[Rev 01] RFR: Add a limition to allocate object heap in aarch64 with tsan enabled.

[Jie He]

On Fri, 24 Apr 2020 04:36:44 GMT, Man Cao <manc at openjdk.org> wrote:

>> Jie He has refreshed the contents of this pull request, and previous commits have been removed. The incremental views
>> will show differences compared to the previous content of the PR. The pull request contains one new commit since the
>> last revision:
>>   Add a limit to allocate object heap in aarch64 with tsan enabled.
>>   
>>   Vm will calculate max heap size according to the physical ram if user
>>   doesn't specify the size(-Xmx). However, with tsan 48-bits memory
>>   mapping of aarch64, it's hard to allocate the large continuous vm
>>   space. Heap allocation often fail in this case, thus, add a limit
>>   at os::has_allocatable_memory_limit to prevent to exceed the
>>   avaiable size. Currently, only support 48-bits address space.
>>   
>>   Can reproduce the behavior by launching "java --version" in a machine
>>   with huge physical memory.
>>   
>>   Change-Id: Iabff934444091436bc657978e2e947c270398f29
>
> src/hotspot/os/posix/os_posix.cpp line 550:
> 
>> 549:   u8 vm_addr_u8 = reinterpret_cast<u8>(&__FUNCTION__);
>> 550:   // High address in 48bits user space is like 0x0000ffffxxxxxxxx.
>> 551:   assert((vm_addr_u8  >> msb_in_aarch64) == 0x1, "warning: allocation could fail in non 48-bit address space.");
> 
> Is this based on kHiAppMemBeg = 0x0ffff00000000ull and kHiAppMemEnd = 0x1000000000000ull?
> Mentioning these two constants like kMidAppMemBeg and kMidAppMemEnd above would be a more helpful comment.

dynamic libraries will be loaded between kHiAppMemBeg and kHiAppMemEnd . it makes hard to allocate large continuous
space.

-------------

PR: https://git.openjdk.java.net/tsan/pull/6