Failing an IRP

NTDEV
1

Hi,

I have implemented a filter driver to handle a custom IOCTL called by an
user mode application.My idea for implementaion is that the driver should
return the size required for the buffer when a NULL input and output buffer
is provided.
I find that if I complete the IRP with a status other than STATUS_SUCCESS ,
the Io Manager updates the IoStatus.Information with some default value.
WDK help says that

[quote] "Unless a driver’s dispatch routine completes an IRP with an error
status value, the lowest-level driver in the chain frequently sets the IRP’s
I/O status block to the values that will be returned to the original
requester of the I/O operation.

Now I want to return an ERROR STATUS to the user app and have the required
the size returned in IoStatus.Information. http: Is
there any way to do this without the IoManager modifying the contents of
IoStatus.Information http:</http:> ?

Thanks
Abdul</http:>

2

STATUS_BUFFER_TOO_SMALL does what you want.

When user mode calls DeviceIoControl(), the required buffer size is
returned in dwBytesReturned.

At 13:25 21/09/2010, Abdul Qader wrote:

Hi,

I have implemented a filter driver to handle a custom IOCTL called
by an user mode application.My idea for implementaion is that the
driver should return the size required for the buffer when a NULL
input and output buffer is provided.
I find that if I complete the IRP with a status other than
STATUS_SUCCESS , the Io Manager updates the IoStatus.Information
with some default value.
WDK help says that

[quote] "Unless a driver’s dispatch routine completes an IRP with an
error status value, the lowest-level driver in the chain frequently
sets the IRP’s I/O status block to the values that will be returned
to the original requester of the I/O operation.
>
>Now I want to return an ERROR STATUS to the user app and have the
>required the size returned in
>http:IoStatus.Information. Is there any
>way to do this without the IoManager modifying the contents of
>http:</http:>IoStatus.Information ?
>
>Thanks
>Abdul
></http:>

3

Thanks Mark…

4

Mark,

Even on returning such a status I find that the IoManager updates the
Information with value 148.

Here is my code:

case IOCTL_KEYBOARD_TRANSFER_DATA:
/*Buffered IO: */
if(NULL == Irp->AssociatedIrp.SystemBuffer)
{
DebugPrint((“System Buffer NULL”));
Irp->IoStatus.Information = 4 ;
status = STATUS_BUFFER_TOO_SMALL;
}
else
{
…extra code
}
break;
}

if (!NT_SUCCESS(status)) {
IoCompleteRequest(Irp, IO_NO_INCREMENT);
return status;
}

On debugging I find Irp->IoStatus.Information = 148 once IoCompleteRequest
is called.

Thanks
Abdul

5

Set “ba w4” (on 32bit builds) on the field before IoCompleteRequest, and look at where and how it is updated.

–
Maxim S. Shatskih
Windows DDK MVP
xxxxx@storagecraft.com
http://www.storagecraft.com

“Abdul Qader” wrote in message news:xxxxx@ntdev…
Mark,

Even on returning such a status I find that the IoManager updates the Information with value 148.

Here is my code:

case IOCTL_KEYBOARD_TRANSFER_DATA:
/*Buffered IO: */
if(NULL == Irp->AssociatedIrp.SystemBuffer)
{
DebugPrint((“System Buffer NULL”));
Irp->IoStatus.Information = 4 ;
status = STATUS_BUFFER_TOO_SMALL;
}
else
{
…extra code
}
break;
}

if (!NT_SUCCESS(status)) {
IoCompleteRequest(Irp, IO_NO_INCREMENT);
return status;
}

On debugging I find Irp->IoStatus.Information = 148 once IoCompleteRequest is called.

Thanks
Abdul

6

Did you leave out setting Irp->IoStatus.Status?

Bill Wandel

From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Abdul Qader
Sent: Tuesday, September 21, 2010 10:50 AM
To: Windows System Software Devs Interest List
Subject: Re: [ntdev] Failing an IRP

Mark,

Even on returning such a status I find that the IoManager updates the
Information with value 148.

Here is my code:

case IOCTL_KEYBOARD_TRANSFER_DATA:

/*Buffered IO: */

if(NULL == Irp->AssociatedIrp.SystemBuffer)

{

DebugPrint((“System Buffer NULL”));

Irp->IoStatus.Information = 4 ;

status = STATUS_BUFFER_TOO_SMALL;

}

else

{

…extra code

}

break;

}

if (!NT_SUCCESS(status)) {

IoCompleteRequest(Irp, IO_NO_INCREMENT);

return status;

}

On debugging I find Irp->IoStatus.Information = 148 once IoCompleteRequest
is called.

Thanks

Abdul

— NTDEV is sponsored by OSR For our schedule of WDF, WDM, debugging and
other seminars visit: http://www.osr.com/seminars To unsubscribe, visit the
List Server section of OSR Online at
http://www.osronline.com/page.cfm?name=ListServer

7

I don’t think dwBytesReturned is updated on failed status.

8

Sorry Maxim … Can you elaborate…?
what is “ba w4”?

Thanks

On Tue, Sep 21, 2010 at 8:47 PM, Maxim S. Shatskih
wrote:

> Set “ba w4” (on 32bit builds) on the field before IoCompleteRequest, and
> look at where and how it is updated.
>
>
>
>

9

The usual way to do this for win32 style applications is for the IOCTL to
have a output data structure with a default non-zero length that contains
the required additional (buffer) length, and have the driver indicate a
STATUS_BUFFER_TOO_SMALL status as noted elsewhere.

For example:

struct foo
{
ULONG bytesNeeded;
UCHAR buffer[1];
}

Your driver fills in the 'bytesNeeded value and returns sizeof(foo) bytes
(Irp->IoStatus.Information = sizeof(foo);). The app then allocates a big
enough block of data and resends the IOCTL.

What you are trying to do actually can be done but is non-standard for
windows application programming.

Mark Roddy

On Tue, Sep 21, 2010 at 8:25 AM, Abdul Qader wrote:

> Hi,
>
> I have implemented a filter driver to handle a custom IOCTL called by an
> user mode application.My idea for implementaion is that the driver should
> return the size required for the buffer when a NULL input and output buffer
> is provided.
> I find that if I complete the IRP with a status other than STATUS_SUCCESS ,
> the Io Manager updates the IoStatus.Information with some default value.
> WDK help says that
>

[quote]
"Unless a driver’s dispatch routine completes an IRP with an error
> status value, the lowest-level driver in the chain frequently sets the IRP’s
> I/O status block to the values that will be returned to the original
> requester of the I/O operation.
>
> Now I want to return an ERROR STATUS to the user app and have the required
> the size returned in IoStatus.Information.http:</http:> Is
> there any way to do this without the IoManager modifying the contents of
> IoStatus.Information http:</http:> ?
>
> Thanks
> Abdul
>
>
>
> — NTDEV is sponsored by OSR For our schedule of WDF, WDM, debugging and
> other seminars visit: http://www.osr.com/seminars To unsubscribe, visit
> the List Server section of OSR Online at
> http://www.osronline.com/page.cfm?name=ListServer

10

At 16:34 21/09/2010, xxxxx@broadcom.com wrote:

I don’t think dwBytesReturned is updated on failed status.

if (ulBufLength < ulTotalLength)
{
pIrp->IoStatus.Status = ntStatus = STATUS_BUFFER_TOO_SMALL;
pIrp->IoStatus.Information = ulTotalLength;
IoCompleteRequest (pIrp, IO_NO_INCREMENT);
return (ntStatus);
}

Does exactly that. As Bill Wandel noted, the OP doesn’t appear to
have set IoStatus.Status.

Mark.

11

http://blogs.msdn.com/b/doronh/archive/2006/12/12/how-to-return-the-number-of-bytes-required-for-a-subsequent-operation.aspx

d

From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of Abdul Qader
Sent: Tuesday, September 21, 2010 7:50 AM
To: Windows System Software Devs Interest List
Subject: Re: [ntdev] Failing an IRP

Mark,

Even on returning such a status I find that the IoManager updates the Information with value 148.

Here is my code:

case IOCTL_KEYBOARD_TRANSFER_DATA:
/*Buffered IO: */
if(NULL == Irp->AssociatedIrp.SystemBuffer)
{
DebugPrint((“System Buffer NULL”));
Irp->IoStatus.Information = 4 ;
status = STATUS_BUFFER_TOO_SMALL;
}
else
{
…extra code
}

break;
}

if (!NT_SUCCESS(status)) {
IoCompleteRequest(Irp, IO_NO_INCREMENT);
return status;
}

On debugging I find Irp->IoStatus.Information = 148 once IoCompleteRequest is called.

Thanks
Abdul
— NTDEV is sponsored by OSR For our schedule of WDF, WDM, debugging and other seminars visit: http://www.osr.com/seminars To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer

12

>what is “ba w4”?

WinDbg’s “set a breakpoint for writes” command.

–
Maxim S. Shatskih
Windows DDK MVP
xxxxx@storagecraft.com
http://www.storagecraft.com

13

Thanks Doron…

Cool blog you got there…

On Tue, Sep 21, 2010 at 10:20 PM, Doron Holan wrote:

>
> http://blogs.msdn.com/b/doronh/archive/2006/12/12/how-to-return-the-number-of-bytes-required-for-a-subsequent-operation.aspx
>
>
>
> d
>
>
>
> From: xxxxx@lists.osr.com [mailto:
> xxxxx@lists.osr.com] *On Behalf Of *Abdul Qader
> Sent: Tuesday, September 21, 2010 7:50 AM
>
> To: Windows System Software Devs Interest List
> Subject: Re: [ntdev] Failing an IRP
>
>
>
> Mark,
>
>
>
> Even on returning such a status I find that the IoManager updates the
> Information with value 148.
>
>
>
> Here is my code:
>
>
>
> case IOCTL_KEYBOARD_TRANSFER_DATA:
>
> /*Buffered IO: */
>
> if(NULL == Irp->AssociatedIrp.SystemBuffer)
>
> {
>
> DebugPrint((“System Buffer NULL”));
>
> Irp->IoStatus.Information = 4 ;
>
> status = STATUS_BUFFER_TOO_SMALL;
>
> }
>
> else
>
> {
>
> …extra code
>
> }
>
>
>
>
>
>
>
> break;
>
> }
>
>
>
> if (!NT_SUCCESS(status)) {
>
> IoCompleteRequest(Irp, IO_NO_INCREMENT);
>
> return status;
>
> }
>
>
>
>
>
> On debugging I find Irp->IoStatus.Information = 148 once IoCompleteRequest
> is called.
>
>
>
> Thanks
>
> Abdul
>
> — NTDEV is sponsored by OSR For our schedule of WDF, WDM, debugging and
> other seminars visit: http://www.osr.com/seminars To unsubscribe, visit
> the List Server section of OSR Online at
> http://www.osronline.com/page.cfm?name=ListServer
>
> —
> NTDEV is sponsored by OSR
>
> For our schedule of WDF, WDM, debugging and other seminars visit:
> http://www.osr.com/seminars
>
> To unsubscribe, visit the List Server section of OSR Online at
> http://www.osronline.com/page.cfm?name=ListServer
>