Finding Problems With Exceptions, No not Casper

[James Tabor]

Hi all,

Most common, when exiting a app by clicking the "X" in the upper right corner,

(./subsys/win32k/objects/gdiobj.c:591) Attempted to free global gdi handle 0xe04
04b4, caller needs to get ownership first!!!
(lib/rtl/exception.c:75) RtlRaiseStatus(Status 0xc0000005)
(./subsys/win32k/ntuser/message.c:1121) Failed to copy message to kernel: invali
d usermode buffer
(lib/comctl32/toolbar.c:417) bitmap for ID 0, index 0 is not valid, number of bi
tmaps in imagelist: -2
(./ntoskrnl/ke/queue.c:420) Unwaiting Thread: 5
(lib/rtl/exception.c:75) RtlRaiseStatus(Status 0xc0000005)
(./subsys/win32k/ntuser/message.c:1121) Failed to copy message to kernel: invali
d usermode buffer

C0000005 is an Access error, due to MmCopyToCaller being called inside kernel space
when copying one object to another with in kernel space, sometimes but it is common.

In win32k/ntuser/message.c there is and it's my best guess,
Status = MmCopyFromCaller(KernelMem, (PVOID) UserModeMsg->lParam, Size);
at line ~1118, in function CopyMsgToKernelMem. If that function is used to copy data
inside kernel space it will through and exception.


This one is from OpenOffice, Some where in kernel32.dll.

(lib/rtl/exception.c:29) RtlRaiseException(Status 0054f404)
(lib/rtl/exception.c:36) ExceptionAddress 7c801217
(./ntoskrnl/ke/exception.c:94) KiRaiseException
(lib/rtl/exception.c:29) RtlRaiseException(Status 0054f41c)
(lib/rtl/exception.c:36) ExceptionAddress 7c801217
(./ntoskrnl/ke/exception.c:94) KiRaiseException
0 bytes requested - initiating pool verification
0 bytes requested - initiating pool verification

Sorry I haven't looked at this one yet, I just started!
8^>
James

[Hervé Poussineau]

Hi,

You might want to to take a look at bug 1075.

Hervé 

-----Message d'origine-----
De : ros-dev-bounces(a)reactos.org [mailto:[e-mail]] De la
part de James Tabor
Envoyé : vendredi 6 janvier 2006 16:09
À : ReactOS Development List
Objet : [ros-dev] Finding Problems With Exceptions, No not Casper

Hi all,

Most common, when exiting a app by clicking the "X" in the upper right
corner,

(./subsys/win32k/objects/gdiobj.c:591) Attempted to free global gdi handle
0xe04 04b4, caller needs to get ownership first!!!
(lib/rtl/exception.c:75) RtlRaiseStatus(Status 0xc0000005)
(./subsys/win32k/ntuser/message.c:1121) Failed to copy message to kernel:
invali d usermode buffer
(lib/comctl32/toolbar.c:417) bitmap for ID 0, index 0 is not valid, number
of bi tmaps in imagelist: -2
(./ntoskrnl/ke/queue.c:420) Unwaiting Thread: 5
(lib/rtl/exception.c:75) RtlRaiseStatus(Status 0xc0000005)
(./subsys/win32k/ntuser/message.c:1121) Failed to copy message to kernel:
invali d usermode buffer

C0000005 is an Access error, due to MmCopyToCaller being called inside
kernel space when copying one object to another with in kernel space,
sometimes but it is common.

In win32k/ntuser/message.c there is and it's my best guess, Status =
MmCopyFromCaller(KernelMem, (PVOID) UserModeMsg->lParam, Size); at line
~1118, in function CopyMsgToKernelMem. If that function is used to copy data
inside kernel space it will through and exception.


This one is from OpenOffice, Some where in kernel32.dll.

(lib/rtl/exception.c:29) RtlRaiseException(Status 0054f404)
(lib/rtl/exception.c:36) ExceptionAddress 7c801217
(./ntoskrnl/ke/exception.c:94) KiRaiseException
(lib/rtl/exception.c:29) RtlRaiseException(Status 0054f41c)
(lib/rtl/exception.c:36) ExceptionAddress 7c801217
(./ntoskrnl/ke/exception.c:94) KiRaiseException 0 bytes requested -
initiating pool verification 0 bytes requested - initiating pool
verification

Sorry I haven't looked at this one yet, I just started!
8^>
James

_______________________________________________
Ros-dev mailing list
Ros-dev(a)reactos.org
http://www.reactos.org/mailman/listinfo/ros-dev

[James Tabor]

Hi!
Hervé Poussineau wrote:
>   Hi,
>  
>  You might want to to take a look at bug 1075.
>  
>  Hervé 
>   
Yes, its the same thing! But I found out you can not use MmCopyFromCaller inside
kernel space. I guess someone will get it and change it. I guess it would be best
to figure out "if inside K space" it would be just a RtlCopyMemory instead.

I'll work on it, but everyone should know if you see it during coding you will know
what caused it. LOL! That's how I found it. Between two debug prints so it was easy
to catch it. 8^D

Thanks,
James

[Ge van Geldorp]

>   From: James Tabor
>  
>  Yes, its the same thing! But I found out you can not use 
>  MmCopyFromCaller inside kernel space. I guess someone will 
>  get it and change it. I guess it would be best to figure out 
>  "if inside K space" it would be just a RtlCopyMemory instead.
>  
>  I'll work on it, but everyone should know if you see it 
>  during coding you will know what caused it. LOL! That's how I 
>  found it. Between two debug prints so it was easy to catch it. 8^D 
The reason I've left it alone is that the IntSendOrPostMessage (or something
like that...) function that causes this is a hack and should die.

GvG