[Acquire|Release]ForModWrite callbacks

Hi All,

I have tried setting these pseudo-Fast I/O callbacks in my file-system
driver
(for Windows 2000) but never see them getting invoked by the
modified page-writer. I have simply set them in the fast-io-dispatch table
(with proper sizeof(table)). I am not setting any Flags in the
COMMON_FCB_HEADER.

Am I missing something?

Thanks,
Vipul.

Vipul,

The mechanism you describe is correct. That your entry is not being called
suggests that either your file is not being flushed by the modified page
writer or that you have not set up the header properly.

Can you describe the steps you are taking to test this path? Perhaps you
have a problem in the implementation of the test.

Regards,

Tony

Tony Mason
Consulting Partner
OSR Open Systems Resources, Inc.
http://www.osr.com

-----Original Message-----
From: Vipul [mailto:xxxxx@yahoo.com]
Sent: Tuesday, June 17, 2003 1:22 PM
To: File Systems Developers
Subject: [ntfsd] [Acquire|Release]ForModWrite callbacks

Hi All,

I have tried setting these pseudo-Fast I/O callbacks in my file-system
driver
(for Windows 2000) but never see them getting invoked by the
modified page-writer. I have simply set them in the fast-io-dispatch table
(with proper sizeof(table)). I am not setting any Flags in the
COMMON_FCB_HEADER.

Am I missing something?

Thanks,
Vipul.


You are currently subscribed to ntfsd as: xxxxx@osr.com
To unsubscribe send a blank email to xxxxx@lists.osr.com

Hi Tony,

I was trying out simple (various sizes) memory-mapped writes.

I also don’t think there is any problem with the sizeof(table) since
Acquire/Release
ForCcFlush do get called (and they are at larger offsets than the callbacks
for ModWrite).

I am also setting IsFastIoPossible=FastIoIsNotPossible in the header, since
we
don’t support “real” fast i/o. Could that be the culprit? Apart from that,
I’m
initializing the Main and PagingIo resources. Rest is all zeroed.

“Tony Mason” wrote in message news:xxxxx@ntfsd…
>
> Vipul,
>
> The mechanism you describe is correct. That your entry is not being
called
> suggests that either your file is not being flushed by the modified page
> writer or that you have not set up the header properly.
>
> Can you describe the steps you are taking to test this path? Perhaps you
> have a problem in the implementation of the test.
>
> Regards,
>
> Tony
>
> Tony Mason
> Consulting Partner
> OSR Open Systems Resources, Inc.
> http://www.osr.com
>
>
> -----Original Message-----
> From: Vipul [mailto:xxxxx@yahoo.com]
> Sent: Tuesday, June 17, 2003 1:22 PM
> To: File Systems Developers
> Subject: [ntfsd] [Acquire|Release]ForModWrite callbacks
>
> Hi All,
>
> I have tried setting these pseudo-Fast I/O callbacks in my file-system
> driver
> (for Windows 2000) but never see them getting invoked by the
> modified page-writer. I have simply set them in the fast-io-dispatch
table
> (with proper sizeof(table)). I am not setting any Flags in the
> COMMON_FCB_HEADER.
>
> Am I missing something?
>
> Thanks,
> Vipul.
>
>
>
> —
> You are currently subscribed to ntfsd as: xxxxx@osr.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
>

Vipul,

OK, so what are you doing that would kick off the modified page writer?
Surely you don’t think it will start writing pages based upon some TIME
based algorithm, right?

If you aren’t exhausting the pool of clean pages, the modified page writer
will happily allow your dirty data to sit in memory until the end of
eternity.

DIRTY MORE MEMORY - that will trigger the modified page writer.

Regards,

Tony

Tony Mason
Consulting Partner
OSR Open Systems Resources, Inc.
http://www.osr.com

-----Original Message-----
From: Vipul [mailto:xxxxx@yahoo.com]
Sent: Tuesday, June 17, 2003 2:46 PM
To: File Systems Developers
Subject: [ntfsd] Re: [Acquire|Release]ForModWrite callbacks

Hi Tony,

I was trying out simple (various sizes) memory-mapped writes.

I also don’t think there is any problem with the sizeof(table) since
Acquire/Release
ForCcFlush do get called (and they are at larger offsets than the callbacks
for ModWrite).

I am also setting IsFastIoPossible=FastIoIsNotPossible in the header, since
we
don’t support “real” fast i/o. Could that be the culprit? Apart from that,
I’m
initializing the Main and PagingIo resources. Rest is all zeroed.

“Tony Mason” wrote in message news:xxxxx@ntfsd…
>
> Vipul,
>
> The mechanism you describe is correct. That your entry is not being
called
> suggests that either your file is not being flushed by the modified page
> writer or that you have not set up the header properly.
>
> Can you describe the steps you are taking to test this path? Perhaps you
> have a problem in the implementation of the test.
>
> Regards,
>
> Tony
>
> Tony Mason
> Consulting Partner
> OSR Open Systems Resources, Inc.
> http://www.osr.com
>
>
> -----Original Message-----
> From: Vipul [mailto:xxxxx@yahoo.com]
> Sent: Tuesday, June 17, 2003 1:22 PM
> To: File Systems Developers
> Subject: [ntfsd] [Acquire|Release]ForModWrite callbacks
>
> Hi All,
>
> I have tried setting these pseudo-Fast I/O callbacks in my file-system
> driver
> (for Windows 2000) but never see them getting invoked by the
> modified page-writer. I have simply set them in the fast-io-dispatch
table
> (with proper sizeof(table)). I am not setting any Flags in the
> COMMON_FCB_HEADER.
>
> Am I missing something?
>
> Thanks,
> Vipul.
>
>
>
> —
> You are currently subscribed to ntfsd as: xxxxx@osr.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
>
>


You are currently subscribed to ntfsd as: xxxxx@osr.com
To unsubscribe send a blank email to xxxxx@lists.osr.com