reactos/ntoskrnl/kd64/kdprint.c

552 lines
15 KiB
C
Raw Normal View History

/*
* PROJECT: ReactOS Kernel
* LICENSE: GPL - See COPYING in the top level directory
* FILE: ntoskrnl/kd64/kdprint.c
* PURPOSE: KD64 Trap Handler Routines
* PROGRAMMERS: Alex Ionescu (alex.ionescu@reactos.org)
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
* Stefan Ginsberg (stefan.ginsberg@reactos.org)
*/
/* INCLUDES ******************************************************************/
#include <ntoskrnl.h>
#define NDEBUG
#include <debug.h>
#define KD_PRINT_MAX_BYTES 512
/* FUNCTIONS *****************************************************************/
[NTOS:KD64] Implement KdLogDbgPrint() for the WinDbg !dbgprint command. See this command's documentation: https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/-dbgprint and the section "DbgPrint buffer and the debugger" https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/reading-and-filtering-debugging-messages#dbgprint-buffer-and-the-debugger for more details. - Loosely implement the function, based on our existing circular printout buffers in kdio.c. - Enable its usage in the KdpPrint() and KdpPrompt() functions. Notice that this function will *only* capture the strings being sent **to** the debugger, and not the strings the debugger itself produce. (This means that we cannot use the KdPrintCircularBuffer as a replacement for our KDBG dmesg one, for example...) How to test: Run ReactOS under WinDbg, and use the !dbgprint command to view the buffer. You can also use the Memory Window, place yourself at the address pointed by KdPrintCircularBuffer and KdPrintWritePointer, and read its contents. What you should observe: Prior notice: The circular buffer in debug builds of ReactOS and Windows is 0x8000 bytes large. In release builds, its size is down to 0x1000. 1- When you start e.g. the 2nd-stage GUI installation of ReactOS, going past the initial "devices installation" and letting it stabilize on the Welcome page, break into WinDbg and run the !dbgprint command. You should notice that the end of its output is weirdly truncated, compared to what has been actually emitted to the debug output. Comparing this with the actual contents of the circular buffer (via Memory Window), shows that the buffer contents is actually correct. 2- Copy all the text that has been output by the !dbgprint command and paste it in an editor; count the number of all characters appearing + newlines (only CR or LF), and observe that this number is "mysteriously" equal to 16384 == 0x4000. 3- Continue running ReactOS installation for a little while, breaking back back into WinDbg and looking at !dbgprint again. Its output seems to be still stopping at the same place as before (but the actual buffer memory contents shows otherwise). Continue running ROS installation, and break into the debugger when ROS is about to restart. You should now observe that the dbgprint buffer rolled over: dd nt!KdPrintRolloverCount shows 1. Carefully analysing the output of !dbgprint, however, you will notice that it looks a bit garbage-y: the first part of the output is actually truncated after 16384 characters, then you get a second part of the buffer showing what ReactOS was printing while shutting down. Then you get again what was shown at the top of the !dbgprint output. (Of course, comparing with the actual contents of the circular buffer in memory shows that its contents are fine...) The reason of these strange observations, is because there is an intrinsic bug in the !dbgprint command implementation (in kdexts.dll). Essentially, it displays the contents of the circular buffer in two single dprintf() calls: one for the "older" (bottom) part of the buffer: [WritePointer, EndOfBuffer] and one for the "newer" (upper) part of the buffer: [CircularBuffer, WritePointer[ . The first aspect of the bug (causing observation 3), is that those two parts are not necessarily NULL-terminated strings (especially after rollover), so for example, displaying the upper part of the buffer, will potentially also display part of the buffer's bottom part. The second aspect of the bug (explaining observations 1 and 2), is due to the implementation of the dprintf() function (callback in dbgenv.dll). There, it uses a fixed-sized buffer of size 0x4000 == 16384 characters. Since the output of the circular buffer is not done by little chunks, but by the two large parts, if any of those are larger than 0x4000 they get truncated on display. (This last observation is confirmed in a completely different context by https://community.osr.com/discussion/112439/dprintf-s-max-string-length .)
2022-11-23 22:24:59 +00:00
KIRQL
NTAPI
KdpAcquireLock(
_In_ PKSPIN_LOCK SpinLock)
{
KIRQL OldIrql;
/* Acquire the spinlock without waiting at raised IRQL */
while (TRUE)
{
/* Loop until the spinlock becomes available */
while (!KeTestSpinLock(SpinLock));
/* Spinlock is free, raise IRQL to high level */
KeRaiseIrql(HIGH_LEVEL, &OldIrql);
/* Try to get the spinlock */
if (KeTryToAcquireSpinLockAtDpcLevel(SpinLock))
break;
/* Someone else got the spinlock, lower IRQL back */
KeLowerIrql(OldIrql);
}
return OldIrql;
}
VOID
NTAPI
KdpReleaseLock(
_In_ PKSPIN_LOCK SpinLock,
_In_ KIRQL OldIrql)
{
/* Release the spinlock */
KiReleaseSpinLock(SpinLock);
// KeReleaseSpinLockFromDpcLevel(SpinLock);
/* Restore the old IRQL */
KeLowerIrql(OldIrql);
}
VOID
NTAPI
KdLogDbgPrint(
_In_ PSTRING String)
{
SIZE_T Length, Remaining;
KIRQL OldIrql;
/* If the string is empty, bail out */
if (!String->Buffer || (String->Length == 0))
return;
/* If no log buffer available, bail out */
if (!KdPrintCircularBuffer /*|| (KdPrintBufferSize == 0)*/)
return;
/* Acquire the log spinlock without waiting at raised IRQL */
OldIrql = KdpAcquireLock(&KdpPrintSpinLock);
Length = min(String->Length, KdPrintBufferSize);
Remaining = KdPrintCircularBuffer + KdPrintBufferSize - KdPrintWritePointer;
if (Length < Remaining)
{
KdpMoveMemory(KdPrintWritePointer, String->Buffer, Length);
KdPrintWritePointer += Length;
}
else
{
KdpMoveMemory(KdPrintWritePointer, String->Buffer, Remaining);
Length -= Remaining;
if (Length > 0)
KdpMoveMemory(KdPrintCircularBuffer, String->Buffer + Remaining, Length);
KdPrintWritePointer = KdPrintCircularBuffer + Length;
/* Got a rollover, update count (handle wrapping, must always be >= 1) */
++KdPrintRolloverCount;
if (KdPrintRolloverCount == 0)
++KdPrintRolloverCount;
}
/* Release the spinlock */
KdpReleaseLock(&KdpPrintSpinLock, OldIrql);
}
BOOLEAN
NTAPI
KdpPrintString(
_In_ PSTRING Output)
{
STRING Data, Header;
DBGKD_DEBUG_IO DebugIo;
USHORT Length;
/* Copy the string */
KdpMoveMemory(KdpMessageBuffer,
Output->Buffer,
Output->Length);
/* Make sure we don't exceed the KD Packet size */
Length = Output->Length;
if ((sizeof(DBGKD_DEBUG_IO) + Length) > PACKET_MAX_SIZE)
{
/* Normalize length */
Length = PACKET_MAX_SIZE - sizeof(DBGKD_DEBUG_IO);
}
/* Build the packet header */
DebugIo.ApiNumber = DbgKdPrintStringApi;
DebugIo.ProcessorLevel = (USHORT)KeProcessorLevel;
DebugIo.Processor = KeGetCurrentPrcb()->Number;
DebugIo.u.PrintString.LengthOfString = Length;
Header.Length = sizeof(DBGKD_DEBUG_IO);
Header.Buffer = (PCHAR)&DebugIo;
/* Build the data */
Data.Length = Length;
Data.Buffer = KdpMessageBuffer;
/* Send the packet */
KdSendPacket(PACKET_TYPE_KD_DEBUG_IO, &Header, &Data, &KdpContext);
/* Check if the user pressed CTRL+C */
return KdpPollBreakInWithPortLock();
}
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
BOOLEAN
NTAPI
KdpPromptString(
_In_ PSTRING PromptString,
_In_ PSTRING ResponseString)
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
{
STRING Data, Header;
DBGKD_DEBUG_IO DebugIo;
ULONG Length;
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
KDSTATUS Status;
/* Copy the string to the message buffer */
KdpMoveMemory(KdpMessageBuffer,
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
PromptString->Buffer,
PromptString->Length);
/* Make sure we don't exceed the KD Packet size */
Length = PromptString->Length;
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
if ((sizeof(DBGKD_DEBUG_IO) + Length) > PACKET_MAX_SIZE)
{
/* Normalize length */
Length = PACKET_MAX_SIZE - sizeof(DBGKD_DEBUG_IO);
}
/* Build the packet header */
DebugIo.ApiNumber = DbgKdGetStringApi;
DebugIo.ProcessorLevel = (USHORT)KeProcessorLevel;
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
DebugIo.Processor = KeGetCurrentPrcb()->Number;
DebugIo.u.GetString.LengthOfPromptString = Length;
DebugIo.u.GetString.LengthOfStringRead = ResponseString->MaximumLength;
Header.Length = sizeof(DBGKD_DEBUG_IO);
Header.Buffer = (PCHAR)&DebugIo;
/* Build the data */
Data.Length = Length;
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
Data.Buffer = KdpMessageBuffer;
/* Send the packet */
KdSendPacket(PACKET_TYPE_KD_DEBUG_IO, &Header, &Data, &KdpContext);
/* Set the maximum lengths for the receive */
Header.MaximumLength = sizeof(DBGKD_DEBUG_IO);
Data.MaximumLength = sizeof(KdpMessageBuffer);
/* Enter receive loop */
do
{
/* Get our reply */
Status = KdReceivePacket(PACKET_TYPE_KD_DEBUG_IO,
&Header,
&Data,
&Length,
&KdpContext);
/* Return TRUE if we need to resend */
if (Status == KdPacketNeedsResend) return TRUE;
/* Loop until we succeed */
} while (Status != KdPacketReceived);
/* Don't copy back a larger response than there is room for */
Length = min(Length,
ResponseString->MaximumLength);
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Copy back the string and return the length */
KdpMoveMemory(ResponseString->Buffer,
KdpMessageBuffer,
Length);
ResponseString->Length = (USHORT)Length;
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Success; we don't need to resend */
return FALSE;
}
VOID
NTAPI
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
KdpCommandString(IN PSTRING NameString,
IN PSTRING CommandString,
IN KPROCESSOR_MODE PreviousMode,
IN PCONTEXT ContextRecord,
IN PKTRAP_FRAME TrapFrame,
IN PKEXCEPTION_FRAME ExceptionFrame)
{
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
BOOLEAN Enable;
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
PKPRCB Prcb = KeGetCurrentPrcb();
/* Check if we need to do anything */
if ((PreviousMode != KernelMode) || (KdDebuggerNotPresent)) return;
/* Enter the debugger */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
Enable = KdEnterDebugger(TrapFrame, ExceptionFrame);
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
/* Save the CPU Control State and save the context */
KiSaveProcessorControlState(&Prcb->ProcessorState);
KdpMoveMemory(&Prcb->ProcessorState.ContextFrame,
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
ContextRecord,
sizeof(CONTEXT));
/* Send the command string to the debugger */
KdpReportCommandStringStateChange(NameString,
CommandString,
&Prcb->ProcessorState.ContextFrame);
/* Restore the processor state */
KdpMoveMemory(ContextRecord,
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
&Prcb->ProcessorState.ContextFrame,
sizeof(CONTEXT));
KiRestoreProcessorControlState(&Prcb->ProcessorState);
/* Exit the debugger and return */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
KdExitDebugger(Enable);
}
VOID
NTAPI
KdpSymbol(IN PSTRING DllPath,
- DBGKD_WAIT_STATE_CHANGE64 is used in KD protocol 5, not number 6 that we use. Protocol 6 uses the DBGKD_ANY_WAIT_STATE_CHANGE structure which is sized according to the largest control-report structure (AMD64_DBGKD_CONTROL_REPORT currently), and is larger than DBGKD_WAIT_STATE_CHANGE64 on x86. This worked because our DBGKD_WAIT_STATE_CHANGE32/64 structures contained incorrect DBGKD_CONTROL_REPORT (used) and CONTEXT (unused) members that sized up the wait-state structure to pass WinDbg's length verification! It actually becomes larger than DBGKD_ANY_WAIT_STATE_CHANGE, but WinDbg only seems bail out only if the structure is too small. Remove the incorrect members from the protocol 5 structures and change to DBGKD_ANY_WAIT_STATE_CHANGE everywhere. - Correct the value of SIZE_OF_FX_REGISTERS -- it was 4 times too low which resulted in KeContextToTrapFrame not properly clearing out the XMM register area. Correct the define and move it out from ke.h to x86's ketypes.h and use it in the FXSAVE format structure. Also remove the IOPM definitions from ke.h as they have been in the NDK for a while. - KD uses STRINGs, not ANSI_STRINGs -- they are the same thing, but let's be consistent. - ExceptionRecord32To64 should be available for both 32 and 64 bit builds (and it shouldn't be a forceinline). Get rid of CopyExceptionRecord and determine if we need to convert or can just copy it directly instead. - Use _WIN64 instead of _M_AMD64 when determining if we need to set the DBGKD_VERS_FLAG_PTR64 flag. - Don't check Nt/DbgQueryDebugFilterState for zero or nonzero -- it actually returns TRUE, FALSE or STATUS_INVALID_PARAMETER_1! Check for != TRUE in preparation for proper implementation of NtSet/QueryDebugFilterState. - Fix Format parameter of DbgPrintReturnControlC -- it is const like the other DbgPrint* routines. - Be consistent with the types used in debug.c and don't set local variables to zero if we are going to return to caller -- this doesn't seem to be required anymore. - Fix DebugService and DebugService2: DebugService should take a ULONG followed by 4 pointers and DebugService2 doesn't return anything. - Use ZwCurrentProcess() instead of -1 or 0xFFFFFFFF (which is incorrect for 64-bit) for the ProcessId parameter of DbgLoad/UnloadImageSymbols to clarify what is being passed. Don't use ZwCurrentProcess() in KeBugCheckWithTf for the pointer parameter of DbgUnLoadImageSymbols either. Use MAXULONG_PTR casted to PVOID instead. - Use better named and sized variables in KdpTrap for setting the "return register" in the caller's CONTEXT. - Correct and clarify the comment documenting under what conditions we pass user mode exceptions to the kernel debugger. svn path=/trunk/; revision=43741
2009-10-25 15:56:38 +00:00
IN PKD_SYMBOLS_INFO SymbolInfo,
IN BOOLEAN Unload,
IN KPROCESSOR_MODE PreviousMode,
IN PCONTEXT ContextRecord,
IN PKTRAP_FRAME TrapFrame,
IN PKEXCEPTION_FRAME ExceptionFrame)
{
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
BOOLEAN Enable;
PKPRCB Prcb = KeGetCurrentPrcb();
/* Check if we need to do anything */
if ((PreviousMode != KernelMode) || (KdDebuggerNotPresent)) return;
/* Enter the debugger */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
Enable = KdEnterDebugger(TrapFrame, ExceptionFrame);
/* Save the CPU Control State and save the context */
KiSaveProcessorControlState(&Prcb->ProcessorState);
KdpMoveMemory(&Prcb->ProcessorState.ContextFrame,
ContextRecord,
sizeof(CONTEXT));
/* Report the new state */
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
KdpReportLoadSymbolsStateChange(DllPath,
SymbolInfo,
Unload,
&Prcb->ProcessorState.ContextFrame);
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
/* Restore the processor state */
KdpMoveMemory(ContextRecord,
&Prcb->ProcessorState.ContextFrame,
sizeof(CONTEXT));
KiRestoreProcessorControlState(&Prcb->ProcessorState);
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
/* Exit the debugger and return */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
KdExitDebugger(Enable);
}
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
USHORT
NTAPI
KdpPrompt(
_In_reads_bytes_(PromptLength) PCHAR PromptString,
_In_ USHORT PromptLength,
_Out_writes_bytes_(MaximumResponseLength) PCHAR ResponseString,
_In_ USHORT MaximumResponseLength,
_In_ KPROCESSOR_MODE PreviousMode,
_In_ PKTRAP_FRAME TrapFrame,
_In_ PKEXCEPTION_FRAME ExceptionFrame)
{
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
STRING PromptBuffer, ResponseBuffer;
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
BOOLEAN Enable, Resend;
PCHAR SafeResponseString;
CHAR CapturedPrompt[KD_PRINT_MAX_BYTES];
CHAR SafeResponseBuffer[KD_PRINT_MAX_BYTES];
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Normalize the lengths */
PromptLength = min(PromptLength,
sizeof(CapturedPrompt));
MaximumResponseLength = min(MaximumResponseLength,
sizeof(SafeResponseBuffer));
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Check if we need to verify the string */
if (PreviousMode != KernelMode)
{
/* Handle user-mode buffers safely */
_SEH2_TRY
{
/* Probe and capture the prompt */
ProbeForRead(PromptString, PromptLength, 1);
KdpMoveMemory(CapturedPrompt, PromptString, PromptLength);
PromptString = CapturedPrompt;
/* Probe and make room for the response */
ProbeForWrite(ResponseString, MaximumResponseLength, 1);
SafeResponseString = SafeResponseBuffer;
}
_SEH2_EXCEPT(EXCEPTION_EXECUTE_HANDLER)
{
/* Bad string pointer, bail out */
_SEH2_YIELD(return 0);
}
_SEH2_END;
}
else
{
SafeResponseString = ResponseString;
}
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Setup the prompt and response buffers */
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
PromptBuffer.Buffer = PromptString;
[NTOS:KD/KD64/KDBG] Share some code between our legacy KD/KDBG and KD64. Our legacy KD module is slowly being phased out for the more recent KD64 Kernel Debugger that supports WinDbg, but at the same time we must retain support for GCC debugging and the KDBG interface. For the time being few #ifdef _WINKD_ have been introduced in KD64 so that some of its code/data does not completely get shared yet with the legacy KD, until the latter becomes phased out. KD Modifications: ================= - Remove the implementation of NtQueryDebugFilterState() / NtSetDebugFilterState() that now comes entirely from KD64. - Remove KD variables that are now shared with KD64. - Share common code with KD64: KdpMoveMemory(), KdpZeroMemory(), KdpCopyMemoryChunks(), KdpPrint(), KdpPrompt(). - KDBG: Remove the duplicated KdpCopyMemoryChunks() function. - In KdpServiceDispatcher() and KdpEnterDebuggerException(), call the KdpPrint() worker function that correctly probes and captures its arguments. - Temporarily stub out KdEnterDebugger() and KdExitDebugger() that is used by the shared code, until KD is removed and only the KD64 version of these functions remain. - Re-implement the KD/KDBG KdpPrompt() function using a custom KdpPromptString() helper compatible with KD64, that is called by the KD64 implementation of KdpPrompt(). This KdpPromptString() helper now issues the prompt on all the KD loggers: e.g. if you use both at the same time COM-port and SCREEN debugging, the prompt will appear on both. Before that the prompt was always being displayed on COM port even if e.g. a SCREEN-only debug session was used... - ppc_irq.c: Fix the prototype of KdpServiceDispatcher(). KD64 Fixes: =========== - Initialize the MaximumLength member of the counted STRING variables before using them elsewhere. - Get rid of alloca() within SEH block in KdpPrint() (addendum to 7b95fcf9). - Add the ROS-specific handy dump commands in KdSystemDebugControl().
2019-11-17 21:55:36 +00:00
PromptBuffer.Length = PromptBuffer.MaximumLength = PromptLength;
ResponseBuffer.Buffer = SafeResponseString;
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
ResponseBuffer.Length = 0;
ResponseBuffer.MaximumLength = MaximumResponseLength;
/* Log the print */
[NTOS:KD64] Implement KdLogDbgPrint() for the WinDbg !dbgprint command. See this command's documentation: https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/-dbgprint and the section "DbgPrint buffer and the debugger" https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/reading-and-filtering-debugging-messages#dbgprint-buffer-and-the-debugger for more details. - Loosely implement the function, based on our existing circular printout buffers in kdio.c. - Enable its usage in the KdpPrint() and KdpPrompt() functions. Notice that this function will *only* capture the strings being sent **to** the debugger, and not the strings the debugger itself produce. (This means that we cannot use the KdPrintCircularBuffer as a replacement for our KDBG dmesg one, for example...) How to test: Run ReactOS under WinDbg, and use the !dbgprint command to view the buffer. You can also use the Memory Window, place yourself at the address pointed by KdPrintCircularBuffer and KdPrintWritePointer, and read its contents. What you should observe: Prior notice: The circular buffer in debug builds of ReactOS and Windows is 0x8000 bytes large. In release builds, its size is down to 0x1000. 1- When you start e.g. the 2nd-stage GUI installation of ReactOS, going past the initial "devices installation" and letting it stabilize on the Welcome page, break into WinDbg and run the !dbgprint command. You should notice that the end of its output is weirdly truncated, compared to what has been actually emitted to the debug output. Comparing this with the actual contents of the circular buffer (via Memory Window), shows that the buffer contents is actually correct. 2- Copy all the text that has been output by the !dbgprint command and paste it in an editor; count the number of all characters appearing + newlines (only CR or LF), and observe that this number is "mysteriously" equal to 16384 == 0x4000. 3- Continue running ReactOS installation for a little while, breaking back back into WinDbg and looking at !dbgprint again. Its output seems to be still stopping at the same place as before (but the actual buffer memory contents shows otherwise). Continue running ROS installation, and break into the debugger when ROS is about to restart. You should now observe that the dbgprint buffer rolled over: dd nt!KdPrintRolloverCount shows 1. Carefully analysing the output of !dbgprint, however, you will notice that it looks a bit garbage-y: the first part of the output is actually truncated after 16384 characters, then you get a second part of the buffer showing what ReactOS was printing while shutting down. Then you get again what was shown at the top of the !dbgprint output. (Of course, comparing with the actual contents of the circular buffer in memory shows that its contents are fine...) The reason of these strange observations, is because there is an intrinsic bug in the !dbgprint command implementation (in kdexts.dll). Essentially, it displays the contents of the circular buffer in two single dprintf() calls: one for the "older" (bottom) part of the buffer: [WritePointer, EndOfBuffer] and one for the "newer" (upper) part of the buffer: [CircularBuffer, WritePointer[ . The first aspect of the bug (causing observation 3), is that those two parts are not necessarily NULL-terminated strings (especially after rollover), so for example, displaying the upper part of the buffer, will potentially also display part of the buffer's bottom part. The second aspect of the bug (explaining observations 1 and 2), is due to the implementation of the dprintf() function (callback in dbgenv.dll). There, it uses a fixed-sized buffer of size 0x4000 == 16384 characters. Since the output of the circular buffer is not done by little chunks, but by the two large parts, if any of those are larger than 0x4000 they get truncated on display. (This last observation is confirmed in a completely different context by https://community.osr.com/discussion/112439/dprintf-s-max-string-length .)
2022-11-23 22:24:59 +00:00
KdLogDbgPrint(&PromptBuffer);
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Enter the debugger */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
Enable = KdEnterDebugger(TrapFrame, ExceptionFrame);
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Enter prompt loop */
do
{
/* Send the prompt and receive the response */
Resend = KdpPromptString(&PromptBuffer, &ResponseBuffer);
/* Loop while we need to resend */
} while (Resend);
/* Exit the debugger */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
KdExitDebugger(Enable);
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Copy back the response if required */
if (PreviousMode != KernelMode)
{
_SEH2_TRY
{
/* Safely copy back the response to user mode */
KdpMoveMemory(ResponseString,
ResponseBuffer.Buffer,
ResponseBuffer.Length);
}
_SEH2_EXCEPT(EXCEPTION_EXECUTE_HANDLER)
{
/* String became invalid after we exited, fail */
_SEH2_YIELD(return 0);
}
_SEH2_END;
}
- Replace RtlpGetExceptionAddress by the _ReturnAddress intrinsic and add it to ARM intrin.h as it was missing. - Simplify RtlpCheckForActiveDebugger: Remove the BOOLEAN parameter as we would always pass it FALSE. Always return FALSE false from kernel mode for simplicity. - Fix a critical flaw in our exception support: RtlRaiseException and RtlRaiseStatus were implemented in C on x86. This lead to unpredictable register corruption because the compiler could not know that it had to preserve non-volatile registers before calling RtlCaptureContext as the saved context is later used to restore the caller in case the exception is handled and execution is continued. This made the functions unsafe to return from as any non-volatile register could be corrupted. Implement them in assembly for x86 to safely capture the context using only EBP and ESP. The C versions of those routines are still used and shared for the other architectures we support -- needs to be determined if this is safe and correct for those architectures. - The ntdll exception Wine exposed this issue, and all tests now pass. The remaining failures on the build server are caused by missing or incomplete debug register support in KVM/QEMU. Run the test in another VM or on real hardware and all the tests will pass. - Implement Debug Prompt (DbgPrompt) support for KD and KDBG. The KDBG implementation reads the prompt from keyboard or serial depending on the mode so that sysreg and rosdbg can support it too. - Properly implement RtlAssert using DbgPrompt to prompt for the action to take instead of always doing a breakpoint. The new implementation is disabled until sysreg can support this. Also move RtlAssert to its own file as it has nothing to do with the error routines (nor does it belong in exception.c). - Note that DbgPrompt was already used in PspCatchCriticalBreak, and this would have resulted in a silent hang as BREAKPOINT_PROMPT wasn't handled at all by KDBG. - Implement KiRaiseAssertion (10 lines of code with the trap macros) and thus support NT_ASSERT. Add partial support for it to KDBG to print out a warning and the address of the failure, but don't do anything else. Also add NT_ASSERT to the DDK headers so that we can use it, but don't use it yet as the ARM method of performing this has not been decided nor implemented. - KiTrap3 doesn't set STATUS_SUCCESS but BREAKPOINT_BREAK. They have the same numerical value but very different meaning -- BREAKPOINT_BREAK means that the exception is a software breakpoint and not a debug service call. Fix some comments to document that this is what is checked for. - Fix inverted and broken logic in KdpReport. It would never pass second chance exceptions to the debugger, didn't respect the stop-on-exception flag properly and would always fail to handle some special exceptions in both first and second chance instead of just failing to handle it in first chance. Clean up, reformat and document what is going on. - The DebugPrint and DebugPrompt support routines only perform a 2D interrupt on x86; use more portable comments. - Add Alex to the programmer section of x86's kdsup.c -- he wrote KdpGetStateChange, KdpSetContextState and the code that was previously in KdpRead/WriteControlSpace. - Add my name to the parts of KD where I have made significant work on getting KD/WinDbg support up and running. - KD debugging is now quite functional and stable. Some bugs and stubs remain to be flushed out, but overall KD is now much better and easier to use than KDBG. svn path=/trunk/; revision=43705
2009-10-23 22:51:39 +00:00
/* Return the number of characters received */
return ResponseBuffer.Length;
}
static
NTSTATUS
NTAPI
KdpPrintFromUser(
_In_ ULONG ComponentId,
_In_ ULONG Level,
_In_reads_bytes_(Length) PCHAR String,
_In_ USHORT Length,
_In_ KPROCESSOR_MODE PreviousMode,
_In_ PKTRAP_FRAME TrapFrame,
_In_ PKEXCEPTION_FRAME ExceptionFrame,
_Out_ PBOOLEAN Handled)
{
CHAR CapturedString[KD_PRINT_MAX_BYTES];
ASSERT(PreviousMode == UserMode);
ASSERT(Length <= sizeof(CapturedString));
/* Capture user-mode buffers */
_SEH2_TRY
{
/* Probe and capture the string */
ProbeForRead(String, Length, 1);
KdpMoveMemory(CapturedString, String, Length);
String = CapturedString;
}
_SEH2_EXCEPT(EXCEPTION_EXECUTE_HANDLER)
{
/* Bad string pointer, bail out */
_SEH2_YIELD(return STATUS_ACCESS_VIOLATION);
}
_SEH2_END;
/* Now go through the kernel-mode code path */
return KdpPrint(ComponentId,
Level,
String,
Length,
KernelMode,
TrapFrame,
ExceptionFrame,
Handled);
}
NTSTATUS
NTAPI
KdpPrint(
_In_ ULONG ComponentId,
_In_ ULONG Level,
_In_reads_bytes_(Length) PCHAR String,
_In_ USHORT Length,
_In_ KPROCESSOR_MODE PreviousMode,
_In_ PKTRAP_FRAME TrapFrame,
_In_ PKEXCEPTION_FRAME ExceptionFrame,
_Out_ PBOOLEAN Handled)
{
NTSTATUS Status;
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
BOOLEAN Enable;
- DBGKD_WAIT_STATE_CHANGE64 is used in KD protocol 5, not number 6 that we use. Protocol 6 uses the DBGKD_ANY_WAIT_STATE_CHANGE structure which is sized according to the largest control-report structure (AMD64_DBGKD_CONTROL_REPORT currently), and is larger than DBGKD_WAIT_STATE_CHANGE64 on x86. This worked because our DBGKD_WAIT_STATE_CHANGE32/64 structures contained incorrect DBGKD_CONTROL_REPORT (used) and CONTEXT (unused) members that sized up the wait-state structure to pass WinDbg's length verification! It actually becomes larger than DBGKD_ANY_WAIT_STATE_CHANGE, but WinDbg only seems bail out only if the structure is too small. Remove the incorrect members from the protocol 5 structures and change to DBGKD_ANY_WAIT_STATE_CHANGE everywhere. - Correct the value of SIZE_OF_FX_REGISTERS -- it was 4 times too low which resulted in KeContextToTrapFrame not properly clearing out the XMM register area. Correct the define and move it out from ke.h to x86's ketypes.h and use it in the FXSAVE format structure. Also remove the IOPM definitions from ke.h as they have been in the NDK for a while. - KD uses STRINGs, not ANSI_STRINGs -- they are the same thing, but let's be consistent. - ExceptionRecord32To64 should be available for both 32 and 64 bit builds (and it shouldn't be a forceinline). Get rid of CopyExceptionRecord and determine if we need to convert or can just copy it directly instead. - Use _WIN64 instead of _M_AMD64 when determining if we need to set the DBGKD_VERS_FLAG_PTR64 flag. - Don't check Nt/DbgQueryDebugFilterState for zero or nonzero -- it actually returns TRUE, FALSE or STATUS_INVALID_PARAMETER_1! Check for != TRUE in preparation for proper implementation of NtSet/QueryDebugFilterState. - Fix Format parameter of DbgPrintReturnControlC -- it is const like the other DbgPrint* routines. - Be consistent with the types used in debug.c and don't set local variables to zero if we are going to return to caller -- this doesn't seem to be required anymore. - Fix DebugService and DebugService2: DebugService should take a ULONG followed by 4 pointers and DebugService2 doesn't return anything. - Use ZwCurrentProcess() instead of -1 or 0xFFFFFFFF (which is incorrect for 64-bit) for the ProcessId parameter of DbgLoad/UnloadImageSymbols to clarify what is being passed. Don't use ZwCurrentProcess() in KeBugCheckWithTf for the pointer parameter of DbgUnLoadImageSymbols either. Use MAXULONG_PTR casted to PVOID instead. - Use better named and sized variables in KdpTrap for setting the "return register" in the caller's CONTEXT. - Correct and clarify the comment documenting under what conditions we pass user mode exceptions to the kernel debugger. svn path=/trunk/; revision=43741
2009-10-25 15:56:38 +00:00
STRING OutputString;
if (NtQueryDebugFilterState(ComponentId, Level) == (NTSTATUS)FALSE)
{
/* Mask validation failed */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
*Handled = TRUE;
- Fix support for /CRASHDEBUG and /NODEBUG; we didn't respect those settings properly and would initialize KD at boot even if they were set. - Re-enable the breakpoint in vDbgPrintExWithPrefixInternal() as this works properly now. Without this breakpoint some break-in requests got lost if the break-in occurred when handling a debug print (happened a lot at boot). - Implement Command String support for DbgCommandString() -- we now handle every debug service call. - Implement NtSetDebugFilterState() and NtQueryDebugFilterState() for KD, meaning we now support debug filters properly. - Implement KdRefreshDebuggerNotPresent(), KdChangeOption() and KdPowerTransition(). Stub KdSystemDebugControl() to return error status instead of hanging the system. - Stub the rest of the KD API to print a warning and return a failure packet instead of hanging. - Set and respect KdpContextSent when getting and setting the thread context -- WinDbg doesn't seem to rely on this, but better safe than sorry. - Support MP when getting and setting the thread context too -- if the context is operation is for another processor than the current, just get it through the KiProcessorBlock array. - Initialize the MajorVersion in the KD version block more properly -- the high byte is the major identifier (0 for NT). Add the required DBGKD_MAJOR_TYPES enumeration to wdbgexts.h. - Simplify setting and clearing the InDbgPrint flag in the TEB to minimize the impact on kernel execution; use 2 dedicated routines instead of a generic one. - KdpSymbol doesn't return anything, so don't return an ignore status from KdpReportLoadSymbolsStateChange. - Expose the KdpDefaultRetries and Kd_WIN2000_Mask variables to the registry and add them to KDBG too (unused there). - No reason to implement KdpSysGetVersion per architecture; move it back to the generic code. - Add some ARM offsets to the debugger data block that (N/A on other architectures). - Fix the default size of the DbgPrint log buffer for free builds to save some space. It should be 4 KB for a free build and 32 KB for a checked build. - Move KeDisableInterrupts to cpu.c as it fits here more than in the IRQ support code in irqobj.c. - Use KeDisableInterrupts in KeFreezeExecution instead of checking the x86 EFLAG directly. svn path=/trunk/; revision=43912
2009-11-02 17:45:51 +00:00
return STATUS_SUCCESS;
}
/* Assume failure */
*Handled = FALSE;
/* Normalize the length */
Length = min(Length, KD_PRINT_MAX_BYTES);
/* Check if we need to verify the string */
if (PreviousMode != KernelMode)
{
/* This case requires a 512 byte stack buffer.
* We don't want to use that much stack in the kernel case, but we
* can't use _alloca due to PSEH. So the buffer exists in this
* helper function instead.
*/
return KdpPrintFromUser(ComponentId,
Level,
String,
Length,
PreviousMode,
TrapFrame,
ExceptionFrame,
Handled);
}
- DBGKD_WAIT_STATE_CHANGE64 is used in KD protocol 5, not number 6 that we use. Protocol 6 uses the DBGKD_ANY_WAIT_STATE_CHANGE structure which is sized according to the largest control-report structure (AMD64_DBGKD_CONTROL_REPORT currently), and is larger than DBGKD_WAIT_STATE_CHANGE64 on x86. This worked because our DBGKD_WAIT_STATE_CHANGE32/64 structures contained incorrect DBGKD_CONTROL_REPORT (used) and CONTEXT (unused) members that sized up the wait-state structure to pass WinDbg's length verification! It actually becomes larger than DBGKD_ANY_WAIT_STATE_CHANGE, but WinDbg only seems bail out only if the structure is too small. Remove the incorrect members from the protocol 5 structures and change to DBGKD_ANY_WAIT_STATE_CHANGE everywhere. - Correct the value of SIZE_OF_FX_REGISTERS -- it was 4 times too low which resulted in KeContextToTrapFrame not properly clearing out the XMM register area. Correct the define and move it out from ke.h to x86's ketypes.h and use it in the FXSAVE format structure. Also remove the IOPM definitions from ke.h as they have been in the NDK for a while. - KD uses STRINGs, not ANSI_STRINGs -- they are the same thing, but let's be consistent. - ExceptionRecord32To64 should be available for both 32 and 64 bit builds (and it shouldn't be a forceinline). Get rid of CopyExceptionRecord and determine if we need to convert or can just copy it directly instead. - Use _WIN64 instead of _M_AMD64 when determining if we need to set the DBGKD_VERS_FLAG_PTR64 flag. - Don't check Nt/DbgQueryDebugFilterState for zero or nonzero -- it actually returns TRUE, FALSE or STATUS_INVALID_PARAMETER_1! Check for != TRUE in preparation for proper implementation of NtSet/QueryDebugFilterState. - Fix Format parameter of DbgPrintReturnControlC -- it is const like the other DbgPrint* routines. - Be consistent with the types used in debug.c and don't set local variables to zero if we are going to return to caller -- this doesn't seem to be required anymore. - Fix DebugService and DebugService2: DebugService should take a ULONG followed by 4 pointers and DebugService2 doesn't return anything. - Use ZwCurrentProcess() instead of -1 or 0xFFFFFFFF (which is incorrect for 64-bit) for the ProcessId parameter of DbgLoad/UnloadImageSymbols to clarify what is being passed. Don't use ZwCurrentProcess() in KeBugCheckWithTf for the pointer parameter of DbgUnLoadImageSymbols either. Use MAXULONG_PTR casted to PVOID instead. - Use better named and sized variables in KdpTrap for setting the "return register" in the caller's CONTEXT. - Correct and clarify the comment documenting under what conditions we pass user mode exceptions to the kernel debugger. svn path=/trunk/; revision=43741
2009-10-25 15:56:38 +00:00
/* Setup the output string */
OutputString.Buffer = String;
[NTOS:KD/KD64/KDBG] Share some code between our legacy KD/KDBG and KD64. Our legacy KD module is slowly being phased out for the more recent KD64 Kernel Debugger that supports WinDbg, but at the same time we must retain support for GCC debugging and the KDBG interface. For the time being few #ifdef _WINKD_ have been introduced in KD64 so that some of its code/data does not completely get shared yet with the legacy KD, until the latter becomes phased out. KD Modifications: ================= - Remove the implementation of NtQueryDebugFilterState() / NtSetDebugFilterState() that now comes entirely from KD64. - Remove KD variables that are now shared with KD64. - Share common code with KD64: KdpMoveMemory(), KdpZeroMemory(), KdpCopyMemoryChunks(), KdpPrint(), KdpPrompt(). - KDBG: Remove the duplicated KdpCopyMemoryChunks() function. - In KdpServiceDispatcher() and KdpEnterDebuggerException(), call the KdpPrint() worker function that correctly probes and captures its arguments. - Temporarily stub out KdEnterDebugger() and KdExitDebugger() that is used by the shared code, until KD is removed and only the KD64 version of these functions remain. - Re-implement the KD/KDBG KdpPrompt() function using a custom KdpPromptString() helper compatible with KD64, that is called by the KD64 implementation of KdpPrompt(). This KdpPromptString() helper now issues the prompt on all the KD loggers: e.g. if you use both at the same time COM-port and SCREEN debugging, the prompt will appear on both. Before that the prompt was always being displayed on COM port even if e.g. a SCREEN-only debug session was used... - ppc_irq.c: Fix the prototype of KdpServiceDispatcher(). KD64 Fixes: =========== - Initialize the MaximumLength member of the counted STRING variables before using them elsewhere. - Get rid of alloca() within SEH block in KdpPrint() (addendum to 7b95fcf9). - Add the ROS-specific handy dump commands in KdSystemDebugControl().
2019-11-17 21:55:36 +00:00
OutputString.Length = OutputString.MaximumLength = Length;
/* Log the print */
[NTOS:KD64] Implement KdLogDbgPrint() for the WinDbg !dbgprint command. See this command's documentation: https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/-dbgprint and the section "DbgPrint buffer and the debugger" https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/reading-and-filtering-debugging-messages#dbgprint-buffer-and-the-debugger for more details. - Loosely implement the function, based on our existing circular printout buffers in kdio.c. - Enable its usage in the KdpPrint() and KdpPrompt() functions. Notice that this function will *only* capture the strings being sent **to** the debugger, and not the strings the debugger itself produce. (This means that we cannot use the KdPrintCircularBuffer as a replacement for our KDBG dmesg one, for example...) How to test: Run ReactOS under WinDbg, and use the !dbgprint command to view the buffer. You can also use the Memory Window, place yourself at the address pointed by KdPrintCircularBuffer and KdPrintWritePointer, and read its contents. What you should observe: Prior notice: The circular buffer in debug builds of ReactOS and Windows is 0x8000 bytes large. In release builds, its size is down to 0x1000. 1- When you start e.g. the 2nd-stage GUI installation of ReactOS, going past the initial "devices installation" and letting it stabilize on the Welcome page, break into WinDbg and run the !dbgprint command. You should notice that the end of its output is weirdly truncated, compared to what has been actually emitted to the debug output. Comparing this with the actual contents of the circular buffer (via Memory Window), shows that the buffer contents is actually correct. 2- Copy all the text that has been output by the !dbgprint command and paste it in an editor; count the number of all characters appearing + newlines (only CR or LF), and observe that this number is "mysteriously" equal to 16384 == 0x4000. 3- Continue running ReactOS installation for a little while, breaking back back into WinDbg and looking at !dbgprint again. Its output seems to be still stopping at the same place as before (but the actual buffer memory contents shows otherwise). Continue running ROS installation, and break into the debugger when ROS is about to restart. You should now observe that the dbgprint buffer rolled over: dd nt!KdPrintRolloverCount shows 1. Carefully analysing the output of !dbgprint, however, you will notice that it looks a bit garbage-y: the first part of the output is actually truncated after 16384 characters, then you get a second part of the buffer showing what ReactOS was printing while shutting down. Then you get again what was shown at the top of the !dbgprint output. (Of course, comparing with the actual contents of the circular buffer in memory shows that its contents are fine...) The reason of these strange observations, is because there is an intrinsic bug in the !dbgprint command implementation (in kdexts.dll). Essentially, it displays the contents of the circular buffer in two single dprintf() calls: one for the "older" (bottom) part of the buffer: [WritePointer, EndOfBuffer] and one for the "newer" (upper) part of the buffer: [CircularBuffer, WritePointer[ . The first aspect of the bug (causing observation 3), is that those two parts are not necessarily NULL-terminated strings (especially after rollover), so for example, displaying the upper part of the buffer, will potentially also display part of the buffer's bottom part. The second aspect of the bug (explaining observations 1 and 2), is due to the implementation of the dprintf() function (callback in dbgenv.dll). There, it uses a fixed-sized buffer of size 0x4000 == 16384 characters. Since the output of the circular buffer is not done by little chunks, but by the two large parts, if any of those are larger than 0x4000 they get truncated on display. (This last observation is confirmed in a completely different context by https://community.osr.com/discussion/112439/dprintf-s-max-string-length .)
2022-11-23 22:24:59 +00:00
KdLogDbgPrint(&OutputString);
/* Check for a debugger */
if (KdDebuggerNotPresent)
{
/* Fail */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
*Handled = TRUE;
return STATUS_DEVICE_NOT_CONNECTED;
}
/* Enter the debugger */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
Enable = KdEnterDebugger(TrapFrame, ExceptionFrame);
/* Print the string */
- DBGKD_WAIT_STATE_CHANGE64 is used in KD protocol 5, not number 6 that we use. Protocol 6 uses the DBGKD_ANY_WAIT_STATE_CHANGE structure which is sized according to the largest control-report structure (AMD64_DBGKD_CONTROL_REPORT currently), and is larger than DBGKD_WAIT_STATE_CHANGE64 on x86. This worked because our DBGKD_WAIT_STATE_CHANGE32/64 structures contained incorrect DBGKD_CONTROL_REPORT (used) and CONTEXT (unused) members that sized up the wait-state structure to pass WinDbg's length verification! It actually becomes larger than DBGKD_ANY_WAIT_STATE_CHANGE, but WinDbg only seems bail out only if the structure is too small. Remove the incorrect members from the protocol 5 structures and change to DBGKD_ANY_WAIT_STATE_CHANGE everywhere. - Correct the value of SIZE_OF_FX_REGISTERS -- it was 4 times too low which resulted in KeContextToTrapFrame not properly clearing out the XMM register area. Correct the define and move it out from ke.h to x86's ketypes.h and use it in the FXSAVE format structure. Also remove the IOPM definitions from ke.h as they have been in the NDK for a while. - KD uses STRINGs, not ANSI_STRINGs -- they are the same thing, but let's be consistent. - ExceptionRecord32To64 should be available for both 32 and 64 bit builds (and it shouldn't be a forceinline). Get rid of CopyExceptionRecord and determine if we need to convert or can just copy it directly instead. - Use _WIN64 instead of _M_AMD64 when determining if we need to set the DBGKD_VERS_FLAG_PTR64 flag. - Don't check Nt/DbgQueryDebugFilterState for zero or nonzero -- it actually returns TRUE, FALSE or STATUS_INVALID_PARAMETER_1! Check for != TRUE in preparation for proper implementation of NtSet/QueryDebugFilterState. - Fix Format parameter of DbgPrintReturnControlC -- it is const like the other DbgPrint* routines. - Be consistent with the types used in debug.c and don't set local variables to zero if we are going to return to caller -- this doesn't seem to be required anymore. - Fix DebugService and DebugService2: DebugService should take a ULONG followed by 4 pointers and DebugService2 doesn't return anything. - Use ZwCurrentProcess() instead of -1 or 0xFFFFFFFF (which is incorrect for 64-bit) for the ProcessId parameter of DbgLoad/UnloadImageSymbols to clarify what is being passed. Don't use ZwCurrentProcess() in KeBugCheckWithTf for the pointer parameter of DbgUnLoadImageSymbols either. Use MAXULONG_PTR casted to PVOID instead. - Use better named and sized variables in KdpTrap for setting the "return register" in the caller's CONTEXT. - Correct and clarify the comment documenting under what conditions we pass user mode exceptions to the kernel debugger. svn path=/trunk/; revision=43741
2009-10-25 15:56:38 +00:00
if (KdpPrintString(&OutputString))
{
/* User pressed CTRL-C, breakpoint on return */
Status = STATUS_BREAKPOINT;
}
else
{
/* String was printed */
Status = STATUS_SUCCESS;
}
/* Exit the debugger and return */
- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg. - Add another hack to freeldr.c to make sure that HAL is the second entry in the Load Order list -- detect if it isn't and insert it as the second entry manually if it isn't. - KdbInitialize can now assume that the 2nd entry in the Load Order list is HAL, just like ExpLoadBootSymbols and KD does, so get the Loader Entry directly instead of searching for it. - Move KiBootTss and KiBootGdt to freeldr.c as this is where they belong -- they are not required for NTLDR/WINLDR boot style and are only used directly in freeldr.c. - Get rid of the AcpiTableDetected variable from freeldr.c. Instead, set the AcpiTable entry to something and make PoInitSystem check for that instead to preserve the old behavior. - Implement KdpGetFirstParameter and KdpGetSecondParameter for ARM too -- just retrieve R3/R4 here. Also rename those macros to clarify what parameters we are retrieving. - Add MmIsSessionAddress stub and use it from KD handle session space properly in the Memory Query API, and ASSERT that we are not trying to do a copy to/from session space in MmDbgCopyMemory as we don't handle it properly. Put this in mmdbg.c for now as we don't implement session space, and it is only called from KD right now. - Rename the 3 kdsup.c files to kdx86.c, kdx64.c and kdarm.c to differ them from each other. - Implement KdpAllowDisable -- just check if any processor breakpoints are set on any processor in the system and disallow the disable if so. The routine is now architecture dependant, so move it to the appropriate files. - Get rid of the MmFreeLdr* variables too. These have been deprecated for some time now. - The ModuleObject and ImageBaseAddress parameters of MmLoadSystemImage are not optional so don't treat them as such, and don't zero initialize them as callers shouldn't rely on this. - Set LDRP_ENTRY_NATIVE instead of LDRP_COMPAT_DATABASE_PROCESSED to mark the image as a native image. Also fix the value of LDRP_ENTRY_NATIVE. - Fix definition of LDR_DATA_TABLE_ENTRY -- the Checksum member should be in the union too. - Remove some unnecessary externs for stuff we now define globally in the kernel headers. - Rename some variables in KD to better match the logic. - Move some x86 only stuff from global ke.h and ke_x.h to the x86 dependent ke.h. Remove DR_ACTIVE_MASK as it has been deprecated/unused for a while now. svn path=/trunk/; revision=44023
2009-11-08 01:13:49 +00:00
KdExitDebugger(Enable);
*Handled = TRUE;
return Status;
}
VOID
__cdecl
KdpDprintf(
_In_ PCHAR Format,
...)
{
STRING String;
USHORT Length;
va_list ap;
CHAR Buffer[512];
/* Format the string */
va_start(ap, Format);
Length = (USHORT)_vsnprintf(Buffer,
sizeof(Buffer),
Format,
ap);
[NTOS:KD/KD64/KDBG] Share some code between our legacy KD/KDBG and KD64. Our legacy KD module is slowly being phased out for the more recent KD64 Kernel Debugger that supports WinDbg, but at the same time we must retain support for GCC debugging and the KDBG interface. For the time being few #ifdef _WINKD_ have been introduced in KD64 so that some of its code/data does not completely get shared yet with the legacy KD, until the latter becomes phased out. KD Modifications: ================= - Remove the implementation of NtQueryDebugFilterState() / NtSetDebugFilterState() that now comes entirely from KD64. - Remove KD variables that are now shared with KD64. - Share common code with KD64: KdpMoveMemory(), KdpZeroMemory(), KdpCopyMemoryChunks(), KdpPrint(), KdpPrompt(). - KDBG: Remove the duplicated KdpCopyMemoryChunks() function. - In KdpServiceDispatcher() and KdpEnterDebuggerException(), call the KdpPrint() worker function that correctly probes and captures its arguments. - Temporarily stub out KdEnterDebugger() and KdExitDebugger() that is used by the shared code, until KD is removed and only the KD64 version of these functions remain. - Re-implement the KD/KDBG KdpPrompt() function using a custom KdpPromptString() helper compatible with KD64, that is called by the KD64 implementation of KdpPrompt(). This KdpPromptString() helper now issues the prompt on all the KD loggers: e.g. if you use both at the same time COM-port and SCREEN debugging, the prompt will appear on both. Before that the prompt was always being displayed on COM port even if e.g. a SCREEN-only debug session was used... - ppc_irq.c: Fix the prototype of KdpServiceDispatcher(). KD64 Fixes: =========== - Initialize the MaximumLength member of the counted STRING variables before using them elsewhere. - Get rid of alloca() within SEH block in KdpPrint() (addendum to 7b95fcf9). - Add the ROS-specific handy dump commands in KdSystemDebugControl().
2019-11-17 21:55:36 +00:00
va_end(ap);
/* Set it up */
String.Buffer = Buffer;
[NTOS:KD/KD64/KDBG] Share some code between our legacy KD/KDBG and KD64. Our legacy KD module is slowly being phased out for the more recent KD64 Kernel Debugger that supports WinDbg, but at the same time we must retain support for GCC debugging and the KDBG interface. For the time being few #ifdef _WINKD_ have been introduced in KD64 so that some of its code/data does not completely get shared yet with the legacy KD, until the latter becomes phased out. KD Modifications: ================= - Remove the implementation of NtQueryDebugFilterState() / NtSetDebugFilterState() that now comes entirely from KD64. - Remove KD variables that are now shared with KD64. - Share common code with KD64: KdpMoveMemory(), KdpZeroMemory(), KdpCopyMemoryChunks(), KdpPrint(), KdpPrompt(). - KDBG: Remove the duplicated KdpCopyMemoryChunks() function. - In KdpServiceDispatcher() and KdpEnterDebuggerException(), call the KdpPrint() worker function that correctly probes and captures its arguments. - Temporarily stub out KdEnterDebugger() and KdExitDebugger() that is used by the shared code, until KD is removed and only the KD64 version of these functions remain. - Re-implement the KD/KDBG KdpPrompt() function using a custom KdpPromptString() helper compatible with KD64, that is called by the KD64 implementation of KdpPrompt(). This KdpPromptString() helper now issues the prompt on all the KD loggers: e.g. if you use both at the same time COM-port and SCREEN debugging, the prompt will appear on both. Before that the prompt was always being displayed on COM port even if e.g. a SCREEN-only debug session was used... - ppc_irq.c: Fix the prototype of KdpServiceDispatcher(). KD64 Fixes: =========== - Initialize the MaximumLength member of the counted STRING variables before using them elsewhere. - Get rid of alloca() within SEH block in KdpPrint() (addendum to 7b95fcf9). - Add the ROS-specific handy dump commands in KdSystemDebugControl().
2019-11-17 21:55:36 +00:00
String.Length = String.MaximumLength = Length;
/* Send it to the debugger directly */
KdpPrintString(&String);
}