mbox series

[V2,0/4] genirq/affinity: add .calc_sets for improving IRQ allocation & spread

Message ID 20190212130439.14501-1-ming.lei@redhat.com (mailing list archive)
Headers show
Series genirq/affinity: add .calc_sets for improving IRQ allocation & spread | expand

Message

Ming Lei Feb. 12, 2019, 1:04 p.m. UTC
Hi,

Currently pre-caculated set vectors are provided by driver for
allocating & spread vectors. This way only works when drivers passes
same 'max_vecs' and 'min_vecs' to pci_alloc_irq_vectors_affinity(),
also requires driver to retry the allocating & spread.

As Bjorn and Keith mentioned, the current usage & interface for irq sets
is a bit awkward because the retrying should have been avoided by providing
one resonable 'min_vecs'. However, if 'min_vecs' isn't same with
'max_vecs', number of the allocated vectors is unknown before calling
pci_alloc_irq_vectors_affinity(), then each set's vectors can't be
pre-caculated.

Add a new callback of .calc_sets into 'struct irq_affinity' so that
driver can caculate set vectors after IRQ vector is allocated and
before spread IRQ vectors. Add 'priv' so that driver may retrieve
its private data via the 'struct irq_affinity'.


V2:
	- add .calc_sets instead of .setup_affinity() which is easy to
	be abused by drivers

Ming Lei (4):
  genirq/affinity: store irq set vectors in 'struct irq_affinity'
  genirq/affinity: add new callback for caculating set vectors
  nvme-pci: avoid irq allocation retrying via .calc_sets
  genirq/affinity: Document .calc_sets as required in case of multiple
    sets

 drivers/nvme/host/pci.c   | 63 ++++++++++++-----------------------------------
 drivers/pci/msi.c         |  4 +--
 include/linux/interrupt.h | 13 +++++++---
 kernel/irq/affinity.c     | 19 +++++++++-----
 4 files changed, 41 insertions(+), 58 deletions(-)

Comments

Thomas Gleixner Feb. 12, 2019, 1:42 p.m. UTC | #1
On Tue, 12 Feb 2019, Ming Lei wrote:

> Hi,
> 
> Currently pre-caculated set vectors are provided by driver for
> allocating & spread vectors. This way only works when drivers passes
> same 'max_vecs' and 'min_vecs' to pci_alloc_irq_vectors_affinity(),
> also requires driver to retry the allocating & spread.
> 
> As Bjorn and Keith mentioned, the current usage & interface for irq sets
> is a bit awkward because the retrying should have been avoided by providing
> one resonable 'min_vecs'. However, if 'min_vecs' isn't same with
> 'max_vecs', number of the allocated vectors is unknown before calling
> pci_alloc_irq_vectors_affinity(), then each set's vectors can't be
> pre-caculated.
> 
> Add a new callback of .calc_sets into 'struct irq_affinity' so that
> driver can caculate set vectors after IRQ vector is allocated and
> before spread IRQ vectors. Add 'priv' so that driver may retrieve
> its private data via the 'struct irq_affinity'.
> 
> 
> V2:
> 	- add .calc_sets instead of .setup_affinity() which is easy to
> 	be abused by drivers

This looks really well done. If you can address the minor nitpicks, then
this is good to go, unless someone has objections.

Thanks,

	tglx