Ros-dev

ros-dev@reactos.org

5790 discussions

Re: [ros-dev] [ros-diffs] [sir_richard] 45162: [RBUILD]: Use correct kernel entrypoint symbol name. Forgot to commit this earlier. Please, next time before you harass us with e-mails, try to read your logs. LD choses a random entrypoint if the symbol cannot be foun
by Ged Murphy 20 Jan '10

20 Jan '10

sir_richard(a)svn.reactos.org wrote: > Please, next time before you harass us with e-mails, try to read your logs. > LD choses a random entrypoint if the symbol cannot be found. I hope you aren't referring to the 2 emails stating that you broke the build? The 2 guys who so rudely "harassed" you for highlighting your mistake are testers, they can't spot and fix your errors. The fact that you broke the build stops them from testing. Please be a little more considerate. Ged.

1 0

Re: [ros-dev] [ros-diffs] [sir_richard] 45152: [FREELDR]: Jump to a standard 1-parameter STDCALL kernel entrypoint instead of a FASTCALL double-parameter entrypoint. [NTOS]: Make KiSystemStartup the real C entrypoint of the kernel, and move the "Am I being booted by FreeLDR" logic inside it -- it will then call KiRosPrepareForSystemStartup as earlier. [NTOS]: Move the Double Fault and Boot Stack declaration in C code, with the proper alignment attribute. [NTOS]: Although the concern that KiSystemStartup can
by niski 20 Jan '10

20 Jan '10

Hi Sir Richard, just FYI, this rev introduced freeldr crash (happens after loading initial drivers): http://bjauy.com/images/45155-freeloader.png Best regards, Maciej ---------------------------------------------------------------------- Rezerwując wczesniej kupisz bilety lotnicze taniej! Sprawdz >>> http://link.interia.pl/f256c

1 0

[ros-diffs] [sir_richard] 45152
by Dmitry Gorbachev 20 Jan '10

20 Jan '10

Hi, sir_richard wrote: > /* GLOBALS *******************************************************************/ > + > +/* Boot and double-fault/NMI/DPC stack */ > +UCHAR P0BootStackData[KERNEL_STACK_SIZE] __attribute__((aligned (16))); > +UCHAR KiDoubleFaultStackData[KERNEL_STACK_SIZE] __attribute__((aligned (16))); MinGW GCC 4.4.x can ignore in some situations attribute aligned on uninitialized global variables because of a bug. A workaround is to initialize a variable or to compile with -fno-common.

2 1

20 Jan '10

Hi Sir Richard, Build's broken: ntoskrnl/ke/i386/context.c: In function 'KiSwapProcess': ntoskrnl/ke/i386/context.c:30: warning: passing argument 1 of '_InterlockedXor' makes pointer from integer without a cast ntoskrnl/ke/i386/context.c:31: warning: passing argument 1 of '_InterlockedXor' makes pointer from integer without a cast make: *** [/mnt/ramdisk/buildbot/debug/obj/ntoskrnl/ke/i386/context_ntkrnlmp.o] Error 1 make: *** Waiting for unfinished jobs.... Command exited with non-zero status 2 from: http://build.reactos.org:8010/builders/x86_%28Debug%29/builds/3789/steps/co… Gabriel. _________________________________________________________________ 25 Gygabite gratis online. Archivia e condividi i tuoi file! http://www.windowslive.it/skyDrive.aspx

1 0

Re: [ros-dev] [ros-diffs] [sir_richard] 45156: [NTOS]: Implement KiSwapProcess in C. [NTOS]: Implement KiIsNpxPresent and KiIsNpxErrataPresent in C. It's much clearer what these are doing now. [NTOS]: Implement KiFlushNPXState and fix some bugs that were present in the ASM version, such as a wrong NPX state check. [NTOS]: Implement working intrinsics for fxrstor, fxsave, fnsave and enable them for flushing. We'll update the FPU trap code to use these later.
by niski 20 Jan '10

20 Jan '10

Hi Sir Richard, Could you fix the build, please? http://build.reactos.org:8010/builders/x86_%28Debug%29/builds/3789/steps/co… Best regards, Maciej "niski" ---------------------------------------------------------------------- Kup wlasne mieszkanie za 72 tys. zl. Sprawdz najlepsze oferty >>> http://link.interia.pl/f2570

1 0

reactos and linux
by Max 19 Jan '10

19 Jan '10

if it is not sold, I suggest to build it into reactos http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=200427881447#ht_500wt_10… then users can install any kind of app. thoughts?

7 10

Re: [ros-dev] [ros-diffs] [sir_richard] 45148: [NTOS]: Implement GUI thread promotion during the first GUI system call in C. This is tricky due to EBP, and actually requires some tiny inline ASM magic to make it work right. [NTOS]: Implement SYSENTER system calls
by Gabriel ilardi 19 Jan '10

19 Jan '10

Hi Sir Richard, Just for the record, trunk's broken in 1st stage: http://build.reactos.org:8010/builders/x86_%28Test%29/builds/1208/steps/tes… Regards, Gabriel. > Date: Tue, 19 Jan 2010 09:45:30 +0000 > To: ros-diffs(a)reactos.org > From: sir_richard(a)svn.reactos.org > Subject: [ros-diffs] [sir_richard] 45148: [NTOS]: Implement GUI thread promotion during the first GUI system call in C. This is tricky due to EBP, and actually requires some tiny inline ASM magic to make it work right. [NTOS]: Implement SYSENTER system calls in C as well. All system calls are now handled in C. This code will be further optimized/refined soon. > > Author: sir_richard > Date: Tue Jan 19 10:45:30 2010 > New Revision: 45148 > > URL: http://svn.reactos.org/svn/reactos?rev=45148&view=rev > Log: > [NTOS]: Implement GUI thread promotion during the first GUI system call in C. This is tricky due to EBP, and actually requires some tiny inline ASM magic to make it work right. > [NTOS]: Implement SYSENTER system calls in C as well. > > All system calls are now handled in C. This code will be further optimized/refined soon. > > Modified: > trunk/reactos/ntoskrnl/include/internal/ke.h > trunk/reactos/ntoskrnl/include/internal/trap_x.h > trunk/reactos/ntoskrnl/ke/i386/trap.s > trunk/reactos/ntoskrnl/ke/i386/traphdlr.c > > Modified: trunk/reactos/ntoskrnl/include/internal/ke.h > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/ntoskrnl/include/internal/… > ============================================================================== > --- trunk/reactos/ntoskrnl/include/internal/ke.h [iso-8859-1] (original) > +++ trunk/reactos/ntoskrnl/include/internal/ke.h [iso-8859-1] Tue Jan 19 10:45:30 2010 > @@ -138,6 +138,7 @@ > extern ULONG_PTR KiBugCheckData[5]; > extern ULONG KiFreezeFlag; > extern ULONG KiDPCTimeout; > +extern PGDI_BATCHFLUSH_ROUTINE KeGdiFlushUserBatch; > > /* MACROS *************************************************************************/ > > > Modified: trunk/reactos/ntoskrnl/include/internal/trap_x.h > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/ntoskrnl/include/internal/… > ============================================================================== > --- trunk/reactos/ntoskrnl/include/internal/trap_x.h [iso-8859-1] (original) > +++ trunk/reactos/ntoskrnl/include/internal/trap_x.h [iso-8859-1] Tue Jan 19 10:45:30 2010 > @@ -436,3 +436,42 @@ > > return Result; > } > + > +NTSTATUS > +FORCEINLINE > +KiConvertToGuiThread(VOID) > +{ > + NTSTATUS Result; > + PVOID StackFrame; > + > + /* > + * Converting to a GUI thread safely updates ESP in-place as well as the > + * current Thread->TrapFrame and EBP when KeSwitchKernelStack is called. > + * > + * However, PsConvertToGuiThread "helpfully" restores EBP to the original > + * caller's value, since it is considered a nonvolatile register. As such, > + * as soon as we're back after the conversion and we try to store the result > + * which will probably be in some stack variable (EBP-based), we'll crash as > + * we are touching the de-allocated non-expanded stack. > + * > + * Thus we need a way to update our EBP before EBP is touched, and the only > + * way to guarantee this is to do the call itself in assembly, use the EAX > + * register to store the result, fixup EBP, and then let the C code continue > + * on its merry way. > + * > + */ > + __asm__ __volatile__ > + ( > + "movl %%ebp, %1\n" > + "subl %%esp, %1\n" > + "call _PsConvertToGuiThread@0\n" > + "addl %%esp, %1\n" > + "movl %1, %%ebp\n" > + "movl %%eax, %0\n" > + : "=r"(Result), "=r"(StackFrame) > + : > + : "%esp", "%ecx", "%edx" > + ); > + > + return Result; > +} > > Modified: trunk/reactos/ntoskrnl/ke/i386/trap.s > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/ntoskrnl/ke/i386/trap.s?re… > ============================================================================== > --- trunk/reactos/ntoskrnl/ke/i386/trap.s [iso-8859-1] (original) > +++ trunk/reactos/ntoskrnl/ke/i386/trap.s [iso-8859-1] Tue Jan 19 10:45:30 2010 > @@ -129,260 +129,30 @@ > .endfunc > > .func KiFastCallEntry > -TRAP_FIXUPS FastCallDrSave, FastCallDrReturn, DoNotFixupV86, DoNotFixupAbios > _KiFastCallEntry: > > - /* Enter the fast system call prolog */ > - FASTCALL_PROLOG FastCallDrSave, FastCallDrReturn > - > -SharedCode: > - > - /* > - * Find out which table offset to use. Converts 0x1124 into 0x10. > - * The offset is related to the Table Index as such: Offset = TableIndex x 10 > - */ > - mov edi, eax > - shr edi, SERVICE_TABLE_SHIFT > - and edi, SERVICE_TABLE_MASK > - mov ecx, edi > - > - /* Now add the thread's base system table to the offset */ > - add edi, [esi+KTHREAD_SERVICE_TABLE] > - > - /* Get the true syscall ID and check it */ > - mov ebx, eax > - and eax, SERVICE_NUMBER_MASK > - cmp eax, [edi+SERVICE_DESCRIPTOR_LIMIT] > - > - /* Invalid ID, try to load Win32K Table */ > - jnb KiBBTUnexpectedRange > - > - /* Check if this was Win32K */ > - cmp ecx, SERVICE_TABLE_TEST > - jnz NotWin32K > - > - /* Get the TEB */ > - mov ecx, PCR[KPCR_TEB] > - > - /* Check if we should flush the User Batch */ > - xor ebx, ebx > -_ReadBatch: > - or ebx, [ecx+TEB_GDI_BATCH_COUNT] > - jz NotWin32K > - > - /* Flush it */ > - push edx > - push eax > - call [_KeGdiFlushUserBatch] > - pop eax > - pop edx > - > -NotWin32K: > - /* Increase total syscall count */ > - inc dword ptr PCR[KPCR_SYSTEM_CALLS] > - > -#if DBG > - /* Increase per-syscall count */ > - mov ecx, [edi+SERVICE_DESCRIPTOR_COUNT] > - jecxz NoCountTable > - inc dword ptr [ecx+eax*4] > -#endif > - > - /* Users's current stack frame pointer is source */ > -NoCountTable: > - mov esi, edx > - > - /* Allocate room for argument list from kernel stack */ > - mov ebx, [edi+SERVICE_DESCRIPTOR_NUMBER] > - xor ecx, ecx > - mov cl, [eax+ebx] > - > - /* Get pointer to function */ > - mov edi, [edi+SERVICE_DESCRIPTOR_BASE] > - mov ebx, [edi+eax*4] > - > - /* Allocate space on our stack */ > - sub esp, ecx > - > - /* Set the size of the arguments and the destination */ > - shr ecx, 2 > - mov edi, esp > - > - /* Make sure we're within the User Probe Address */ > - cmp esi, _MmUserProbeAddress > - jnb AccessViolation > - > -_CopyParams: > - /* Copy the parameters */ > - rep movsd > - > - /* Do the System Call */ > - call ebx > - > -AfterSysCall: > -#if DBG > - /* Make sure the user-mode call didn't return at elevated IRQL */ > - test byte ptr [ebp+KTRAP_FRAME_CS], MODE_MASK > - jz SkipCheck > - mov esi, eax /* We need to save the syscall's return val */ > - call _KeGetCurrentIrql@0 > - or al, al > - jnz InvalidIrql > - mov eax, esi /* Restore it */ > - > - /* Get our temporary current thread pointer for sanity check */ > - mov ecx, PCR[KPCR_CURRENT_THREAD] > - > - /* Make sure that we are not attached and that APCs are not disabled */ > - mov dl, [ecx+KTHREAD_APC_STATE_INDEX] > - or dl, dl > - jnz InvalidIndex > - mov edx, [ecx+KTHREAD_COMBINED_APC_DISABLE] > - or edx, edx > - jnz InvalidIndex > -#endif > - > -SkipCheck: > - > - /* Deallocate the kernel stack frame */ > - mov esp, ebp > - > -KeReturnFromSystemCall: > - > - /* Get the Current Thread */ > - mov ecx, PCR[KPCR_CURRENT_THREAD] > - > - /* Restore the old trap frame pointer */ > - mov edx, [ebp+KTRAP_FRAME_EDX] > - mov [ecx+KTHREAD_TRAP_FRAME], edx > + /* Sane FS segment */ > + mov ecx, KGDT_R0_PCR > + mov fs, cx > > - /* Exit the system call */ > - mov ecx, ebp > - mov edx, eax > - jmp @KiServiceExit@8 > -.endfunc > - > -KiBBTUnexpectedRange: > - > - /* If this isn't a Win32K call, fail */ > - cmp ecx, SERVICE_TABLE_TEST > - jne InvalidCall > - > - /* Set up Win32K Table */ > - push edx > - push ebx > - call _PsConvertToGuiThread@0 > - > - /* Check return code */ > - or eax, eax > - > - /* Restore registers */ > - pop eax > - pop edx > - > - /* Reset trap frame address */ > - mov ebp, esp > - mov [esi+KTHREAD_TRAP_FRAME], ebp > - > - /* Try the Call again, if we suceeded */ > - jz SharedCode > - > - /* > - * The Shadow Table should have a special byte table which tells us > - * whether we should return FALSE, -1 or STATUS_INVALID_SYSTEM_SERVICE. > - */ > - > - /* Get the table limit and base */ > - lea edx, _KeServiceDescriptorTableShadow + SERVICE_TABLE_TEST > - mov ecx, [edx+SERVICE_DESCRIPTOR_LIMIT] > - mov edx, [edx+SERVICE_DESCRIPTOR_BASE] > - > - /* Get the table address and add our index into the array */ > - lea edx, [edx+ecx*4] > - and eax, SERVICE_NUMBER_MASK > - add edx, eax > - > - /* Find out what we should return */ > - movsx eax, byte ptr [edx] > - or eax, eax > - > - /* Return either 0 or -1, we've set it in EAX */ > - jle KeReturnFromSystemCall > - > - /* Set STATUS_INVALID_SYSTEM_SERVICE */ > - mov eax, STATUS_INVALID_SYSTEM_SERVICE > - jmp KeReturnFromSystemCall > - > -InvalidCall: > - > - /* Invalid System Call */ > - mov eax, STATUS_INVALID_SYSTEM_SERVICE > - jmp KeReturnFromSystemCall > - > -AccessViolation: > - > - /* Check if this came from kernel-mode */ > - test byte ptr [ebp+KTRAP_FRAME_CS], MODE_MASK > - > - /* It's fine, go ahead with it */ > - jz _CopyParams > - > - /* Caller sent invalid parameters, fail here */ > - mov eax, STATUS_ACCESS_VIOLATION > - jmp AfterSysCall > - > -BadStack: > - > - /* Restore ESP0 stack */ > - mov ecx, PCR[KPCR_TSS] > - mov esp, ss:[ecx+KTSS_ESP0] > - > - /* Generate V86M Stack for Trap 6 */ > - push 0 > - push 0 > - push 0 > - push 0 > - > - /* Generate interrupt stack for Trap 6 */ > - push KGDT_R3_DATA + RPL_MASK > - push 0 > - push 0x20202 > - push KGDT_R3_CODE + RPL_MASK > - push 0 > - jmp _KiTrap06 > - > -#if DBG > -InvalidIrql: > - /* Save current IRQL */ > - push PCR[KPCR_IRQL] > - > - /* Set us at passive */ > - mov dword ptr PCR[KPCR_IRQL], 0 > - cli > - > - /* Bugcheck */ > - push 0 > - push 0 > - push eax > - push ebx > - push IRQL_GT_ZERO_AT_SYSTEM_SERVICE > - call _KeBugCheckEx@20 > - > -InvalidIndex: > - > - /* Get the index and APC state */ > - movzx eax, byte ptr [ecx+KTHREAD_APC_STATE_INDEX] > - mov edx, [ecx+KTHREAD_COMBINED_APC_DISABLE] > - > - /* Bugcheck */ > - push 0 > - push edx > - push eax > - push ebx > - push APC_INDEX_MISMATCH > - call _KeBugCheckEx@20 > - ret > -#endif > + /* Sane stack and frame */ > + mov esp, PCR[KPCR_TSS] > + mov esp, [esp+KTSS_ESP0] > + > + /* Make space for trap frame on the stack */ > + sub esp, KTRAP_FRAME_V86_ES > + > + /* Save EBP, EBX, ESI, EDI only! */ > + mov [esp+KTRAP_FRAME_EBX], ebx > + mov [esp+KTRAP_FRAME_ESI], esi > + mov [esp+KTRAP_FRAME_EDI], edi > + mov [esp+KTRAP_FRAME_EBP], ebp > + > + /* Call C handler -- note that EDX is the user stack, and EAX the syscall */ > + mov ecx, esp > + add edx, 8 > + jmp _KiFastCallEntryHandler > +.endfunc > > .func Kei386EoiHelper@0 > _Kei386EoiHelper@0: > > Modified: trunk/reactos/ntoskrnl/ke/i386/traphdlr.c > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/ntoskrnl/ke/i386/traphdlr.… > ============================================================================== > --- trunk/reactos/ntoskrnl/ke/i386/traphdlr.c [iso-8859-1] (original) > +++ trunk/reactos/ntoskrnl/ke/i386/traphdlr.c [iso-8859-1] Tue Jan 19 10:45:30 2010 > @@ -1465,6 +1465,7 @@ > } > > /* Check for syscall fault */ > +#if 0 > if ((TrapFrame->Eip == (ULONG_PTR)CopyParams) || > (TrapFrame->Eip == (ULONG_PTR)ReadBatch)) > { > @@ -1472,7 +1473,7 @@ > UNIMPLEMENTED; > while (TRUE); > } > - > +#endif > /* Check for VDM trap */ > ASSERT((KiVdmTrap(TrapFrame)) == FALSE); > > @@ -1725,9 +1726,14 @@ > goto ExitCall; > } > > - /* GUI calls are not yet supported */ > - UNIMPLEMENTED; > - while (TRUE); > + /* Convert us to a GUI thread -- must wrap in ASM to get new EBP */ > + Result = KiConvertToGuiThread(); > + if (__builtin_expect(!NT_SUCCESS(Result), 0)) > + { > + /* Figure out how we should fail to the user */ > + UNIMPLEMENTED; > + while (TRUE); > + } > > /* Try the call again */ > continue; > @@ -1741,8 +1747,7 @@ > if (__builtin_expect(Offset & SERVICE_TABLE_TEST, 0)) > { > /* Get the batch count and flush if necessary */ > - UNIMPLEMENTED; > - while (TRUE); > + if (NtCurrentTeb()->GdiBatchCount) KeGdiFlushUserBatch(); > } > > /* Increase system call count */ > @@ -1816,6 +1821,39 @@ > /* Enable interrupts and make the call */ > _enable(); > KiSystemCall(ServiceNumber, Arguments); > +} > + > +VOID > +__attribute__((regparm(3))) > +KiFastCallEntryHandler(IN ULONG ServiceNumber, > + IN PVOID Arguments, > + IN PKTRAP_FRAME TrapFrame) > +{ > + PKTHREAD Thread; > + > + /* Fixup segments */ > + Ke386SetDs(KGDT_R3_DATA | RPL_MASK); > + Ke386SetEs(KGDT_R3_DATA | RPL_MASK); > + > + /* Set up a fake INT Stack and enable interrupts */ > + TrapFrame->HardwareSegSs = KGDT_R3_DATA | RPL_MASK; > + TrapFrame->HardwareEsp = (ULONG_PTR)Arguments - 8; // Stack is 2 frames down > + TrapFrame->EFlags = __readeflags() | EFLAGS_INTERRUPT_MASK; > + TrapFrame->SegCs = KGDT_R3_CODE | RPL_MASK; > + TrapFrame->Eip = SharedUserData->SystemCallReturn; > + __writeeflags(0x2); > + > + /* Get the current thread */ > + Thread = KeGetCurrentThread(); > + > + /* Call the shared handler (inline) */ > + KiSystemCallHandler(TrapFrame, > + ServiceNumber, > + Arguments, > + Thread, > + UserMode, > + Thread->PreviousMode, > + KGDT_R3_TEB | RPL_MASK); > } > > VOID > > _________________________________________________________________ 25 Gygabite gratis online. Archivia e condividi i tuoi file! http://www.windowslive.it/skyDrive.aspx

1 0

Re: [ros-dev] Request for a project
by sudheer kumar 19 Jan '10

19 Jan '10

hello guys, I am sorry for not mentioning a specific area that I am interested. I am interested in filesystems. Thank you Aleksey , victor and others for providing the list of projects. From the list I am interested in the partition manger as arty shed some light it. Any suggestions and help are most welcome. Thanks & Regards, sudheer.

6 11

CreateFileW doesn't support relative path???
by Xiaoming Gao 18 Jan '10

18 Jan '10

Hi,guys: it seems that CreateFileW doesn't support relative path to create a file,,,weird.... the code belows: InitializeObjectAttributes(&ObjectAttributes, &NtPathU, 0, NULL, NULL); if (lpSecurityAttributes) { if(lpSecurityAttributes->bInheritHandle) ObjectAttributes.Attributes |= OBJ_INHERIT; ObjectAttributes.SecurityDescriptor = lpSecurityAttributes->lpSecurityDescriptor; } if(!(dwFlagsAndAttributes & FILE_FLAG_POSIX_SEMANTICS)) ObjectAttributes.Attributes |= OBJ_CASE_INSENSITIVE; /* perform the call */ Status = NtCreateFile (&FileHandle, dwDesiredAccess, &ObjectAttributes, &IoStatusBlock, NULL, FileAttributes, dwShareMode, dwCreationDisposition, Flags, EaBuffer, EaLength);

1 0

Update on Wikipedia page
by Joshua Bailey 17 Jan '10

17 Jan '10

I read the PDF and other information earlier today. I've just updated the Wikipedia page (in English, at least) regarding Arwinss with the newer, more detailed information. Hopefully, once more people know about Arwinss and what it will do for ReactOS, more people will appreciate the potential of this project and join in. I felt you guys should know, since I brought up the issue earlier. Feel free to check on it and correct any errors I've made. :) -Joshua Bailey

4 6

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Ros-dev