| Message ID | 1365098790-9078-1-git-send-email-jon-hunter@ti.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thursday 04 April 2013 11:36 PM, Jon Hunter wrote: > The L3 interrupt numbers are incorrect for OMAP4+ and are conflicting > with some of the timer interrupts causing the allocation of timer > interrupts to fail. > > The problem is caused by adding 32 to the interrupt number for the L3 > interrupts to account for per processor interrupts (PPI) and software > generated interrupts (SGI) which typically are mapped to the first 32 > interrupts in the ARM GIC. This is not necessary because the first > parameter of the ARM GIC interrupt property specifies the GIC interrupt > type (ie. SGI, PPI, etc). Hence, fix the interrupt number fo the L3 > interrupts by substracting 32. > > Cc: Santosh Shilimkar <santosh.shilimkar@ti.com> > Signed-off-by: Jon Hunter <jon-hunter@ti.com> > --- > > Please note that this problem is observed in Benoit's for_3.10/dts branch [1]. > > [1] http://git.kernel.org/cgit/linux/kernel/git/bcousson/linux-omap-dt.git > Thats correct. I overlooked the 32 addition part. This patch should also be pulled into Benoit's 3.10 tree. For the patch, Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
On 04/05/2013 08:26 AM, Santosh Shilimkar wrote: > On Thursday 04 April 2013 11:36 PM, Jon Hunter wrote: >> The L3 interrupt numbers are incorrect for OMAP4+ and are conflicting >> with some of the timer interrupts causing the allocation of timer >> interrupts to fail. >> >> The problem is caused by adding 32 to the interrupt number for the L3 >> interrupts to account for per processor interrupts (PPI) and software >> generated interrupts (SGI) which typically are mapped to the first 32 >> interrupts in the ARM GIC. This is not necessary because the first >> parameter of the ARM GIC interrupt property specifies the GIC interrupt >> type (ie. SGI, PPI, etc). Hence, fix the interrupt number fo the L3 >> interrupts by substracting 32. >> >> Cc: Santosh Shilimkar <santosh.shilimkar@ti.com> >> Signed-off-by: Jon Hunter <jon-hunter@ti.com> >> --- >> >> Please note that this problem is observed in Benoit's for_3.10/dts branch [1]. >> >> [1] http://git.kernel.org/cgit/linux/kernel/git/bcousson/linux-omap-dt.git >> > Thats correct. I overlooked the 32 addition part. This patch should > also be pulled into Benoit's 3.10 tree. Done. I've just applied it. But I will probably merge it with the original patch, because having a broken patch and the fix in the same pull request does not look right. > For the patch, > Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com> Thanks, Benoit
On Friday 05 April 2013 01:52 PM, Benoit Cousson wrote: > Santosh and Jon, > > On 04/05/2013 10:08 AM, Benoit Cousson wrote: >> On 04/05/2013 08:26 AM, Santosh Shilimkar wrote: >>> On Thursday 04 April 2013 11:36 PM, Jon Hunter wrote: >>>> The L3 interrupt numbers are incorrect for OMAP4+ and are conflicting >>>> with some of the timer interrupts causing the allocation of timer >>>> interrupts to fail. >>>> >>>> The problem is caused by adding 32 to the interrupt number for the L3 >>>> interrupts to account for per processor interrupts (PPI) and software >>>> generated interrupts (SGI) which typically are mapped to the first 32 >>>> interrupts in the ARM GIC. This is not necessary because the first >>>> parameter of the ARM GIC interrupt property specifies the GIC interrupt >>>> type (ie. SGI, PPI, etc). Hence, fix the interrupt number fo the L3 >>>> interrupts by substracting 32. >>>> >>>> Cc: Santosh Shilimkar <santosh.shilimkar@ti.com> >>>> Signed-off-by: Jon Hunter <jon-hunter@ti.com> >>>> --- >>>> >>>> Please note that this problem is observed in Benoit's for_3.10/dts branch [1]. >>>> >>>> [1] http://git.kernel.org/cgit/linux/kernel/git/bcousson/linux-omap-dt.git >>>> >>> Thats correct. I overlooked the 32 addition part. This patch should >>> also be pulled into Benoit's 3.10 tree. >> >> Done. I've just applied it. But I will probably merge it with the >> original patch, because having a broken patch and the fix in the same >> pull request does not look right. > > Please find below the fixed version. > Thanks Benoit. Patch looks fine to me. Regards, Santosh
diff --git a/arch/arm/boot/dts/omap4.dtsi b/arch/arm/boot/dts/omap4.dtsi index 3329140..3ae6a3f 100644 --- a/arch/arm/boot/dts/omap4.dtsi +++ b/arch/arm/boot/dts/omap4.dtsi @@ -97,8 +97,8 @@ reg = <0x44000000 0x1000>, <0x44800000 0x2000>, <0x45000000 0x1000>; - interrupts = <0 41 0x4>, - <0 42 0x4>; + interrupts = <0 9 0x4>, + <0 10 0x4>; counter32k: counter@4a304000 { compatible = "ti,omap-counter32k"; diff --git a/arch/arm/boot/dts/omap5.dtsi b/arch/arm/boot/dts/omap5.dtsi index f4c71d9..86f784e 100644 --- a/arch/arm/boot/dts/omap5.dtsi +++ b/arch/arm/boot/dts/omap5.dtsi @@ -90,8 +90,8 @@ reg = <0x44000000 0x2000>, <0x44800000 0x3000>, <0x45000000 0x4000>; - interrupts = <0 41 0x4>, - <0 42 0x4>; + interrupts = <0 9 0x4>, + <0 10 0x4>; counter32k: counter@4ae04000 { compatible = "ti,omap-counter32k";
The L3 interrupt numbers are incorrect for OMAP4+ and are conflicting with some of the timer interrupts causing the allocation of timer interrupts to fail. The problem is caused by adding 32 to the interrupt number for the L3 interrupts to account for per processor interrupts (PPI) and software generated interrupts (SGI) which typically are mapped to the first 32 interrupts in the ARM GIC. This is not necessary because the first parameter of the ARM GIC interrupt property specifies the GIC interrupt type (ie. SGI, PPI, etc). Hence, fix the interrupt number fo the L3 interrupts by substracting 32. Cc: Santosh Shilimkar <santosh.shilimkar@ti.com> Signed-off-by: Jon Hunter <jon-hunter@ti.com> --- Please note that this problem is observed in Benoit's for_3.10/dts branch [1]. [1] http://git.kernel.org/cgit/linux/kernel/git/bcousson/linux-omap-dt.git arch/arm/boot/dts/omap4.dtsi | 4 ++-- arch/arm/boot/dts/omap5.dtsi | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-)