This patchset is a replacement for: http://marc.info/?l=linux-scsi&m=120753407109400&w=2 The problem is that we call blk_cleanup_queue too early for bsg queues for SAS. We need to call blk_cleanup_queue in the release function as we do with scsi devices. A sas_host device doesn't have the release function so the previous patch adds struct device to struct sas_host_attrs. On IRC, James tolds me that kref would make things simpler. This patchset adds release callback support to bsg, which is called when a bsg device goes away. bsg_register_queue() takes a pointer to a callback function. Objects like sas_host that can't use the release callback in struct device can use this feature. bsg internally uses kref for refcounts on bsg devices instead of get/put_device in fops->open/release. bsg calls put_device and the caller's release callback (if it was registered) in kref_put's release. This patchset against scsi-post-merge since there are some changes to scsi_transport_sas in scsi-post-merge. This patchset also depends on the bsg fixes for scsi devices: http://marc.info/?l=linux-scsi&m=120692552424155&w=2 The bsg fixes for scsi devices needs to be applied to scsi-misc first, scsi-post-merge needs to be rebased, then this patch can be applied. James, let me know if you want me send all the floating bsg patches against scsi-post-merge. -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
