Re: [PATCH 1/4] PCI: Replace dev_*() printk wrappers with pci_*() printk wrappers

[Frederick Lawler]

Bjorn,

Bjorn Helgaas wrote on 4/28/19 7:52 PM:
> On Sun, Apr 28, 2019 at 07:02:58PM -0500, Bjorn Helgaas wrote:
> > On Sat, Apr 27, 2019 at 02:13:01PM -0500, fred@xxxxxxxxxxxx wrote:
> > > From: Frederick Lawler <fred@xxxxxxxxxxxx>
> > >
> > > Replace remaining instances of dev_*() printk wrappers with pci_*()
> > > printk wrappers. No functional change intended.
> > >
> > > Signed-off-by: Frederick Lawler <fred@xxxxxxxxxxxx>
> > > ---
> > >   drivers/pci/pcie/aer.c        | 13 ++++++-------
> > >   drivers/pci/pcie/aer_inject.c |  4 ++--
> > >   drivers/pci/pcie/dpc.c        | 27 ++++++++++++---------------
> > >   3 files changed, 20 insertions(+), 24 deletions(-)
>
> > >   	aer_enable_rootport(rpc);
> > > -	dev_info(device, "AER enabled with IRQ %d\n", dev->irq);
> > > +	pci_info(pdev, "AER enabled with IRQ %d\n", dev->irq);
> >
> > And this, and many others below.  *This* patch should only convert
> >
> >    - pci_printk(KERN_DEBUG, pdev, ...)
> >    + pci_info(pdev, ...)
> >
> > and
> >
> >    - dev_printk(KERN_DEBUG, pcie_dev, ...)
> >    + dev_info(pcie_dev, ...)
>
> Just to clarify, I do *want* both changes, just in separate patches.
> So we'd have
>
>    1) Convert KERN_DEBUG uses to pci_info() for pci_dev usage and to
>       dev_info() for pcie_device usage.  I think pciehp is probably an
>       exception to this; this patch shouldn't touch ctrl_dbg().
>
>    2) Convert "dev_info(pcie_device)" to "pci_info(pci_dev)".  It might
>       be worth doing this in separate patches for each service.  If we
>       decide they're simple enough to combine, that's trivial for me to
>       do.  It's a little more hassle to split things up afterwards.
>
>       In pciehp, if you do this in the ctrl_*() definitions, it will
>       make the patch much smaller.
>
>    3) In pciehp, ctrl_dbg() could probably be changed to use pci_dbg()
>       so we'd use the standard kernel dynamic debug stuff instead of
>       having the pciehp-specific module parameter.
>
> Thanks a lot for working on all this.  I think it will make the user
> experience significantly simpler.
>
> Bjorn

Will do, thanks!

Frederick Lawler