This patch avoids that Valgrind complains about the memset() call
that was used to initialize the array that it triggers an invalid
write.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
With SPC-4, the 0xE4 descriptor LU ID TYPE field should be ignored, and
the NUL bit is obsolete. Update the ValidTgtDescr accordingly.
Signed-off-by: David Disseldorp <ddiss@suse.de>
ExtendedCopy.DescrLimits currently attempts to test Maximum Descriptor
List Length bounds checking by issuing a request with
(max_target_desc_count + 1 + max_segment_desc_count + 1) descriptors.
Maximum Descriptor List Length is a separately advertised field, that
may not be exceeded with the above logic, e.g. LIO advertises:
- Max Target Desc Count = 2
- Max Segment Desc Count = 1
- Max Desc List Len = 1024
Calculate the number of descriptors using the advertised Maximum
Descriptor List Length value, to explicitly test bounds checking.
Signed-off-by: David Disseldorp <ddiss@suse.de>
We need to use iovectors for iSER to work as it requires that the vectors
are attached to the task before we queue the task.
Signed-off-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Dispatch a write and subsequently reset the logical unit via iSCSI TMF.
This is based on the AbortTaskSimpleAsync test.
Signed-off-by: David Disseldorp <ddiss@suse.de>
When run against an iSCSI target sd->sgio_dev may be NULL, so the
strncmp() coredumps.
Fixes: 698ae7a649
Signed-off-by: David Disseldorp <ddiss@suse.de>
As BLKSECTGET is so unreliable, skip using this to detect max transfer sizes
for any device path starting with "/dev/sg"
instead force the max to be 120k
Signed-off-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Not all targets support cold/warm reset task management functions.
Abort these tests if the support is missing and the commands fail.
Signed-off-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
The iscsi cmdsn tests are noisy and generate a lot of [FAILURE]
warnings for things that are expected.
Fix this by do the plumbing to allow using a EXPECT_STATUS_TIMEOUT
for commands we expect will never be replied to and will thus time out.
Signed-off-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
The current WRITE SAME (WS) 0 blocks test at LBA 0 isn't always correct
because if:
- The target's WSNZ bit set is 0
- The target's MAXIMUM WRITE SAME LENGTH (MWSL) is greater than 0
- We are issuing a WS with a NUMBER OF LOGICAL BLOCKS of 0
then whether the command should return success or failure depends on
whether the starting LBA is within MWSL blocks of the end of the LUN
(see http://www.t10.org/pipermail/t10/2016-May/017988.html for details).
Replace the vestigial code which tried to handle this (but no longer
worked since commit 8585e4509b ) and add
some additional MAXIMUM WRITE SAME LENGTH tests.
Signed-off-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Based on the corresponding simple tests. Issue a WRITESAME request and
confirm that the data written matches the data read back.
Signed-off-by: David Disseldorp <ddiss@suse.de>
There is technically no reason why a SSD with (assumingly with thin
provisioning) can not support SANITIZE/OVERWRITE instead or or in addition
to the more expected SANITIZE/BLOCK-ERASE so remove the warning for this.
Signed-off-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Tests such as ProutPreempt presume each path they have is sending a
different initiator name. When using multipath this did not happen and
initiatorname1 was used on all paths.
Fix this by creating and initializing an array (currently size 2) of
initiator names and iterating it in step with each new path we make
(hopefully the compiler will complain if iscsi-test-cu ever supports
more than two multipaths and there isn't a corresponding initiatorname).
Signed-off-by: Sitsofe Wheeler <sitsofe@yahoo.com>
When libiscsi is built with clang several spurious
missing-field-initializers warnings like the following are reported:
test_async_write.c:61:45: error: missing field 'completed' initializer
[-Werror,-Wmissing-field-initializers]
struct tests_async_write_state state = { 0 };
These appear to be due to clang being overzealous in its checking
because the code is legal C (see
https://llvm.org/bugs/show_bug.cgi?id=21689 ) but it causes compilation
to be aborted. Workaround the problem by specifying a 0 or NULL for each
structure member.
Dispatch a number of Compare And Write requests simultaneously via
separate iSCSI sessions and await all responses.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Dispatch a number of READ10 requests simultaneously, and await all
responses.
The iscsi-support test helper functions are currently all synchronous,
so this test mostly uses the bare libiscsi API. As new async tests are
added, we can make a more informed decision on which boilerplate code
can be split out.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Ceph RADOS block devices split images into 4MB objects. This test
ensures that cross object IOs are successful.
Signed-off-by: Daniel Oliveira <doliveira@suse.com>
Signed-off-by: David Disseldorp <ddiss@suse.de>
