@@ -141,7 +141,6 @@ struct tcmu_dev {
struct radix_tree_root data_blocks;
struct idr commands;
- spinlock_t commands_lock;
struct timer_list timeout;
unsigned int cmd_time_out;
@@ -1076,10 +1075,7 @@ static unsigned int tcmu_handle_completions(struct tcmu_dev *udev)
}
WARN_ON(tcmu_hdr_get_op(entry->hdr.len_op) != TCMU_OP_CMD);
- spin_lock(&udev->commands_lock);
cmd = idr_remove(&udev->commands, entry->hdr.cmd_id);
- spin_unlock(&udev->commands_lock);
-
if (!cmd) {
pr_err("cmd_id not found, ring is broken\n");
set_bit(TCMU_DEV_BIT_BROKEN, &udev->flags);
@@ -1175,7 +1171,6 @@ static struct se_device *tcmu_alloc_device(struct se_hba *hba, const char *name)
INIT_LIST_HEAD(&udev->timedout_entry);
idr_init(&udev->commands);
- spin_lock_init(&udev->commands_lock);
setup_timer(&udev->timeout, tcmu_device_timedout,
(unsigned long)udev);
@@ -1404,16 +1399,14 @@ static void tcmu_dev_kref_release(struct kref *kref)
spin_unlock_bh(&timed_out_udevs_lock);
/* Upper layer should drain all requests before calling this */
- spin_lock_irq(&udev->commands_lock);
+ mutex_lock(&udev->cmdr_lock);
idr_for_each_entry(&udev->commands, cmd, i) {
if (tcmu_check_and_free_pending_cmd(cmd) != 0)
all_expired = false;
}
idr_destroy(&udev->commands);
- spin_unlock_irq(&udev->commands_lock);
WARN_ON(!all_expired);
- mutex_lock(&udev->cmdr_lock);
tcmu_blocks_release(&udev->data_blocks, 0, udev->dbi_max + 1);
mutex_unlock(&udev->cmdr_lock);
@@ -2066,9 +2059,9 @@ static void check_timedout_devices(void)
list_del_init(&udev->timedout_entry);
spin_unlock_bh(&timed_out_udevs_lock);
- spin_lock(&udev->commands_lock);
+ mutex_lock(&udev->cmdr_lock);
idr_for_each(&udev->commands, tcmu_check_expired_cmd, NULL);
- spin_unlock(&udev->commands_lock);
+ mutex_unlock(&udev->cmdr_lock);
spin_lock_bh(&timed_out_udevs_lock);
}
No need for the commands_lock. The cmdr_lock is already held during idr addition and deletion, so just grab it during traversal. Note: This also fixes 2 issues. There was no reason to use spin_lock_irq. We only were locking against the submitting, completion and timer context. The first two are done in process context (even though tcmu_irqcontrol is named irq it is run from the write() calls context) and the timer was in softirq, so we just needed to do _bh locking. We should have been using either one tcmu_handle_completions to prevent the case where tcmu_handle_completions could be interrupted a timer softirq while the commands_lock is held. Signed-off-by: Mike Christie <mchristi@redhat.com> --- drivers/target/target_core_user.c | 13 +++---------- 1 file changed, 3 insertions(+), 10 deletions(-)