> From: Richard Earnshaw
> Sent: Monday, August 18, 2014 6:34 PM
> >
> > The problem is that in the pattern for constable_4 we don't have the
> information
> > about the access mode for this entry. In the testcase along this patch the
> rtx
> > manipulated in the pattern is VOIDmode while the acces
On 13/08/14 08:49, Thomas Preud'homme wrote:
>> From: Richard Earnshaw
>> Sent: Monday, August 11, 2014 4:54 PM
>>
>> I think this is the wrong place for this sort of fix up. HFmode values
>> are fixed up in the consttable_4 pattern and it looks wrong to be that
>> HImode values are then fixed up
> From: Richard Earnshaw
> Sent: Monday, August 11, 2014 4:54 PM
>
> I think this is the wrong place for this sort of fix up. HFmode values
> are fixed up in the consttable_4 pattern and it looks wrong to be that
> HImode values are then fixed up in a different place. We should be
> consistent a
On 11/08/14 09:53, Richard Earnshaw wrote:
> On 11/08/14 08:49, Thomas Preud'homme wrote:
>> Being 32-bit wide in size, constant pool entries are always filled as 32-bit
>> quantities. This works fine for little endian system but gives some incorrect
>> results for big endian system when the value
On 11/08/14 08:49, Thomas Preud'homme wrote:
> Being 32-bit wide in size, constant pool entries are always filled as 32-bit
> quantities. This works fine for little endian system but gives some incorrect
> results for big endian system when the value is *accessed* with a mode smaller
> than 32-bit
Being 32-bit wide in size, constant pool entries are always filled as 32-bit
quantities. This works fine for little endian system but gives some incorrect
results for big endian system when the value is *accessed* with a mode smaller
than 32-bit in size. Suppose the case of the value 0x1234 that is
