| Message ID | YxatO5OaI2RpxQ2M@google.com (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Lorenzo Pieralisi |
| Headers | show |
| Series | PCI: apple: do not leak reset GPIO on unbind/unload/error | expand |
On Tue, 06 Sep 2022 03:15:23 +0100, Dmitry Torokhov <dmitry.torokhov@gmail.com> wrote: > > The driver allocates reset GPIO in apple_pcie_setup_port() but neither > releases the resource, nor uses devm API to have it released > automatically. > > Let's fix this by switching to devm API. While at it let's use generic > devm_fwnode_gpiod_get() instead of OF-specific gpiod_get_from_of_node() > - this will allow us top stop exporting the latter down the road. > > Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up") > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com> Acked-by: Marc Zyngier <maz@kernel.org> M.
On 06/09/2022 11.15, Dmitry Torokhov wrote: > The driver allocates reset GPIO in apple_pcie_setup_port() but neither > releases the resource, nor uses devm API to have it released > automatically. > > Let's fix this by switching to devm API. While at it let's use generic > devm_fwnode_gpiod_get() instead of OF-specific gpiod_get_from_of_node() > - this will allow us top stop exporting the latter down the road. > > Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up") > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com> > --- > > This patch has been pulled out of the series > https://lore.kernel.org/all/20220903-gpiod_get_from_of_node-remove-v1-0-b29adfb27a6c@gmail.com/ > with updated justification (leak fix vs pure API deprecation). > > drivers/pci/controller/pcie-apple.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/pci/controller/pcie-apple.c b/drivers/pci/controller/pcie-apple.c > index a2c3c207a04b..66f37e403a09 100644 > --- a/drivers/pci/controller/pcie-apple.c > +++ b/drivers/pci/controller/pcie-apple.c > @@ -516,8 +516,8 @@ static int apple_pcie_setup_port(struct apple_pcie *pcie, > u32 stat, idx; > int ret, i; > > - reset = gpiod_get_from_of_node(np, "reset-gpios", 0, > - GPIOD_OUT_LOW, "PERST#"); > + reset = devm_fwnode_gpiod_get(pcie->dev, of_fwnode_handle(np), "reset", > + GPIOD_OUT_LOW, "PERST#"); > if (IS_ERR(reset)) > return PTR_ERR(reset); > Reviewed-by: Hector Martin <marcan@marcan.st> I actually caught this one a while back, just didn't get around to submitting it yet since it's part of my WIP PCIe power management branch (that I'm getting back to next week or so!) :) https://github.com/AsahiLinux/linux/commit/04f5628fc73ea0369f66c83ba473cb6f8187d2b3 - Hector
Nit: please capitalize beginning of the sentence in the subject - that's what we do for PCI controllers commits. "PCI: apple: Do not leak reset GPIO on unbind/unload/error" On Mon, Sep 05, 2022 at 07:15:23PM -0700, Dmitry Torokhov wrote: > The driver allocates reset GPIO in apple_pcie_setup_port() but neither > releases the resource, nor uses devm API to have it released > automatically. > > Let's fix this by switching to devm API. While at it let's use generic > devm_fwnode_gpiod_get() instead of OF-specific gpiod_get_from_of_node() > - this will allow us top stop exporting the latter down the road. > > Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up") Should I take it via the PCI tree ? Usually we send fixes through -rcX only if the fix applies to code merged last merge window, which is not the case here, so I would queue if for v6.1. Lorenzo > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com> > --- > > This patch has been pulled out of the series > https://lore.kernel.org/all/20220903-gpiod_get_from_of_node-remove-v1-0-b29adfb27a6c@gmail.com/ > with updated justification (leak fix vs pure API deprecation). > > drivers/pci/controller/pcie-apple.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/pci/controller/pcie-apple.c b/drivers/pci/controller/pcie-apple.c > index a2c3c207a04b..66f37e403a09 100644 > --- a/drivers/pci/controller/pcie-apple.c > +++ b/drivers/pci/controller/pcie-apple.c > @@ -516,8 +516,8 @@ static int apple_pcie_setup_port(struct apple_pcie *pcie, > u32 stat, idx; > int ret, i; > > - reset = gpiod_get_from_of_node(np, "reset-gpios", 0, > - GPIOD_OUT_LOW, "PERST#"); > + reset = devm_fwnode_gpiod_get(pcie->dev, of_fwnode_handle(np), "reset", > + GPIOD_OUT_LOW, "PERST#"); > if (IS_ERR(reset)) > return PTR_ERR(reset); > > -- > 2.37.2.789.g6183377224-goog > > > -- > Dmitry
On 2022-09-09 10:16, Lorenzo Pieralisi wrote: > Nit: please capitalize beginning of the sentence in the subject - > that's > what we do for PCI controllers commits. > > "PCI: apple: Do not leak reset GPIO on unbind/unload/error" > > On Mon, Sep 05, 2022 at 07:15:23PM -0700, Dmitry Torokhov wrote: >> The driver allocates reset GPIO in apple_pcie_setup_port() but neither >> releases the resource, nor uses devm API to have it released >> automatically. >> >> Let's fix this by switching to devm API. While at it let's use generic >> devm_fwnode_gpiod_get() instead of OF-specific >> gpiod_get_from_of_node() >> - this will allow us top stop exporting the latter down the road. >> >> Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up") > > Should I take it via the PCI tree ? Usually we send fixes through -rcX > only if the fix applies to code merged last merge window, which is not > the case here, so I would queue if for v6.1. I think 6.1 is perfectly fine. Thanks, M.
On Mon, 5 Sep 2022 19:15:23 -0700, Dmitry Torokhov wrote: > The driver allocates reset GPIO in apple_pcie_setup_port() but neither > releases the resource, nor uses devm API to have it released > automatically. > > Let's fix this by switching to devm API. While at it let's use generic > devm_fwnode_gpiod_get() instead of OF-specific gpiod_get_from_of_node() > - this will allow us top stop exporting the latter down the road. > > [...] Applied to pci/apple, thanks! [1/1] PCI: apple: do not leak reset GPIO on unbind/unload/error https://git.kernel.org/lpieralisi/pci/c/a6b9ede1f3df Thanks, Lorenzo
diff --git a/drivers/pci/controller/pcie-apple.c b/drivers/pci/controller/pcie-apple.c index a2c3c207a04b..66f37e403a09 100644 --- a/drivers/pci/controller/pcie-apple.c +++ b/drivers/pci/controller/pcie-apple.c @@ -516,8 +516,8 @@ static int apple_pcie_setup_port(struct apple_pcie *pcie, u32 stat, idx; int ret, i; - reset = gpiod_get_from_of_node(np, "reset-gpios", 0, - GPIOD_OUT_LOW, "PERST#"); + reset = devm_fwnode_gpiod_get(pcie->dev, of_fwnode_handle(np), "reset", + GPIOD_OUT_LOW, "PERST#"); if (IS_ERR(reset)) return PTR_ERR(reset);
The driver allocates reset GPIO in apple_pcie_setup_port() but neither releases the resource, nor uses devm API to have it released automatically. Let's fix this by switching to devm API. While at it let's use generic devm_fwnode_gpiod_get() instead of OF-specific gpiod_get_from_of_node() - this will allow us top stop exporting the latter down the road. Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up") Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com> --- This patch has been pulled out of the series https://lore.kernel.org/all/20220903-gpiod_get_from_of_node-remove-v1-0-b29adfb27a6c@gmail.com/ with updated justification (leak fix vs pure API deprecation). drivers/pci/controller/pcie-apple.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)