bdev/raid: remove split counting

Now that I/O never span a strip boundary, we don't
need to count splits.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Iea5df5ca055fe4b338fb5cb0828a45502c239458

Reviewed-on: https://review.gerrithub.io/423325
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This commit is contained in:
Jim Harris 2018-08-23 13:32:02 -07:00
parent bfcb7f64c1
commit 29f0c4c76d
2 changed files with 17 additions and 91 deletions

View File

@ -270,54 +270,30 @@ static void
raid_bdev_io_completion(struct spdk_bdev_io *bdev_io, bool success, void *cb_arg)
{
struct spdk_bdev_io *parent_io = cb_arg;
struct raid_bdev_io *raid_io = (struct raid_bdev_io *)parent_io->driver_ctx;
assert(raid_io->splits_comp_outstanding);
raid_io->splits_comp_outstanding--;
if (raid_io->status == SPDK_BDEV_IO_STATUS_SUCCESS) {
/*
* Store failure status if any of the child bdev io. If any of the child
* fails, overall parent bdev_io is considered failed but parent bdev io
* status is only communicated to above layers on all child completions
*/
raid_io->status = success;
}
/* Free child bdev io */
spdk_bdev_free_io(bdev_io);
if (!raid_io->splits_pending && !raid_io->splits_comp_outstanding) {
/*
* If all childs are submitted and all childs are completed, process
* parent bdev io completion and complete the parent bdev io with
* appropriate status. If any of the child bdev io is failed, parent
* bdev io is considered failed.
*/
if (raid_io->status) {
spdk_bdev_io_complete(parent_io, SPDK_BDEV_IO_STATUS_SUCCESS);
} else {
spdk_bdev_io_complete(parent_io, SPDK_BDEV_IO_STATUS_FAILED);
}
if (success) {
spdk_bdev_io_complete(parent_io, SPDK_BDEV_IO_STATUS_SUCCESS);
} else {
spdk_bdev_io_complete(parent_io, SPDK_BDEV_IO_STATUS_FAILED);
}
}
/*
* brief:
* raid_bdev_submit_children function is used to split the parent io and submit
* the childs to bdev layer. bdev layer redirects the childs to appropriate base
* bdev nvme module
* raid_bdev_submit_children function is used to submit I/O to the correct
* member disk
* params:
* bdev_io - parent bdev io
* start_strip - start strip number of this io
* end_strip - end strip number of this io
* cur_strip - current strip number of this io to start processing
* buf - pointer to buffer for this io
* returns:
* 0 - success
* non zero - failure
*/
static int
raid_bdev_submit_children(struct spdk_bdev_io *bdev_io,
uint64_t start_strip, uint64_t end_strip, uint64_t cur_strip, uint8_t *buf)
raid_bdev_submit_children(struct spdk_bdev_io *bdev_io, uint64_t start_strip, uint8_t *buf)
{
struct raid_bdev_io *raid_io = (struct raid_bdev_io *)bdev_io->driver_ctx;
struct raid_bdev_io_channel *raid_ch = spdk_io_channel_get_ctx(raid_io->ch);
@ -329,19 +305,11 @@ raid_bdev_submit_children(struct spdk_bdev_io *bdev_io,
uint32_t pd_idx;
int ret = 0;
if (start_strip != end_strip) {
SPDK_ERRLOG("I/O spans strip boundary\n");
assert(false);
}
pd_strip = start_strip / raid_bdev->num_base_bdevs;
pd_idx = start_strip % raid_bdev->num_base_bdevs;
offset_in_strip = bdev_io->u.bdev.offset_blocks & (raid_bdev->strip_size - 1);
pd_lba = (pd_strip << raid_bdev->strip_size_shift) + offset_in_strip;
pd_blocks = bdev_io->u.bdev.num_blocks;
raid_io->splits_comp_outstanding++;
assert(raid_io->splits_pending);
raid_io->splits_pending--;
if (raid_bdev->base_bdev_info[pd_idx].desc == NULL) {
SPDK_ERRLOG("base bdev desc null for pd_idx %u\n", pd_idx);
assert(0);
@ -367,17 +335,7 @@ raid_bdev_submit_children(struct spdk_bdev_io *bdev_io,
assert(0);
}
if (ret != 0) {
/*
* If failed to submit child io to bdev layer then queue the parent
* bdev io with current active split information in the wait queue
* for that core. This will get resume from this point only. Assume
* if 4 splits are required and 2 childs are submitted, then parent
* io is queued to io waitq of this core and it will get resumed and
* try to submit the remaining 3 and 4 childs
*/
raid_io->buf = buf;
raid_io->splits_comp_outstanding--;
raid_io->splits_pending++;
return ret;
}
@ -387,7 +345,6 @@ raid_bdev_submit_children(struct spdk_bdev_io *bdev_io,
/*
* brief:
* get_curr_base_bdev_index function calculates the base bdev index
* which should be processed next based on splits_pending parameter
* params:
* raid_bdev - pointer to pooled bdev
* raid_io - pointer to parent io context
@ -399,23 +356,16 @@ get_curr_base_bdev_index(struct raid_bdev *raid_bdev, struct raid_bdev_io *raid_
{
struct spdk_bdev_io *bdev_io;
uint64_t start_strip;
uint64_t end_strip;
uint64_t cur_strip;
bdev_io = SPDK_CONTAINEROF(raid_io, struct spdk_bdev_io, driver_ctx);
start_strip = bdev_io->u.bdev.offset_blocks >> raid_bdev->strip_size_shift;
end_strip = (bdev_io->u.bdev.offset_blocks + bdev_io->u.bdev.num_blocks - 1) >>
raid_bdev->strip_size_shift;
cur_strip = start_strip + ((end_strip - start_strip + 1) - raid_io->splits_pending);
return (cur_strip % raid_bdev->num_base_bdevs);
return (start_strip % raid_bdev->num_base_bdevs);
}
/*
* brief:
* raid_bdev_io_terminate function terminates the execution of the IO. If
* any outstanding children are there it waits for completion, otherwise it
* immediately completes the IO with failure.
* raid_bdev_io_terminate function terminates the execution of the IO.
* params:
* bdev_io - pointer to parent io
* raid_io - pointer to parent io context
@ -425,15 +375,7 @@ get_curr_base_bdev_index(struct raid_bdev *raid_bdev, struct raid_bdev_io *raid_
static void
raid_bdev_io_terminate(struct spdk_bdev_io *bdev_io, struct raid_bdev_io *raid_io)
{
if (raid_io->splits_comp_outstanding == 0) {
/* If no children is outstanding, immediately fail the parent IO */
spdk_bdev_io_complete(bdev_io, SPDK_BDEV_IO_STATUS_FAILED);
} else {
/* If any children is outstanding,
* wait for them to complete but don't send further Ios */
raid_io->splits_pending = 0;
raid_io->status = SPDK_BDEV_IO_STATUS_FAILED;
}
spdk_bdev_io_complete(bdev_io, SPDK_BDEV_IO_STATUS_FAILED);
}
/*
@ -490,8 +432,6 @@ raid_bdev_waitq_io_process(void *ctx)
struct raid_bdev *raid_bdev;
int ret;
uint64_t start_strip;
uint64_t end_strip;
uint64_t cur_strip;
bdev_io = SPDK_CONTAINEROF(raid_io, struct spdk_bdev_io, driver_ctx);
/*
@ -500,10 +440,7 @@ raid_bdev_waitq_io_process(void *ctx)
*/
raid_bdev = (struct raid_bdev *)bdev_io->bdev->ctxt;
start_strip = bdev_io->u.bdev.offset_blocks >> raid_bdev->strip_size_shift;
end_strip = (bdev_io->u.bdev.offset_blocks + bdev_io->u.bdev.num_blocks - 1) >>
raid_bdev->strip_size_shift;
cur_strip = start_strip + ((end_strip - start_strip + 1) - raid_io->splits_pending);
ret = raid_bdev_submit_children(bdev_io, start_strip, end_strip, cur_strip, raid_io->buf);
ret = raid_bdev_submit_children(bdev_io, start_strip, raid_io->buf);
if (ret != 0) {
raid_bdev_io_submit_fail_process(raid_bdev, bdev_io, raid_io, ret);
}
@ -543,14 +480,12 @@ _raid_bdev_submit_rw_request(struct spdk_io_channel *ch, struct spdk_bdev_io *bd
start_strip = bdev_io->u.bdev.offset_blocks >> raid_bdev->strip_size_shift;
end_strip = (bdev_io->u.bdev.offset_blocks + bdev_io->u.bdev.num_blocks - 1) >>
raid_bdev->strip_size_shift;
/*
* IO parameters used during io split and io completion
*/
raid_io->splits_pending = (end_strip - start_strip + 1);
raid_io->splits_comp_outstanding = 0;
raid_io->status = SPDK_BDEV_IO_STATUS_SUCCESS;
ret = raid_bdev_submit_children(bdev_io, start_strip, end_strip, start_strip,
bdev_io->u.bdev.iovs->iov_base);
if (start_strip != end_strip) {
assert(false);
SPDK_ERRLOG("I/O spans strip boundary!\n");
raid_bdev_io_submit_fail_process(raid_bdev, bdev_io, raid_io, -EINVAL);
}
ret = raid_bdev_submit_children(bdev_io, start_strip, bdev_io->u.bdev.iovs->iov_base);
if (ret != 0) {
raid_bdev_io_submit_fail_process(raid_bdev, bdev_io, raid_io, ret);
}

View File

@ -136,15 +136,6 @@ struct raid_bdev_io {
/* current buffer location, used in queueing logic */
uint8_t *buf;
/* outstanding child completions */
uint16_t splits_comp_outstanding;
/* pending splits yet to happen */
uint16_t splits_pending;
/* status of parent io */
bool status;
};
/*