[coreboot] [PATCH v2] AMD DBM690T IRQ cleanup

[Carl-Daniel Hailfinger]

On 01.10.2008 19:53, Marc Jones wrote:
> Carl-Daniel Hailfinger wrote:
>> Hi,
>>
>> I decided to prepare a patch for the stuff I mentioned in the DBM690T
>> review.
>>
>> Use easily readable macros to setup interrupt routing.
>> Change a few PCI bus/dev/fn to use hexadecimal numbers.
>>
>> Signed-off-by: Carl-Daniel Hailfinger
>> <c-d.hailfinger.devel.2006 at gmx.net>
>
> The cleanup is nice but you didn't translate correctly. See below.

Thanks for checking! It seems I mistyped the regular expression for
search and replace.


>> Index: LinuxBIOSv2-irq_macros/src/mainboard/amd/dbm690t/mptable.c
>> ===================================================================
>> --- LinuxBIOSv2-irq_macros/src/mainboard/amd/dbm690t/mptable.c   
>> (Revision 3624)
>> +++ LinuxBIOSv2-irq_macros/src/mainboard/amd/dbm690t/mptable.c   
>> (Arbeitskopie)
>> @@ -122,94 +122,72 @@
>>      smp_write_intsrc(mc, mp_ExtINT,
>>               MP_IRQ_TRIGGER_EDGE | MP_IRQ_POLARITY_HIGH, bus_isa,
>>               0x0, apicid_sb600, 0x0);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0x1, apicid_sb600, 0x1);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0x0, apicid_sb600, 0x2);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0x3, apicid_sb600, 0x3);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0x4, apicid_sb600, 0x4);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0x6, apicid_sb600, 0x6);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0x7, apicid_sb600, 0x7);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0xc, apicid_sb600, 0xc);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0xd, apicid_sb600, 0xd);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_EDGE |
>> MP_IRQ_POLARITY_HIGH,
>> -             bus_isa, 0xe, apicid_sb600, 0xe);
>>  
>> +    /* ISA ints are edge-triggered, and usually originate from the
>> ISA bus,
>> +     * or its remainings.
>> +     */
>> +#define ISA_INT(intr, pin) \
>> +    smp_write_intsrc(mc, mp_INT,
>> MP_IRQ_TRIGGER_EDGE|MP_IRQ_POLARITY_HIGH,  bus_isa, (intr),
>> apicid_sb600, (pin))
>> +
>> +    ISA_INT(0x1, 0x1);
>> +    ISA_INT(0x0, 0x2);
>> +    ISA_INT(0x3, 0x3);
>> +    ISA_INT(0x4, 0x4);
>> +    ISA_INT(0x6, 0x6);
>> +    ISA_INT(0x7, 0x7);
>> +    ISA_INT(0xc, 0xc);
>> +    ISA_INT(0xd, 0xd);
>> +    ISA_INT(0xe, 0xe);
>> +
>> +    /* PCI interrupts are level triggered, and are
>> +     * associated with a specific bus/device/function tuple.
>> +     */
>> +#define PCI_INT(bus, dev, fn, pin) \
>> +        smp_write_intsrc(mc, mp_INT,
>> MP_IRQ_TRIGGER_LEVEL|MP_IRQ_POLARITY_LOW, (bus), (((dev)<<2)|(fn)),
>> apicid_sb600, (pin))
>> +
>>      /* usb */
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_LEVEL |
>> MP_IRQ_POLARITY_LOW,
>> -             0, 19 << 2 | 0, apicid_sb600, 0x10);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_LEVEL |
>> MP_IRQ_POLARITY_LOW,
>> -             0, 19 << 2 | 1, apicid_sb600, 0x11);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_LEVEL |
>> MP_IRQ_POLARITY_LOW,
>> -             0, 19 << 2 | 2, apicid_sb600, 0x12);
>> -    smp_write_intsrc(mc, mp_INT, MP_IRQ_TRIGGER_LEVEL |
>> MP_IRQ_POLARITY_LOW,
>> -             0, 19 << 2 | 3, apicid_sb600, 0x13);
>> +    PCI_INT(0, 0x13, 2, 0);
>> +    PCI_INT(0, 0x13, 2, 1);
>> +    PCI_INT(0, 0x13, 2, 2);
>> +    PCI_INT(0, 0x13, 2, 3);
>
> This should be:
>     PCI_INT(0, 0x13, 0, 0x10);
>     PCI_INT(0, 0x13, 1, 0x11);
>     PCI_INT(0, 0x13, 2, 0x12);
>     PCI_INT(0, 0x13, 3, 0x13);
>
> I think you put the << 2 in the fn throughout the patch.

The regexp grabbed the wrong parts of the code. I'll fix and resend.

Regards,
Carl-Daniel

-- 
http://www.hailfinger.org/