Blue Swirl <blauwir...@gmail.com> writes:
> On 5/11/10, Markus Armbruster <arm...@redhat.com> wrote:
>> Blue Swirl <blauwir...@gmail.com> writes:
>>
>> > On 5/9/10, chen huacai <zltjiang...@gmail.com> wrote:
>> >> This patch add initial support of VIA IDE controller used by fulong mini
>> pc
>> >>
>> >> Signed-off-by: Huacai Chen <zltjiang...@gmail.com>
>> >> -----
>> [...]
>>
>> >> diff --git a/hw/ide/via.c b/hw/ide/via.c
>> >> new file mode 100644
>> >> index 0000000..9adc5b5
>> >> --- /dev/null
>> >> +++ b/hw/ide/via.c
>> >> @@ -0,0 +1,189 @@
>>
>> [...]
>>
>> >> +static void bmdma_writeb(void *opaque, uint32_t addr, uint32_t val)
>> >> +{
>> >> + BMDMAState *bm = opaque;
>> >> +#ifdef DEBUG_IDE
>> >> + printf("bmdma: writeb 0x%02x : 0x%02x\n", addr, val);
>> >> +#endif
>> >> + switch(addr & 3) {
>> >> + case 2:
>> >> + bm->status = (val & 0x60) | (bm->status & 1) | (bm->status &
>> >> ~val & 0x06);
>> >> + break;
>> >
>> > Some gccs complain if there is no default case.
>>
>>
>> Are you sure? Which version?
>>
>> It warns when the switch expression's type is an enumeration, and the
>> switch doesn't handle all enumeration constants, and has no default
>> case.
>>
>> I can be made to warn whenever there's no default, but we don't do that.
>
> This problem came up with 299b520cd4092be3c53f8380b81315c33927d9d3
> (fixed before commit) and 4450521668471c7685551d8c5bcc582d754e9843,
> with mingw gcc and OpenBSD gccs.
[...]
>
> Maybe those cases were different from this.
The first commit you quoted is indeed different: the default prevents a
spurious warning: ‘context’ may be used uninitialized in this function
from some versions of gcc. No such confusion can arise here.
In the second one, the switch expression is of enumeration type.
