implement
live insertion/removal, questionable by itself. Plus, as number of people
commented, GEOM is not the best place for I/O scheduler, since it has
limited information about layers both above and below it, required for
efficient scheduling. Plus with the modern shift to SSDs there is just no
more significant need for this kind of scheduling.
Approved by: imp, phk, luigi
Relnotes: yes
If we postpone consumer destruction till close, then the close calls should
not be ignored. Delay geom withering till the last close too.
MFC after: 2 weeks
X-MFC-with: r356162
Sponsored by: iXsystems, Inc.
Previous code closed and destroyed consumer even with I/O in progress.
This patch postpones the destruction till the last close.
MFC after: 2 weeks
Sponsored by: iXsystems, Inc.
Previous code closed and detached consumer even with I/O still in progress.
This patch adds locking and request counting to postpone the close till
the last of running requests completes.
MFC after: 2 weeks
Sponsored by: iXsystems, Inc.
Change the "count_until_fail" option of gnop, now it enables the failing
rating instead of setting them to 100%.
The original patch introduced the new flag, which sets the fail/rate to 100%
after N requests. In some cases, we don't want to have 100% of failure
probabilities. We want to start failing at some point.
For example, on the early stage, we may like to allow some read/writes requests
before having some requests delayed - when we try to mount the partition,
or when we are trying to import the pool.
Another case may be to check how scrub in ZFS will behave on different stages.
This allows us to cover more cases.
The previous behavior still may be configured.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D22632
Thanks to this option we can create more then one gnop provider from
single provider. This may be useful for temporary labeling some data
on the disk.
Reviewed by: markj, allanjude, bcr
Differential Revision: https://reviews.freebsd.org/D22304
Previous code closed and destroyed consumer even with I/O in progress.
This patch postpones the destruction till the last close, identical to
GEOM_STRIPE, since they seem to have common origin.
MFC after: 2 weeks
Sponsored by: iXsystems, Inc.
Previous code closed and destroyed direct read consumer even with I/O still
in progress. This patch adds locking and request counting to postpone the
close till the last of running requests completes.
MFC after: 2 weeks
Sponsored by: iXsystems, Inc.
Previous code destroyed softc even with provider still open, that resulted
in panic under load. This change postpones the free till the final close,
when we know for sure there will be no more I/O requests.
MFC after: 2 weeks
Sponsored by: iXsystems, Inc.
gvinum was the only GEOM class, using consumer nstart/nend fields. Making
it do its own accounting for orphanization purposes allows in perspective
to remove burden of that expensive for SMP accounting from GEOM.
Also the previous implementation spinned in a tight event loop, waiting
for all active BIOs to complete, while the new one knows exactly when it
is possible to close the consumer.
MFC after: 1 month
Sponsored by: iXsystems, Inc.
Add BIO_SPEEDUP bio command and g_io_speedup wrapper. It tells the
lower layers that the upper layers are dealing with some shortage
(dirty pages and/or disk blocks). The lower layers should do what they
can to speed up anything that's been delayed.
The first use will be to tell the CAM I/O scheduler that any TRIM
shaping should be short-circuited because the system needs
blocks. We'll also call it when there's too many resources used by
UFS.
Reviewed by: kirk, kib
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D18351
to specify an optional separator to insert before partition name;
eg if it's set to "c/", you'll get "ada0c/s1" instead of "ada0s1".
(It cannot be set to just “/“, since ada0 is a device node, not
a directory.)
Reviewed by: imp
MFC after: 2 weeks
Sponsored by: Klara Inc.
Differential Revision: https://reviews.freebsd.org/D22193
pp->private just can not be NULL in those places.
In g_disk_start() and g_disk_ioctl() both dp != NULL and !dp->d_destroyed
should always be true if disk_gone() and disk_destroy() are used properly,
since GEOM does not send requests to errored providers. If the protocol is
not followed, then no amount of additional checks here give real safety.
In g_disk_access() though the checks are useful, since GEOM blocks only
new opens for errored providers, but allows closes. It should not happen
if disk_gone() and disk_destroy() are used properly, but may otherwise.
To improve cases when disk_gone() is not used, call it from disk_destroy().
It does not give full guaranties, but it errors the provider and makes
GEOM block unwanted requests at least after some race.
MFC after: 2 weeks
For normal I/Os consumer and provider statuses are checked by g_io_check().
But ioctl calls often do not go through it, being dispatched directly. This
change makes their semantics more alike, protecting lower levels.
MFC after: 2 weeks
Add ATF tests for most gmultipath operations. Add some dtrace probes too,
primarily for configuration changes that happen in response to provider
errors.
PR: 178473
MFC after: 2 weeks
Sponsored by: Axcient
Differential Revision: https://reviews.freebsd.org/D22235
In most cases with debug disabled this function does nothing, but argument
passing and the call still cost measurable time due to cache misses, etc.
MFC after: 2 weeks
Sponsored by: iXsystems, Inc.
It closes the race condition and so allows to remove few NULL checks.
Also while there, use dev->si_drv1 in addition to cp->private to store
softc pointer. For calls coming from the dev side it gives reliable cache
hit instead of often miss before.
MFC after: 2 weeks
Sponsored by: iXsystems, Inc.
to geom, and nothing we call requires it to be held. It's left over
from a time when the latter wasn't the case. Retire it.
Reviewed in concept: scottl@
via 'diskinfo -v'. This avoids the need to track it down via CAM,
and should also work for disks that don't use CAM. And since it's
inherited thru the GEOM hierarchy, in most cases one doesn't need
to walk the GEOM graph either, eg you can use it on a partition
instead of disk itself.
Reviewed by: allanjude, imp
Sponsored by: Klara Inc
Differential Revision: https://reviews.freebsd.org/D22249
filling in the same defaults that the current userland module uses.
This allows an old geom_nop.so userland module to work with a new kernel.
Approved by: imp (mentor)
Reviewed by: cem
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21972
the request. It is the same as gctl_get_paraml() except that the request
is not marked with an error if the parameter is not present.
Approved by: imp (mentor)
Reviewed by: cem
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D21972
