Hello, First, I'm not very sure if this mailing list allows to ask technical questions. I apologize if it's not a right place to ask questions. I use one LSI SAS 9207-8e controllers to attach 8 SSDs to my server. They together can deliver over 300,000 IO per second. I use mpt2sas driver and MPI-X is enabled. In order to handle so many IOs, I assign different IRQs to different CPU cores. I see very high CPU consumption, so I hope to enable interrupt coalescing to reduce CPU consumption. I monitored interrupts by using systemtap to count the number of invocations of handle_irq, and found the number of interrupts delivered by the LSI HBA is more or less the same as the number of IOs performed on the SSDs. I assume interrupt coalescing doesn't work in my case. If so, how do I enable interrupt coalescing in the LSI HBA. I checked all mpt2sas parameters and none of them are used to control it. The tool lsiutil doesn't work for this controller. I'm pretty sure the LSI HBA supports interrupt coalescing. Do I have to use some special configuration (the Linux kernel or hardware) in order to make it work? Thanks, Da -- 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
