Hi everyone
I’ve seen people mentioning a online version of this book a few times in the archives here. I understand if you don’t want to post a link to it here, but after spending many hours searching, the only copy of this book I can find is a really bad scanned copy with like every 4th line missing…
Anyone know if there is a decent rip out there so I’m not searching in vain?
Thanks in advance
/Johan Bj?rk
Why would you want to steal someone’s work? Write the publisher and ask
for a copy. If enough people do it, they will print more and might even
work to get it updated. O’Reilly is the publisher.
“Johan Björk” wrote in message news:xxxxx@ntfsd…
Hi everyone
I’ve seen people mentioning a online version of this book a few times in
the archives here. I understand if you don’t want to post a link to it
here, but after spending many hours searching, the only copy of this
book I can find is a really bad scanned copy with like every 4th line
missing…
Anyone know if there is a decent rip out there so I’m not searching in
vain?
Thanks in advance
/Johan Björk
just for your FYI the ISBN no is 1-56592-249-2.
Regards,
Vikramaditya.
“Life is not measured by the number of breaths we take, but by the moments
that take our breath away”
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com]On Behalf Of David J. Craig
Sent: Friday, January 02, 2004 11:30 AM
To: Windows File Systems Devs Interest List
Subject: Re:[ntfsd] Windows file system internals - Rajeev Nagar
Why would you want to steal someone’s work? Write the publisher and ask
for a copy. If enough people do it, they will print more and might even
work to get it updated. O’Reilly is the publisher.
“Johan Bj?rk” wrote in message news:xxxxx@ntfsd…
Hi everyone
I’ve seen people mentioning a online version of this book a few times in
the archives here. I understand if you don’t want to post a link to it
here, but after spending many hours searching, the only copy of this
book I can find is a really bad scanned copy with like every 4th line
missing…
Anyone know if there is a decent rip out there so I’m not searching in
vain?
Thanks in advance
/Johan Bj?rk
—
Questions? First check the IFS FAQ at
https://www.osronline.com/article.cfm?id=17
You are currently subscribed to ntfsd as: xxxxx@tacitnetworks.com
To unsubscribe send a blank email to xxxxx@lists.osr.com
Hi all,
don’t sue me, BUT I had the same problem a while ago. After finding a “used
copy” at a price of more than $200 I decided to write an email to Mr. Nagar
(whose email address I happily found using Google). I asked him wether and
where I could find a copy of his book at a normal price (I am still a student
and cannot afford just $200 for a single book of medium size). I also mentioned
the “online copy” (probably the same one Johan stumbled about, it is very
bad quality). He replied I should use the online copy since O’Reilly is not
willing to re-release (or even review) the book for a further edition.
It’s still called a standard work in the area of NTFSD-Development, but it’s
not worth a dime if you can’t find it.
BTW: I would not call it stealing if you get an electronic version of some
book - you cannot even read it everywhere you want … you are bound to some
computer to read it. That’s why I try to get printed copies for all the books
I like (and which I can afford).
Oliver
May the source be with you, stranger 
Contacts / Kontakte
eMail: xxxxx@gmx.net|info|de|com
ICQ UIN #281645
http://assarbad.org & http://assarbad.net
You can buy it on Amazon pretty easily…but wow, it costs $200+, up to
$450! Wow.
You could also just BUY a used copy. There are a bunch available on
Amazon.com, Bibliofind, and other used book sellers. (However, expect to
pay what the book is worth!)
> You could also just BUY a used copy. There are a bunch available on
Amazon.com, Bibliofind, and other used book sellers. (However, expect to
pay what the book is worth!)
Sorry, even if I import an 150 year old book from Israel it is 8 times
cheaper than the $200 at Amazon (including shipping!). BTW: I found only one
USED(!!!) copy at the time when I searched and this used copy was $200. No
question, this price is connected with demand (and a very high demand already if you
ask me), but this price of $200 is not only unrealistic but unfair. I
believe O’reilly is already aware of it - but don’t ask me why they don’t publish
another edition of it (maybe some troubles between author and publisher but
the publisher blocks any chance for the author to find another publisher?!).
And furthermore if you are familiar with some basics of economics you should
know that “what the book is worth” is a very relative measure. I guess some
companies would even by it at $2000 if they get only one copy … but I want
to support the author when buying a book, not the “broker” who re-sells it at
tremendous prices.
Oliver
May the source be with you, stranger
Contacts / Kontakte
eMail: xxxxx@gmx.net|info|de|com
ICQ UIN #281645
http://assarbad.org & http://assarbad.net
The most probably reason why they don’t publish it again is that it needs a lot of
editing.
Remember that the book does not cover Windows 2000 and above features - a lot of
which FS driver writers need to be aware of.
It’s a good reference for starters. The current IFS Kit is, IMO, worth more than the
book - in terms of material not the price.
Regards, Dejan.
Oliver Schneider wrote:
Sorry, even if I import an 150 year old book from Israel it is 8 times
cheaper than the $200 at Amazon (including shipping!). BTW: I found only one
USED(!!!) copy at the time when I searched and this used copy was $200. No
question, this price is connected with demand (and a very high demand already if you
ask me), but this price of $200 is not only unrealistic but unfair. I
believe O’reilly is already aware of it - but don’t ask me why they don’t publish
another edition of it (maybe some troubles between author and publisher but
the publisher blocks any chance for the author to find another publisher?!).
And furthermore if you are familiar with some basics of economics you should
know that “what the book is worth” is a very relative measure. I guess some
companies would even by it at $2000 if they get only one copy … but I want
to support the author when buying a book, not the “broker” who re-sells it at
tremendous prices.Oliver
–
May the source be with you, stranger
Contacts / Kontakte
eMail: xxxxx@gmx.net|info|de|com
ICQ UIN #281645
http://assarbad.org & http://assarbad.net
Questions? First check the IFS FAQ at https://www.osronline.com/article.cfm?id=17
You are currently subscribed to ntfsd as: xxxxx@alfasp.com
To unsubscribe send a blank email to xxxxx@lists.osr.com
–
Kind regards, Dejan M. MVP for DDK
http://www.alfasp.com E-mail: xxxxx@alfasp.com
Alfa Transparent File Encryptor - Transparent file encryption services.
Alfa File Protector - File protection and hiding library for Win32 developers.
Alfa File Monitor - File monitoring library for Win32 developers.
> Remember that the book does not cover Windows 2000 and above features - a
lot of
which FS driver writers need to be aware of.
w2k added nearly nothing new to FSDs, only the w2k3’s Filter Manager was a real
new strain there.
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
FastIo. Or was that XP?
w2k added nearly nothing new to FSDs, only the w2k3’s Filter Manager was a real new
strain there.
–
Kind regards, Dejan M. MVP for DDK
http://www.alfasp.com E-mail: xxxxx@alfasp.com
Alfa Transparent File Encryptor - Transparent file encryption services.
Alfa File Protector - File protection and hiding library for Win32 developers.
Alfa File Monitor - File monitoring library for Win32 developers.
I’m sure Microsoft would provide help to anyone who wants to write a new
book on the subject (current IFS architecture plus info on Filter
Manager), but given that LongHorn is pushing 2007 now, I unfortunately
don’t see anyone rushing to do this.
I’ve just started developing on Linux BTW; boy, what a breath of fresh
air it is to have the source code right in front of you. I can see how
it might make sense to claim that somebody developing in VB or COM does
not require access to the source code behind their tools or libraries,
but for kernel development there is no excuse. The number of kernel
developers is far smaller than the number of VB, COM, or Win32
developers, making comprehensive support and documentation less
cost-effective, and the kernel-mode development environment is much more
tightly bound to the operating system, making it that much more
important to know how the OS is behaving, not just what the inputs and
outputs are for a particular API.
Maxim S. Shatskih wrote:
> Remember that the book does not cover Windows 2000 and above features - a
lot of
>which FS driver writers need to be aware of.
w2k added nearly nothing new to FSDs, only the w2k3’s Filter Manager was a real
new strain there.Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
–
Nick Ryan
(Microsoft Windows MVP for DDK)
If the ‘help’ that Microsoft would provide is the source code to the
file system ‘stuff’ and the storage stacks, it might be possible to do
something more than the current IFS Kit provides. Without that it is
very difficult. The other problem is that would be a lot of source code
when you have NT4, 2000, XP, 2003, and LongHorn with all service packs
that had changes. OSR knows the file system very well, but I don’t
think they want to waste the time for that little money.
Just the analysis would be a major undertaking and from what I have
heard writing a book doesn’t pay very well. Very few authors have
continued to write updates to their books. In fact, the only author I
can think of is Walter Oney who has continued to write books, but not
very many - 95, WDM #1, and WDM #2. In the application arena I have
seen several books by the same author as new versions of the OS have
been released, but kernel stuff if fairly rare. If you work for two
years and only get a $50,000 advance on royalties of $2 per book that
means you have to sell 25,000 to get more than your advance. A kernel
book that sells 50,000 copies would be a rare bird, especially how I see
so many people asking for places to download a copy of this or that. Or
the samples from some book - I used to see that one a lot. If we could
see most of the file systems and filters, I bet we would be surprised to
see how many are based upon an old version of FileMon when the source
was available.
Compared to the time when Nagar’s book was released, the current IFS Kit
is a big help, except for NTFS. I haven’t looked at the NTFS drivers
available with ‘other’ operating systems, but I keep using FAT32 just
because documentation on NTFS is not available. That is one reason why
I wrote a format utility that doesn’t care about the 32GB limit imposed
by Windows. I may have a place to sell it as it handles SmartMedia as
Toshiba recommends. I am going to try and keep the rights to it so I
can sell it to others, since the contract is mostly for a support
utility. Doing it was an interesting experience, but I still have a
little more to do to finish it. I still hate GUI, but I did learn a lot
about that area too. Context sensitive help can be so much fun.
“Nick Ryan” wrote in message news:xxxxx@ntfsd…
> I’m sure Microsoft would provide help to anyone who wants to write a
new
> book on the subject (current IFS architecture plus info on Filter
> Manager), but given that LongHorn is pushing 2007 now, I unfortunately
> don’t see anyone rushing to do this.
>
> I’ve just started developing on Linux BTW; boy, what a breath of fresh
> air it is to have the source code right in front of you. I can see how
> it might make sense to claim that somebody developing in VB or COM
does
> not require access to the source code behind their tools or libraries,
> but for kernel development there is no excuse. The number of kernel
> developers is far smaller than the number of VB, COM, or Win32
> developers, making comprehensive support and documentation less
> cost-effective, and the kernel-mode development environment is much
more
> tightly bound to the operating system, making it that much more
> important to know how the OS is behaving, not just what the inputs and
> outputs are for a particular API.
>
> Maxim S. Shatskih wrote:
> >> Remember that the book does not cover Windows 2000 and above
features - a
> >
> > lot of
> >
> >>which FS driver writers need to be aware of.
> >
> >
> > w2k added nearly nothing new to FSDs, only the w2k3’s Filter Manager
was a real
> > new strain there.
> >
> > Maxim Shatskih, Windows DDK MVP
> > StorageCraft Corporation
> > xxxxx@storagecraft.com
> > http://www.storagecraft.com
> >
> >
>
> –
> Nick Ryan
> (Microsoft Windows MVP for DDK)
>
I don’t think anybody writes computer books to make money, they do it to
enhance their reputation in the industry. Which in fact may make them
more money indirectly as a result of being able to charge higher rates
or get a better job (or simply to get a more satisfactory job). Whether
this is worth a year or so of time spent on a book is a tradeoff the
potential author must consider himself.
David J. Craig wrote:
If the ‘help’ that Microsoft would provide is the source code to the
file system ‘stuff’ and the storage stacks, it might be possible to do
something more than the current IFS Kit provides. Without that it is
very difficult. The other problem is that would be a lot of source code
when you have NT4, 2000, XP, 2003, and LongHorn with all service packs
that had changes. OSR knows the file system very well, but I don’t
think they want to waste the time for that little money.Just the analysis would be a major undertaking and from what I have
heard writing a book doesn’t pay very well. Very few authors have
continued to write updates to their books. In fact, the only author I
can think of is Walter Oney who has continued to write books, but not
very many - 95, WDM #1, and WDM #2. In the application arena I have
seen several books by the same author as new versions of the OS have
been released, but kernel stuff if fairly rare. If you work for two
years and only get a $50,000 advance on royalties of $2 per book that
means you have to sell 25,000 to get more than your advance. A kernel
book that sells 50,000 copies would be a rare bird, especially how I see
so many people asking for places to download a copy of this or that. Or
the samples from some book - I used to see that one a lot. If we could
see most of the file systems and filters, I bet we would be surprised to
see how many are based upon an old version of FileMon when the source
was available.Compared to the time when Nagar’s book was released, the current IFS Kit
is a big help, except for NTFS. I haven’t looked at the NTFS drivers
available with ‘other’ operating systems, but I keep using FAT32 just
because documentation on NTFS is not available. That is one reason why
I wrote a format utility that doesn’t care about the 32GB limit imposed
by Windows. I may have a place to sell it as it handles SmartMedia as
Toshiba recommends. I am going to try and keep the rights to it so I
can sell it to others, since the contract is mostly for a support
utility. Doing it was an interesting experience, but I still have a
little more to do to finish it. I still hate GUI, but I did learn a lot
about that area too. Context sensitive help can be so much fun.“Nick Ryan” wrote in message news:xxxxx@ntfsd…
>
>>I’m sure Microsoft would provide help to anyone who wants to write a
>
> new
>
>>book on the subject (current IFS architecture plus info on Filter
>>Manager), but given that LongHorn is pushing 2007 now, I unfortunately
>>don’t see anyone rushing to do this.
>>
>>I’ve just started developing on Linux BTW; boy, what a breath of fresh
>>air it is to have the source code right in front of you. I can see how
>>it might make sense to claim that somebody developing in VB or COM
>
> does
>
>>not require access to the source code behind their tools or libraries,
>>but for kernel development there is no excuse. The number of kernel
>>developers is far smaller than the number of VB, COM, or Win32
>>developers, making comprehensive support and documentation less
>>cost-effective, and the kernel-mode development environment is much
>
> more
>
>>tightly bound to the operating system, making it that much more
>>important to know how the OS is behaving, not just what the inputs and
>>outputs are for a particular API.
>>
>>Maxim S. Shatskih wrote:
>>
>>>> Remember that the book does not cover Windows 2000 and above
>
> features - a
>
>>>lot of
>>>
>>>
>>>>which FS driver writers need to be aware of.
>>>
>>>
>>>w2k added nearly nothing new to FSDs, only the w2k3’s Filter Manager
>
> was a real
>
>>>new strain there.
>>>
>>>Maxim Shatskih, Windows DDK MVP
>>>StorageCraft Corporation
>>>xxxxx@storagecraft.com
>>>http://www.storagecraft.com
>>>
>>>
>>
>>–
>>Nick Ryan
>>(Microsoft Windows MVP for DDK)
>>
>
>
>
>
–
Nick Ryan
(Microsoft Windows MVP for DDK)
> If the ‘help’ that Microsoft would provide is the source code to the
file system ‘stuff’ and the storage stacks, it might be possible to do
something more than the current IFS Kit provides.
At least a clear documentation on how the FCB locks are used would be good.
The suggestion in Rajeev’s book was a bit contradicting to FASTFAT source IIRC.
Also - why use ValidDataLength at all? Why not mandate it to be always equal to
EOF? Handling sparse tails? Sorry, but if the FSD supports sparse tails, then
it supports sparse files in general, in the middle of the file too, and this
support must be done by means of low-level IO routines and not the VDL.
So, why ever use VDL? Why not mandate it to be == EOF? Will it cause any
undesirable effects on Cache Manager?
How the Cache Manager handles the space which is >= VDL and < EOF?
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
Neither w2k nor XP added anything to FAST_IO_DISPATCH. It is the same as
NT4 had.
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
----- Original Message -----
From: “Dejan Maksimovic”
To: “Windows File Systems Devs Interest List”
Sent: Monday, January 05, 2004 3:15 AM
Subject: Re: [ntfsd] Windows file system internals - Rajeev Nagar
>
> FastIo. Or was that XP?
>
> > w2k added nearly nothing new to FSDs, only the w2k3’s Filter Manager was a
real new
> > strain there.
>
> –
> Kind regards, Dejan M. MVP for DDK
> http://www.alfasp.com E-mail: xxxxx@alfasp.com
> Alfa Transparent File Encryptor - Transparent file encryption services.
> Alfa File Protector - File protection and hiding library for Win32
developers.
> Alfa File Monitor - File monitoring library for Win32 developers.
>
>
>
> —
> Questions? First check the IFS FAQ at
https://www.osronline.com/article.cfm?id=17
>
> You are currently subscribed to ntfsd as: xxxxx@storagecraft.com
> To unsubscribe send a blank email to xxxxx@lists.osr.com
Valid data length makes sense in a file system where you forcibly zero
out the regions of the file on disk. While you won’t notice much
difference for small or medium sized files, you will see a difference
for large files.
Also, you need to keep the historical perspective here (something that
sometimes gets lost). Compression did not show up in the original
version of NTFS (I believe it materialized in NT 3.51) and sparse file
support did not show up until Windows 2000.
Thus, Valid Data Length makes perfect sense for a file system that
supports large files (e.g., NTFS) in the absence of compression (which
eliminates the zero-filled regions anyway) or sparse file support (which
encodes them as part of the file layout description). Even now you can
observe this: create a large NTFS file system (I’d recommend > 1 TB to
really get the full effect here) and then create a large NTFS file
(again, 1 TB will give you a good sense of this) by setting the EOF to
be 1TB. So long as you don’t write data out to 1TB you won’t see a long
pause.
Now repeat this experiment but instead of moving EOF, write a single
byte at the 1TB location. Be prepared for a long pause (at a rate of
40MB/s you’d expect to take at least 7.5 hours to do this) as NTFS zeros
out all of the blocks on the disk up to that last single byte location.
In 1990, 1 TB was a huge amount of data storage (we were using “big” 1GB
disks at the time) and probably near the upper end of their design
center (I know that was the case for the journaling file system I helped
design and implement in that time period). But everyone knew disk
subsystems were slow, and demand-zeroing out large regions of files was
a performance bottleneck (we actually used a different technique for
ensuring zero-filled regions specifically to avoid the same hit because
otherwise it shows up on certain benchmarks.)
I very seldom find decisions in Windows that don’t make sense when
considered in their original context - and generally when they don’t
make sense to me it is because I do not understand the context in which
the design decision was made.
Regards,
Tony
Tony Mason
Consulting Partner
OSR Open Systems Resources, Inc.
http://www.osr.com
Hope to see you at the next OSR file systems class in Boston, MA,
February 23, 2003!
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Maxim S. Shatskih
Sent: Monday, January 05, 2004 5:48 AM
To: ntfsd redirect
Subject: Re: Re:[ntfsd] Windows file system internals - Rajeev Nagar
If the ‘help’ that Microsoft would provide is the source code to the
file system ‘stuff’ and the storage stacks, it might be possible to do
something more than the current IFS Kit provides.
At least a clear documentation on how the FCB locks are used would be
good.
The suggestion in Rajeev’s book was a bit contradicting to FASTFAT
source IIRC.
Also - why use ValidDataLength at all? Why not mandate it to be always
equal to
EOF? Handling sparse tails? Sorry, but if the FSD supports sparse tails,
then
it supports sparse files in general, in the middle of the file too, and
this
support must be done by means of low-level IO routines and not the VDL.
So, why ever use VDL? Why not mandate it to be == EOF? Will it cause any
undesirable effects on Cache Manager?
How the Cache Manager handles the space which is >= VDL and < EOF?
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
Questions? First check the IFS FAQ at
https://www.osronline.com/article.cfm?id=17
You are currently subscribed to ntfsd as: xxxxx@osr.com
To unsubscribe send a blank email to xxxxx@lists.osr.com
I thought "MVP"s got access to some source, including
NTFS. At least I read that MS was going to start
doing this. No?
— Nick Ryan wrote:
> what
> a breath of fresh
> air it is to have the source code right in front of
> you.
__________________________________
Do you Yahoo!?
Find out what made the Top Yahoo! Searches of 2003
http://search.yahoo.com/top2003
Possibly, but Microsoft keeps delaying implementation of the program.
Also many people have many concerns over the legal terms surrounding the
program, but we can’t say more due to NDA.
Randy Cook wrote:
I thought "MVP"s got access to some source, including
NTFS. At least I read that MS was going to start
doing this. No?— Nick Ryan wrote:
>
>>what
>>a breath of fresh
>>air it is to have the source code right in front of
>>you.
>
>
> __________________________________
> Do you Yahoo!?
> Find out what made the Top Yahoo! Searches of 2003
> http://search.yahoo.com/top2003
>
–
Nick Ryan
(Microsoft Windows MVP for DDK)
Ahh, the filters were touched by this - if needed.
“Maxim S. Shatskih” wrote:
Neither w2k nor XP added anything to FAST_IO_DISPATCH. It is the same as NT4
had.
–
Kind regards, Dejan M. MVP for DDK
http://www.alfasp.com E-mail: xxxxx@alfasp.com
Alfa Transparent File Encryptor - Transparent file encryption services.
Alfa File Protector - File protection and hiding library for Win32 developers.
Alfa File Monitor - File monitoring library for Win32 developers.
If my observations are correct then, valid data length and EOF use to have
same values till Windows 2000. There was no API ( Win32 or Nt)to set VDL and
MS provided Win32 API and NT API( IFS KIT) for setting EOF. Please correct
me, if I am wrong. One more thing, when one has set the EOF, FS fill up zero
and there was no way to avoid this.
With Window XP and Windows2003, these two values are used in a smarter ways.
in the same ways as Tony describes in his mail). Now, there are APIs to set
the VDL.
My understanding is that motivation comes from SAN. One of the usage is
following:
IBM and Veritas have implemented File Servers which are different than
typical file servers based NTFS & Windows technology. Instead of
transferring the data on LAN, client commits the data directly on physical
location bypassing the FS layer and FS server is in fact “just” the meta
data server.
Setting up VDL, without zeroing the regions, also has some utility for
backup & restore vendors. Think of server-less restores.
Regards
Pash
-----Original Message-----
From: Tony Mason [mailto:xxxxx@osr.com]
Sent: Monday, January 05, 2004 7:02 AM
To: Windows File Systems Devs Interest List
Subject: RE: Re:[ntfsd] Windows file system internals - Rajeev Nagar
Valid data length makes sense in a file system where you forcibly zero
out the regions of the file on disk. While you won’t notice much
difference for small or medium sized files, you will see a difference
for large files.
Also, you need to keep the historical perspective here (something that
sometimes gets lost). Compression did not show up in the original
version of NTFS (I believe it materialized in NT 3.51) and sparse file
support did not show up until Windows 2000.
Thus, Valid Data Length makes perfect sense for a file system that
supports large files (e.g., NTFS) in the absence of compression (which
eliminates the zero-filled regions anyway) or sparse file support (which
encodes them as part of the file layout description). Even now you can
observe this: create a large NTFS file system (I’d recommend > 1 TB to
really get the full effect here) and then create a large NTFS file
(again, 1 TB will give you a good sense of this) by setting the EOF to
be 1TB. So long as you don’t write data out to 1TB you won’t see a long
pause.
Now repeat this experiment but instead of moving EOF, write a single
byte at the 1TB location. Be prepared for a long pause (at a rate of
40MB/s you’d expect to take at least 7.5 hours to do this) as NTFS zeros
out all of the blocks on the disk up to that last single byte location.
In 1990, 1 TB was a huge amount of data storage (we were using “big” 1GB
disks at the time) and probably near the upper end of their design
center (I know that was the case for the journaling file system I helped
design and implement in that time period). But everyone knew disk
subsystems were slow, and demand-zeroing out large regions of files was
a performance bottleneck (we actually used a different technique for
ensuring zero-filled regions specifically to avoid the same hit because
otherwise it shows up on certain benchmarks.)
I very seldom find decisions in Windows that don’t make sense when
considered in their original context - and generally when they don’t
make sense to me it is because I do not understand the context in which
the design decision was made.
Regards,
Tony
Tony Mason
Consulting Partner
OSR Open Systems Resources, Inc.
http://www.osr.com
Hope to see you at the next OSR file systems class in Boston, MA,
February 23, 2003!
-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of Maxim S. Shatskih
Sent: Monday, January 05, 2004 5:48 AM
To: ntfsd redirect
Subject: Re: Re:[ntfsd] Windows file system internals - Rajeev Nagar
If the ‘help’ that Microsoft would provide is the source code to the
file system ‘stuff’ and the storage stacks, it might be possible to do
something more than the current IFS Kit provides.
At least a clear documentation on how the FCB locks are used would be
good.
The suggestion in Rajeev’s book was a bit contradicting to FASTFAT
source IIRC.
Also - why use ValidDataLength at all? Why not mandate it to be always
equal to
EOF? Handling sparse tails? Sorry, but if the FSD supports sparse tails,
then
it supports sparse files in general, in the middle of the file too, and
this
support must be done by means of low-level IO routines and not the VDL.
So, why ever use VDL? Why not mandate it to be == EOF? Will it cause any
undesirable effects on Cache Manager?
How the Cache Manager handles the space which is >= VDL and < EOF?
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com
Questions? First check the IFS FAQ at
https://www.osronline.com/article.cfm?id=17
You are currently subscribed to ntfsd as: xxxxx@osr.com
To unsubscribe send a blank email to xxxxx@lists.osr.com
Questions? First check the IFS FAQ at
https://www.osronline.com/article.cfm?id=17
You are currently subscribed to ntfsd as: xxxxx@legato.com
To unsubscribe send a blank email to xxxxx@lists.osr.com