| Message ID | 20230117172450.2938962-1-arnd@kernel.org |
|---|---|
| State | Accepted |
| Headers | show |
| Series | rtc: sunplus: fix format string for printing resource | expand |
On 17/01/2023 18:24:44+0100, Arnd Bergmann wrote: > From: Arnd Bergmann <arnd@arndb.de> > > On 32-bit architectures with 64-bit resource_size_t, sp_rtc_probe() > causes a compiler warning: > > drivers/rtc/rtc-sunplus.c: In function 'sp_rtc_probe': > drivers/rtc/rtc-sunplus.c:243:33: error: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=] > 243 | dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > The best way to print a resource is the special %pR format string, > and similarly to print a pointer we can use %p and avoid the cast. > I got this one this morning, which one is more correct? :) https://lore.kernel.org/all/20230117054232.24023-1-rdunlap@infradead.org/ > Fixes: fad6cbe9b2b4 ("rtc: Add driver for RTC in Sunplus SP7021") > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > drivers/rtc/rtc-sunplus.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/rtc/rtc-sunplus.c b/drivers/rtc/rtc-sunplus.c > index e8e2ab1103fc..4b578e4d44f6 100644 > --- a/drivers/rtc/rtc-sunplus.c > +++ b/drivers/rtc/rtc-sunplus.c > @@ -240,8 +240,8 @@ static int sp_rtc_probe(struct platform_device *plat_dev) > if (IS_ERR(sp_rtc->reg_base)) > return dev_err_probe(&plat_dev->dev, PTR_ERR(sp_rtc->reg_base), > "%s devm_ioremap_resource fail\n", RTC_REG_NAME); > - dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", > - sp_rtc->res->start, (unsigned long)sp_rtc->reg_base); > + dev_dbg(&plat_dev->dev, "res = %pR, reg_base = %p\n", > + sp_rtc->res, sp_rtc->reg_base); > > sp_rtc->irq = platform_get_irq(plat_dev, 0); > if (sp_rtc->irq < 0) > -- > 2.39.0 >
On 1/17/23 09:55, Alexandre Belloni wrote: > On 17/01/2023 18:24:44+0100, Arnd Bergmann wrote: >> From: Arnd Bergmann <arnd@arndb.de> >> >> On 32-bit architectures with 64-bit resource_size_t, sp_rtc_probe() >> causes a compiler warning: >> >> drivers/rtc/rtc-sunplus.c: In function 'sp_rtc_probe': >> drivers/rtc/rtc-sunplus.c:243:33: error: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=] >> 243 | dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", >> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> >> The best way to print a resource is the special %pR format string, >> and similarly to print a pointer we can use %p and avoid the cast. >> > > I got this one this morning, which one is more correct? :) > https://lore.kernel.org/all/20230117054232.24023-1-rdunlap@infradead.org/ I prefer my handling of res->start and Arnd's no-cast handling of reg_base. IMO using "%pR" prints too much info, but that's more up to the file's author or maintainer... How's that? :) >> Fixes: fad6cbe9b2b4 ("rtc: Add driver for RTC in Sunplus SP7021") >> Signed-off-by: Arnd Bergmann <arnd@arndb.de> >> --- >> drivers/rtc/rtc-sunplus.c | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/rtc/rtc-sunplus.c b/drivers/rtc/rtc-sunplus.c >> index e8e2ab1103fc..4b578e4d44f6 100644 >> --- a/drivers/rtc/rtc-sunplus.c >> +++ b/drivers/rtc/rtc-sunplus.c >> @@ -240,8 +240,8 @@ static int sp_rtc_probe(struct platform_device *plat_dev) >> if (IS_ERR(sp_rtc->reg_base)) >> return dev_err_probe(&plat_dev->dev, PTR_ERR(sp_rtc->reg_base), >> "%s devm_ioremap_resource fail\n", RTC_REG_NAME); >> - dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", >> - sp_rtc->res->start, (unsigned long)sp_rtc->reg_base); >> + dev_dbg(&plat_dev->dev, "res = %pR, reg_base = %p\n", >> + sp_rtc->res, sp_rtc->reg_base); >> >> sp_rtc->irq = platform_get_irq(plat_dev, 0); >> if (sp_rtc->irq < 0) >> -- >> 2.39.0 >> >
On Tue, Jan 17, 2023, at 19:24, Randy Dunlap wrote: > On 1/17/23 09:55, Alexandre Belloni wrote: >> On 17/01/2023 18:24:44+0100, Arnd Bergmann wrote: >>> From: Arnd Bergmann <arnd@arndb.de> >>> >>> On 32-bit architectures with 64-bit resource_size_t, sp_rtc_probe() >>> causes a compiler warning: >>> >>> drivers/rtc/rtc-sunplus.c: In function 'sp_rtc_probe': >>> drivers/rtc/rtc-sunplus.c:243:33: error: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=] >>> 243 | dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", >>> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>> >>> The best way to print a resource is the special %pR format string, >>> and similarly to print a pointer we can use %p and avoid the cast. >>> >> >> I got this one this morning, which one is more correct? :) >> https://lore.kernel.org/all/20230117054232.24023-1-rdunlap@infradead.org/ Both are equally correct, it's just a preference. > I prefer my handling of res->start and Arnd's no-cast handling of reg_base. > IMO using "%pR" prints too much info, but that's more up to the file's author > or maintainer... Right, I could have equally well picked the %pap version, and just went for brevity in the source. It's only pr_debug(), so very few users are going to actually see the output. Arnd
On 1/17/23 11:36, Arnd Bergmann wrote: > On Tue, Jan 17, 2023, at 19:24, Randy Dunlap wrote: >> On 1/17/23 09:55, Alexandre Belloni wrote: >>> On 17/01/2023 18:24:44+0100, Arnd Bergmann wrote: >>>> From: Arnd Bergmann <arnd@arndb.de> >>>> >>>> On 32-bit architectures with 64-bit resource_size_t, sp_rtc_probe() >>>> causes a compiler warning: >>>> >>>> drivers/rtc/rtc-sunplus.c: In function 'sp_rtc_probe': >>>> drivers/rtc/rtc-sunplus.c:243:33: error: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=] >>>> 243 | dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", >>>> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>> >>>> The best way to print a resource is the special %pR format string, >>>> and similarly to print a pointer we can use %p and avoid the cast. >>>> >>> >>> I got this one this morning, which one is more correct? :) >>> https://lore.kernel.org/all/20230117054232.24023-1-rdunlap@infradead.org/ > > Both are equally correct, it's just a preference. > >> I prefer my handling of res->start and Arnd's no-cast handling of reg_base. >> IMO using "%pR" prints too much info, but that's more up to the file's author >> or maintainer... > > Right, I could have equally well picked the %pap version, and just > went for brevity in the source. It's only pr_debug(), so very few > users are going to actually see the output. Alexandre, sounds like you should just go with Arnd's patch.
On Tue, 17 Jan 2023 18:24:44 +0100, Arnd Bergmann wrote: > On 32-bit architectures with 64-bit resource_size_t, sp_rtc_probe() > causes a compiler warning: > > drivers/rtc/rtc-sunplus.c: In function 'sp_rtc_probe': > drivers/rtc/rtc-sunplus.c:243:33: error: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=] > 243 | dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > [...] Applied, thanks! [1/1] rtc: sunplus: fix format string for printing resource commit: 08279468a294d8c996a657ecc9e51bd5c084c75d
diff --git a/drivers/rtc/rtc-sunplus.c b/drivers/rtc/rtc-sunplus.c index e8e2ab1103fc..4b578e4d44f6 100644 --- a/drivers/rtc/rtc-sunplus.c +++ b/drivers/rtc/rtc-sunplus.c @@ -240,8 +240,8 @@ static int sp_rtc_probe(struct platform_device *plat_dev) if (IS_ERR(sp_rtc->reg_base)) return dev_err_probe(&plat_dev->dev, PTR_ERR(sp_rtc->reg_base), "%s devm_ioremap_resource fail\n", RTC_REG_NAME); - dev_dbg(&plat_dev->dev, "res = 0x%x, reg_base = 0x%lx\n", - sp_rtc->res->start, (unsigned long)sp_rtc->reg_base); + dev_dbg(&plat_dev->dev, "res = %pR, reg_base = %p\n", + sp_rtc->res, sp_rtc->reg_base); sp_rtc->irq = platform_get_irq(plat_dev, 0); if (sp_rtc->irq < 0)