Re: oops in ioapic_write_entry

[Yinghai Lu <yinghai@kernel.org>]

On 08/02/2010 08:13 PM, Eric W. Biederman wrote:
> Yinghai Lu <yinghai@kernel.org> writes:
> 
>> On 08/02/2010 06:32 PM, Yinghai Lu wrote:
>>> On 08/02/2010 04:17 PM, Dave Airlie wrote:
>>>>>
>>>>> the kernel is using mptable, and the  system have mcp55, so how come
>>>>> with irq 35?
>>>>> assume we should only have ioapic irq 0 - 23 ...
>>>>>
>>>>> Can you send out boot log with "debug apic=debug pci=routeirq" with
>>>>> 2.6.32 and 2.6.35?
>>>>
>>>> Okay el6log is from a RHEL6 2.6.32 kernel, but it should give a good
>>>> baseline, the 2.6.35 oops even earlier with all those options and is
>>>> in the second attachment.
>>>
>>
> 
> This patch is wrong and there is no reason to even suspect it will
> affect this problem.  At best this patch will trade one set of bugs
> for another because at least on some platforms we always did something
> like this.  Having an irq 35 is odd and certainly a result of recent
> changes, but in this case it doesn't look like it has anything to do
> with the problem.
> 
> Nacked-by: "Eric W. Biederman" <ebiederm@xmission.com>
> 
>> please use this one instead..., forget to run quilt refresh before sending it.
>>
>> [PATCH -v2] x86: fix pin_2_irq mapping
>>
>> We should not twist gsi to irq mapping if acpi is not used.
>>
>> -v2 remove not used irq_to_gsi()
>>
>> Signed-off-by: Yinghai Lu <yinghai@kernel.org>
>>
>> ---
>>  arch/x86/include/asm/io_apic.h |   10 ++++++++++
>>  arch/x86/kernel/acpi/boot.c    |    4 ++--
>>  arch/x86/kernel/apic/io_apic.c |    5 +----
>>  3 files changed, 13 insertions(+), 6 deletions(-)
>>
>> Index: linux-2.6/arch/x86/include/asm/io_apic.h
>> ===================================================================
>> --- linux-2.6.orig/arch/x86/include/asm/io_apic.h
>> +++ linux-2.6/arch/x86/include/asm/io_apic.h
>> @@ -185,6 +185,16 @@ int mp_find_ioapic_pin(int ioapic, u32 g
>>  void __init mp_register_ioapic(int id, u32 address, u32 gsi_base);
>>  extern void __init pre_init_apic_IRQ0(void);
>>  
>> +#ifdef CONFIG_ACPI
>> +unsigned int gsi_to_irq(unsigned int gsi);
>> +u32 irq_to_gsi(int irq);
>> +#else
>> +static inline unsigned int gsi_to_irq(unsigned int gsi)
>> +{
>> +	return gsi;
>> +}
>> +#endif
>> +
>>  #else  /* !CONFIG_X86_IO_APIC */
>>  
>>  #define io_apic_assign_pci_irqs 0
>> Index: linux-2.6/arch/x86/kernel/acpi/boot.c
>> ===================================================================
>> --- linux-2.6.orig/arch/x86/kernel/acpi/boot.c
>> +++ linux-2.6/arch/x86/kernel/acpi/boot.c
>> @@ -100,7 +100,7 @@ static u32 isa_irq_to_gsi[NR_IRQS_LEGACY
>>  	0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
>>  };
>>  
>> -static unsigned int gsi_to_irq(unsigned int gsi)
>> +unsigned int gsi_to_irq(unsigned int gsi)
>>  {
>>  	unsigned int irq = gsi + NR_IRQS_LEGACY;
>>  	unsigned int i;
>> @@ -123,7 +123,7 @@ static unsigned int gsi_to_irq(unsigned
>>  	return irq;
>>  }
>>  
>> -static u32 irq_to_gsi(int irq)
>> +u32 irq_to_gsi(int irq)
>>  {
>>  	unsigned int gsi;
>>  
>> Index: linux-2.6/arch/x86/kernel/apic/io_apic.c
>> ===================================================================
>> --- linux-2.6.orig/arch/x86/kernel/apic/io_apic.c
>> +++ linux-2.6/arch/x86/kernel/apic/io_apic.c
>> @@ -1029,10 +1029,7 @@ static int pin_2_irq(int idx, int apic,
>>  	} else {
>>  		u32 gsi = mp_gsi_routing[apic].gsi_base + pin;
>>  
>> -		if (gsi >= NR_IRQS_LEGACY)
>> -			irq = gsi;
>> -		else
>> -			irq = gsi_top + gsi;
>> +		irq = gsi_to_irq(gsi);
>>  	}
>>  
>>  #ifdef CONFIG_X86_32

what is the point for making irq = gsi_top + gsi when mptable is used instead of acpi?

Yinghai