[Bug tree-optimization/66012] Sub-optimal 64bit load is generated instead of zero-extension

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66012

Andrew Pinski  changed:

   What|Removed |Added

   Last reconfirmed|2015-12-23 00:00:00 |2016-9-21

--- Comment #4 from Andrew Pinski  ---
On the trunk we generate one 32bit load and one 64bit load (at least for
aarch64):
test:
adrpx0, l
add x1, x0, :lo12:l
ldr w0, [x0, #:lo12:l]   ; 32bit load to w0
ldr x1, [x1, 8]  ; 64bit load to x1
orr x0, x0, x1, lsl 32
ret

[Bug tree-optimization/66012] Sub-optimal 64bit load is generated instead of zero-extension

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66012

--- Comment #2 from Andrew Pinski  ---
Happens on aarch64 also:
test:
adrpx0, l
add x1, x0, :lo12:l
ldr x1, [x1, 8]
ldr w0, [x0, #:lo12:l]
orr x0, x0, x1, lsl 32
ret

[Bug tree-optimization/66012] Sub-optimal 64bit load is generated instead of zero-extension

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66012

Andrew Pinski  changed:

   What|Removed |Added

   Keywords||missed-optimization
 Status|UNCONFIRMED |NEW
   Last reconfirmed||2015-12-23
 Ever confirmed|0   |1
   Severity|normal  |enhancement

--- Comment #3 from Andrew Pinski  ---
Confirmed.

[Bug tree-optimization/66012] Sub-optimal 64bit load is generated instead of zero-extension

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66012

Jakub Jelinek jakub at gcc dot gnu.org changed:

   What|Removed |Added

 CC||jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek jakub at gcc dot gnu.org ---
In GIMPLE that masking is generally useless though, you shift the bits away,
and without the extra BIT_AND_EXPR the expression is more canonical and
shorter.
So, supposedly during expansion or combine you could figure this from the left
shift with large shift count.