OSRLogoOSRLogoOSRLogo x Seminar Ad

Everything Windows Driver Development

GoToHomePage xLoginx

    Thu, 14 Mar 2019     118020 members


  Online Dump Analyzer
OSR Dev Blog
The NT Insider
The Basics
File Systems
ListServer / Forum
  Express Links
  · The NT Insider Digital Edition - May-June 2016 Now Available!
  · Windows 8.1 Update: VS Express Now Supported
  · HCK Client install on Windows N versions
  · There's a WDFSTRING?
  · When CAN You Call WdfIoQueueP...ously

Duplicate Disk Writes

Well, will wonders never cease!  While perusing the KB articles today (Zzzzzzzz...) I noticed an interesting article entitled "Write operations to an external storage device take a long time to complete".  The gist of the article is that Microsoft has finally gotten around to eliminating the duplicate write operations that file systems developers sometimes see.

The problem is due to the fact that the Modified Page Writer, Mapped Page Writer, and Lazy Writer didn't attempt to synchronize their lists of pages to be written. I first saw this problem many years ago, and when I discussed it with a dev team member I remember being told "yeah, that's just the way it works."  No big deal.

Now that there's a proliferation of slow external devices (like, USB key drives and such) this decision has apparently been rethought.

The KB article can be found at http://support.microsoft.com/default.aspx?scid=kb;EN-US;828012 and a hotfix is available.

BTW, this KB article applies to XP Home SP1, XP SP1, and XP 64-bit Version 2002 SP1.

Good news is welcome, no matter WHEN it comes!   Hooray!

Related Articles
A Common Topic Explained - Sharing Memory Between Drivers and Applications
Tracking State and Context - Reference Counting for File System Filter Drivers
Check the Knowledge Base!
Files Opened as a result of a Remote Request
Back to Basics - An Introduction to Transactions

User Comments
Rate this article and give us feedback. Do you find anything missing? Share your opinion with the community!
Post Your Comment

"I need this hotfix"
Dear All,

I'm also suffering by the same problem, my pendrive is very slow on my desktop PC (I can write data to is by 0,2-0,6MB/sec only, but with my notebook is works with 4-5MB/sec!)

I need this hotfix, I hope that someone has it!!! Any URL where I get get this?

(I tried to disable write caching but it did not help at all. Installed the latest USB drivers, etc...)

Thank you

18-Dec-04, Zoltan Zeman

"slow write to USB hotfix"
Dear All,

I'm suffering with the same problem: XP SP1 and all of my USB2 peripheries are very slow, when I write data to them (0,2-0,6MB/sec). The a.m. hotfix would solve my problem, however it is not available anymore :( at MS. They recommend to install SP2, but I don't want to do that, I need this simple and fast hotfix.

Does anyone of you knows an URL where I can download this? Or anyone has it and send me by email?

Thanks for your help!

18-Dec-04, Zoltan Zeman

"I'm betting you're right, but the article limits the scope to XP SP1"
From the end of the KB article . ..

"This problem only affects Windows XP with SP1 installed or integrated. It does not affect Windows XP without SP1 installed, Microsoft Windows 2000, or Microsoft Windows Server 2003."

This implies that the issue they addressed was introduced with XP SP1 and doesn't exists in any other version of Windows. Which means either they're wrong, you're wrong, or the issue you've seen has yet to be addressed. I'm betting on either the first or the last option!

27-May-04, Toby Ovod-Everett

""Re: Duplicate Disk Writes""
Good article. Regardless of who's at fault and why, I appreciate the heads up that articles like this give. As a matter of fact I just bumped into a situation that this info and hotfix will help resolve. Thanks!

27-May-04, Ron Arndt Jr.

"Re: Duplicate Disk Writes"
JD wrote:

> Multiple writes may have been the least of my worries.

Apparently, MS thought so too. They probably focused their efforts on real bugs... like address mapping bits.


27-May-04, Richard Masoner

"Re: Duplicate Disk Writes"
Excuse me while I barf. The fact that there were three "Writers" with seperate page lists that did not have synchronization of a Page Modified bit makes me wonder about the other context bits a-hum... like address mapping bits? Multiple writes may have been the least of my worries. Also the dev team members comment doesn't leave me with much confidence. (S)He could have easily said "It works, don't know why, screw it, I've got more important things to worry about."

How many man hours have been lost? How much wear and tear on equipment. And, something for MS to ponder about, how many sales did it loose due to poor disk benchmarks?


27-May-04, James Dempsey

Post Your Comments.
Print this article.
Email this article.
bottom nav links