Now backported to GCC 4.6 branch as attached.
Ramana
2012-01-30 Ramana Radhakrishnan
Backport from mainline.
2012-01-20 Ramana Radhakrishnan
PR target/50313
* config/arm/arm.c (arm_load_pic_register): Use
gen_pic_load_addr_unified. Delete calls to g
This is what I finally committed to trunk. I'll backport this to 4.6
once I have finished testing it there.
2012-01-20 Ramana Radhakrishnan
PR target/50313
* config/arm/arm.c (arm_load_pic_register): Use
gen_pic_load_addr_unified. Delete calls to gen_pic_load_addr_32
On 18 January 2012 15:26, Jakub Jelinek wrote:
> On Wed, Jan 18, 2012 at 03:03:47PM +, Ramana Radhakrishnan wrote:
>> +;; operand1 is the memory address to go into pic_load_addr_32bit.
>> +;; operand2 is the PIC label to be emitted from pic_add_dot_plus_*.
>> +;; We do this to allow hoisting o
On Wed, Jan 18, 2012 at 03:03:47PM +, Ramana Radhakrishnan wrote:
> +;; operand1 is the memory address to go into pic_load_addr_32bit.
> +;; operand2 is the PIC label to be emitted from pic_add_dot_plus_*.
> +;; We do this to allow hoisting of the entire c
> +(define_insn_and_split "pic_load_ad
Hi ,
PR50313 is a case where having the 2 patterns "pic_load_addr_*" and
"pic_add_dot_plus_eight/four" from expand time becomes a problem as
discussed by Jakub in his comment in the audit trail for PR48308.
There is a separate problem in combine as explained by my comment in
the audit trail for PR
