Lpc2000

Hi Dmitry,

Thanks, this makes some sense now. I wish Keil had documented the __packed
better (I looked at the manual on their web site but it doesn't discuss a
__packed typecast and I would never have guessed that the typecast
actually rearranges the array memory!) I'll give this a go next. Thanks
also for your info bdmlpc.

Kind regards,

Nic

> Hi Nic,
> I had the same problem...
> It seems that Keil's __packed keyword in the front of a pointer
> effectively
> means
> memcpy(packed_p, source_p, sizeof(source));
>
> In your case your can write
> 	EP0Buf[0] = 0;
> 	EP0Buf[1] = 0;
> or just
> *(short *)EP0Buf = 0;
> if you add __atribute__((aligned(2))) to the EP0Buf declaration:
> 	char __atribute__((aligned(2))) EP0Buf[USB_MAX_PACKET0];
> This will in result write to assembly:
> 	.comm EP0Buf, USB_MAX_PACKET0, 2
> where last digit '2' is an optional align passed to linker.
>
> Have fun,
> Dmitry.
>
> P.S. in my case I couldn't run Keil's example on my board with gcc anyway.
>
>
>
> On Tuesday 28 February 2006 05:54, weartronics wrote:
>> Dear lpc2000,
>>
>> In the Keil LPC2148 USB examples, EP0 buffer is declared as a byte array
>> BYTE  EP0Buf[USB_MAX_PACKET0];
>>
>> but later it is type cast to the packed word array
>> *((__packed WORD *)EP0Buf) = 0;
>>
>> I don't understand how the pointer EP0Buf would be changed by casting
>> it to the __packed word pointer. Is anyone able to explain the
>> function of the type cast to me so I can write a GCC equivalent (where
>> the type cast to a packed type is not supported)?
>>
>> Thanks,
>>
>> Nic
>>
>>
>>
>>
>>
>>
>> Yahoo! Groups Links
>>
>>
>>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>
>