Am 17.06.2013 22:59, schrieb Richard Henderson:
> On 06/16/2013 08:57 AM, Andreas Färber wrote:
>> +enum CPUInterruptFlags {
>> + CPU_INTERRUPT_HARD = 0x0002,
>> + CPU_INTERRUPT_EXITTB = 0x0004,
>> + CPU_INTERRUPT_HALT = 0x0020,
>> + CPU_INTERRUPT_DEBUG = 0x0080,
>> +
>> +#if defined(CONFIG_SOFTMMU) || defined(CONFIG_USER_ONLY)
>> + /* Several target-specific external hardware interrupts.
>> + * Each target/cpu.h should define proper names based on them.
>> + */
>> + CPU_INTERRUPT_TGT_EXT_0 = 0x0008,
>> + CPU_INTERRUPT_TGT_EXT_1 = 0x0010,
>> + CPU_INTERRUPT_TGT_EXT_2 = 0x0040,
>> + CPU_INTERRUPT_TGT_EXT_3 = 0x0200,
>> + CPU_INTERRUPT_TGT_EXT_4 = 0x1000,
>
> I don't like this. It's a subtle point between C and C++ that the later can
> change the implementation type of the enum based on the contents of the enum.
> While we don't currently ever build with C++, I'd rather not have something
> like this hiding.
>
> Is the ifdef really better than the poison?
With today's poison we can't use it in qom/cpu.c nor intc/sh_intc.c.
Andreas
--
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg
