This makes sure that system APC, such as APC_BREAK_PROCESS do not get
interrupted in the middle of their execution, and that the APC
completion notification is always correctly sent back to the caller.
Otherwise DbgBreakProcess sometimes did not return until
WaitForDebugEvent/ContinueDebugEvent are called, because of a race
condition between the APC servicing thread, and the newly created
exception thread.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Jacek Caban <jacek@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The wait_select_reply call may return STATUS_USER_APC/STATUS_KERNEL_APC,
depending on which APC is about to be returned but the apc call will
always be APC_NONE right after the wait. It needs an additional select
request to actually return the call.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Jacek Caban <jacek@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This makes the next iteration to immediately return STATUS_SUCCESS
because of compare_addr returning FALSE with size == 0.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The "broken" function was applied to the wrong condition in the "ok"
statement (RtlIpv6StringToAddress is supposed to set the terminator to
the first character after the address, not the second-to-last character
of the address). However, since this test is already being skipped on XP
and Vista and we really don't need a test for how exactly XP and Vista
are broken, we can just delete it.
Signed-off-by: Alex Henrie <alexhenrie24@gmail.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The search was initiated with base == 0, which returns NULL immediately
if MEM_TOP_DOWN is not used. Use address_space_start instead.
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=47974
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This introduces map_free_area function which tries mapping the expected
free areas until it finds one that succeeds. It now also works for
memory regions outside of the reserved region, but could probably be
improved.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We used this function to find free areas outside of the reserved range,
and it's obviously incorrect as there can be some system or external
memory mapping we don't know about.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
On Windows 10 version 1607, a process called "Memory Compression" violates this
invariant.
Signed-off-by: Zebediah Figura <z.figura12@gmail.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This helps to keep compatibility with old prefixes being reused for wine-5.0.
Old prefixes in their registry have
[System\\CurrentControlSet\\Control\\Nls\\Codepage]
"37"=""
[System\\CurrentControlSet\\Control\\Nls\\Language]
"0409"=""
and this leads to LCMapString(LCMAP_LOWERCASE/LCMAP_UPPERCASE) return garbage.
This is a regression caused by 94a3add0ea.
Signed-off-by: Dmitry Timoshkov <dmitry@baikal.ru>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Takes care of one more use of "long double".
Signed-off-by: Erich E. Hoover <erich.e.hoover@gmail.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
If 0 is returned, the caller has no way of determining this. This fixes a
test failure in kernel32:change introduced by f46fa9c92.
Signed-off-by: Zebediah Figura <z.figura12@gmail.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Alexandre Julliard