RFR: 8245000: Windows GDI functions don't support large pages

[Stefan Karlsson]

Thanks for reviewing!

StefanK

On 2020-05-18 13:56, Kim Barrett wrote:
>> On May 18, 2020, at 4:17 AM, Stefan Karlsson <stefan.karlsson at oracle.com> wrote:
>>
>> On 2020-05-15 20:26, Kim Barrett wrote:
>>>> On May 15, 2020, at 1:47 PM, Stefan Karlsson <stefan.karlsson at oracle.com> wrote:
>>>>
>>>> Hi all,
>>>>
>>>> Please review this patch to turn off -XX:+UseLargePages as long as Windows GDI functions don't work with large pages.
>>>>
>>>> https://cr.openjdk.java.net/~stefank/8245000/webrev.00.cleanup
>>>> https://cr.openjdk.java.net/~stefank/8245000/webrev.01.workaround
>>>> https://cr.openjdk.java.net/~stefank/8245000/webrev.all/
>>> A couple of things that may or may not need to be addressed.
>>> Otherwise, looks good to me.
>>>
>>> ------------------------------------------------------------------------------
>>> src/hotspot/os/windows/os_windows.cpp
>>> 3118   if (_large_page_size > default_page_size) {
>>> ...
>>> 3124   UseLargePages = _large_page_size != 0;
>>>
>>> Should UseLargePages be true if _large_page_size > 0 but < default_page_size?
>>> (Assuming that's actually possible, but I see nothing in this code
>>> that makes that impossible.  The OS might...)
>>>
>>> The old code did the same thing, so if it's a bug, it's been there all along.
>> The code that sets up _lage_page_size uses GetLargePageMinimum, which sets the size to the system's large page size or 0. It's not explicitly stated that it's not < default_page_size (4 * 1024), but it would really surprise me if it was:
>>
>> https://docs.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-getlargepageminimum
>>
>> It also uses LargePageSizeInBytes, but only if it is larger than, and aligned to, the given GetLargePageMinimum value.
>>
>> So, I think we are OK here.
> Thanks for the explanation and the doc pointer.  Agreed this is okay.
>
>>> src/hotspot/os/windows/os_windows.cpp
>>> 2920   if (!gdi_can_use_split_reservation_memory(UseLargePages, min_interleave_granularity)) {
>>> 2921     WARN("Windows GDI cannot handle split reservations.");
>>> 2922     WARN("...Ignoring UseNUMAInterleaving flag.");
>>>
>>> If the failure is due to UseLargePages, it might be worth mentioning
>>> that in the warning.
>> We've already done the large pages checks when this code runs. If we can't use large pages, the UseLargePages flag will have been turned off at this point.
> Oh, of course.
>
> Looks good.
>