[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <PS1PR0601MB14659C69D22D403547B9F1B4C5B10@PS1PR0601MB1465.apcprd06.prod.outlook.com>
Date: Thu, 3 Aug 2017 05:07:47 +0000
From: HARUNOBU KUROKAWA <harunobu.kurokawa.dn@...esas.com>
To: Lorenzo Pieralisi <lorenzo.pieralisi@....com>
CC: "horms@...ge.net.au" <horms@...ge.net.au>,
"bhelgaas@...gle.com" <bhelgaas@...gle.com>,
"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
"linux-renesas-soc@...r.kernel.org"
<linux-renesas-soc@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH] PCI: rcar-pcie: Fix memory leak when no PCIe card is
inserted
Hi
> -----Original Message-----
> From: Lorenzo Pieralisi [mailto:lorenzo.pieralisi@....com]
> Sent: Wednesday, August 2, 2017 6:03 PM
> To: HARUNOBU KUROKAWA <harunobu.kurokawa.dn@...esas.com>
> Cc: horms@...ge.net.au; bhelgaas@...gle.com; linux-pci@...r.kernel.org; linux-renesas-soc@...r.kernel.org;
> linux-kernel@...r.kernel.org
> Subject: Re: [PATCH] PCI: rcar-pcie: Fix memory leak when no PCIe card is inserted
>
> On Wed, Aug 02, 2017 at 04:13:20PM +0900, Harunobu Kurokawa wrote:
> > When no PCIe card is inserted, there is a memory leak as
> > pci_free_resource_list is not called before returning.
> >
> > Signed-off-by: Harunobu Kurokawa <harunobu.kurokawa.dn@...esas.com>
> > ---
> > drivers/pci/host/pcie-rcar.c | 7 ++++---
> > 1 file changed, 4 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/pci/host/pcie-rcar.c
> > b/drivers/pci/host/pcie-rcar.c index ac80fbb..9b06de6 100644
> > --- a/drivers/pci/host/pcie-rcar.c
> > +++ b/drivers/pci/host/pcie-rcar.c
> > @@ -1190,14 +1190,15 @@ static int rcar_pcie_probe(struct
> > platform_device *pdev)
> >
> > return 0;
> >
> > -err_free_bridge:
> > - pci_free_host_bridge(bridge);
> > -
> > err_pm_put:
> > pm_runtime_put(dev);
> >
> > err_pm_disable:
> > pm_runtime_disable(dev);
> > +
> > +err_free_bridge:
> > + pci_free_host_bridge(bridge);
> > + pci_free_resource_list(&pcie->resources);
> > return err;
> > }
>
> The error code path needs fixing but not like this (with a single patch lumping all fixes - well, actually it does not - together).
>
> Step one, patch below. Step two free the &pci->resources list (which IIUC was not freed even before commit 90634e854079) in a separate
> patch.
>
> Am I missing something ?
>
> Please apply, check and repost the patch if it is ok, thanks.
Yes, I'll sprit the patch and repost them.
Best regards,
Kurokawa.
>
> -- >8 --
> Subject: [PATCH] PCI: rcar: Fix error exit path
>
> Commit 90634e854079 ("PCI: rcar: Convert PCI scan API to
> pci_scan_root_bus_bridge()") converted PCI root bus scan API to the new pci_scan_root_bus_bridge() API; in the process some error
> paths were not updated correctly which may cause memory leaks.
>
> Fix the driver error exit path reinstating the previous correct error exit behaviour.
>
> Fixes: 90634e854079 ("PCI: rcar: Convert PCI scan API to pci_scan_root_bus_bridge()")
> Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@....com>
> ---
> drivers/pci/host/pcie-rcar.c | 11 +++++------
> 1 file changed, 5 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/pci/host/pcie-rcar.c b/drivers/pci/host/pcie-rcar.c index 246d485..007523e 100644
> --- a/drivers/pci/host/pcie-rcar.c
> +++ b/drivers/pci/host/pcie-rcar.c
> @@ -471,10 +471,8 @@ static int rcar_pcie_enable(struct rcar_pcie *pcie)
> bridge->msi = &pcie->msi.chip;
>
> ret = pci_scan_root_bus_bridge(bridge);
> - if (ret < 0) {
> - kfree(bridge);
> + if (ret < 0)
> return ret;
> - }
>
> bus = bridge->bus;
>
> @@ -1190,14 +1188,15 @@ static int rcar_pcie_probe(struct platform_device *pdev)
>
> return 0;
>
> -err_free_bridge:
> - pci_free_host_bridge(bridge);
> -
> err_pm_put:
> pm_runtime_put(dev);
>
> err_pm_disable:
> pm_runtime_disable(dev);
> +
> +err_free_bridge:
> + pci_free_host_bridge(bridge);
> +
> return err;
> }
>
> --
> 2.10.0
Powered by blists - more mailing lists