When blob is deleted or reduced in size, the trimmed clusters
are cleared using bs_batch_clear_dev().
blob_persist_clear_clusters() iterated over the clusters
and batched each contigous set of LBA.
The lba_count is of the uint32_t type, long enough
contigous set of LBA could overflow it.
As a result that range would only be cleared with the
overflown value.
This patch prevents that overflow from occuring by
verifying against UINT32_MAX.
This is already addressed with an API change in SPDK 21.10:
(f01146ae)blob: use uint64_t for unmap and write_zeroes lba count
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I17282a2a81587c26b7e89a74dcc6a9c8a017ce15
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9971
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
During spdk_bs_init() data present on the device can be cleared
depending on the clear_method used.
Picking BS_CLEAR_WITH_UNMAP results in deallocate command sent
when NVMe device is used.
The size of logical blocks do not have to match the indirection unit
size (IU) of the device. The SSDs can only unmap full physical blocks.
Since the deallocate command is advisory, the controller can complete
the command without deallocating any blocks.
In practice this results in some controllers not deallocating any blocks
if number of logical blocks is not aligned to the IU size.
Make sure that spdk_bs_init() handles this case.
The size of the unmap is now chunked at UINT32_MAX - 127,
which is an even multiple of 512B blocks with a 64KiB IU.
This is already addressed with an API change in SPDK 21.10:
(f01146ae)blob: use uint64_t for unmap and write_zeroes lba count
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ie00bd5ae7cdaec4f1486f16fceb6cb18e30bf563
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9970
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7564 (master)
(cherry picked from commit 330e9f779c)
Change-Id: I72da312e0d7fe71ae40b582501b52be615de4b63
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9992
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This is done in order to detect cpu times in a more deterministic
fashion.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7563 (master)
(cherry picked from commit 0ea8956025)
Change-Id: I005d6fececdc15500bcf637e96f2fbecbd4273f1
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9989
Reviewed-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Issues like https://github.com/spdk/spdk/issues/1784 showed that
governor may need some more time to properly adjust cpufreq's sysfs
knobs. In case we see that frequency setting hasn't changed after
given iteration, wait for 0.5s to make sure it finally settles.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6283 (master)
(cherry picked from commit 92e49c0044)
Change-Id: Ibf25d0d1962bf2b07b13d60d6096c5cc185c1279
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9967
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
spdk_vtophys() takes a mapping_length parameter, so
it can return the length for which the returned
virtual address is valid.
But spdk_vtophys() will only return the max
between the valid length and the input mapping_length
parameter.
So the nvme SGL building code for contiguous buffers
was broken, since it would only set the mapping_length
once, before the loop started. Worst case, if a buffer
started just before (maybe 256 bytes) before a huge page
boundary, each time through the loop we would create
a new SGL for only 256 bytes at a time, very quickly
running out of SGL entries for a large buffer.
Fixes#1852.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7659 (master)
(cherry picked from commit 5354d0c63f)
Change-Id: Ib1000d8b130e8e4bfeacccd6e60f8109428dfc1e
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9958
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8646 (master)
(cherry picked from commit 117e555d9d)
Change-Id: I8f21f6f9a1cf837931646a7458f0ad8147cf6810
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9951
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Signed-off-by: Maciej Wawryk <maciejx.wawryk@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6940 (master)
(cherry picked from commit 1fc0c2d8b9)
Change-Id: I8d75f5fa22e79c93d1f8929b6df048b440d9fb8a
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9934
Reviewed-by: Maciej Wawryk <maciejx.wawryk@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Replace old '/home/sys_sgsw' location by new created
'/home/sys_sgci/spdk_dependencies'
Also update readme with current user name.
Signed-off-by: Maciej Wawryk <maciejx.wawryk@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6927 (master)
(cherry picked from commit 1fc0c2d8b9)
Change-Id: Ifcea15f704183c7ae0044ee02bd28b6bd891b262
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9933
Reviewed-by: Maciej Wawryk <maciejx.wawryk@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Currently, crypto job also executes tests from the SPDK_TEST_BLOCKDEV
suite, including spdk_dd tests. The dd tests under a bare-metal system
can take up to 4 minutes alone and that brings the crypto job to the
very edge of timing out.
Since these tests are extensively covered under separate vg jobs we
won't lose any coverage by excluding them from crypto's - the basic
set of SPDK_TEST_BLOCKDEV tests would still remain as part of this
job.
Fixes: #2017
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8545 (master)
(cherry picked from commit c6db5f41e5)
Change-Id: Ic245c7a1711fbb95131df977b81dbdac8af1c15a
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9915
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Simply fail. It's expected to have this image already installed on
the system.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6375 (master)
(cherry picked from commit ee191fa807)
Change-Id: Ie7bbde84fe633c8b5a4d41bd6dcbee66d6f4a0c5
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9914
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This is needed since fio, by default, links to libtcmalloc if found
during the build. See:
01fe773df4Fixes#2192
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9822 (master)
(cherry picked from commit bec627fcb0)
Change-Id: I1de8fe9d489f92acb79ef0f169229a6c09ce1179
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9859
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
These are currently prone to false-positives especially in terms of
scoping. Consider the following:
foo() {
local bar=(42)
echo "${bar[*]}"
}
bar=43
foo
echo "$bar"
Some versions of shellcheck, including the latest, 0.7.2, complain
about $bar being reused as a plain string here. This is incorrect
since foo() holds its own copy of bar[@] hence the assignment which
takes place outside of it doesn't affect its content.
SC2178 can be mitigated be reversing the order of declaration:
bar=43
foo() { ... }
...
but the SC2128 still remains.
Currently, in our code majority of these warnings are coming from
false-positives due to initial source'ing which most of our test
scripts do (e.g. they fetch a function where local bar=() is used
and in the test itself $bar happens to be assigned a plain string.
This is still valid code).
To mitigate, disable these directives untill shellcheck is capable
of properly interpreting scoping when checking them.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8503 (master)
(cherry-picked from commit 9d48662103)
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Change-Id: Ifbde973eae6e261d79e1c340eb28644bce5f4e45
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9801
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Older versions of fio depend on raw driver which recently got removed
from the linux kernel:
git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=603e4922f1c
Due to that, these fio versions won't build anymore under latest
kernels available, e.g., for fedora (due to missing raw.h from the
kernel-headers package). Details here:
https://github.com/axboe/fio/commit/382975557
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9718 (master)
(cherry picked from commit 65f01122b2)
Change-Id: Ie63769395c482f1965c6664e8f5143c71ef80253
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9838
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
SC2268: Avoid x-prefix in comparisons as it no longer serves a purpose
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8505 (master)
(cherry picked from commit dd2b935dc4)
Change-Id: Ica5ddfa8c39f34741c7344906abe802ff7451b1f
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9837
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
SC2251: This ! is not on a condition and skips errexit.
Use && exit 1 instead, or make sure $? is checked.
This is critical since our test suites heavily depend on errexit
to catch all abnormal conditions. Replace ! foo ... with a call to
NOT() wherever possible. For the test/ocf pieces, use [[ ]] instead
to define the condition for the existance of the ocf bdev.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8504 (master)
(cherry picked from commit 742f04b3f7)
Change-Id: Ied3f99b9f1dc0594e4aef64fc21e51498f19ac23
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9836
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
SC2269: This variable is assigned to itself, so the assignment does
nothing.
Since the --id check doesn't change the value of $id, remove it
completely. This simplifies the process_shm() so the --id type is
considered to a be default and $id changed only in case --pid was
explicitly requested.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8502 (master)
(cherry picked from commit 63fd943282)
Change-Id: If8a4a43c6822fa1b32217a0b783cb96d0fee2644
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9835
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
SC2030: Modification of var is local (to subshell caused by pipeline)
This directive is raised since the for loop (or rather its contents)
is piped through to another process. In this context it's not harmful
since we are interested only in sending the stdout but since we operate
on variables that happen to be local to the process that handles the
loop shellcheck sees that as a potential issue. To make it happy
remove the pipe and sort pci addresses prior running the loop.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8501 (master)
(cherry picked from commit 8c176db2ca)
Change-Id: I835c5ac1da9012129c5d01d62880307f70caab1b
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9834
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This is a first commit in the series addressing potential Bash issues
as discovered by the latest shellcheck release (0.7.2, shipped with
the very latest fedora33 and fedora34). The goal is to either fix,
locally or globally disable given directive(s).
SC2155: Declare and assign separately to avoid masking return values
Simplify the setting of the variable pointing at the root of the repo.
Also, keep it consistent with the rest of the scripts and declare it as
$rootdir.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8500 (master)
(cherry picked from commit 36f5f0df17)
Change-Id: I63e0b1a85ce16f7983e9ba6dd985046e8a39a650
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9833
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
The define ZBD_IGNORE has been removed from upstream fio.
fio will now return an error if --zonemode=zbd is used with a
--filename that points to something that is unsupported by either
the zbd code in fio or by the ioengine itself.
ioengines are now supposed to return -EINVAL for unsupported files
(instead of returning 0 with ZBD_IGNORE for unsupported files).
This change does not need any special ifdef FIO_IOOPS_VERSION check.
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8326 (master)
(cherry picked from commit bd53d09c31)
Change-Id: I7bec0b1f5dc8f166ebf683f6f3937b2ef295a21e
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9832
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
open_blobids holds bit array of currently open blobs,
this is a way for quicker determination than iterating
over all blobs. See patch introducing it:
(30ee8137)blob: Add a bitmask for quickly checking which blobs are open
That patch added resizes of this bit array to bs init
and bs recovery path (not shut down cleanly).
But that patch skipped over bs load from a clean shutdown.
This resulted in blob open having multiple blob pointers that
target the same blob id.
Fixes#1937
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7998 (master)
(cherry picked from commit 42ad9d0379)
Change-Id: I3c42a63d168d1f5b013b449f010c5b207936045b
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9831
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Patch f95e331be3a1f856b816948990dd2afc67ea4020 is no longer needed,
since it was solved at upstream level.
Signed-off-by: Pawel Piatek <pawelx.piatek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8698 (master)
(cherry picked from 93c4adda4d)
Change-Id: Icf064b121351566988c26390f45286e7a456ca39
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8731
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Caught by clang 12.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8854 (master)
(cherry picked from d673fcee3d)
Change-Id: Ibed23c3aab80508aebf06e4ba44e14e385a3256b
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8852
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Caught by clang 12 and UBSAN.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I2fbb82a9225ab0cb8616bd671f8b81987cb3d708
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8853
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
We cannot rely on DSM/DEALLOCATE as a write zeroes
alternative, even if DLFEAT reports that deallocated
blocks will be read as all zeroes. DEALLOCATE is
advisory, meaning that blocks may not actually be
deallocated. In cases where they are not deallocated,
they will not be read back later as zeroes.
QEMU 6.0 started reporting DLFEAT as returning zeroes
for deallocated blocks but for some of our write
zeroes tests, blocks aren't actually deallocated.
We may be able to add quirks in the future if we know
that a controller reliably deallocates blocks, but
for now we need to revert this completely.
Note that since bdev/nvme module now does not support
write zeroes in any cases, we need to disable the
write zeroes call in the unit tests.
Fixes issue #1932.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7723 (master)
(cherry picked from 3faf457f56
with minor modifications)
Change-Id: I79f0673774b621a9ffcc46891728cc7719e34cdb
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8879
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
QEMU 6.0 by default uses a RedHat dev/vendor ID rather
than the Intel one that has always been used to date.
We need the NVME_QUIRK_MAXIMUM_PCI_ACCESS_WIDTH quirk
so that we do not use wide instructions to copy SQEs
to a virtualized CMB, since QEMU does not support
that.
The NVME_INTEL_QUIRK_NO_LOG_PAGES quirk is only needed
for devices with SPDK_PCI_VID_INTEL, so we do not need
to carry this one over to the new REDHAT entry.
Fixes issue #1986.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8226 (master)
(cherry picked from 130c94c489)
Change-Id: I3d339b3525e7c6ceb792eb9d143e7a922c19344d
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8878
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
In case the addr argument was not an existing unix socket file the rpc
client would consider it to be an actual ip address. As a result
connect() would be called with improper set of arguments. This could
cause the rpc.py to block for undesired amount of time until connect()
finally decided to return (seen on some fedora33 builds).
This was affecting sh wrapper functions like waitforlisten() which
use rpc.py to determine if given app is ready to be talk to blocking
execution of the tests for way too long then intendent.
To avoid such a scenario determine the format of the address and use
routines proper for given address family.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7777 (master)
(cherry picked from 6c1a1a3dca)
Change-Id: Iaac701d72c772629fa7c6478ff4781b0c5d485d5
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8877
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Fixes issue #1847
Signed-off-by: paul luse <paul.e.luse@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7016 (master)
(cherry picked from 48c057814b
with minor changes)
Change-Id: I34109935b5e8187192239fa34ff43946d0e7632f
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8876
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This is relevant mostly for the VMs. Allocating that much memory puts
a lot of pressure on remaining pieces of the system as currently we
allocate total of 12GB memory per VM instance in the CI pool. Default of
4GB hp should be enough to cover majority of the tests hence if there's
any suite that requires more memory, setup.sh should be called there
directly with an optimal value.
This is done to make sure we can accomodate proper resources in the CI
by knowing what are the actual requirements rather than just blindly
allocating "as much as possible".
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8436 (master)
(cherry-picked from 227428c3a4 with
minor changes)
Change-Id: Ie958c518c7ed702b068c7fc9042183fdf41aa816
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8875
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
There is a special case when using 8-byte metadata + PI + PRACT
where no metadata is transferred to/from controller.
Since _nvme_ns_cmd_rw() already calculates the proper sector size
using _nvme_get_host_buffer_sector_size(), which takes PRACT into
account, change the sectors_per_max_io calculation to also take
PRACT into account.
This will avoid certain requests that don't need splitting getting
split.
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6247 (master)
(cherry picked from commit 4249dc1010)
Change-Id: I8d450d37c2458453701189f0e0eca4b8fe71173b
Signed-off-by: John Kariuki <John.K.Kariuki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8023
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This particular version reports plenty of false positives causing the
test to fail. Example:
nvme.md:2: warning: @copybrief or @copydoc target 'spdk_nvme_ctrlr_alloc_io_qpair()\ilinebr' not found
nvme.md:2: warning: @copybrief or @copydoc target 'spdk_nvme_ctrlr_get_ns()\ilinebr' not found
nvme.md:2: warning: @copybrief or @copydoc target 'spdk_nvme_ns_cmd_read()\ilinebr' not found
nvme.md:2: warning: @copybrief or @copydoc target 'spdk_nvme_ns_cmd_readv()\ilinebr' not found
nvme.md:2: warning: @copybrief or @copydoc target 'spdk_nvme_ns_cmd_read_with_md()\ilinebr' not found
nvme.md:2: warning: @copybrief or @copydoc target 'spdk_nvme_ns_cmd_write()\ilinebr' not found
nvme.md:2: warning: @copybrief or @copydoc target 'spdk_nvme_ns_cmd_writev()\ilinebr' not found
This is potentially a regression in caused by the following issue:
https://github.com/doxygen/doxygen/issues/7948
The above seems to be resolved in latest releases, however, until
they are shipped as part of distros we support, similar mitigation is
in order.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Change-Id: I1e39c94f9f388de99e1106bcf15750e4b39384e6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7160
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7948
Reviewed-by: Karol Latecki <karol.latecki@intel.com>
Signed-off-by: Simon A. F. Lund <simon.lund@samsung.com>
Change-Id: I039a2e22a665e0d52082e6876f2c7c1cd9a336c1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6386
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
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7931
Reviewed-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Patches stopped applying cleanly because of dpdk/dpdk changes:
7d5cfaa7508de0fd248b05effbf421a98317006a
4ad4b20a79052d9c8062b64eaf0170c16a333ff8
Needed to rebase custom patches.
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Change-Id: I41b34c5c58070720c4010632e89468042e470613
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7938
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Commit e9ebbdaefc updated
custom patches for mainline DPDK, but did not keep their
previous versions. We still need the old patches to build
with older DPDK versions.
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7610 (master)
(cherry picked from commit c06df11513)
Change-Id: Ibe222a444bed2fdee63c8dcfcebc8402bb991fbe
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7637
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Previous patches stopped applying cleanly at DPDK
21.05-rc1, needed to update.
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7538 (master)
(cherry picked from commit e9ebbdaefc)
Change-Id: If276f608c8394439064e9dfe30fed6d5531805d2
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7636
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
DPDK recently clarified some semantics on the rte_devargs 'data'
and 'args' fields. This actually breaks our use of the 'data'
field to store the 2 second timeout timestamp for delaying
attach to newly inserted devices. Investigating this further,
it does not seem our use of the 'data' field was valid - it just
happened to work until now.
We could use the 'args' field now. But knowing whether to use
'args' or 'data' would then be dependent on the DPDK version.
We cannot use RTE_VERSION_NUM to decide, because this is a
compile time decision, and it is possible in shared library
use cases that we could actually link and execute against a
different version of DPDK than we built against.
So instead we will create our own env_devargs structure that
will store these allowed_at timestamps. Currently it's just
a linked list (which is exactly how DPDK does it) - we could
make it more optimal with a hash table down the road, but this
code only executes when we are doing PCI enumeration so it is
not performance critical.
Fixes#1904.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7506 (master)
(cherry picked from commit 6c75f0a266)
Change-Id: I3ee5d65ba90635b5a96b97dd0f4ab72a093fe8f7
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7635
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7505 (master)
(cherry picked from commit de9f4857d0)
Change-Id: I5265f99fc12e134bf156de75de433a739b9ef1ef
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7634
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Fio 3.19 fails to build with Clang 11 because of
implicit-const-int-float-conversion error.
Disable -Werror checking for now.
Fixes#1848
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6991 (master)
(cherry picked from commit 7366e569fb)
Change-Id: I0f3922d7af412de35d9f79f3f5413af146996bf3
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7633
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Facing similar issue as with 9ab2855d - can't build SPDK against fio
3.19.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6805 (master)
(cherry picked from commit f58afe93b8)
Change-Id: I4ea22e4f068a19d8d87be836454ec8deeeb2d48a
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7649
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Use -Wno-stringop-overflow only for GCC > 10.
Change made to use the same condition as in
dpdkbuild/Makefile.
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7537 (master)
(cherry picked from commit e1fe11875b)
Change-Id: I6a396e250807f46720ff5dcaf21f13d802404a1b
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7652
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
DPDK now officially supports only GCC > 4.9.0.
Some older systems like Centos7 (gcc 4.8.5) can still
build DPDK, but sometimes fail due to -Werror enabled.
Disable this option to increase build success chance.
Fixes#1891
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7536 (master)
(cherry picked from commit 71b4e2a968)
Change-Id: I0950ad01e2d8361851e9ac42c556004980e00ff4
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7651
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This is done in preparation for the RPM tests.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7142 (master)
(cherry picked from commit d22233ac59)
Change-Id: I93cd0d3d7ddeff54387e2a270e9713b2415845b0
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7650
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Added nvmf_tcp_init() to run tcp through physical E810 network card.
If there is not E810 in system, tcp will still run on virtual eth.
Signed-off-by: wanghailiangx <hailiangx.e.wang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6932 (master)
(cherry picked from commit 460bacc914)
Change-Id: I9a983c92175b27c45457682d42d72ef307fef260
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7655
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6994 (master)
(cherry picked from commit c09b139e61)
Change-Id: Iafc067f880525f25363db1adf9a94a9fc9215455
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7747
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6993 (master)
(cherry picked from commit 4d52dae32a)
Change-Id: Ibc3149b5b9591826ff9aeed38457c2aa9ff747e4
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7746
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
collect_devices() is split into two functions:
- collect_net_devices(): Collect ethernet net devs from the
net class.
- collect_rxe_devices(): Collect all rxe devices from the
infiniband class.
This is done in order to make handling of some conditions easier.
Case and point, in newer kernels, device/net link is not anymore
created for the soft roce devices, instead only ./parent attribute
is available. collect_rxe_devices() is adjusted to handle such
a condition.
Signed-off-by: Michal Berger <michalx.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6992
(cherry picked from commit e99cfdb6c9)
Change-Id: Idefa39c4a62c9e650a03e237f49940461e9782a6
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7745
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Older losetup versions loaded "loop" in background, but
with 2.36.1 this doesn't happen. We know we're going to
use this module so let's load it ourselves.
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7086 (master)
(cherry picked from commit f88416a1de)
Change-Id: If3364e92817ae9f09e9897872bd61ca9167184a1
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7744
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Check test flags before calling rdma_device_init if
it's really needed, so that we don't call it when doing
TCP tests.
Signed-off-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6998 (master)
(cherry picked from commit b7423187a2)
Change-Id: I2675a008cf84c7c4084cb83b44c279d2a497c3cb
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7805
Tested-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I473719d4142702572ff4a51858bcf8cb86ef65ea
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6684
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Fixes spdk_top crashing problem upon selecting a row in one tab,
then switching with TAB key to a different tab with fewer rows
and calling pop-up details window with ENTER key.
Signed-off-by: Krzysztof Karas <krzysztof.karas@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6367 (master)
(cherry picked from commit 6a34348219)
Change-Id: I52c5729eca0ce0eaad88767ff3add471a780140a
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6729
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Allow to pass configuration file as first parameter with
fallback to previous behaviour where config file is read
from home directory.
Signed-off-by: Pawel Piatek <pawelx.piatek@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6361 (master)
(cherry picked from commit 23c8f667a2)
Change-Id: I45972b04fe5fc978bf3b2bede45aae363be3069c
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6545
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Pawel Piatek <pawelx.piatek@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Documentation for vhost target states that CPU mask must be a subset
of application CPU mask. This wasn't enforced right now and allowed
the cpumask on controller creation so long as at least single
CPU core overlapped with application's CPU mask.
This might have been misleading and covered up user configuration errors.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6212 (master)
(cherry picked from commit b74363bcdf)
Change-Id: I03f959ec37efd0be9b98cff9c93c5f996b04af35
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6390
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
`--cpumask` argument for vhost controller creation was
untested. This patch specifies this argument and verifies output
under following conditions:
- no cpumask, resulting in it being equal to the app cpumask
- cpumask equal to app cpumask
- single core
- single non-main core
- two cores
It was motivated by recent changes to reactor code affecting
vhost without it being caught by CI. See commit 7cc83b62.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6211 (master)
(cherry picked from commit daab1abc4f)
Change-Id: I92c603061619092007bc27a189d3d8787461a16f
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6389
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
After setting io timeout, host can avoid nbd io
stuck or kernel hang occasionally caused by nbd
stop or underlying bdev removal.
Signed-off-by: Liu Xiaodong <xiaodong.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6191 (master)
(cherry picked from commit b7c33b5e61)
Change-Id: I4ba2a0af7ff7bed369cdaf86121f082136dc1a0b
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6394
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: <dongx.yi@intel.com>
The spec does not disallow TEXT PDUs with no data. In that
case, just return immediately from iscsi_parse_params.
This avoids a NULL pointer dereference with a TEXT PDU that has
no data, but CONTINUE flag is set.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6319 (master)
(cherry picked from commit f3fd56fc3c)
Change-Id: I2605293daf171633a45132d7b5532fdfc9128aff
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6601
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Recent refactoring added a critical bug that no media event is pushed
to the target bdev.
Fix the bug by changing return to break.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6455 (master)
(cherry picked from commit bd8de45801)
Change-Id: Ia674ea2c9c21d08dd23b50a0f726da55011d4be4
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6616
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
When use rpm -i to install spdk-tools, it requires python3-configshell
and python3-pexpect. Add this requires to spec to avoid install failure.
Signed-off-by: sunshihao <sunshihao@huawei.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6336 (master)
(cherry picked from commit a4ee18886d)
Change-Id: I7fc7e9aab35c3fd730401282b72b1604f765a73f
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6431
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: sunshihao <sunshihao@huawei.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Vhost is enabled by default, so rte_net was always included.
When disabled, rte_power failed as it depends on rte_ethdev and rte_net.
rte_vhost was only possible to enable on Linux, so there
is no conflict with adding it next to rte_power under this condition.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6398 (master)
(cherry picked from commit 232f34d754)
Change-Id: I2e183004d6457e404471740a0540dcb08aa738d8
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6430
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Fixes#1777.
When a qpair cannot be allocated because the transport connection fails,
the qpair was freed without unlinking it from the other structures.
This was leading to a segfault when attempting to create and free other
qpairs.
Also added a unit test to cover this case.
Signed-off-by: Vasuki Manikarnike <vasuki.manikarnike@hpe.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6272 (master)
(cherry picked from commit d92c2f118a)
Change-Id: I74b78d1847f90117248b07203b43a11ff5cfa5d6
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6393
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
In dpdk 19.11 version, RTE_VHOST_USER_ASYNC_COPY is not define.
After dpdk 20.08, we can use RTE_VHOST_USER_ASYNC_COPY.
Use version check to avoid this problem.
Signed-off-by: sunshihao <sunshihao@huawei.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6274 (master)
(cherry picked from commit 9529803fe0)
Change-Id: Iaf9914e8380f3d54cded1e2f16af6a7dc3504f95
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6255
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
In dpdk 19.11, rte_kernel_driver is the old version, add version check before use the members.
Signed-off-by: sunshihao <sunshihao@huawei.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6252 (master)
(cherry picked from commit b1687cd456)
Change-Id: Ic1db37cc0760c7d03692fd2cdcbb6ff1e41f872d
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6321
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: sunshihao <sunshihao@huawei.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
As of DPDK main commit 682a6454 ("power: add ethdev
power management"), rte_power depends on rte_ethdev.
So add rte_ethdev library, but only on Linux since
that is the only place where we use rte_power.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6210 (master)
(cherry picked from commit 710838103b)
Change-Id: Ibc3473d6ec75c14a8c0c5517616950017cb2ea96
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6320
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Removed the suppress for tracking ABI differences.
Maintanance release will need to capture any changes,
rather than ignore some.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ia01fecf872f213b61b9545ff7c476a9312e02109
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6158
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>
This reverts commit 72eed604b4.
spdk_vhost_init() relies on having a thread on each reactor.
Without the revert, every vhost controller is created on the same
core and even passing --cpumask when creating does not affect it.
Proper fix would be to change spdk_vhost_init(),
but would require additional testing. For now revert the patch
that spawned always idle threads.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
(cherry picked from commit 724d50d3e1b8d881575efb0a1643220179c0fe2d)
Change-Id: I0dbd6c9fe1d9d23ada260da8fc7b48086223c632
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6187
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Added changelog entry for dynamic scheduler, along with general
information on scheduler framework and behaviour of particular
scheduler implemenations.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
(cherry picked from commit a691353a2a8d99984875c550af037df983996093)
Change-Id: I9fcef56323c4be136b6b531297b070562981eee5
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6185
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Recently, we discovered that under the SPDK + UIO mode, it is possible that vm may be wrong when doing io.
such as:
[root@VM_6_151_centos ~]# mkfs.xfs /dev/vdc
meta-data=/dev/vdc isize=512 agcount=4, agsize=234421142 blks
= sectsz=4096 attr=2, projid32bit=1
= crc=1 finobt=0, sparse=0
data = bsize=4096 blocks=937684566, imaxpct=5
= sunit=0 swidth=0 blks
naming =version 2 bsize=4096 ascii-ci=0 ftype=1
log =internal log bsize=4096 blocks=457853, version=2
= sectsz=4096 sunit=1 blks, lazy-count=1
realtime =none extsz=4096 blocks=0, rtextents=0
bad magic number
bad magic number
Metadata corruption detected at xfs_sb block 0x0/0x1000
libxfs_writebufr: write verifer failed on xfs_sb bno 0x0/0x1000
the github issue link: https://github.com/spdk/spdk/issues/1738
After investigation, we found as below:
The qemu set to VHOST_USER_SET_MEM_TABLE then call vhost_user_set_mem_table(dpdk version spdk-20.11)->mmap(without MAP_POPULATE)
the mmap will only return the virtual address not map the virtual address to the physical address actually.
after call vtophys_notify->vtophys_get_paddr_pagemap, we will get a wrong physical(the virtual address we not access)
So this patch is to set MAP_POPULATE during we use SPDK + UIO.
Signed-off-by: andypma <andypma@tencent.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5970 (master)
(cherry picked from commit 7741de6b7d)
Change-Id: Ib921b6d7381f46bb5894ed787cef52e02524197a
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6182
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Users can use "--with-vfio-user" to enable it when testing it.
For CI configuration, we add a new test flag SPDK_TEST_VFIOUSER
to enable CI tests.
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6139 (master)
(cherry picked from commit 3b673b8431)
Change-Id: Id284df721171d01cc52491ebf4088bcc17eee147
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6181
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Since iWARP doesn't support CQ resize, we can't
accept new qpairs when CQ size is insufficient.
In other case if we accept new qpairs we can
et CQ overrun and receive IBV_EVENT_CQ_ERR ibv
async event. In that case we will have to close
all qpairs that use this CQ.
Part of the fix for issue #1747
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6106 (master)
(cherry picked from commit 4a2c27f717)
Change-Id: I30b9163a7c8abb8651f8813c6bbdb80000741c9c
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6179
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When SRQ is enabled, all qpairs share common pool of
rdma requests and we should check that rdma_req which
cid matches belongs to the correct qpair.
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6110 (master)
(cherry picked from commit 137866e573)
Change-Id: I2505ca39c5110e8e67a48abbe34b6df9ff00229f
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6146
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
2021-01-29 08:15:53 +00:00
85 changed files with 986 additions and 399 deletions
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
