[PATCH] Add host feature detection for Qualcomm CPUs

[Tobias Grosser]

Tim, you are right.

If I got the location right, than the linux kernel has just a single
function for all the vendors [1]. James, was this the place you where
looking at?
As Linux seems to give us those flags without checking for the vendor, the
correct solution seems in fact to remove all vendor checks under the
__arm__ && __linux__ condition.

I attached an updated patch? Would that work for you?

Cheers,
Tobi

[1]
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/arch/arm/kernel/setup.c#n868

On Mon, Jun 10, 2013 at 8:41 PM, Tim Northover <t.p.northover at gmail.com>wrote:

> >> It looks like the feature strings will be the same - and in fact we
> >> grabbed them straight from the Linux kernel source which returns the
> same
> >> strings for ARM-like CPUs (QC or ARM).
> >
> > I was unsure if the feature strings will be the same. However, I just
> looked
> > into the Linux code and it seems it always emits the same feature strings
> > for both targets. Here an updated patch.
>
> So why are we checking the implementor at all? This whole code is
> under #ifdef "__arm__" and "__linux__". Looking at Wikipedia there are
> more ARM implementors to come, and I would be very surprised if Linux
> picked a different set of feature names for them. (More surprised than
> if they broke our code by renaming a feature, for example).
>
> Cheers.
>
> Tim.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130611/045329c4/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Make-host-ARM-CPU-feature-detection-independent-of-t.patch
Type: application/octet-stream
Size: 2870 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130611/045329c4/attachment.obj>