Align iscsi_header[] and data[] on an 8-byte boundary instead of on a 4-byte
boundary. With this patch applied pahole produces the following output:
struct iser_rx_desc {
struct iser_hdr iser_header; /* 0 28 */
char pad1[4]; /* 28 4 */
char iscsi_header[48]; /* 32 48 */
/* --- cacheline 1 boundary (64 bytes) was 16 bytes ago --- */
char data[128]; /* 80 128 */
/* --- cacheline 3 boundary (192 bytes) was 16 bytes ago --- */
struct ibv_sge rx_sg; /* 208 16 */
struct ibv_mr * hdr_mr; /* 224 8 */
char pad2[24]; /* 232 24 */
/* size: 256, cachelines: 4, members: 7 */
};
Additionally, this patch fixes the following build errors:
iser.c: In function 'iser_alloc_rx_descriptors':
iser.c:916:11: error: taking address of packed member of 'struct iser_rx_desc' may result in an unaligned pointer value [-Werror=address-of-packed-member]
916 | rx_sg = &rx_desc->rx_sg;
| ^~~~~~~~~~~~~~~
iser.c: In function 'iser_post_recvm':
iser.c:955:20: error: taking address of packed member of 'struct iser_rx_desc' may result in an unaligned pointer value [-Werror=address-of-packed-member]
955 | rx_wr->sg_list = &rx_desc->rx_sg;
| ^~~~~~~~~~~~~~~
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
This patch fixes the following linker error:
/usr/bin/ld: ../lib/.libs/libiscsipriv.a(libiscsipriv_la-iser.o): undefined reference to symbol 'sem_post@@GLIBC_2.2.5'
/usr/bin/ld: //lib/x86_64-linux-gnu/libpthread.so.0: error adding symbols: DSO missing from command line
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
The iSCSICHAP.Simple test ensures that CHAP_A comma separated lists
are correctly handled. iSCSICHAP.Invalid tests a range of invalid
CHAP_A parameters.
Signed-off-by: David Disseldorp <ddiss@suse.de>
The buffer is memcopied into the PDU. const makes it a little clearer
that the caller isn't handing over ownership.
Signed-off-by: David Disseldorp <ddiss@suse.de>
For now this only changes iSCSIDataSnInvalid to use
iscsi_pdu_set_datasn() instead of scsi_set_uint32().
iSCSICmdSnTooHigh/Low changes should also be possible, as noted, but
haven't been converted yet.
Signed-off-by: David Disseldorp <ddiss@suse.de>
libiscsipriv.la exposes all symbols, so allows tests to perform low
level PDU manipulation, etc. without duplicating code.
As a trade-off, this results in a larger iscsi-test-cu, with libiscsi.so
linking removed.
Link: https://github.com/sahlberg/libiscsi/issues/297
Signed-off-by: David Disseldorp <ddiss@suse.de>
All users have now been changed to overload via the iscsi_ctx transport
callback, so the dlsym based overloading can now be dropped.
Signed-off-by: David Disseldorp <ddiss@suse.de>
In preparation for removing the test-tool iscsi_queue_pdu() dlsym
overload hack, convert this test to use the regular transport driver
callback for this purpose.
Signed-off-by: David Disseldorp <ddiss@suse.de>
In preparation for removing the test-tool iscsi_queue_pdu() dlsym
overload hack, convert this test to use the regular transport driver
callback for this purpose.
Signed-off-by: David Disseldorp <ddiss@suse.de>
In preparation for removing the test-tool iscsi_queue_pdu() dlsym
overload hack, convert this test to use the regular transport driver
callback for this purpose.
Signed-off-by: David Disseldorp <ddiss@suse.de>
In preparation for removing the test-tool iscsi_queue_pdu() dlsym
overload hack, convert this test to use the regular transport driver
callback for this purpose.
This commit also removes a number of unnecessary "WRITE10 is not
implemented" checks, which should all be caught by the first check.
Signed-off-by: David Disseldorp <ddiss@suse.de>
In preparation for removing the test-tool iscsi_queue_pdu() dlsym
overload hack, convert this test to use the regular transport driver
callback for this purpose.
Signed-off-by: David Disseldorp <ddiss@suse.de>
In preparation for removing the test-tool iscsi_queue_pdu() dlsym
overload hack, convert this test to use the regular transport driver
callback for this purpose.
Signed-off-by: David Disseldorp <ddiss@suse.de>
In preparation for removing the test-tool iscsi_queue_pdu() dlsym
overload hack, convert the Compare&Write InvalidDataOutSize test to use
the regular transport driver callback for this purpose.
Signed-off-by: David Disseldorp <ddiss@suse.de>
rfc3720 indicates that SendTargets on discovery *and* normal operational
sessions must be supported by targets:
A system that contains targets MUST support discovery sessions on
each of its iSCSI IP address-port pairs, and MUST support the
SendTargets command on the discovery session.
...
A target MUST support the SendTargets command on operational
sessions...
Signed-off-by: David Disseldorp <ddiss@suse.de>
wr_buf was hardcoded with blocksize 512, resulting in incorrect behavior
for larger block sizes, and buffer overflow. Use the block_size global.
Signed-off-by: Khazhismel Kumykov <khazhy@google.com>
Added support for negotiating below keys:
RDMAExtensions, TargetRecvDataSegmentLength, and
InitiatorRecvDataSegmentLength.
These are required to support iSER. See RFC5046 Section 6.
clear_pr() can be called before connect_scsi_device() has been invoked
on each MPIO sdev. This results in a NULL pointer dereference in the
send_scsi_command() MPIO NOP-responder code-path.
Fix this by invoking clear_pr() after the MPIO connect_scsi_device()
loop has run. clear_pr() also only needs to be invoked via a single
path, so do so via the first path only.
Link: https://github.com/sahlberg/libiscsi/issues/289
Fixes: 16435a9 ("iscsi-test-cu: Improve persistent reservation clearing")
Signed-off-by: David Disseldorp <ddiss@suse.de>
Clear persistent reservations before sending the first SCSI command to the
target to avoid that that command fails due to persistent reservations left
behind by a previous run of the test tool.
Avoid that a subsequent test fails if a test did not remove the persistent
reservations it obtained.
Make sure that clear_pr() returns -1 if clearing persistent reservations
fails.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
