[vector api] RFR: Fix AArch64 build failure and jtreg long64/double64 failures for AArch64 NEON

[Yang Zhang (Arm Technology China)]

Hi Jp

When you patch is merged, I will update my patch based on yours.

Regards
Yang

-----Original Message-----
From: Halimi, Jean-Philippe <jean-philippe.halimi at intel.com> 
Sent: Thursday, October 25, 2018 2:57 AM
To: Halimi, Jean-Philippe <jean-philippe.halimi at intel.com>; Vladimir Ivanov <vladimir.x.ivanov at oracle.com>; Yang Zhang (Arm Technology China) <Yang.Zhang at arm.com>; panama-dev at openjdk.java.net
Cc: nd <nd at arm.com>
Subject: RE: [vector api] RFR: Fix AArch64 build failure and jtreg long64/double64 failures for AArch64 NEON

Hi,

Please review the following webrev. It should fix the problem.
http://cr.openjdk.java.net/~jphalimi/webrev_x86_check_fix/webrev/

Jp

-----Original Message-----
From: panama-dev [mailto:panama-dev-bounces at openjdk.java.net] On Behalf Of Halimi, Jean-Philippe
Sent: Wednesday, October 24, 2018 11:33 AM
To: Vladimir Ivanov <vladimir.x.ivanov at oracle.com>; Yang Zhang (Arm Technology China) <Yang.Zhang at arm.com>; panama-dev at openjdk.java.net
Cc: nd <nd at arm.com>
Subject: RE: [vector api] RFR: Fix AArch64 build failure and jtreg long64/double64 failures for AArch64 NEON

Working on it now. Apologies for the inconvenience.

Jp

-----Original Message-----
From: Vladimir Ivanov [mailto:vladimir.x.ivanov at oracle.com] 
Sent: Wednesday, October 24, 2018 11:20 AM
To: Yang Zhang (Arm Technology China) <Yang.Zhang at arm.com>; panama-dev at openjdk.java.net; Halimi, Jean-Philippe <jean-philippe.halimi at intel.com>
Cc: nd <nd at arm.com>
Subject: Re: [vector api] RFR: Fix AArch64 build failure and jtreg long64/double64 failures for AArch64 NEON



On 22/10/2018 18:40, Yang Zhang (Arm Technology China) wrote:
> Hi,
> 
> I'm working on supporting vector api for aarch64 NEON.
> I have a patch which can fix AArch64 build failure and jtreg long64/double64 failures for AArch64 NEON. Could you please help to review this?
> 
> Webrev:
> http://cr.openjdk.java.net/~yzhang/vectorapi.bugfix/webrev.00/

src/hotspot/share/opto/library_call.cpp:
+#ifdef X86
    int bits = num_elem * BitsPerByte * type2aelembytes(bt);
    if (bits == 512 && !VM_Version::supports_avx512vlbwdq()) {
      return NULL;
    }
+#endif

That's ok as a temporary workaround, but shared code shouldn't access platform-specific API and VM_Version::supports_avx512vlbwdq() is x86-specific.

So far, platform-specific logic is hidden behind Matcher.

Jp, I see you added the problematic code [1]. Please, fix it.

Best regards,
Vladimir Ivanov

[1] http://hg.openjdk.java.net/panama/dev/rev/a6e1c1fd5e27