Note that we originally allocates irqs this way, and Keith changed it a while ago for good reasons. So I'd really like to see good reasons for moving away from this, and some heuristics to figure out which way to use. E.g. if the device supports more irqs than I/O queues your scheme might always be fine.
- [PATCH V2] nvme-pci: assign separate irq vectors for adm... Jianchao Wang
- Re: [PATCH V2] nvme-pci: assign separate irq vector... Andy Shevchenko
- Re: [PATCH V2] nvme-pci: assign separate irq ve... jianchao.wang
- Re: [PATCH V2] nvme-pci: assign separate irq vector... Christoph Hellwig
- Re: [PATCH V2] nvme-pci: assign separate irq ve... Sagi Grimberg
- Re: [PATCH V2] nvme-pci: assign separate ir... jianchao.wang
- Re: [PATCH V2] nvme-pci: assign separat... Keith Busch
- Re: [PATCH V2] nvme-pci: assign se... jianchao.wang
- Re: [PATCH V2] nvme-pci: assign separate irq ve... Ming Lei
- Re: [PATCH V2] nvme-pci: assign separate ir... Keith Busch
- Re: [PATCH V2] nvme-pci: assign separate irq ve... jianchao.wang