| Message ID | 20221122060816.1187260-1-yangyingliang@huawei.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net,resend] ixgbe: fix pci device refcount leak | expand |
On 11/21/2022 10:08 PM, Yang Yingliang wrote: > As comment of pci_get_domain_bus_and_slot() says, it returns > a pci device with refcount increment, when finish using it, > the caller must decrement the reference count by calling > pci_dev_put(). > > In ixgbe_get_first_secondary_devfn() and ixgbe_x550em_a_has_mii(), > pci_dev_put() is called to avoid leak. > > Fixes: 8fa10ef01260 ("ixgbe: register a mdiobus") > Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> > --- <snip> > @@ -882,6 +884,7 @@ static bool ixgbe_x550em_a_has_mii(struct ixgbe_hw *hw) > * of those two root ports > */ > func0_pdev = ixgbe_get_first_secondary_devfn(PCI_DEVFN(0x16, 0)); > + pci_dev_put(func0_pdev); > if (func0_pdev) { > if (func0_pdev == pdev) > return true; > @@ -889,6 +892,7 @@ static bool ixgbe_x550em_a_has_mii(struct ixgbe_hw *hw) > return false; > } > func0_pdev = ixgbe_get_first_secondary_devfn(PCI_DEVFN(0x17, 0)); > + pci_dev_put(func0_pdev); > if (func0_pdev == pdev) > return true; It would probably be better to defer these puts until after the checks and values are set. I'd think some local vars and gotos may be cleaner than placing puts before all the returns. Thanks, Tony
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.c index 24aa97f993ca..ed0d6a8f239c 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.c @@ -855,9 +855,11 @@ static struct pci_dev *ixgbe_get_first_secondary_devfn(unsigned int devfn) rp_pdev = pci_get_domain_bus_and_slot(0, 0, devfn); if (rp_pdev && rp_pdev->subordinate) { bus = rp_pdev->subordinate->number; + pci_dev_put(rp_pdev); return pci_get_domain_bus_and_slot(0, bus, 0); } + pci_dev_put(rp_pdev); return NULL; } @@ -882,6 +884,7 @@ static bool ixgbe_x550em_a_has_mii(struct ixgbe_hw *hw) * of those two root ports */ func0_pdev = ixgbe_get_first_secondary_devfn(PCI_DEVFN(0x16, 0)); + pci_dev_put(func0_pdev); if (func0_pdev) { if (func0_pdev == pdev) return true; @@ -889,6 +892,7 @@ static bool ixgbe_x550em_a_has_mii(struct ixgbe_hw *hw) return false; } func0_pdev = ixgbe_get_first_secondary_devfn(PCI_DEVFN(0x17, 0)); + pci_dev_put(func0_pdev); if (func0_pdev == pdev) return true;
As comment of pci_get_domain_bus_and_slot() says, it returns a pci device with refcount increment, when finish using it, the caller must decrement the reference count by calling pci_dev_put(). In ixgbe_get_first_secondary_devfn() and ixgbe_x550em_a_has_mii(), pci_dev_put() is called to avoid leak. Fixes: 8fa10ef01260 ("ixgbe: register a mdiobus") Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> --- Cc all pepole in the maintainer list. --- drivers/net/ethernet/intel/ixgbe/ixgbe_phy.c | 4 ++++ 1 file changed, 4 insertions(+)