Full generated stub save and restore PR now in draft

[Vladimir Kozlov]

I added comments to PR.

Thanks,
Vladimir K

On 11/20/25 10:25 AM, Andrew Dinn wrote:
> I have a draft PR against mainline which implements save and restore for all generated stubs:
> 
>    https://github.com/openjdk/jdk/pull/28433
> 
> It appears to be functioning correctly on both aarch64 and x86 but a run of the full Oracle test suite would be helpful. 
> The patch switches on stub save and restore by default and I have updated test AOTCodeFlags so it exercises stub save 
> and restore with and without adapter save and restore and this passes on both arches.
> 
> This PR not only saves and restores code it also sorts out a lot of the mess of publishing and updating external and 
> stub addresses as a side effect of loading or generating stubs. This is an important step for us to be able to make 
> nmethod save and restore reliable when we add it (which partly explains why I targeted this to mainline).
> 
> regards,
> 
> 
> Andrew Dinn
> -----------
>