| Message ID | 20230321193208.366561-2-u.kleine-koenig@pengutronix.de (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Krzysztof Wilczyński |
| Headers | show |
| Series | PCI: Convert to platform remove callback returning void | expand |
On Tuesday 21 March 2023 20:31:54 Uwe Kleine-König wrote: > The .remove() callback for a platform driver returns an int which makes > many driver authors wrongly assume it's possible to do error handling by > returning an error code. However the value returned is (mostly) ignored > and this typically results in resource leaks. To improve here there is a > quest to make the remove callback return void. In the first step of this > quest all drivers are converted to .remove_new() which already returns > void. > > Trivially convert this driver from always returning zero in the remove > callback to the void returning variant. There are more important fixes for this driver waiting on the list, so I do not see reason for sending such unimportant change for this driver which does not fix any issue. I would suggest to put this change at the end of the pending queue of aardvark patches to prevent any rebasing of the important fixes patches and possible merge conflicts. > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > --- > drivers/pci/controller/pci-aardvark.c | 6 ++---- > 1 file changed, 2 insertions(+), 4 deletions(-) > > diff --git a/drivers/pci/controller/pci-aardvark.c b/drivers/pci/controller/pci-aardvark.c > index 513d8edf3a5c..71ecd7ddcc8a 100644 > --- a/drivers/pci/controller/pci-aardvark.c > +++ b/drivers/pci/controller/pci-aardvark.c > @@ -1927,7 +1927,7 @@ static int advk_pcie_probe(struct platform_device *pdev) > return 0; > } > > -static int advk_pcie_remove(struct platform_device *pdev) > +static void advk_pcie_remove(struct platform_device *pdev) > { > struct advk_pcie *pcie = platform_get_drvdata(pdev); > struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie); > @@ -1989,8 +1989,6 @@ static int advk_pcie_remove(struct platform_device *pdev) > > /* Disable phy */ > advk_pcie_disable_phy(pcie); > - > - return 0; > } > > static const struct of_device_id advk_pcie_of_match_table[] = { > @@ -2005,7 +2003,7 @@ static struct platform_driver advk_pcie_driver = { > .of_match_table = advk_pcie_of_match_table, > }, > .probe = advk_pcie_probe, > - .remove = advk_pcie_remove, > + .remove_new = advk_pcie_remove, > }; > module_platform_driver(advk_pcie_driver); > > -- > 2.39.2 >
On Tue, Mar 21, 2023 at 08:36:04PM +0100, Pali Rohár wrote: > On Tuesday 21 March 2023 20:31:54 Uwe Kleine-König wrote: > > The .remove() callback for a platform driver returns an int which makes > > many driver authors wrongly assume it's possible to do error handling by > > returning an error code. However the value returned is (mostly) ignored > > and this typically results in resource leaks. To improve here there is a > > quest to make the remove callback return void. In the first step of this > > quest all drivers are converted to .remove_new() which already returns > > void. > > > > Trivially convert this driver from always returning zero in the remove > > callback to the void returning variant. > > There are more important fixes for this driver waiting on the list, so I > do not see reason for sending such unimportant change for this driver > which does not fix any issue. I would suggest to put this change at the > end of the pending queue of aardvark patches to prevent any rebasing of > the important fixes patches and possible merge conflicts. I read some frustration out of your reply. However I don't think I'm to blame for anything here. A recommendation to check floating patches on the respective mailing list before sending out a patch would be news to me, and I'd consider such a requirement a too big burden on submitters. Browsing a bit in the linux-pci archives I see I'm not the first to get a similar reply by you[1]. For me as a contributor who rarely does PCI stuff such a feedback is not exactly welcoming and I'd wish for me and others a more friendly interaction. Instead of calling other people's patches unimportant and blaming them for sending cleanup patches, I suggest you resend the patches you care about and highlight why they are important. At least if I were the responsible maintainer, you'd have more success with such a strategy. Having said that, I don't have a problem if the aardvark patch is postponed in favour of some more important changes. If a conflict occurs during application, I happily adapt my patch and send it at a later time. In such a case, just tell me, ideally by making the problem reproduce on next. Best regards Uwe [1] I found: https://lore.kernel.org/linux-pci/20221207075750.6usm4mgejtpcrktw@pali/ https://lore.kernel.org/linux-pci/20221216182524.s6a4uihgavji7bti@pali/
diff --git a/drivers/pci/controller/pci-aardvark.c b/drivers/pci/controller/pci-aardvark.c index 513d8edf3a5c..71ecd7ddcc8a 100644 --- a/drivers/pci/controller/pci-aardvark.c +++ b/drivers/pci/controller/pci-aardvark.c @@ -1927,7 +1927,7 @@ static int advk_pcie_probe(struct platform_device *pdev) return 0; } -static int advk_pcie_remove(struct platform_device *pdev) +static void advk_pcie_remove(struct platform_device *pdev) { struct advk_pcie *pcie = platform_get_drvdata(pdev); struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie); @@ -1989,8 +1989,6 @@ static int advk_pcie_remove(struct platform_device *pdev) /* Disable phy */ advk_pcie_disable_phy(pcie); - - return 0; } static const struct of_device_id advk_pcie_of_match_table[] = { @@ -2005,7 +2003,7 @@ static struct platform_driver advk_pcie_driver = { .of_match_table = advk_pcie_of_match_table, }, .probe = advk_pcie_probe, - .remove = advk_pcie_remove, + .remove_new = advk_pcie_remove, }; module_platform_driver(advk_pcie_driver);
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is (mostly) ignored and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new() which already returns void. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> --- drivers/pci/controller/pci-aardvark.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)