opus - Sep 2019 - Int overflow fix not present in tree.

Hi! I'm Yves, bit fiddler, hoping to roll a fresh version of libopus into
chromium.

This fix
<https://chromium-review.googlesource.com/c/chromium/src/+/1072869/> never
made it to opus repository.

Was it intentional? If so, what is the rational?
If not, could the fix be applied on the libopus master branch?

Thanks for the amazing work!

-- 
Warm regards,
Yves
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/opus/attachments/20190910/ef501b45/attachment.html>

Hi Yves,

At that time, the issue wasn't reproducible in opus master, but there had
been some significant changes since the last tagged release that made it
hard to pinpoint a single patch to cherry pick to Chromium, hence this
local fix.

Cheers,
Felicia

On Tue, Sep 10, 2019 at 7:56 AM Yves Gerey <yvesg at google.com> wrote:
> Hi! I'm Yves, bit fiddler, hoping to roll a fresh version of libopus
into
> chromium.
>
> This fix
> <https://chromium-review.googlesource.com/c/chromium/src/+/1072869/>
never
> made it to opus repository.
>
> Was it intentional? If so, what is the rational?
> If not, could the fix be applied on the libopus master branch?
>
> Thanks for the amazing work!
>
> --
> Warm regards,
> Yves
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/opus/attachments/20190910/20f279c9/attachment.html>

Felicia Lim wrote:
> Hi Yves,
> 
> At that time, the issue wasn't reproducible in opus master, but there 
> had been some significant changes since the last tagged release that 
> made it hard to pinpoint a single patch to cherry pick to Chromium, 
> hence this local fix.
I'm not convinced that using a saturating add here is the correct thing 
to do, either (as opposed to, say, multiplying out by the coefficients 
individually).

Either way, there are exact analogs of this code in NSQ.c, as well as 
all of the SIMD optimizations (x86/NSQ_sse4_1.c and 
x86/NSQ_del_dec_sse4_1.c, arm/NSQ_del_dec_neon_intr.c, and 
mips/NSQ_del_dec_mipsr1.h). If it is a possible issue in master, the fix 
looks incomplete to me, at least.