| Message ID | 63f6d2e7ea17f6522f36abd6cf5e6a7f25f215c6.1665033267.git.christophe.leroy@csgroup.eu (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 8836402d4b208b2211fc60538ff45d6bb3b73a64 |
| Headers | show |
| Series | usb: Replace NO_IRQ by 0 | expand |
On Thu, Oct 06, 2022 at 07:15:44AM +0200, Christophe Leroy wrote: > NO_IRQ is used to check the return of irq_of_parse_and_map(). > > On some architecture NO_IRQ is 0, on other architectures it is -1. > > irq_of_parse_and_map() returns 0 on error, independent of NO_IRQ. This isn't clear. Does absence of an irq count as an error? In other words, will irq_of_parse_and_map() sometimes return 0 and other times return NO_IRQ? What about architectures on which 0 is a valid irq number? > So use 0 instead of using NO_IRQ. > > Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> > --- > drivers/usb/host/ehci-grlib.c | 2 +- > drivers/usb/host/ehci-ppc-of.c | 2 +- > drivers/usb/host/fhci-hcd.c | 2 +- > drivers/usb/host/ohci-ppc-of.c | 2 +- > drivers/usb/host/uhci-grlib.c | 2 +- > 5 files changed, 5 insertions(+), 5 deletions(-) > > diff --git a/drivers/usb/host/ehci-grlib.c b/drivers/usb/host/ehci-grlib.c > index a2c3b4ec8a8b..0717f2ccf49d 100644 > --- a/drivers/usb/host/ehci-grlib.c > +++ b/drivers/usb/host/ehci-grlib.c > @@ -99,7 +99,7 @@ static int ehci_hcd_grlib_probe(struct platform_device *op) > hcd->rsrc_len = resource_size(&res); > > irq = irq_of_parse_and_map(dn, 0); > - if (irq == NO_IRQ) { > + if (!irq) { > dev_err(&op->dev, "%s: irq_of_parse_and_map failed\n", > __FILE__); > rv = -EBUSY; Since NO_IRQ is sometimes set to -1, shouldn't this test (and all the other ones you changed) really be doing: if (!irq || irq == NO_IRQ) { ... ? Alan Stern
Le 06/10/2022 à 15:50, Alan Stern a écrit : > On Thu, Oct 06, 2022 at 07:15:44AM +0200, Christophe Leroy wrote: >> NO_IRQ is used to check the return of irq_of_parse_and_map(). >> >> On some architecture NO_IRQ is 0, on other architectures it is -1. >> >> irq_of_parse_and_map() returns 0 on error, independent of NO_IRQ. > > This isn't clear. Does absence of an irq count as an error? In other > words, will irq_of_parse_and_map() sometimes return 0 and other times > return NO_IRQ? What about architectures on which 0 is a valid irq > number? NO_IRQ doesn't exist anywhere in core functions. Only some drivers and some architectures have relics of it. irq_of_parse_and_map() will always return 0 on error. 0 can't be a valid logical IRQ number. It may only be a valid hwirq number but it will always be translated to a non-zero logical irq number. I'm trying to get rid of NO_IRQ completely in powerpc code, therefore trying to clean-up all drivers used by powerpc architecture. Long time ago Linus advocated for not using NO_IRQ, see https://lkml.org/lkml/2005/11/21/221 Thanks Christophe > >> So use 0 instead of using NO_IRQ. >> >> Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> >> --- >> drivers/usb/host/ehci-grlib.c | 2 +- >> drivers/usb/host/ehci-ppc-of.c | 2 +- >> drivers/usb/host/fhci-hcd.c | 2 +- >> drivers/usb/host/ohci-ppc-of.c | 2 +- >> drivers/usb/host/uhci-grlib.c | 2 +- >> 5 files changed, 5 insertions(+), 5 deletions(-) >> >> diff --git a/drivers/usb/host/ehci-grlib.c b/drivers/usb/host/ehci-grlib.c >> index a2c3b4ec8a8b..0717f2ccf49d 100644 >> --- a/drivers/usb/host/ehci-grlib.c >> +++ b/drivers/usb/host/ehci-grlib.c >> @@ -99,7 +99,7 @@ static int ehci_hcd_grlib_probe(struct platform_device *op) >> hcd->rsrc_len = resource_size(&res); >> >> irq = irq_of_parse_and_map(dn, 0); >> - if (irq == NO_IRQ) { >> + if (!irq) { >> dev_err(&op->dev, "%s: irq_of_parse_and_map failed\n", >> __FILE__); >> rv = -EBUSY; > > Since NO_IRQ is sometimes set to -1, shouldn't this test (and all the > other ones you changed) really be doing: > > if (!irq || irq == NO_IRQ) { ... > > ? > > Alan Stern
Le 06/10/2022 à 15:50, Alan Stern a écrit : > On Thu, Oct 06, 2022 at 07:15:44AM +0200, Christophe Leroy wrote: >> diff --git a/drivers/usb/host/ehci-grlib.c b/drivers/usb/host/ehci-grlib.c >> index a2c3b4ec8a8b..0717f2ccf49d 100644 >> --- a/drivers/usb/host/ehci-grlib.c >> +++ b/drivers/usb/host/ehci-grlib.c >> @@ -99,7 +99,7 @@ static int ehci_hcd_grlib_probe(struct platform_device *op) >> hcd->rsrc_len = resource_size(&res); >> >> irq = irq_of_parse_and_map(dn, 0); >> - if (irq == NO_IRQ) { >> + if (!irq) { >> dev_err(&op->dev, "%s: irq_of_parse_and_map failed\n", >> __FILE__); >> rv = -EBUSY; > > Since NO_IRQ is sometimes set to -1, shouldn't this test (and all the > other ones you changed) really be doing: > > if (!irq || irq == NO_IRQ) { ... > No, because irq_of_parse_and_map() will never return -1 as an error. It will always be 0. The below call gives no result. git grep -w NO_IRQ drivers/of/ include/
On Thu, Oct 06, 2022 at 02:01:57PM +0000, Christophe Leroy wrote: > > > Le 06/10/2022 à 15:50, Alan Stern a écrit : > > On Thu, Oct 06, 2022 at 07:15:44AM +0200, Christophe Leroy wrote: > >> NO_IRQ is used to check the return of irq_of_parse_and_map(). > >> > >> On some architecture NO_IRQ is 0, on other architectures it is -1. > >> > >> irq_of_parse_and_map() returns 0 on error, independent of NO_IRQ. > > > > This isn't clear. Does absence of an irq count as an error? In other > > words, will irq_of_parse_and_map() sometimes return 0 and other times > > return NO_IRQ? What about architectures on which 0 is a valid irq > > number? > > NO_IRQ doesn't exist anywhere in core functions. Only some drivers and > some architectures have relics of it. > > irq_of_parse_and_map() will always return 0 on error. > > 0 can't be a valid logical IRQ number. It may only be a valid hwirq > number but it will always be translated to a non-zero logical irq number. > > I'm trying to get rid of NO_IRQ completely in powerpc code, therefore > trying to clean-up all drivers used by powerpc architecture. > > Long time ago Linus advocated for not using NO_IRQ, see > https://lkml.org/lkml/2005/11/21/221 Okay, good. Please resubmit the patch and include some of these things in the patch description. Alan Stern
Le 06/10/2022 à 16:24, Alan Stern a écrit : > On Thu, Oct 06, 2022 at 02:01:57PM +0000, Christophe Leroy wrote: >> >> >> Le 06/10/2022 à 15:50, Alan Stern a écrit : >>> On Thu, Oct 06, 2022 at 07:15:44AM +0200, Christophe Leroy wrote: >>>> NO_IRQ is used to check the return of irq_of_parse_and_map(). >>>> >>>> On some architecture NO_IRQ is 0, on other architectures it is -1. >>>> >>>> irq_of_parse_and_map() returns 0 on error, independent of NO_IRQ. >>> >>> This isn't clear. Does absence of an irq count as an error? In other >>> words, will irq_of_parse_and_map() sometimes return 0 and other times >>> return NO_IRQ? What about architectures on which 0 is a valid irq >>> number? >> >> NO_IRQ doesn't exist anywhere in core functions. Only some drivers and >> some architectures have relics of it. >> >> irq_of_parse_and_map() will always return 0 on error. >> >> 0 can't be a valid logical IRQ number. It may only be a valid hwirq >> number but it will always be translated to a non-zero logical irq number. >> >> I'm trying to get rid of NO_IRQ completely in powerpc code, therefore >> trying to clean-up all drivers used by powerpc architecture. >> >> Long time ago Linus advocated for not using NO_IRQ, see >> https://lkml.org/lkml/2005/11/21/221 > > Okay, good. Please resubmit the patch and include some of these things > in the patch description. I sent v2 with more details in the commit message. Thanks Christophe
diff --git a/drivers/usb/host/ehci-grlib.c b/drivers/usb/host/ehci-grlib.c index a2c3b4ec8a8b..0717f2ccf49d 100644 --- a/drivers/usb/host/ehci-grlib.c +++ b/drivers/usb/host/ehci-grlib.c @@ -99,7 +99,7 @@ static int ehci_hcd_grlib_probe(struct platform_device *op) hcd->rsrc_len = resource_size(&res); irq = irq_of_parse_and_map(dn, 0); - if (irq == NO_IRQ) { + if (!irq) { dev_err(&op->dev, "%s: irq_of_parse_and_map failed\n", __FILE__); rv = -EBUSY; diff --git a/drivers/usb/host/ehci-ppc-of.c b/drivers/usb/host/ehci-ppc-of.c index 28a19693c19f..62a0a193798c 100644 --- a/drivers/usb/host/ehci-ppc-of.c +++ b/drivers/usb/host/ehci-ppc-of.c @@ -119,7 +119,7 @@ static int ehci_hcd_ppc_of_probe(struct platform_device *op) hcd->rsrc_len = resource_size(&res); irq = irq_of_parse_and_map(dn, 0); - if (irq == NO_IRQ) { + if (!irq) { dev_err(&op->dev, "%s: irq_of_parse_and_map failed\n", __FILE__); rv = -EBUSY; diff --git a/drivers/usb/host/fhci-hcd.c b/drivers/usb/host/fhci-hcd.c index 2ba09c3fbc2f..6d3215a3ef81 100644 --- a/drivers/usb/host/fhci-hcd.c +++ b/drivers/usb/host/fhci-hcd.c @@ -693,7 +693,7 @@ static int of_fhci_probe(struct platform_device *ofdev) /* USB Host interrupt. */ usb_irq = irq_of_parse_and_map(node, 0); - if (usb_irq == NO_IRQ) { + if (!usb_irq) { dev_err(dev, "could not get usb irq\n"); ret = -EINVAL; goto err_usb_irq; diff --git a/drivers/usb/host/ohci-ppc-of.c b/drivers/usb/host/ohci-ppc-of.c index 591f675cc930..f2f6c832ec98 100644 --- a/drivers/usb/host/ohci-ppc-of.c +++ b/drivers/usb/host/ohci-ppc-of.c @@ -120,7 +120,7 @@ static int ohci_hcd_ppc_of_probe(struct platform_device *op) } irq = irq_of_parse_and_map(dn, 0); - if (irq == NO_IRQ) { + if (!irq) { dev_err(&op->dev, "%s: irq_of_parse_and_map failed\n", __FILE__); rv = -EBUSY; diff --git a/drivers/usb/host/uhci-grlib.c b/drivers/usb/host/uhci-grlib.c index 3ef6d52839e5..907d5f01edfd 100644 --- a/drivers/usb/host/uhci-grlib.c +++ b/drivers/usb/host/uhci-grlib.c @@ -116,7 +116,7 @@ static int uhci_hcd_grlib_probe(struct platform_device *op) hcd->rsrc_len = resource_size(&res); irq = irq_of_parse_and_map(dn, 0); - if (irq == NO_IRQ) { + if (!irq) { printk(KERN_ERR "%s: irq_of_parse_and_map failed\n", __FILE__); rv = -EBUSY; goto err_usb;
NO_IRQ is used to check the return of irq_of_parse_and_map(). On some architecture NO_IRQ is 0, on other architectures it is -1. irq_of_parse_and_map() returns 0 on error, independent of NO_IRQ. So use 0 instead of using NO_IRQ. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> --- drivers/usb/host/ehci-grlib.c | 2 +- drivers/usb/host/ehci-ppc-of.c | 2 +- drivers/usb/host/fhci-hcd.c | 2 +- drivers/usb/host/ohci-ppc-of.c | 2 +- drivers/usb/host/uhci-grlib.c | 2 +- 5 files changed, 5 insertions(+), 5 deletions(-)