Hi Timo et al.,
To make the issue wider known than just me comment on GitHub. That commit introduced a regression.
Le 20/08/2018 à 18:07, Timo Kreuzer a écrit :
https://git.reactos.org/?p=reactos.git;a=commitdiff;h=c219be9409af429909b143...
commit c219be9409af429909b1433da536516b3c3ab823 Author: Timo Kreuzer timo.kreuzer@reactos.org AuthorDate: Sun Feb 4 19:20:56 2018 +0100 Commit: Timo Kreuzer timo.kreuzer@reactos.org CommitDate: Mon Aug 20 18:07:35 2018 +0200
[NTOS:MM] Fix paged pool expansion
ntoskrnl/mm/ARM3/pool.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/ntoskrnl/mm/ARM3/pool.c b/ntoskrnl/mm/ARM3/pool.c index 7cebabe818..113a7aa981 100644 --- a/ntoskrnl/mm/ARM3/pool.c +++ b/ntoskrnl/mm/ARM3/pool.c @@ -561,10 +565,11 @@ MiAllocatePoolPages(IN POOL_TYPE PoolType, MiInitializePfnForOtherProcess(PageFrameNumber, (PMMPTE)PointerPde, MmSystemPageDirectory[(PointerPde - MiAddressToPde(NULL)) / PDE_COUNT]); +#endif
/* Write the actual PDE now */-// MI_WRITE_VALID_PDE(PointerPde, TempPde); -#endif
MI_WRITE_VALID_PDE(PointerPde, TempPde);
That little change for x86 actually breaks VBox support. It hits an ASSERT during 3rd stage: *** Assertion failed: PointerPde->u.Hard.Valid == 0 *** Source File: /srv/buildbot/Build_GCCLin_x86/build/ntoskrnl/mm/ARM3/miarm.h, line 975 See testbots for instance: https://build.reactos.org/builders/Test%20VBox/builds/8625/steps/test/logs/s...
Cheers, Pierre