GCC has some functions, variables & type attributes which can be used as aliases
for some of the SAL annotations. Although it's not as rich & precise, it's still useful
since we actually enable -Werror on GCC builds whereas we don't use such an option
on MSVC builds.
For now, _Must_inspect_result_ is aliased to warn_result_unused attribute.
Spinlocks are not reentrant (and this is done a lot), using them forces us to have
an horrible hack in the kernel, which unschedules threads which are at DISPATCH_LEVEL
thus allowing another thread to take ownership of the spinlock while the unscheduled
thread should already hold it.
CORE-6473
NEVER DO THIS! It is guaranteed to be wrong. Instead always individually pack single structures that need packing.
This fixes USB mouse on 64 bit builds.
RtlpCaptureNonVolatileContextPointers walks the stack and captures the addresses of all non-volatile registers on the stack, when they have been saved first. This is needed to be able to fix up the non-volatile on a system call, which doesn't capture non-volatiles, but relies on them to be restored by the callees.
Instead of only checking for the TargetFrame, also check for a mode change, i.e. RIP went from kernel to user, in which case the target frame was not reached yet, because it was too large, but processing can't continue here.
RtlSetUnwindContext uses RtlpCaptureNonVolatileContextPointers to set the non-volatile registers in the the stack. They will be picked up, when returning back or unwinding, e.g. to the system call handler.
CORE-17601
Dynamically load SetThreadUILanguage(), so as to support systems where this API is not present.
Hopefully implemented in a thread-safe manner.
Also, put include directory next to the library and use
target_include_directories(.. INTERFACE ..) to get this right.
This is because :
- Having includes & implementation in two different places buggers me
- This makes sure that there is no "if it compiles everything is fine" behaviour from anyone
because now even static libraries need it for GCC amd64 build
Also add __USE_PSEH2__ define for the non SEH-aware compilers out there and use it in a few headers
where we define macros involving __try
The array is there for the entries smaller than the decommit threshold, the rationale
being that entries which are larger will likely be split for honoring other allocations
or be coalesced and eventually decommitted.
This with the previous commits make a huge perf boost to memory-intensive applications like cmake
CORE-15793
We are not ready for enabling ATLASSERT. Enabling ATL assertions takes time to realize. CORE-17505
- Disable ATLASSERT by undefining _DEBUG.
- Revert currently non-fixable codes.
- Fix generic text mapping for GetWindowText and GetWindowTextLength functions.
- Fix the position.
- Fix the length.
- Fail elegantly if necessary.
CORE-9281
Addendum to 0.4.15-dev-1453-g 4ad7b6d
Looks like CMake on master does not longer understand the syntax of
replace_compile_flags().
Will investigate that later again, when other issues have been
solved that do currently prevent me from testing that locally at all.
Most likely remove_target_compile_option() has to be used instead now.
[AUTOCHK] Add also support for scanning FATX volumes.
The Format(), FormatEx(), Chkdsk(), ChkdskEx() functions exposed by the
U*.DLL user-mode FS library dlls are different (and have different
prototypes) than the similarly-named functions exported by FMIFS.DLL .
In particular, what we used to call "xxxChkdskEx()" and "xxxFormatEx()"
in our U*.DLL libraries actually correspond more, from their arguments,
to the "Chkdsk()" and "Format()" functions in Windows' U*.DLL . Their
*Ex() counterparts instead take most of the parameters through a
structure passed by pointer.
On FMIFS.DLL side, while FMIFS!Chkdsk() calls U*.DLL!Chkdsk() and
FMIFS!ChkdskEx() calls U*.DLL!ChkdskEx() (and we do not implement these
*Ex() functions at the moment), both FMIFS!Format() and FMIFS!FormatEx()
call U*.DLL!Format() instead, while FMIFS!FormatEx2() calls
U*.DLL!FormatEx() (that we do not implement yet either) !!
To improve that, refactor the calls to these U*.DLL functions so as to
respect the more compatible prototypes: They contain the correct number
of parameters in a compatible order. However, some of the parameters do
not have the same types yet: the strings are kept here in PUNICODE_STRINGS,
while on Windows they are passed via an undocumented DSTRING struct, and
the FMIFS callback is instead a MESSAGE struct/class on Windows.
Finally, the MEDIA_TYPE parameter in U*.DLL!Format() is equivalent, yet
not fully 100% in 1-to-1 correspondence, with the FMIFS_MEDIA_FLAG used
in the corresponding FMIFS.DLL functions.
One thing to notice is that the U*.DLL!Format() (and the Ex) functions
support a BOOLEAN (a flag resp.) for telling that a backwards-compatible
FS version should be used instead of the (default) latest FS version.
This is used e.g. by the FAT FS, where by default FAT32 is selected
(depending also on other constraints like, the disk and the partition
sizes), unless that bit is set in which case, FAT16 (or 12) is used.
- Reduce the diff with upstream by aliasing malloc() and free()
with RtlAllocateHeap() and RtlFreeHeap() respectively.
- Fix a memory leak in the failure code path of the do-while
memory reallocation loop for IOCTL_BTRFS_QUERY_FILESYSTEMS
in is_mounted_multi_device().