Messages in this thread | | | From | Grant Likely <> | Date | Thu, 20 Dec 2012 21:56:17 +0000 | Subject | Re: [PATCH 1/3] gpio-langwell: cleanup driver |
| |
On Thu, Dec 20, 2012 at 9:35 PM, David Cohen <david.a.cohen@intel.com> wrote: > On 12/19/2012 05:18 PM, Grant Likely wrote: >> On Tue, 18 Dec 2012 17:52:11 -0800, David Cohen <david.a.cohen@intel.com> >>> @@ -215,19 +217,14 @@ static int lnw_irq_type(struct irq_data *d, >>> unsigned type) >>> return 0; >>> } >>> -static void lnw_irq_unmask(struct irq_data *d) >>> -{ >>> -} >>> - >>> -static void lnw_irq_mask(struct irq_data *d) >>> -{ >>> -} >>> +static void lnw_irq_noop(struct irq_data *d) {} >> >> Umm, this looks entirely wrong. There needs to be either mask/unmask or >> enable/disable ops for irq_chips. Yes I see that this patch is just >> consolidating two empty functions, but they are two empty functions that >> appear to be completely wrong. > > > I see your point. The solution does not belong to a clean up patch, > so I'll just remove it from here.
Right, that is appropriate.
>>> static int __devinit lnw_gpio_probe(struct pci_dev *pdev, >>> @@ -349,7 +327,7 @@ static int __devinit lnw_gpio_probe(struct pci_dev >>> *pdev, >>> retval = pci_request_regions(pdev, "langwell_gpio"); >>> if (retval) { >>> dev_err(&pdev->dev, "error requesting resources\n"); >>> - goto err2; >>> + goto err; >> >> Renaming the labels just increases the noise in the diff. I prefer to >> see labels in the form "err-what-i-just-tried-to-do:" so they don't need >> to get reshuffled every time the code logic changes. >> >> There is no good reason for this change. Please drop it. > > > Maybe instead of drop it I'd prefer to fix the labels. They > are wrong currently.
If there is real breakage here, then split it off into a separate bug fix patch. Otherwise I really wouldn't bother.
g.
| |