Christoph Hellwig <hch@infradead.org> writes:

quoted text
> On Wed, Jun 18, 2008 at 02:09:51PM -0400, Jeff Moyer wrote:
>> +		/* For async O_DIRECT writes, we need to invalidate the 
>> +		 * page cache after the write completes.  Kick off a
>> +		 * workqueue to do this and issue the completion in process
>> +		 * context.
>> +		 */
>> +		if (dio->rw == READ) {
>> +			int ret = dio_complete(dio, dio->iocb->ki_pos, 0);
>> +			aio_complete(dio->iocb, ret, 0);
>> +			kfree(dio);
>> +		} else {
>> +			unsigned long flags;
>> +			spin_lock_irqsave(&iocb_completion_list_lock, flags);
>> +			list_add(&dio->done_list, &iocb_completion_list);
>> +			spin_unlock_irqrestore(&iocb_completion_list_lock, flags);
>> +			schedule_work(&aio_complete_work);
>> +		}
>
> Can we please move all these aio_complete calls to user context?  Having
> AIO contexts completing from irq context is a major pain for complex
> filesystems like XFS.

Can you help me understand why this is a pain?  I'm having trouble
making the connection.

Thanks!

Jeff
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/