RFR (trivial): 8219888: aarch64 : add CPU detection code for HiSilicon TSV110

Yangfei (Felix) felix.yang at huawei.com
Fri Mar 1 03:07:48 UTC 2019

I am afraid we are duplicating code having one if-clause for each CPU.
Here, we have three different CPUs sharing this if-cause: CAVIUM, BROADCOM and HISILICON.
It’s OK to place all vendor specific description together at the top, but I am not clear about the model of BRAODCOM.
So I am going this way: separate description for each model.

The top comment “says” this if-clause was for ThunderX2, which looks a little strange.

I would suggest either we place all vendor specific descriptions together at the top, or have a new if-clause for each (much cleaner).

  // ThunderX2
  if ((_cpu == CPU_CAVIUM && (_model == 0xAF)) ||
      (_cpu == CPU_BROADCOM && (_model == 0x516)) ||
  // HiSilicon TSV110
      (_cpu == CPU_HISILICON && (_model == 0xd01))) {
    if (FLAG_IS_DEFAULT(AvoidUnalignedAccesses)) {
      FLAG_SET_DEFAULT(AvoidUnalignedAccesses, true);
    if (FLAG_IS_DEFAULT(UseSIMDForMemoryOps)) {
      FLAG_SET_DEFAULT(UseSIMDForMemoryOps, true);



-----Original Message-----

From: aarch64-port-dev <aarch64-port-dev-bounces at openjdk.java.net> On Behalf Of Yangfei (Felix)

Sent: Friday, March 1, 2019 9:23 AM

To: hotspot-runtime-dev at openjdk.java.net

Cc: aarch64-port-dev <aarch64-port-dev at openjdk.java.net>

Subject: [aarch64-port-dev ] RFR (trivial): 8219888: aarch64 : add CPU detection code for HiSilicon TSV110


    Please review this trivial patch adding support for HiSilicon TSV110.

    Bug: https://bugs.openjdk.java.net/browse/JDK-8219888

    Webrev: http://cr.openjdk.java.net/~fyang/8219888/webrev.00/

    Tested on Huawei Kunpeng 920 server platform.  OK to push?



More information about the hotspot-runtime-dev mailing list