| Message ID | 20210831095728.2447598-1-liu.yun@linux.dev (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | ARM: s3c: irq-s3c24xx: Fix return value check for s3c24xx_init_intc() | expand |
On 31/08/2021 11:57, Jackie Liu wrote: > From: Jackie Liu <liuyun01@kylinos.cn> > > The s3c24xx_init_intc() returns an error pointer upon failure, not NULL. > let's add an error pointer check in s3c24xx_handle_irq. > > Fixes: 1f629b7a3ced ("ARM: S3C24XX: transform irq handling into a declarative form") > Signed-off-by: Jackie Liu <liuyun01@kylinos.cn> Please use scripts/get_maintainers.pl to get list of mailing lists to CC. You skipped two - arm and LKML. > --- > arch/arm/mach-s3c/irq-s3c24xx.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/arm/mach-s3c/irq-s3c24xx.c b/arch/arm/mach-s3c/irq-s3c24xx.c > index 0c631c14a817..d58bf0f9bf9a 100644 > --- a/arch/arm/mach-s3c/irq-s3c24xx.c > +++ b/arch/arm/mach-s3c/irq-s3c24xx.c > @@ -362,11 +362,11 @@ static inline int s3c24xx_handle_intc(struct s3c_irq_intc *intc, > static asmlinkage void __exception_irq_entry s3c24xx_handle_irq(struct pt_regs *regs) > { > do { > - if (likely(s3c_intc[0])) > + if (likely(!IS_ERR_OR_NULL(s3c_intc[0]))) > if (s3c24xx_handle_intc(s3c_intc[0], regs, 0)) Thanks for the patch but it does not look entirely correct. For platform based machines, neither ERR nor NULL can happen here. The s3c24xx_handle_irq() will be set as IRQ handler iff this succeeds: s3c_intc[0] = s3c24xx_init_intc() If this fails, the next calls to s3c24xx_init_intc() won't be executed. For DT machine, s3c_init_intc_of() could set the IRQ handler without setting s3c_intc[0] only if it was called with num_ctrl=0. There is no such code path, so again the s3c_intc[0] will have a valid pointer if set_handle_irq() is called. Therefore in s3c24xx_handle_irq(), the s3c_intc[0] is always something. The code can be simplified by removing if(), if we really wanted and were sure about it. > continue; > > - if (s3c_intc[2]) > + if (!IS_ERR_OR_NULL(s3c_intc[2])) For the non-DT case, this seems ugly but proper solution. The s3c_intc[2] could be NULL (not set at all) or set as ERR (if s3c24xx_init_intc() fails). > if (s3c24xx_handle_intc(s3c_intc[2], regs, 64)) > continue; > > Best regards, Krzysztof
Hi Krzysztof, Thanks for you message. 在 2021/9/1 下午8:09, Krzysztof Kozlowski 写道: > On 31/08/2021 11:57, Jackie Liu wrote: >> From: Jackie Liu <liuyun01@kylinos.cn> >> >> The s3c24xx_init_intc() returns an error pointer upon failure, not NULL. >> let's add an error pointer check in s3c24xx_handle_irq. >> >> Fixes: 1f629b7a3ced ("ARM: S3C24XX: transform irq handling into a declarative form") >> Signed-off-by: Jackie Liu <liuyun01@kylinos.cn> > > Please use scripts/get_maintainers.pl to get list of mailing lists to > CC. You skipped two - arm and LKML. > >> --- >> arch/arm/mach-s3c/irq-s3c24xx.c | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/arch/arm/mach-s3c/irq-s3c24xx.c b/arch/arm/mach-s3c/irq-s3c24xx.c >> index 0c631c14a817..d58bf0f9bf9a 100644 >> --- a/arch/arm/mach-s3c/irq-s3c24xx.c >> +++ b/arch/arm/mach-s3c/irq-s3c24xx.c >> @@ -362,11 +362,11 @@ static inline int s3c24xx_handle_intc(struct s3c_irq_intc *intc, >> static asmlinkage void __exception_irq_entry s3c24xx_handle_irq(struct pt_regs *regs) >> { >> do { >> - if (likely(s3c_intc[0])) >> + if (likely(!IS_ERR_OR_NULL(s3c_intc[0]))) >> if (s3c24xx_handle_intc(s3c_intc[0], regs, 0)) > > Thanks for the patch but it does not look entirely correct. > > For platform based machines, neither ERR nor NULL can happen here. > The s3c24xx_handle_irq() will be set as IRQ handler iff this succeeds: > s3c_intc[0] = s3c24xx_init_intc() > > If this fails, the next calls to s3c24xx_init_intc() won't be executed. > > For DT machine, s3c_init_intc_of() could set the IRQ handler without > setting s3c_intc[0] only if it was called with num_ctrl=0. There is no > such code path, so again the s3c_intc[0] will have a valid pointer if > set_handle_irq() is called. > > Therefore in s3c24xx_handle_irq(), the s3c_intc[0] is always something. > > The code can be simplified by removing if(), if we really wanted and > were sure about it. In fact, I didn't study his underlying logic in depth, but found that this place was not particularly perfect based on the return value of the function, because I happened to encounter a similar problem elsewhere. > > >> continue; >> >> - if (s3c_intc[2]) >> + if (!IS_ERR_OR_NULL(s3c_intc[2])) > > For the non-DT case, this seems ugly but proper solution. The > s3c_intc[2] could be NULL (not set at all) or set as ERR (if > s3c24xx_init_intc() fails). > >> if (s3c24xx_handle_intc(s3c_intc[2], regs, 64)) >> continue; >> >> > > > Best regards, > Krzysztof > Would you mind review v2? -------------------------------------------------------------- diff --git a/arch/arm/mach-s3c/irq-s3c24xx.c b/arch/arm/mach-s3c/irq-s3c24xx.c index 0c631c14a817..df471d322493 100644 --- a/arch/arm/mach-s3c/irq-s3c24xx.c +++ b/arch/arm/mach-s3c/irq-s3c24xx.c @@ -362,11 +362,24 @@ static inline int s3c24xx_handle_intc(struct s3c_irq_intc *intc, static asmlinkage void __exception_irq_entry s3c24xx_handle_irq(struct pt_regs *regs) { do { - if (likely(s3c_intc[0])) - if (s3c24xx_handle_intc(s3c_intc[0], regs, 0)) - continue; + /* For platform based machines, neither ERR nor NULL can happen here. + * The s3c24xx_handle_irq() will be set as IRQ handler iff this succeeds: + * + * s3c_intc[0] = s3c24xx_init_intc() + * + * If this fails, the next calls to s3c24xx_init_intc() won't be executed. + * + * For DT machine, s3c_init_intc_of() could set the IRQ handler without + * setting s3c_intc[0] only if it was called with num_ctrl=0. There is no + * such code path, so again the s3c_intc[0] will have a valid pointer if + * set_handle_irq() is called. + * + * Therefore in s3c24xx_handle_irq(), the s3c_intc[0] is always something. + */ + if (s3c24xx_handle_intc(s3c_intc[0], regs, 0)) + continue; - if (s3c_intc[2]) + if (!IS_ERR_OR_NULL(s3c_intc[2])) if (s3c24xx_handle_intc(s3c_intc[2], regs, 64)) continue; --- Thanks, Jackie
On 01/09/2021 14:28, Jackie Liu wrote: > > Hi Krzysztof, Thanks for you message. > > 在 2021/9/1 下午8:09, Krzysztof Kozlowski 写道: >> On 31/08/2021 11:57, Jackie Liu wrote: >>> From: Jackie Liu <liuyun01@kylinos.cn> >>> >>> The s3c24xx_init_intc() returns an error pointer upon failure, not NULL. >>> let's add an error pointer check in s3c24xx_handle_irq. >>> >>> Fixes: 1f629b7a3ced ("ARM: S3C24XX: transform irq handling into a declarative form") >>> Signed-off-by: Jackie Liu <liuyun01@kylinos.cn> >> >> Please use scripts/get_maintainers.pl to get list of mailing lists to >> CC. You skipped two - arm and LKML. >> >>> --- >>> arch/arm/mach-s3c/irq-s3c24xx.c | 4 ++-- >>> 1 file changed, 2 insertions(+), 2 deletions(-) >>> >>> diff --git a/arch/arm/mach-s3c/irq-s3c24xx.c b/arch/arm/mach-s3c/irq-s3c24xx.c >>> index 0c631c14a817..d58bf0f9bf9a 100644 >>> --- a/arch/arm/mach-s3c/irq-s3c24xx.c >>> +++ b/arch/arm/mach-s3c/irq-s3c24xx.c >>> @@ -362,11 +362,11 @@ static inline int s3c24xx_handle_intc(struct s3c_irq_intc *intc, >>> static asmlinkage void __exception_irq_entry s3c24xx_handle_irq(struct pt_regs *regs) >>> { >>> do { >>> - if (likely(s3c_intc[0])) >>> + if (likely(!IS_ERR_OR_NULL(s3c_intc[0]))) >>> if (s3c24xx_handle_intc(s3c_intc[0], regs, 0)) >> >> Thanks for the patch but it does not look entirely correct. >> >> For platform based machines, neither ERR nor NULL can happen here. >> The s3c24xx_handle_irq() will be set as IRQ handler iff this succeeds: >> s3c_intc[0] = s3c24xx_init_intc() >> >> If this fails, the next calls to s3c24xx_init_intc() won't be executed. >> >> For DT machine, s3c_init_intc_of() could set the IRQ handler without >> setting s3c_intc[0] only if it was called with num_ctrl=0. There is no >> such code path, so again the s3c_intc[0] will have a valid pointer if >> set_handle_irq() is called. >> >> Therefore in s3c24xx_handle_irq(), the s3c_intc[0] is always something. >> >> The code can be simplified by removing if(), if we really wanted and >> were sure about it. > > In fact, I didn't study his underlying logic in depth, but found that > this place was not particularly perfect based on the return value of the > function, because I happened to encounter a similar problem elsewhere. > >> >> >>> continue; >>> >>> - if (s3c_intc[2]) >>> + if (!IS_ERR_OR_NULL(s3c_intc[2])) >> >> For the non-DT case, this seems ugly but proper solution. The >> s3c_intc[2] could be NULL (not set at all) or set as ERR (if >> s3c24xx_init_intc() fails). >> >>> if (s3c24xx_handle_intc(s3c_intc[2], regs, 64)) >>> continue; >>> >>> >> >> >> Best regards, >> Krzysztof >> > > Would you mind review v2? Sure, please send it. Best regards, Krzysztof
diff --git a/arch/arm/mach-s3c/irq-s3c24xx.c b/arch/arm/mach-s3c/irq-s3c24xx.c index 0c631c14a817..d58bf0f9bf9a 100644 --- a/arch/arm/mach-s3c/irq-s3c24xx.c +++ b/arch/arm/mach-s3c/irq-s3c24xx.c @@ -362,11 +362,11 @@ static inline int s3c24xx_handle_intc(struct s3c_irq_intc *intc, static asmlinkage void __exception_irq_entry s3c24xx_handle_irq(struct pt_regs *regs) { do { - if (likely(s3c_intc[0])) + if (likely(!IS_ERR_OR_NULL(s3c_intc[0]))) if (s3c24xx_handle_intc(s3c_intc[0], regs, 0)) continue; - if (s3c_intc[2]) + if (!IS_ERR_OR_NULL(s3c_intc[2])) if (s3c24xx_handle_intc(s3c_intc[2], regs, 64)) continue;