If an application passes buffers to libiscsi for reading they are
currently added to a linked list one by one. This leads to a malloc
for each buffer object plus O(n) walks trough the list to add
the buffer to then end of the list. Additionally the buffer read
routine takes up to O(n) iterations to find the right buffer
for a request.
This patch introduces an scsi_iovector struct to pass buffers to
an scsi task. Adding a new buffer is in O(1) and finding the
right buffer to also. Malloc requirements are in O(log(n)).
Additionally the scsi_iovector struct is itended to be binary
compatible to an QEMUIOVector allowing to pass this structure
directly to the library.
Initial tests have been made booting an Ubuntu LTS 12.04.1
Desktop server up to the login prompt. The following observations
have been made with regards to scsi_malloc calls:
original implementation: ~11.500 mallocs
using iovector instead of list: ~ 7.500 mallocs
passing the iovector directly: 0 mallocs
To enable this feature in qemu for testing the following patch might
be used:
diff --git a/block/iscsi.c b/block/iscsi.c
index a6a819d..2809c15 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -390,11 +390,16 @@ iscsi_aio_readv(BlockDriverState *bs, int64_t sector_num,
return NULL;
}
+#if defined(LIBISCSI_FEATURE_IOVECTOR)
+ assert(sizeof(struct QEMUIOVector) == sizeof(struct scsi_iovector));
+ scsi_iovector_assign(acb->task, (struct scsi_iovector*) acb->qiov);
+#else
for (i = 0; i < acb->qiov->niov; i++) {
scsi_task_add_data_in_buffer(acb->task,
acb->qiov->iov[i].iov_len,
acb->qiov->iov[i].iov_base);
}
+#endif
iscsi_set_events(iscsilun);
---
Signed-off-by: Peter Lieven <pl@kamp.de>
added a few tweaks to further remove the need to memory
allocation resizing by preallocating the right buffer size.
Signed-off-by: Peter Lieven <pl@kamp.de>
If there is an error in iscsi_service or on poll we
exit the event_loop. But in this case scsi_sync_cb is
never reached and the status is 0. This will make
the caller think that the sync task has been completed
successfully.
If there is an error during reconnect we should not loop in event_loop
as this will cause a deadlock in case there is actually an error like
connection drop etc. If this reconnect fails the reconnect routine
will retry itself.
Split the INQUIRY data-in unmarshalling into smaller chunks, fixing some
potential memory leaks along the way.
Signed-off-by: Arne Redlich <arne.redlich@googlemail.com>
If BYTCHK is false the target will perform a medium check of the indicated
LBAs only and not compare with anything out of the DATA-OUT buffers.
As such we dont need to/should not transfer any DAT-OUT to the target.
The memory tracking code reports memory allocated by iscsi_allocate_pdu_with_itt_flags_size() as lost.
This memory is allocated by the iscsi part of libiscsi, but later freed by the lowlevel scsi part. We
will fix this later by introducing an iscsi_task object.
These two functions belong in the iscsi layer, not the scsi layer so move them
out from scsi-lowlevel.c so that we can start turning scsi-lowlevel.c to a pure
scsi layer and remove all dependencies to iscsi from it.
We dont need to unlink the in-buffers when cancelling a scsi task
since the whole task structure will go away shortly anyway and all
entries in the link will be automatically freed when the task is freed.
Change the list-head structure for in-task scsi memory allocations to
be private to scsi-lowlevel since is is never accessed from anyehwere else and
it is private to this function.
Remove the pointer to the user data in the list head and replace it with a zero length buffer at the end of the header.
Signed-off-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
