Re: [PATCH] igb_uio: MSI IRQ mode, irq enable/disable refactored

[Bruce Richardson <bruce.richardson@intel.com>]

On Wed, Aug 23, 2017 at 09:47:35AM +0100, Bruce Richardson wrote:
> On Tue, Aug 22, 2017 at 09:55:53AM -0700, Stephen Hemminger wrote:
> > On Mon, 21 Aug 2017 19:33:45 +0200
> > Markus Theil <markus.theil@tu-ilmenau.de> wrote:
> > 
> > > This patch adds MSI IRQ mode and in a way, that should
> > > also work on older kernel versions. The base for my patch
> > > was an attempt to do this in cf705bc36c which was later reverted in
> > > d8ee82745a. Compilation was tested on Linux 3.2, 4.10 and 4.12.
> > > 
> > > MSI(X) setup was already using pci_alloc_irq_vectors before,
> > > but calls to pci_free_irq_vectors were missing and added.
> > > 
> > > Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
> > 
> > I wonder if DPDK should only N-1 Long Term Stable kernel.org kernels?
> > That would mean 4.4.83 or later now, and 4.9 or later starting with 18.XX releases.
> > 
> > If enterprise distro's want to backport more, that is their prerogative but upstream
> > DPDK shouldn't have to worry about it. The current mess with KNI especially is out
> > of hand.
> 
> I agree in principal about limiting ourselves to only supporting a more
> limited set of kernel versions in mainline. However, the exact number of
> versions probably needs some discussion - my initial impression is that
> what you propose is a little too limited.

Actually, having had a bit more caffeine, I need to correct myself. :-(

Since these are out of tree modules, it means that most distro's are not
including them in their packaging. This rules out the possibility of
them getting backported by those distros, leaving the maintainence of
them entirely to the DPDK community. Therefore, I don't think we can
limit ourselves to just LTS kernels, but also need to include kernels of
major distros we need to support. In short, pretty much where we are
today, I think. :-(

/Bruce