[PATCH 1/2] ntdll: Introduce free_ranges indexing sequence.

[Rémi Bernon]

On 4/10/20 12:05 PM, Rémi Bernon wrote:
> On 4/10/20 12:01 PM, Marvin wrote:
>> Hi,
>>
>> While running your changed tests, I think I found new failures.
>> Being a bot and all I'm not very good at pattern recognition, so I 
>> might be
>> wrong, but could you please double-check?
>>
>> Full results can be found at:
>> https://testbot.winehq.org/JobDetails.pl?Key=69304
>>
>> Your paranoid android.
>>
>>
>> === debiant (build log) ===
>>
>> Task: WineTest did not produce the wow32 report
>>
> 
> Looks like my assumptions were incorrect, I'll investigate but I'm still 
> interested in some feedback in the meantime if there's something obvious 
> I missed.

This was simply because I added the fragmentation mitigations before 
actually using the ranges for allocation. The rbtree returned addresses 
are page aligned when the ranges tracking code already uses 64k aligned 
addresses. Moving the mitigations after the patch that starts using the 
ranges makes things work fine (and that's what I was testing).

I can send an updated version with the mitigations added last but all 
the ideas are there already, so I'll wait a bit for any comments.
-- 
Rémi Bernon <rbernon at codeweavers.com>