[Bug target/71235] march=silvermont turns on aes by default

[tjoener at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71235

--- Comment #4 from Jan De Kock  ---
@Jakub, isn't that the whole point? To be the lowest common denominator?

In the case of silvermont, the desktop SKUs (Bay-Trail-D) and mobile SKUs
(Bay-Trail-M), both don't support AES. That's a pretty large group if you ask
me

[Bug target/71235] New: march=silvermont turns on aes by default

[tjoener at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71235

Bug ID: 71235
   Summary: march=silvermont turns on aes by default
   Product: gcc
   Version: unknown
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
 Component: target
  Assignee: unassigned at gcc dot gnu.org
  Reporter: tjoener at gmail dot com
  Target Milestone: ---

march=silvermont automatically enables -maes, but not all silvermonts have aes
support. Avoton has it, but Bay Trail-D does not.

Eg: http://ark.intel.com/products/78867

The documentation does not make this clear, as it states the silvermont option
is for silvermont CPUs with aes support. Disabling this manually is no trouble.
Just wondering if it shouldn't be the least common denominator, ie silvermonts
without aes.

Maybe there should be a differentiation on silvermont with aes and silvermont
without aes?