[11u] RFR (Backport): 8177899: Tests fail due to code cache exhaustion on machines with many cores

[Tobias Hartmann]

Hi Martin,

this looks good to me.

Best regards,
Tobias

On 06.06.19 11:35, Doerr, Martin wrote:
> Hi,
> 
>  
> 
> we noticed that jdk11u is also affected by
> 
> https://bugs.openjdk.java.net/browse/JDK-8177899
> 
>  
> 
> The VM didn’t come up on a SPARC machine due to too high upper limit of compiler threads leading to
> misconfigured code cache.
> 
>  
> 
> Unfortunately, the original change does not apply cleanly and needs manual resolution.
> 
>  
> 
> Original change:
> 
> http://hg.openjdk.java.net/jdk/jdk/rev/0451e0a2f1f5
> 
>  
> 
> My backport:
> 
> http://cr.openjdk.java.net/~mdoerr/8177899_code_cache/jdk11u/webrev.00/
> 
>  
> 
> Here’s the complete list of what I had to integrate manually:
> 
>  
> 
> compile.cpp: Does not apply cleanly because of conflict with
> 
> JDK-8209594: guarantee(this->is8bit(imm8)) failed: Short forward jump exceeds 8-bit offset
> 
> Trivial to resolve: Original JDK-8177899 removes this part of JDK-8209594.
> 
> We just need to take the new line: int size = C2Compiler::initial_code_buffer_size(const_size);
> 
>  
> 
> compilationPolicy.cpp: Does not apply cleanly because neighboring hunk has changed:
> 
> JDK-8214206: Fix for JDK-8213419 is broken on 32-bit
> 
> Just need to insert code manually next to where log2_int was changed to log2_intptr.
> 
>  
> 
> tieredThresholdPolicy.cpp: Does not apply cleanly because the file was renamed:
> 
> JDK-8209186: Rename SimpleThresholdPolicy to TieredThresholdPolicy
> 
> Change needs to get applied to simpleThresholdPolicy.cpp instead. Not difficult.
> 
>  
> 
> Please review. I’m targeting to 11.0.5.
> 
>  
> 
> Best regards,
> 
> Martin
> 
>  
>