Messages in this thread | | | Date | Wed, 7 Sep 2022 10:49:46 -0500 | From | Bjorn Helgaas <> | Subject | Re: [PATCH v3 02/10] PCI/PTM: Cache PTM Capability offset |
| |
On Tue, Sep 06, 2022 at 04:18:23PM -0700, Sathyanarayanan Kuppuswamy wrote: > On 9/6/22 3:23 PM, Bjorn Helgaas wrote:
> > void pci_disable_ptm(struct pci_dev *dev) > > { > > - int ptm; > > + int ptm = dev->ptm_cap; > > I think you don't need to store it. Directly use dev->ptm?
True, no need, but the value is used three times in this function, so I think the variable reduces clutter overall.
> > void pci_restore_ptm_state(struct pci_dev *dev) > > { > > + int ptm = dev->ptm_cap; > > It can be u16?
Done, thanks! I see that in ee8b1c478a9f ("PCI: Return u16 from pci_find_ext_capability() and similar"), I forgot to change the inline stub from int to u16. I'll add a patch to do that. Probably not a prerequisite, since the stub is for !CONFIG_PCI and this code won't be compiled at all in that case.
> > void pci_ptm_init(struct pci_dev *dev) > > { > > - int pos; > > + int ptm; > > Why rename? Also ptm can be u16
"ptm" conveys more information than "pos" and I think it's worth using the same name for all the functions.
Thank you!
Bjorn
| |