| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This typo was in #24989 so would be a new regression in 3.0.
In an x86 build, it causes us to not get the cache size correct,
leading us to use a smaller default cache size and do more GCs.
Tested with GCPerfSim and this PR reduces TotalNumberGCs by 33% using an x86 build.
|
|
* Fall back to CpuId if failed to get cache size from OS
It's possible for GetLogicalProcessorCacheSizeFromOS() to fail;
this happens on alpine linux where it compiles to just `return 0;`.
As a fallback, we can get the cache size from CpuId. Previously that
was specific to x86; this PR preserves the behavior that we never call
GetLogicalProcessorCacheSizeFromOS on x86.
CpuId only works on x86 and amd64; on other systems we may still return
0 from here. Then GC defaults to a cache size of only 0.25MB.
Note: Removed the code in an `#ifdef _WIN64` that was nested inside of
`#if defined (_TARGET_X86_)`. Presuming that is dead code.
* Fix exception handler
|
|
|
|
* Reduce indirect dispatch in hot paths in metadata
* Remove allocation indirections in utilcode for coreclr
- we no longer have a multi-dll distribution of code that needs to share heaps
* Remove unused code in GetCLRFunction
* Remove virtual dispatch around impl/decl methodtable access in MethodData
|
|
* Remove unused code in vm/util.*
|
|
* Remove all uses of CanRunManagedCode()
|
|
* Remove concept of AppDomains from the GC
- Leave constructs allowing for multiple handle tables, as scenarios for that have been proposed
- Remove FEATURE_APPDOMAIN_RESOURCE_MONITORING
|
|
The unmanaged runtime has number of different paths for case-insensitive operations, each with unique issues. This is replacing InternalCasingHelper with SString to reduce number of different paths for case-insensitive operations that need to be maintained.
Contributes to #20616
|
|
|
|
The cost of doing this is showing up in the TechEmpower profiles. There is a lot of thread locale state in unmanaged libraries and OS APIs that the threadpool friendly code has to be aware off. It does not make sense to keep reseting just the Locale.
|
|
GCHeapHardLimit - specifies a hard limit for the GC heap
GCHeapHardLimitPercent - specifies a percentage of the physical memory this process is allowed to use
If both are specified, GCHeapHardLimit is checked first and only when it's not specified
would we check GCHeapHardLimitPercent.
If neither is specified but the process is running inside a container with a memory
limit specified, we will take this as the hard limit:
max (20mb, 75% of the memory limit on the container)
If one of the HardLimit configs is specified, and the process is running inside a container
with a memory limit, the GC heap usage will not exceed the HardLimit but the total memory
is still the memory limit on the container so when we calculate the memory load it's based
off the container memory limit.
An example,
process is running inside a container with 200mb limit
user also specified GCHeapHardLimit as 100mb.
if 50mb out of the 100mb is used for GC, and 100mb is used for other things, the memory load
is (50 + 100)/200 = 75%.
Some notes on these configs -
+ The limit is the commit size.
+ This is only supported on 64-bit.
+ For Server GC the minimum *reserved* segment size is 16mb per heap, this is to avoid the
scenario where the hard limit is small but the process can use many procs and we end up
with tiny segments which doesn't make sense. We then keep track of the committed on the segments
so the total does not exceed the hard limit.
|
|
|
|
GCToEEInterface (#19774)
|
|
Delete ENABLE_DOWNLEVEL_FOR_NLS and everything under it
|
|
GetLargestOnDieCacheSize into GetCacheSizePerLogicalCpu (#15975)
* refactor: combine GetLargestOnDieCacheSize and GetLogicalCpuCount in GetCacheSizePerLogicalCpu
* Perform PhysicalMemoryLimit check also for workstation GC
|
|
* Change jit notifications so that they pass the native code address. This fixes !bpmd so that it will set the correct breakpoint on tiered jitted methods.
* code review feedback
* don't handle OnCodeGenerated
|
|
* Support GDBJIT on NI/IL_STUBS
* Move tls_isSymReaderInProgress into gdbjit.cpp
|
|
Look for cache info in /sys/.../cache/index*/size
If that fails estimate cache size based on processor count
|
|
|
|
Fixes #9321 and deletes CleanupToDoList.cs
Delete unmanaged security implementation
|
|
* [x86/Linux] Use CDECL as STDAPICALLTYPE
* Use system default calling convention for internal functions
|
|
* Remove CLRIoCompletionHosted as it always returns false.
* Remove CLRSyncHosted as it always returns false.
* Remove CLRMemoryHosted as it always returns false.
* Remove CLRTaskHosted as it always returns false.
* Remove CLRAssemblyHosted, CLRGCHosted,and CLRSecurityHosted.
They are not called.
* Remove IsThreadPoolHosted as it always returns false.
* Remove EnterRuntime and LeaveRuntime, as they do nothing.
* Add back calls to RevertIfImpersonated and GCX_PREEMP.
I accidentally deleted the call to RevertIfImpersonated instead of just
removing an extra parameter.
When I removed the HR_LEAVE_RUNTIME_HOLDER macro from windowsruntime.h,
I not only removed a LeaveRuntimeHolder but also a GCX_PREEMP. So I added
it back. The holder and GCX_PREEMP where only inserted when the
FEATURE_LEAVE_RUNTIME_HOLDER macro was defined. Since it is always defined,
I removed it. Also as I understand it, you would always want to have a
GCX_PREEMP before calling into the Windows API as not to block the GC,
so I'm not sure why you would want to disable it.
|
|
|
|
This reverts commit 751771a8976f909af772e35c167bd7e3ffbe44c8.
|
|
|
|
|
|
* [x86/Linux] Fix inconsistency in GetCLRFunction definitions
GetCLRFunction is treated as pfnGetCLRFunction_t which has __stdcall
convention, but is implemented without __stdcall.
This inconsistency causes segmentaion fault while initializing CoreCLR
for x86/Linux.
This commit fixes such inconsistency via adding __stdcall to
GetCLRFunction implementation.
In addition, this commit declares GetCLRFuntion in 'utilcode.h' and
and revises .cpp files to include 'utilcode.h' instead of declaring
'GetCLRFunction'.
* Remove unnecessary includes
* Remove another unnecessay include
|
|
* fix permissive C++ code (MSVC /permissive-)
These were found by the C++ compiler group when doing "Real world code"
build tests using /permissive-. We are sharing these with you to help you clean up
your code before the new version of the compiler comes out. For more information on /permissive-
see https://blogs.msdn.microsoft.com/vcblog/2016/11/16/permissive-switch/.
----------------------------
Under /permissive-, skipping the initialization of a variable is not allowed.
As an extension the compiler allowed this when there was no destructor for the type.
void func(bool b)
{
if(b) goto END;
int value = 0; //error C2362: initialization of 'value' is skipped by 'goto END'
int array[10]; //Okay, not initialized.
//... value used here
END:
return;
}
Fix 1) Limit the scope of value:
{
int value = 0;
//... value used here
}
END:
Fix 2) Initialize/declare value before the 'goto'
int value = 0;
if(b) goto END;
//... value used here
END:
Fix 3) Don't initialize value in the variable declaration.
int value;
value = 0
//... value used here
END:
-------------------
Alternative token representations.
The following are reserved as alternative representations for operators:
and and_eq bitand bitor compl not not_eq or or_eq xor xor_eq
//Can't use reserved names for variables:
static int and = 0; // Change name (possibly to 'and_')
void func()
{
_asm {
xor edx,edx // xor is reserved, change to uppercase XOR
or eax,eax // or is reserved, change to uppercase OR
}
}
* Apply formatting patch.
* fixes from code review.
I addressed @janvorli requests from the pull request code review.
|
|
The "JIT flags" currently passed between the EE and the JIT have traditionally
been bit flags in a 32-bit word. Recently, a second 32-bit word was added to
accommodate additional flags, but that set of flags is definitely "2nd class":
they are not universally passed, and require using a separate set of bit
definitions, and comparing those bits against the proper, 2nd word.
This change replaces all uses of bare DWORD or 'unsigned int' types
representing flags with CORJIT_FLAGS, which is now an opaque type. All
flag names were renamed from CORJIT_FLG_* to CORJIT_FLAG_* to ensure all
cases were changed to use the new names, which are also scoped within the
CORJIT_FLAGS type itself.
Another motivation to do this, besides cleaner code, is to allow enabling the
SSE/AVX flags for x86. For x86, we had fewer bits available in the "first
word", so would have to either put them in the "second word", which, as
stated, was very much 2nd class and not plumbed through many usages, or
we could move other bits to the "second word", with the same issues. Neither
was a good option.
RyuJIT compiles with both COR_JIT_EE_VERSION > 460 and <= 460. I introduced
a JitFlags adapter class in jitee.h to handle both JIT flag types. All JIT
code uses this JitFlags type, which operates identically to the new
CORJIT_FLAGS type.
In addition to introducing the new CORJIT_FLAGS type, the SSE/AVX flags are
enabled for x86.
The JIT-EE interface GUID is changed, as this is a breaking change.
|
|
inline assembly (#7018)
Fix inline assembly implementation of getextcpuid() and strict aliasing violations in its callers
The implementation writes to the unsigned char result array, so
the clobber "memory" is needed.
As well its callers have strict aliasing violations from reading the
result array as a DWORD array. The fix is to memcpy the resulting
buffer onto the DWORD array.
|
|
[Linux] GDB JIT support for source line debug information.
|
|
|
|
Instruction emulation for Break Points
Debugger::FuncEvalSetup
NativeWalker
ExceptionHijack
Exception intercept
Managed Return Value
|
|
|
|
[tfs-changeset: 1575640]
|
|
Added the proper contracts to the DAC notification functions.
|
|
Added a SOS DAC interface (ISOSDacInterface4::GetClrNotification) to get the exception
notification arguments instead of using the GetLastExceptionInformation function from the lldb
sosplugin that depends on coreclr symbols being present.
On the coreclr side, the clr notification arguments are saved in a global variable that is DAC
accessible. A critical section was added to protect this global variable while the special
exception is raised.
Setting the internal COMPlus_DebugBreakOnAssert environment variable causes 3 or 4 breaks in
the debugger with no reason. It was breaking in the function that was determining whether it
should break. I was using COMPlus_BreakOnEELoad=2 to break after coreclr was loaded and initialized
to set managed breakpoints and on a debug build it generates an assert (on release just a DebugBreak).
|
|
|
|
[tfs-changeset: 1529946]
|
|
|
|
|
|
c_runtime/vprintf/test1 is disabled as casting NULL to va_list is
against the C specification.
Fix SetFilePointer tests on 32 bit platforms.
Define _FILE_OFFSET_BITS=64 so that we have long file support on 32 bit
platforms.
Implement context capture/restore for ARM.
Link libgcc_s before libunwind on ARM so C++ exceptions work.
Translate armasm to gas syntax.
Specify Thumb, VFPv3, ARMv7 for the ARM target.
Add ARM configuration to mscorlib build
Implement GetLogicalProcessorCacheSizeFromOS in PAL.
Set UNWIND_CONTEXT_IS_UCONTEXT_T from configure check.
|
|
|
|
|
|
[tfs-changeset: 1407945]
|
