On Thu, Oct 18, 2018 at 05:28:04PM -0600, Keith Busch wrote:
> On Fri, Oct 19, 2018 at 07:18:52AM +0800, kbuild test robot wrote:
> > drivers/pci/pcie/aer.c: In function 'aer_probe':
> > >> drivers/pci/pcie/aer.c:1387:26: error: initialization from incompatible pointer type [-Werror=incompatible-pointer-types]
> > struct device *device = &dev->port->device;
> > ^
> > cc1: some warnings being treated as errors
> >
> > vim +1387 drivers/pci/pcie/aer.c
> >
> > 1376
> > 1377 /**
> > 1378 * aer_probe - initialize resources
> > 1379 * @dev: pointer to the pcie_dev data structure
> > 1380 *
> > 1381 * Invoked when PCI Express bus loads AER service driver.
> > 1382 */
> > 1383 static int aer_probe(struct pcie_device *dev)
> > 1384 {
> > 1385 int status;
> > 1386 struct aer_rpc *rpc;
> > > 1387 struct device *device = &dev->port->device;
>
> You must be building patches incrementally since this was (inadvertently)
> fixed by the following commit 09d260ca3ad04d9b4c4d5e8b0fa1e3501b3e3793
> ("PCI/AER: Use managed resource allocations").
Yes, I think the 0-day service builds each incremental patch.
> Bjorn,
> If you want to replay the history, I quickly rebase the two affected
> patches so this intermediate compile breakage doesn't happen. Up to you.
I fixed this up by removing this from 0459ba0b0d50 ("PCI/AER: Use threaded
IRQ for bottom half"), since I don't think it's relevant to that patch:
@@ -1418,18 +1384,20 @@ static int aer_probe(struct pcie_device *dev)
{
int status;
struct aer_rpc *rpc;
- struct device *device = &dev->port->dev;
+ struct device *device = &dev->port->device;
and making 09d260ca3ad0 ("PCI/AER: Use managed resource allocations") look
like this:
@@ -1384,10 +1380,9 @@ static int aer_probe(struct pcie_device *dev)
{
int status;
struct aer_rpc *rpc;
- struct device *device = &dev->port->dev;
+ struct device *device = &dev->device;
