RFR(S): 8241492: Strip mining not working for test/hotspot/jtreg/compiler/c2/Test6850611.java

[Doerr, Martin]

Hi Roland,

thanks for fixing it.

Tricky, but correct:
limit - iv_phi when counting forwards (and iv_phi - limit otherwise) results in the correct difference when it's interpreted as unsigned.

Looks good to me.

Should we add -Xbatch to the new test?

Please update copyright year.

Best regards,
Martin


> -----Original Message-----
> From: hotspot-compiler-dev <hotspot-compiler-dev-
> bounces at openjdk.java.net> On Behalf Of Robbin Ehn
> Sent: Freitag, 27. März 2020 14:00
> To: Roland Westrelin <rwestrel at redhat.com>; hotspot-compiler-
> dev at openjdk.java.net
> Subject: Re: RFR(S): 8241492: Strip mining not working for
> test/hotspot/jtreg/compiler/c2/Test6850611.java
> 
> Thanks a lot for addressing this Roland!
> 
> (not a review)
> 
> /Robbin
> 
> On 2020-03-27 11:17, Roland Westrelin wrote:
> >
> > https://bugs.openjdk.java.net/browse/JDK-8241492
> > http://cr.openjdk.java.net/~roland/8241492/webrev.00/
> >
> > The loop is strip mined but the computation of the inner loop number of
> > iterations is wrong. It is the min of the number of iterations left and
> > LoopStripMiningIter. In this case, the number of iterations left
> > (guaranteed always positive) is greater than the max signed int but the
> > min() that's used operates on signed ints so even though there's a large
> > number of iterations left, the resulting number of iterations is not
> > LoopStripMiningIter. Switching to a unsigned min() fixes this.
> >
> > Roland.
> >