On Thu, Oct 06, 2011 at 09:56:16AM -0700, Roland Dreier wrote: > Testing in_interrupt() to know when sleeping is allowed is not really > reliable (since eg it won't be true if the caller is holding a spinlock). > Instead have the caller tell core_tmr_alloc_req() what GFP_xxx to use; > every caller except tcm_qla2xxx can use GFP_KERNEL. Looks much better, Reviewed-by: Christoph Hellwig <hch@xxxxxx> Again I'd love to change the model here a bit more for the mid-term future. Just like for actualy command submission I'd like to make sure core_tmr_alloc_req / transport_lookup_tmr_lun / transport_generic_handle_tmr are always called from process context, that is instead of queueing it up to the transport processing thread in transport_generic_handle_tmr use a workqueue to do it earlier on. Your patch made it clear that only qla2xxx doesn't have the basic requirements for that yet. -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
