[RFC][PATCH,reload] PR middle-end/42431, fix latent bug in gen_reload_chain_without_interm_reg_p()

[Jeff Law]

On 03/02/10 03:23, Jakub Jelinek wrote:
> On Tue, Mar 02, 2010 at 11:13:49AM +0100, Richard Guenther wrote:
>    
>>> Bootstrapped and regression tested on x86_64-linux-gnu.
>>>        
>> !   VEC_free (rtx_p , heap, substitute_stack);
>>      
> Extra space after rtx_p.
>    
I'll fix this.

>> !   return result;
>>
>> not freeing the stack should speed up further calls for the cost of
>> a very small leak.
>>
>> Otherwise ok.
>>      
> I guess it can be freed e.g. at the end of reload1.c (reload).
>    
Freeing it at the end of reload is the better solution, if for no other 
reason than it keeps extra crap out of valgrind reports.  I'm doing a 
bootstrap & regerssion test with that change and will install after it 
completes.

jeff