The IOs are never completed with NOMEM from push/sequence callbacks and
NOMEM IOs are already retried in internal callbacks, so there's no point
in calling _bdev_io_handle_no_mem().
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Iecc2a41f2a394836f62d541e6235277f333f226b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17765
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
The functions that were passed as callbacks for the memory domain
pull/push calls were prefixed with an underscore, which doesn't really
explain the difference between the corresponding functions without an
underscore. So, they're now renamed to *_and_track() to emphasize that
they additionally responsible for tracking IOs.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ia9e56230fe244d2c64d729e97445fae105418a76
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17931
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
The IOs don't need to be put onto the io_memory_domain queue if there's
no need for memory domain push. This makes push_data consistent with
other memory domain operations (pull_data, pull_md, push_md).
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I85d95f6ce580a15b23f56ab5101e49236f341cb1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17763
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Write_zeroes command fallback had used its own split logic but multiple
writes had been serialized.
Use the unified split logic also for the write_zeroes command fallback.
This not only improves the performance but also simplifies the code.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I955870947ae036482871453b4870f06f6f7f947b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17902
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
As same as copy command, calculation of max write_zeroes size for
fallback case includes division and is costly. The result is constant
for each bdev. Hence, we can calculate it only once and store it into
bdev->max_write_zeroes at bdev registration. However, in unit tests,
bdev->blocklen and bdev->md_len can be changed dynamically. Hence,
adjust bdev->max_write_zeroes for such changes.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I16e4980e7a283caa6c995a7dc61f7e77585d464e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17911
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
ZERO_BUFFER_SIZE is in bytes but it is easier to calculate max
write_zeroes in blocks first and then get the minimum between max
write_zeroes in blocks and remaining_num_blocks rather than converting
remaining_num_blocks to num_bytes. This is helpful to store the result
into bdev->max_write_zeroes for fallback case.
We have one small fix in this patch. As we recently fixed
bdev_io_get_max_buf_len(), to get aligned length,
spdk_bdev_get_buf_align() - 1 is correct.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I104bc837c9eee1303664bfdb3559b0e840d6f0e5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17910
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
The generic bdev layer has a fallback meachanism for the copy command
used when the backend bdev module does not support it. However, its max
size is limited. To remove the limitation, the fallback supports split by
using the unified split logic rather than following the write zeroes
command.
bdev_copy_should_split() and bdev_copy_split() use spdk_bdev_get_max_copy()
rather then referring bdev->max_copy to include the fallback case.
Then, spdk_bdev_copy_blocks() does the following.
If the copy size is large and should be split, use the generic split
logic regardless of whether copy is supported or not.
If copy is supported, send the copy request, or if copy is not
supported, emulate it using regulard read and write requests.
Add unit test case to verify this addition.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Iaf51db56bb4b95f99a0ea7a0237d8fa8ae039a54
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17073
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
As name suffix, _done has been used more often than _complete for
fallback function names. 100 chars per line is suggested implicitly.
Do these small clean up in this patch.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Id14dd3f09be8fd49b947b7a8f8b87108fb56c346
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17900
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Calculation of max copy size for fallback case includes division and is
costly. The result is constant for each bdev. Hence we can calculate it
only once and store it into bdev->max_copy at bdev registration.
Calculation of max copy size for fallback case is almost same as
calculation of max write zero size for fallback case. To reuse the
calculation, the helper function is named as bdev_get_max_write() and
has a num_bytes parameter.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: Iac83a1f16b908d8b36b51d9c51782de40313b6c8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17909
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
As we recently fixed bdev_io_get_max_buf_len(), to get aligned length,
spdk_bdev_get_buf_align() - 1 is correct.
_bdev_get_block_size_with_md() considers both interleaved metadata and
separate metadata cases. It is simpler to use
_bdev_get_block_size_with_md().
The copy command fallback uses write command. As the write zeroes
fallback does, bdev->write_unit_size should be considered.
Fix all in this patch.
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I88fe1b250289f2bab7b541523e8be931eeb8150c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17899
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This is done in preparation for retrying IOs on ENOMEM when pushing
bounce data. Also, rename md_buffer to md_buf to keep the naming
consistent with other code which uses this abbreviation.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I014f178a45a2a751ecca40d119f45bf323f37d0c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17762
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The IOs will now be retried after ENOMEM is received when doing memory
domain pull or appending an accel copy. The retries are performed using
the mechanism that's already in place for IOs completed with
SPDK_BDEV_IO_STATUS_NOMEM.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I284643bf9971338094e14617974f7511f745f24e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17761
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
The IOs with an outstanding memory domain push/pull or accel sequence
finish operation are now added to the io_outstanding counter. It'll be
necessary to correctly calculate nomem_threshold when handling ENOMEM
from those operations.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ice1fb94f1c9054a3a96312a0960ac5085d0b21bc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17760
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
The leading underscore usually indicate that a function providing the
actual implementation for something that's called from some other
wrapper function without the leading underscore. That is not the case
for these functions, so this patch removes the leading underscores.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I6e1186b156116249ee53a3845ae99ba87db5122b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17868
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
In the next patches we'll need to increment the io_outstanding from a
few more places, so it'll be good to have a dedicated function for that.
Also, move _bdev_io_decrement_outstanding() up, so that both functions
are near each other.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I1af5dbe288f7f701c8ba5e85406f02330ae21a39
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17759
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
There are some common operations that need to be done each time a
sequence is executed (and more will be added in the following patches),
so it makes sense to have a common callback. data_transfer_cpl is used
for executing user's callbacks since it's unused at this point.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I4570acbdbe158512d13c31c0ee0c7bb7bf62d18c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17678
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
The IOs are now kept on the io_memory_domain queue only if they have an
outstanding pull/push operation. It'll make it easier to support
retrying pull/push in case of ENOMEM.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: If5a54fac532206ee8472bacf364a5ef6cde8edea
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17677
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This is a preparation for reusing the code handling nomem_io for
other type of NOMEM errors (e.g. from pull/push/append_copy). This
patch doesn't actually change anything functionally - only IOs completed
by a module with SPDK_BDEV_IO_STATUS_NOMEM status are retried.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I12ecb2efcf2d2cdf75b302f9f766b4c16ac99f3e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17676
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
The actual memory domain push already only happened for successfully
completed requests, but the code would go still go through
_bdev_io_push_bounce_data_buffer(), which could cause issues for IOs
completed with NOMEM, because the bounce buffer would be released in
_bdev_io_complete_push_bounce_done().
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Id1af1e31cb416e91bf11101a5ce7919530245e1e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17866
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
The sequence is associated with parent IO, so that's the IO that should
be used when executing a sequence.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ifcdb06094b38a5eaee1691e5aa8de1c8dc9d01a6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17865
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It always returns 0 and any errors are reported in the callback. Making
it void simplifies error handling.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I0d4299a2789a688eae38d76de46d1baf27cbbd8f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17194
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Mellanox Build Bot
bdev modules should have call spdk_bdev_io_complete twice
for the same IO. We can help find cases where this happens
by adding an assert in spdk_bdev_io_complete - confirming
that the current status is still PENDING, before changing
it to the status passed by the caller.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Id8a044a94113f1ac5e3c8d86e426654bfa8d5c5a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17330
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reset the status for a bdev_io that fails with NOMEM status
back to PENDING immediately when it is put on the nomem_io
list, instead of waiting until it gets submitted again.
This helps keep the bdev_io states consistent, so that if
we need to complete these IO for abort reasons later, that
we aren't completing IO that already have a non-PENDING
state.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I9532095141209ed6f7af362b52c689da62e755ce
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17335
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Make sure UUID is present for every bdev, even ephemeral ones.
Furthermore, this change removes assumption that bdev UUID
may remain empty.
Change-Id: I924c1ba9dedfe88a05044bb1073f28085735b1c1
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17106
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Fixed function is used to determine if it is possible to get iobuf
from the pool. To make sure that buf size alignment requirement is
satisifed value returned shall include alignment value but subtracted
by one.
e.g.
transaction size length = 64k
buffer alignment = 1 byte (no alignment requirement)
metadata length = 0
Without the fix the function returned 64k + 1, now it returns 64k
which is correct behavior and allows to proceed with further command
processing (if max buffer size limit is set to 64k only).
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I09104ad21b3652ba1aa5c3805a04b1c6549d04ac
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17513
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Fixed issue indicated by Klocwork scan. 'desc->bdev'
is assigned to 'bdev' ptr, before verification that
'desc' is not NULL
Change-Id: I36e63c27b4d3220e85524133a0ec0e3521770875
Signed-off-by: Marcin Spiewak <marcin.spiewak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17350
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Requests that have their data pushed/pulled from a memory domain or have
an accel sequence executed aren't handled by a bdev module, so we
shouldn't submit an abort request. Those operations cannot be aborted
either, so the abort request is failed in this case.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Icd185c4a2951a555d321cd037de0af1ab157f37a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17020
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
These operations are handled internally by the bdev layer, so it should
first wait until they're completed before issuing reset to a bdev
module.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I74f0d42dcb9a289aa7c3115ca309cb92870548e2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17019
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Similarly to requests executed by accel, we need to track bdev_ios that
have their data pushed/pulled.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ie6b0d2c058e9f13916a065acf8e05d1484eae535
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16978
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
It will make it possible to check if a request is being processed by
accel when doing resets/aborts.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ice07211df316e1eee9640e750ff8e176c8a3ca6f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16977
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
This patch enables passing accel sequence for read requests. The
handling is pretty similar to writes, but the sequence is executed after
a request is completed by a bdev module.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I79fd7d4873265c81a9f4a66362634a1c4901d0c9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16975
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
It is now possible to submit a write request with a sequence of accel
operations that need to be executed before actually writing the data.
Such requests will be directly passed to a bdev module (so that it can
append subsequent operations to an accel sequence) if that bdev supports
accel sequences and the request doesn't need to be split. If either of
these conditions are not met, bdev layer will execute all the
accumulated accel operations before passing the request to a bdev
module.
The reason for not submitting split IOs with an accel sequence is that
we would need to split that accel sequence too. Currently, there's no
such functionality in accel, so we treat this case in the same way as if
the underlying bdev module didn't support accel sequences (it's executed
before bdev_io is split).
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I66c53b3a1a87a35ea2687292206c899f80aaed4a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16974
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
bdev_io_should_split() adds some non-zero overhead, so checking it
multiple times in an IO path is inefficient. So, to avoid that, call
bdev_io_should_split() once during IO initialization and cache the
result in bdev_io.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I1da6514d409f8a4e4bbb14722dd53b2c88988cac
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17058
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
This channel will be used to execute accel operation sequences.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ied4bb57d14a50a923908ffb13ef4ba34ca65175c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16972
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Modules can now report that they support accel chaining for specific
operations through the accel_sequnce_supported() callback.
The support is reported per IO type. This allows modules to support
accel sequences for some operations, while relying on the bdev layer to
handle them for other IO types.
Only bdevs without separate metadata buffers are allowed to support this
new mode. That's because metadata in separate buffer is expected to use
the same memory domain as data buffers. With an accel sequence, those
data memory domains can change, while metadata's memory domain always
stays the same. To support bdevs with separate metadata buffers, we'd
need to add separate pointers for metadata's memory domain. For now,
simply disallow registering bdevs with separate metadata supporting
accel sequences.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I0c49cc00096837d70681a69b2633c2cb3dfd4e39
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16971
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
If an IO is completed, before submitting it to a module, it isn't put on
the io_submitted list, so we can't use bdev_io_complete() to complete
it, as it'll break that list. To avoid that, a new function was added,
bdev_io_complete_unsubmitted(), that will safely complete the IOs in
such case. For now, it's equivalent to executing user's completion
callback, but it'll serve as a good place to release any resources that
should be freed before an IO is completed.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I1442ead9d272d9210553803bed1d1c989a2bf761
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16970
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Add default copy command support in bdev layer for backing devices that
does not support copy command.
Signed-off-by: Rui Chang <rui.chang@arm.com>
Change-Id: I5632e25544e95ac0c53ff91c4cd135dac53323ae
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16638
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The spdk_bdev_ext_io_opts structure is used to pass extra options when
submitting a bdev IO request, without having to modify/add functions to
handle new options. Additionally, the structure has a size field to
allow adding new fields without breaking the ABI (and thus having to
bump up the major version of a library).
It is also a part of spdk_bdev_io and there are several reasons for
removing it from that structure:
1. The size field only makes sense in structures that are passed
through pointers. And spdk_bdev_ext_io_opts is indeed passed as a
pointer to spdk_bdev_{readv,writev}_blocks_ext(), however it is
also embedded in spdk_bdev_io (internal.ext_opts_copy), which is
also part of the API. It means that each time a new field is added
to spdk_bdev_ext_io_opts, the size of spdk_bdev_io will also
change, so we will need to bump the major version of libspdk_bdev
anyway, thus making spdk_bdev_ext_io_opts.size useless.
2. The size field also makes internal.ext_opts cumbersome to use, as
each time one of its fields is accessed, we need to check the size.
Currently the code doesn't do that, because all of the existing
spdk_bdev_ext_io_opts fields were present when this structure was
initially introduced, but we'd need to do check the size before
accessing any new fields.
3. spdk_bdev_ext_io_opts has a metadata field, while spdk_bdev_io
already has u.bdev.md_buf, which means that we store the same thing
in several different places in spdk_bdev_io (u.bdev.md_buf,
u.bdev.ext_opts->metadata, internal.ext_opts->metadata).
Therefore, this patch removes all references to spdk_bdev_ext_io_opts
from spdk_bdev_io and replaces them with fields (memory_domain,
memory_domain_ctx) that were missing in spdk_bdev_io. Unfortunately,
this change breaks the API and requires changes in bdev modules that
supported spdk_bdev_io.u.bdev.ext_opts.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I49b7524eb84d1d4d7f12b7ab025fec36da1ee01f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16773
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It isn't used in this function and the callers always pass NULL.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I07baa13a25b1e4e0b8832a093a53250392b10f10
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/16682
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Starting in SPDK 23.01, calling spdk_bdev_register() and
spdk_bdev_examine() from a thread other than the app thread was
deprecated. This commit removes the deprecation and as such calling
these functions from a thread other than the app thread is an error.
As a side effect of this commit, all bdev module examine_config() and
examine_disk() callbacks will be called on the app thread.
Change-Id: Idaae06608101e2a513d9312ac5544ffe94effe4a
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15826
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
If multiple claims exist on a bdev, examine_disk() is called for each of
them.
Change-Id: I0a6dc3e4bd1da20bbcbddf97a16e04c62c82354c
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15290
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
This commit has no functional change. It refactors an if statement into
a case statement in preparation for supporting claims v2.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: I1862428c91a7066ad9079878d4c1b690a5ef631c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15289
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This implements the v2 claims API. Compared to the original v1 claims,
v2 claims:
- Support read-write-once, read-write-many, and read-only-many claims.
- Are claimed with spdk_bdev_module_claim_desc().
- Are associated with a bdev descriptor that is passed to
spdk_bdev_module_claim_bdev_desc().
- Are released upon close of the bdev descriptor used to obain the
claim.
- Cannot be taken when a descriptor other than the one passed to
spdk_bdev_module_claim_bdev_desc() has write access.
Later commits in this series are needed to fully integrate them with the
bdev subsystem.
Change-Id: I39a356f5893aa45ac346623ec9ce0ec659b38975
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15288
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
As new claim types are introduced, printing error messages about who
holds a claim will get more complicated. This refactors the error
message code into a function to prevent code duplication.
Signed-off-by: Mike Gerdts <mgerdts@nvidia.com>
Change-Id: Icdc5332214f3974e75baf11ba5ea02172c4275e1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15287
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
