> + /*
> + * Wait for any in process SPEC_I_PT=1 or REGISTER_AND_MOVE
> + * PR operation to complete.
> + */
> + while (atomic_read(&deve->pr_ref_count) != 0)
> + cpu_relax();
A busy loop under an irq disabling spinlock is fairly nasty. Can we just
remove the po
Remove TRANSPORT_LUNFLAGS_INITIATOR_ACCESS and just look at if a
non-NULL value is in the node_acl->device_list for the given lun. Since
usually nowhere near all se_dev_entrys are used, this nets a sizable
reduction in memory use.
In core_disable_device_list_for_node, move all calls to functions
r
2 matches
Mail list logo
