RFR 8214449: Make alt_hash a regular immutable member of inner hash table

[Gerard Ziemski]

hi all,

I’m withdrawing this fix after discussion with Coleen and Kim. This will be addressed later as part of a bigger effort to make the concurrent hashtable more robust.


cheers

> On Dec 4, 2018, at 10:00 AM, Gerard Ziemski <gerard.ziemski at oracle.com> wrote:
> 
> Thank you for the review!
> 
>> On Dec 3, 2018, at 2:07 PM, coleen.phillimore at oracle.com wrote:
>> On 12/1/18 12:28 PM, Gerard Ziemski wrote:
>>> Hi all,
>>> 
>>> Please review this followup to JDK-8195100 (Use a low latency hashtable for SymbolTable)
>>> 
>>> Here we make the “_alt_hash” an immutable and regular field (i.e. non static) of the inner hashtable, to avoid manipulating it from outside and possibly getting it wrong.
>> 
>> +class SymbolTableHash : public SymbolTableHashClass {
>> +
>> +private:
>> + bool _alt_hash;
>> +
>> +public:
>> + SymbolTableHash(size_t log2size, size_t log2size_limit, size_t grow_hint, bool alt_hash = false) :
>> + ConcurrentHashTable(log2size, log2size_limit, grow_hint) { } ;
>> + bool get_alt_hash() { return _alt_hash; } ;
>> +};
>> 
>> 
>> Can you rename _alt_hash member and get_alt_hash to _has_alt_hash and hash_alt_hash() respectively?
> 
> I assume you mean:
> 
> get_alt_hash() -> has_alt_hash()
> 
> Done.
> 
>> 
>> Also indent the ConcurrentHashTable constructor call here.
> 
> Done.
> 
> Issue:   https://bugs.openjdk.java.net/browse/JDK-8214449
> Webrev:  http://cr.openjdk.java.net/~gziemski/8214449_rev2
> Testing: Mach5 hs_tier1,2,3,4,5 in progres...
> 
>