RFR: 8311248: Refactor CodeCache::initialize_heaps to simplify adding new CodeCache segments [v2]

[Evgeny Astigeevich]

On Sat, 6 Jan 2024 14:05:33 GMT, Boris Ulasevich <bulasevich at openjdk.org> wrote:

>> The change simplifies the CodeCache::initialize_heaps segment memory split logic while preserving the existing layout:
>> 
>>   if (!non_nmethod_set && !profiled_set && !non_profiled_set) {
>>     ...
>>   } else if (!non_nmethod_set || !profiled_set || !non_profiled_set) {
>>     if (non_profiled_set) {
>>       if (!profiled_set) {
>>         ...
>>       }
>>     } else if (profiled_set) {
>>       ...
>>     } else if (non_nmethod_set) {
>>       ...
>>     }
>>   }
>> 
>> -->
>> 
>>   if (!profiled.set && !non_profiled.set) {
>>     ..
>>   }
>>   if (profiled.set && !non_profiled.set) {
>>     ..
>>   }
>>   if (!profiled.set && non_profiled.set) {
>>     ..
>>   }
>>   if (!non_nmethod.set && profiled.set && non_profiled.set) {
>>     ..
>>   }
>> 
>> 
>> With this change, PrintFlagsFinal shows the actual segment sizes (not an intermediate value before alignment), and the segments completely fill the ReservedCodeCacheSize (no wasted page due to final down alignment).
>
> Boris Ulasevich 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:
> 
>   cleanup & test udpdate

src/hotspot/share/code/codeCache.hpp line 469:

> 467: typedef CodeBlobIterator<CodeBlob, AllCodeBlobsFilter, false /* is_relaxed */> AllCodeBlobsIterator;
> 468: 
> 469: struct CodeCacheSegment {

Would this create a confusion?
`CodeHeap` consists of blocks and names them `segments`.  See `src/hotspot/share/memory/heap.hpp` and  `src/hotspot/share/memory/heap.cpp`.
There is `CodeCache::allocated_segments()` which returns the total number of segments (memory blocks) code heaps use. Also there is `CodeCacheSegmentSize` which defines the size of a memory block for a code heap.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17244#discussion_r1453776864