On Mon, 24 Jun 2019, Weiping Zhang wrote:
> The driver may implement multiple affinity set, and some of
> are empty, for this case we just skip them.
Why? What's the point of creating empty sets? Just because is not a real
good justification.
Leaving the patch for Ming.
Thanks,
tglx
> Signed-off-by: Weiping Zhang <[email protected]>
> ---
> kernel/irq/affinity.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
> index f18cd5aa33e8..6d964fe0fbd8 100644
> --- a/kernel/irq/affinity.c
> +++ b/kernel/irq/affinity.c
> @@ -295,6 +295,10 @@ irq_create_affinity_masks(unsigned int nvecs, struct irq_affinity *affd)
> unsigned int this_vecs = affd->set_size[i];
> int ret;
>
> + /* skip empty affinity set */
> + if (this_vecs == 0)
> + continue;
> +
> ret = irq_build_affinity_masks(affd, curvec, this_vecs,
> curvec, masks);
> if (ret) {
> --
> 2.14.1
>
>
Hi Thomas,
On Mon, Jun 24, 2019 at 05:42:39PM +0200, Thomas Gleixner wrote:
> On Mon, 24 Jun 2019, Weiping Zhang wrote:
>
> > The driver may implement multiple affinity set, and some of
> > are empty, for this case we just skip them.
>
> Why? What's the point of creating empty sets? Just because is not a real
> good justification.
Patch 5 will add 4 new sets for supporting NVMe's weighted round robin
arbitration. It can be a headache to manage so many irq sets(now the total
sets can become 6) dynamically since size of anyone in the new 4 sets can
be zero, so each particular set is assigned one static index for avoiding
the management trouble, then empty set will be seen by
irq_create_affinity_masks().
So looks skipping the empty set makes sense because the API will become
easier to use than before.
Thanks,
Ming
MIng,
On Tue, 25 Jun 2019, Ming Lei wrote:
> On Mon, Jun 24, 2019 at 05:42:39PM +0200, Thomas Gleixner wrote:
> > On Mon, 24 Jun 2019, Weiping Zhang wrote:
> >
> > > The driver may implement multiple affinity set, and some of
> > > are empty, for this case we just skip them.
> >
> > Why? What's the point of creating empty sets? Just because is not a real
> > good justification.
>
> Patch 5 will add 4 new sets for supporting NVMe's weighted round robin
> arbitration. It can be a headache to manage so many irq sets(now the total
> sets can become 6) dynamically since size of anyone in the new 4 sets can
> be zero, so each particular set is assigned one static index for avoiding
> the management trouble, then empty set will be seen by
> irq_create_affinity_masks().
>
> So looks skipping the empty set makes sense because the API will become
> easier to use than before.
That makes sense, but at least some of that information wants to be in the
change log and not some uninformative description of what the patch does.
I was not Cc'ed on the rest of the patches so I had exactly zero context.
Thanks,
tglx
Thomas Gleixner <[email protected]> 于2019年6月25日周二 下午3:36写道：
>
> MIng,
>
> On Tue, 25 Jun 2019, Ming Lei wrote:
> > On Mon, Jun 24, 2019 at 05:42:39PM +0200, Thomas Gleixner wrote:
> > > On Mon, 24 Jun 2019, Weiping Zhang wrote:
> > >
> > > > The driver may implement multiple affinity set, and some of
> > > > are empty, for this case we just skip them.
> > >
> > > Why? What's the point of creating empty sets? Just because is not a real
> > > good justification.
> >
> > Patch 5 will add 4 new sets for supporting NVMe's weighted round robin
> > arbitration. It can be a headache to manage so many irq sets(now the total
> > sets can become 6) dynamically since size of anyone in the new 4 sets can
> > be zero, so each particular set is assigned one static index for avoiding
> > the management trouble, then empty set will be seen by
> > irq_create_affinity_masks().
> >
> > So looks skipping the empty set makes sense because the API will become
> > easier to use than before.
>
Hello Ming,
Thanks your detail explanation.
> That makes sense, but at least some of that information wants to be in the
> change log and not some uninformative description of what the patch does.
>
> I was not Cc'ed on the rest of the patches so I had exactly zero context.
>
Hello Thomas,
I am sorry I didn't cc you the full patchset, I will add more detail
description in
commit message at V4.
> Thanks,
>
> tglx