[PATCH 2/2] ntdll: allow HeapSize to work with global GMEM_MOVEABLE pointers

[Damjan Jovanovic]

This patch attempts to address the problem by patching RtlSizeHeap()
to subtract 8 (8 being HGLOBAL_STORAGE) from the pointer if it doesn't
seem like a valid pointer, to cater to GMEM_MOVEABLE pointers actually
starting 8 bytes earlier.

I am not familiar with the internals of Windows' memory management so
this might not be the best solution, but all the heap tests pass so it
can't be that wrong either ;). It certainly gets wxWidgets
applications working.

Signed-off-by: Damjan Jovanovic <damjan.jov(a)gmail.com>
---
 dlls/kernel32/tests/heap.c |  2 +-
 dlls/ntdll/heap.c          | 50 ++++++++++++++++++++++++--------------
 2 files changed, 33 insertions(+), 19 deletions(-)

[Dmitry Timoshkov]

Damjan Jovanovic <damjan.jov(a)gmail.com> wrote:
> This patch attempts to address the problem by patching RtlSizeHeap()
> to subtract 8 (8 being HGLOBAL_STORAGE) from the pointer if it doesn't
> seem like a valid pointer, to cater to GMEM_MOVEABLE pointers actually
> starting 8 bytes earlier.

Probably fixing GlobalAlloc() instead would be more correct.

-- 
Dmitry.

[Damjan Jovanovic]

On Sat, Oct 26, 2019 at 8:55 AM Dmitry Timoshkov <dmitry(a)baikal.ru> wrote:
> Damjan Jovanovic <damjan.jov(a)gmail.com> wrote:
> > This patch attempts to address the problem by patching RtlSizeHeap()
> > to subtract 8 (8 being HGLOBAL_STORAGE) from the pointer if it doesn't
> > seem like a valid pointer, to cater to GMEM_MOVEABLE pointers actually
> > starting 8 bytes earlier.
>
> Probably fixing GlobalAlloc() instead would be more correct.

You're right. I am currently investigating global memory layout in
different Windows versions. Please ignore this patch.