[PATCH v3 1/2] ntdll: Introduce free_ranges indexing sequence.

[Rémi Bernon]

On 5/25/20 11:38 PM, Rémi Bernon wrote:
> On 5/25/20 5:57 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=72194
>>
>> Your paranoid android.
>>
>>
>> === debiant (32 bit report) ===
>>
>> ntdll:
>> virtual: Timeout
>>
>> === debiant (32 bit Chinese:China report) ===
>>
>> ntdll:
>> virtual: Timeout
>>
> 
> This fails in the NtMapViewOfSection tests, because in 32bit mode it's 
> possible to use AT_ROUND_TO_PAGE to map memory aligned to 4k boundaries, 
> breaking the 64k alignment assumptions, I'll try to make that work.

This basically means that on 32bit we can have multiple views in the 
same 64kB range. I'll adjust the assert to handle that case and add a 
check of surrounding views to decide whether a range should be unmapped 
or not if it's still in use, but I'll wait a bit for any comment before 
sending the updated version to avoid spamming the list.
-- 
Rémi Bernon <rbernon at codeweavers.com>