Conflicting info between the global Security Bulletin and some SPi Security Bulletin

D

David F

The global detailed Security Bulleting is given in:
http://www.microsoft.com/technet/security/current.aspx

I did find the following problems there.
I normally use Win 2K + SP4.
According the Security Bulletin for the release of SP4,
(http://www.microsoft.com/technet/security/prodtech/windows2000/w2ksp4.mspx)
it contains all new (new - with respect to SP3) patches in the range:
MS01-022 (4.18.2001) till MS03-030 (7.23.2003)
and of course also all prior patches from SP1 to SP3.

But when I looked in that detailed global list, I find for example that as old
MS00--077 (10.13.2000 !) and many other older than MS03-030
should be applied to W2K/SP4. This doesn't make sense.

I found the exact same contradictions with regard to Win XP SP2.
In this case, I found even a worse situation, when patch # MS03-011
showing up in the detailed global list, which according to its date
and patch # should be included in the Security Bulletin list for
WXP/SP2 but it is not!. This list is given in:
http://www.microsoft.com/technet/security/prodtech/windowsXP/xpsp2.mspx


So I don't know what to make of it and how I can rely on that
detailed global list in terms of what to pickup from it and apply to W2K/SP4
(and the same for WXP/SP2).

Any idea?

Thanks,
David
 
R

Roger Abell [MVP]

David F said:
The global detailed Security Bulleting is given in:
http://www.microsoft.com/technet/security/current.aspx

I did find the following problems there.
I normally use Win 2K + SP4.
According the Security Bulletin for the release of SP4,
(http://www.microsoft.com/technet/security/prodtech/windows2000/w2ksp4.mspx)
it contains all new (new - with respect to SP3) patches in the range:
MS01-022 (4.18.2001) till MS03-030 (7.23.2003)
and of course also all prior patches from SP1 to SP3.

But when I looked in that detailed global list, I find for example that as
old
MS00--077 (10.13.2000 !) and many other older than MS03-030
should be applied to W2K/SP4. This doesn't make sense.

I do not see statement that ms00-077 applies to W2k Sp4
Where do you see this? In what you term the Global Sec Bulletin it
states this applies to W2k at Sp1 and Sp2, and if you read into it you
will see that the Sp2 part was added when the specific patch was
reissued to address a further, similar exploit. If you look at the KB
http://support.microsoft.com/?kbid=299796
you will see that the updated patch was first included in SP3

I found the exact same contradictions with regard to Win XP SP2.
In this case, I found even a worse situation, when patch # MS03-011
showing up in the detailed global list, which according to its date
and patch # should be included in the Security Bulletin list for
WXP/SP2 but it is not!. This list is given in:
http://www.microsoft.com/technet/security/prodtech/windowsXP/xpsp2.mspx

ms03-011 is something of a peculiar situation as it deals with the
MS distributed Java VM, which MS had to pull from availability
as part of a settlement with Sun. In fact, this was involved in the
reissue of XP Sp1 as XP Sp1a.
It is possible for even W2k3 to need this patch. It all depends on
the history of the machine. One could install an XP SP2 from an
XP Sp2 CD and not get the VM installed. Later, install of an older
download of some product might install the VM at a level before
3810, causing need for the patch at that point.
Read into the bulletin and you will see there is quite a bit that is
abnormal about this specific patch. Behind the scenes there is a
bit of legal history. One must take the history of a specific machine
into account to determine need, and just being at an OS level mentioned
in the bulletin does not mean one needs the patch (the VM might not be
present).
So I don't know what to make of it and how I can rely on that
detailed global list in terms of what to pickup from it and apply to
W2K/SP4
(and the same for WXP/SP2).

You seem to be going about this the hard way David.
You can either use the Microsoft Update site in Custom mode to
get a list of what a specific machine needs based on what is at
the moment installed; or one can download MBSA and after it
is installed open a cmd prompt at the install directory and issue
mbsacli /?
for syntax, and then use mbsacli with a switch
/n OS+IIS+SQL+Password
in order to only check patches
 
D

David F

Roger Abell said:
I do not see statement that ms00-077 applies to W2k Sp4
Where do you see this? In what you term the Global Sec Bulletin it
states this applies to W2k at Sp1 and Sp2, and if you read into it you
will see that the Sp2 part was added when the specific patch was
reissued to address a further, similar exploit. If you look at the KB
http://support.microsoft.com/?kbid=299796
you will see that the updated patch was first included in SP3

You are right - I copy and paste the wrong patch #. I meant to specify
MS02-032 dated: 6.26.2002, which is listed specifically as already included
in the SP4.
ms03-011 is something of a peculiar situation as it deals with the
MS distributed Java VM, which MS had to pull from availability
as part of a settlement with Sun. In fact, this was involved in the
reissue of XP Sp1 as XP Sp1a.
It is possible for even W2k3 to need this patch. It all depends on

I am confused by "...possible...". This one is ranked "critical" for installing
explicitly on W2K3 (as a matter of fact, on W2K2/3/4). But then again,
like MS02-032, it should have already been included in W2K4. But again,
what is worst, in spite of being released before SP4 was released, I would
expect it to be included in SP4 and me NOT needed to install it on a W2K4.
And like with WXP, it is missing from the Security Bulletin list for SP4.
the history of the machine. One could install an XP SP2 from an
XP Sp2 CD and not get the VM installed. Later, install of an older
download of some product might install the VM at a level before
3810, causing need for the patch at that point.
Read into the bulletin and you will see there is quite a bit that is
abnormal about this specific patch. Behind the scenes there is a
bit of legal history. One must take the history of a specific machine
into account to determine need, and just being at an OS level mentioned
in the bulletin does not mean one needs the patch (the VM might not be
present).

I am confused from the explanation.

My key point is that when I install a new system, I am obviously expected to
install the highest SPi (and for WXP I do have and use indeed the XP SP2 CD
you mentioned to install XP) and then simply add (manually in my case)
each patch listed in that global list that was released after the release of that
SPi, and that that list explicitly says it applies to that SPi.
It is that simply and that straight forward.
Accordingly, I would not expect any patch listed as released before the release
of a specific SPi, being mentioned as still applicable for that SPi.

And in the case of W2K, we have something called Rollup_1_ver_2-for SP4
which is even highier than SP4 (and I did not see an explicit list of what
security patches it contains) and it is not mentioned in that global list.
So the only way I can relate to it is by relaying to the date of the last patch
included as mentioned in its release notes. So I would use any patch listed
in the global list as mentioned above but that its release date is not just highier
than the last one to be included in SP4 but highier than that last date of that
Rollup.

BTW, I also would think that it should not matter, that for the case of XP,
for example, that it would not matter whether one installs the very
original XP he has and add later the downloadable SP2 or use that CD
mentioned above.
You seem to be going about this the hard way David.
You can either use the Microsoft Update site in Custom mode to
get a list of what a specific machine needs based on what is at
the moment installed; or one can download MBSA and after it
is installed open a cmd prompt at the install directory and issue
mbsacli /?
for syntax, and then use mbsacli with a switch
/n OS+IIS+SQL+Password
in order to only check patches
You are right about "the hard way" - to some degree. But - when I looked once
into that MBSA "thing", it looks to me overly cumbersome (like most other
things from MS) and also things being changed all the time.
So I am not sure if in this case it is worth the time invested in it.
I might be wrong of course.
My understanding is that there is something called a "scan file" which is something
like a definition file (like for viruses detection for example), that the executable
MBSA program uses to figure out what my SW product(s) already have and
produces a list of MS0X-0YYs (or similar) I do need still to install. And, if
the executable and "scan file" are installed, it is all dome offline.

It seems that for Office 2K I still have to use MBSA ver. 1.2.1. But I could
NOT find an offline scan file for that ver. and my impression is that this version
works online only - exactly the kind of things I am trying to avoid
Even for MBSA ver 2.0.1, I could not find the scan file, especially the new one.

David
======
Albert Einstein: "Everything should be made as simple as possible, but not simpler."
 
R

Roger Abell [MVP]

David F said:
You are right - I copy and paste the wrong patch #. I meant to specify
MS02-032 dated: 6.26.2002, which is listed specifically as already
included
in the SP4.

That is a Windows Media Player (WMP) patch.
Tell me, what happens in the following scenario . . .
Someone has W2k with WMP 6.4 (which I think is what came with W2k),
and they install SP 4 when it come out.
Then they elect to upgrade to WMP 7.1, but they use a download of the
WMP 7.1 installable obtained before the download was updated to have
the patch of MS02-032 included within it.
So, they have a W2k at SP4 with a vulnerable version of WMP 7.1, no?

You are attempting to make thing too clear-cut; that way does not match
the realities encountered in world.
I am confused by "...possible...". This one is ranked "critical" for
installing
explicitly on W2K3 (as a matter of fact, on W2K2/3/4). But then again,
like MS02-032, it should have already been included in W2K4. But again,

W2k4 ?
Look. At a fixed point in time MS took away from the web all downloads
that included an install of the MS Java VM. This was required by the Sun
lawsuite settlement. That does not mean that people could not use things
obtained before that time to cause it to install.
Hence, it is known that the VM might be installed on certain versions,
that it certainly is installed on some or via specific built histories,
etc..
But, of those, only some can have the statement "the VM is installed"
made with certainty. Hence, how can you expect them to say the patch
is needed with certainty?

You will not find the VM on a fresh W2k3 install nor after SP1.
Yet W2k3 is called out in this bulletin. Why? Because some
histories leading to that instance of a W2k3 build could have
caused the VM to become installed.
what is worst, in spite of being released before SP4 was released, I would
expect it to be included in SP4 and me NOT needed to install it on a W2K4.
And like with WXP, it is missing from the Security Bulletin list for SP4.


I am confused from the explanation.

My key point is that when I install a new system, I am obviously expected
to
install the highest SPi (and for WXP I do have and use indeed the XP SP2
CD
you mentioned to install XP) and then simply add (manually in my case)
each patch listed in that global list that was released after the release
of that
SPi, and that that list explicitly says it applies to that SPi.
It is that simply and that straight forward.
Accordingly, I would not expect any patch listed as released before the
release
of a specific SPi, being mentioned as still applicable for that SPi.

And in the case of W2K, we have something called Rollup_1_ver_2-for SP4
which is even highier than SP4 (and I did not see an explicit list of what
security patches it contains) and it is not mentioned in that global list.
So the only way I can relate to it is by relaying to the date of the last
patch
included as mentioned in its release notes. So I would use any patch
listed
in the global list as mentioned above but that its release date is not
just highier
than the last one to be included in SP4 but highier than that last date of
that
Rollup.

BTW, I also would think that it should not matter, that for the case of
XP,
for example, that it would not matter whether one installs the very
original XP he has and add later the downloadable SP2 or use that CD
mentioned above.

Take the case of the VM patch.
To meet the court mandate MS removed all downloadables the
would install the VM. This included purging the MSDN downloads
for subscribers of W2k with SP4 inlined CD image, etc.
They also released IIRC version 3810 of the VM, which would
install only if the VM was already present on the machine.
Prior to 3810 the VM distributions would update an existing
VM if present or install the VM if not.
Many people wanted to not use the Sun JRE, so you will find
even today downloads of the VM that will install when the VM
is not present, provided only that it is a support OS. So, people
can build a XP SP2, which today will be without the VM, and
then apply one of these older VM distributions (obtained from
third-party archives) and it will install and it will leave the
machine at a vulnerable VM version if they omit downloading
and applying the 3810 VM update.
So, can XP SP2 be in a state that needs the patch? Yes.
Would a XP SP2 installed from a XP SP2 CD need it? No.
Would an XP installed from the original gold XP CD need it? Yes.
Would a gold XP install with SP 2 applied need it? No
Could a gold XP plus SP2 be caused to need it? Perhaps, likely not.
(one could uninstall 3810 and then use outdated download . . .)

Things are just not as clear-cut as you would like.
You are right about "the hard way" - to some degree. But - when I looked
once
into that MBSA "thing", it looks to me overly cumbersome (like most other
things from MS) and also things being changed all the time.
So I am not sure if in this case it is worth the time invested in it.
I might be wrong of course.


If you cannot see the simplicity of mbsacli usage I do not know
how you could be convinced, nor do I wish to try.
My understanding is that there is something called a "scan file" which is
something
like a definition file (like for viruses detection for example), that the
executable
MBSA program uses to figure out what my SW product(s) already have and
produces a list of MS0X-0YYs (or similar) I do need still to install.
And, if
the executable and "scan file" are installed, it is all dome offline.

It seems that for Office 2K I still have to use MBSA ver. 1.2.1. But I
could
NOT find an offline scan file for that ver. and my impression is that this
version
works online only - exactly the kind of things I am trying to avoid

I believe you are correct, that only the old MBSA will detect for O2k
http://support.microsoft.com/kb/895660/en-us
But, IIRC MS will stop providing scan files in the old form as of March,
whence only ver 2.x releases of MBSA will be supported.
Also, is not O2k beyond its service lifetime? so one would not expect
to see further patches soon (like W2k which will soon no longer have
even security patches released to the public).
Originally the MSSecure.xml file used by 1.2.1 was to be discontinued
6 months after 2.0 released, but I believe they extended that (I do not
recall until when)

Even for MBSA ver 2.0.1, I could not find the scan file, especially the
new one.

You are correct that info on where to get the offline file is very,
very thouroughly buried, and since v2 it has no longer been as
simple as grabbing the file from disk on a system attached to the
network after running mbsacli

The download link for the offline WsusScn2.cab file is mentioned in
http://support.microsoft.com/kb/926464
For the more on the change from WsusScan.cab to the new format see
http://support.microsoft.com/kb/927745
 
D

David F

Roger Abell said:
That is a Windows Media Player (WMP) patch.
Tell me, what happens in the following scenario . . .
Someone has W2k with WMP 6.4 (which I think is what came with W2k),
and they install SP 4 when it come out.
Then they elect to upgrade to WMP 7.1, but they use a download of the
WMP 7.1 installable obtained before the download was updated to have
the patch of MS02-032 included within it.
So, they have a W2k at SP4 with a vulnerable version of WMP 7.1, no?

I think the answer is NO - if one behaves reasonably. And here is my rational.



If the downloaded WMP 7.1 is from before the release of MS02-032, then when SP4 would come out
and installed, the WMP would be updated.

If the WMP 7.1 was downloaded after the release of MS02-032, whether before the release of SP4
OR EVEN AFTER the release of SP4, then the WMP should incl. it and the release and installation
of SP4 should have no impact (as far as MS02-032 is concerned).

If there was no updated version of WMP7.1 after MS02-032 release (and actually even if there
was), it is OK for the global list to say that it is applicable to such and such version or
release date or file date of WMP7.1.


Either way, there is no point to say in the global list that MS02-032 applies to W2K4 because it
is not.

Of course, if one downloaded WMP7.1 before MS02-032 released, kept it in the drawer, then
downloads and installs SP4, AND ONLY THEN decided to install that OLD WMP7.1 then it should be
HIS problem - stupidity is indeed costly. It is reasonable to expect that when one installs
something, s/he would use the most recently available version and/or in proper order.

I thought that this is an ABC and elementary logic. Isn't it?

You tell me, am I missing something? Did I leave any scenario uncovered?



But wait, there is more.[J]

MS02-032 was just an example. There are 4 more mentioned in the global list as applicable to
W2K4: MS03-026, MS03-023, MS03-022 & MS02-050 - ALL within the period of time that SP4 is
suppose to cover. The first 3 (MS03-.) though are NOT listed as included in the SP4 release (see
my link above) although they belong there by date according release nots for SP4. MS02-050 (like
MS02-032) IS explicitly incl. in the SP4 release, and unlike MS02-032, it doesn't pertain to
some add-on such as the WMP but purely to the core OS. And there is even more to this MS02-050
patch. It even shows up in the list of patches included with Rollup_1(V2) for SP4 (hereafter
R1V2). Now the release notes for R1V2 say clearly that SP4 must be installed first.



My hunch is that I should not install patches such as MS02-050 on top of SP4 (let alone on top
of R1V2) but what say you in this case?



And what should I make about the other 3? I have no clue whether I should go ahead and add them
on top of SP4 or not (and again, let alone on top of R1V2). They are listed in the R1V2's list
of patches.

See: http://support.microsoft.com/kb/891861



Luckily, nothing in the global list references the R1V2.



All in all, I am in square one of my original thread. I would obviously start with manual
installation of individual patches after installing R1V2 and again the question is whether it is
OK to go by date (in order of course), that is, to pickup from the global list all patches dated
after the last patch included in R1V2's list and ignore anything else.


In general, I agree with you that there could be, for historical, legal, and other reasons some
peculiar scenarios but this doesn't mean that the "Affected Software Service Packs" column, like
any other piece of technical info should not be precise and unequivocal stated.
You are attempting to make thing too clear-cut; that way does not match
the realities encountered in world.


W2k4 ?
Yes.

Look. At a fixed point in time MS took away from the web all downloads
that included an install of the MS Java VM. This was required by the Sun
lawsuite settlement. That does not mean that people could not use things
obtained before that time to cause it to install.
Hence, it is known that the VM might be installed on certain versions,
that it certainly is installed on some or via specific built histories,
etc..
But, of those, only some can have the statement "the VM is installed"
made with certainty. Hence, how can you expect them to say the patch
is needed with certainty?

You will not find the VM on a fresh W2k3 install nor after SP1.
Yet W2k3 is called out in this bulletin. Why? Because some
histories leading to that instance of a W2k3 build could have
caused the VM to become installed.


Take the case of the VM patch.
To meet the court mandate MS removed all downloadables the
would install the VM. This included purging the MSDN downloads
for subscribers of W2k with SP4 inlined CD image, etc.
They also released IIRC version 3810 of the VM, which would
install only if the VM was already present on the machine.
Prior to 3810 the VM distributions would update an existing
VM if present or install the VM if not.
Many people wanted to not use the Sun JRE, so you will find
even today downloads of the VM that will install when the VM
is not present, provided only that it is a support OS. So, people
can build a XP SP2, which today will be without the VM, and
then apply one of these older VM distributions (obtained from
third-party archives) and it will install and it will leave the
machine at a vulnerable VM version if they omit downloading
and applying the 3810 VM update.
So, can XP SP2 be in a state that needs the patch? Yes.
Would a XP SP2 installed from a XP SP2 CD need it? No.
Would an XP installed from the original gold XP CD need it? Yes.
Would a gold XP install with SP 2 applied need it? No
Could a gold XP plus SP2 be caused to need it? Perhaps, likely not.
(one could uninstall 3810 and then use outdated download . . .)

Things are just not as clear-cut as you would like.



If you cannot see the simplicity of mbsacli usage I do not know
how you could be convinced, nor do I wish to try.


I believe you are correct, that only the old MBSA will detect for O2k
http://support.microsoft.com/kb/895660/en-us
But, IIRC MS will stop providing scan files in the old form as of March,
whence only ver 2.x releases of MBSA will be supported.
Also, is not O2k beyond its service lifetime? so one would not expect
to see further patches soon (like W2k which will soon no longer have
even security patches released to the public).

That is OK. I never referred to future expectations but only to what has been
released in the past. And actually, MS promissed to make sure that released
security patches for legacy products will remain available. But even w/o such
a promiss it is OK because one can download and keep them by himself.

I don't jump on every new version MS releases. I do it only on "have to"
basis either because of HW or SW applications requirements. I am fairly
happy with W2K for development (and Office 2K). Others, such as WXP
I use only for testing.
 
R

Roger Abell [MVP]

I really do no know how best to respond to you expect to
say that if it is called out as affected (note that this does not
always mean needed, just that it can be needed) then there
is/are case(s) in which it can be needed.

I have every reason to believe the folks at MSRC know what
they are doing this regard, having been in on email exhanges
previously (nearly monthly) about making sure the released
bulletins are clear and accurate.

Also, keep in mind that it was only somewhere around the
end of 2002 when the patch updater technology started to
see the results from the push for improved strategy and for
unification of the multiple updaters in use. Prior to that time
it was possible to have a patch not take just because it was
applied in the wrong order while applying a specific group
of reboot requiring patches at the same time without reboots.

Anyway, the scanning tools are the best way to determine
what is needed by the current state of a system. Yes, this
does get more difficult for a non-connected system, and yes,
for older, end-of-life products one may have to use a variety
of tools to scan as much as possible. Nevertheless, trying
to wind one's way through the fine details would really
mean looking at the installed binaries on the system and
comparing their versions with the versions called out as
supplied by the patches (in the bulletin or associated KB).
One cannot just say, this is a W2k at SP 4, and since the
high-level statement of affected versions includes that OS
therefore this one needs this. If you really want to do the
decision process you might need to read into the cab files
used in the scans and look at just what detection info they
are keying in on, which you will find in the xml called out
per OS version.

For the VM example you say, well yes that particular install
history would result in a vulnerable system, but it is that
person's fault that they did that. That may be true enough,
but you know where that person will place the blame, right?

There used to be a number of pages on the web that tracked
the patches, trying to organize the info much as you seem to
be attempting. As far as I know, those people have mostly
all stopped keeping their listings updated subsequent to the
release by MS of the scanners. There is a reason, it is just
hard to do better than one gets by running a scan.

Roger

David F said:
Roger Abell said:
That is a Windows Media Player (WMP) patch.
Tell me, what happens in the following scenario . . .
Someone has W2k with WMP 6.4 (which I think is what came with W2k),
and they install SP 4 when it come out.
Then they elect to upgrade to WMP 7.1, but they use a download of the
WMP 7.1 installable obtained before the download was updated to have
the patch of MS02-032 included within it.
So, they have a W2k at SP4 with a vulnerable version of WMP 7.1, no?

I think the answer is NO - if one behaves reasonably. And here is my
rational.



If the downloaded WMP 7.1 is from before the release of MS02-032, then
when SP4 would come out
and installed, the WMP would be updated.

If the WMP 7.1 was downloaded after the release of MS02-032, whether
before the release of SP4
OR EVEN AFTER the release of SP4, then the WMP should incl. it and the
release and installation
of SP4 should have no impact (as far as MS02-032 is concerned).

If there was no updated version of WMP7.1 after MS02-032 release (and
actually even if there
was), it is OK for the global list to say that it is applicable to such
and such version or
release date or file date of WMP7.1.


Either way, there is no point to say in the global list that MS02-032
applies to W2K4 because it
is not.

Of course, if one downloaded WMP7.1 before MS02-032 released, kept it in
the drawer, then
downloads and installs SP4, AND ONLY THEN decided to install that OLD
WMP7.1 then it should be
HIS problem - stupidity is indeed costly. It is reasonable to expect that
when one installs
something, s/he would use the most recently available version and/or in
proper order.

I thought that this is an ABC and elementary logic. Isn't it?

You tell me, am I missing something? Did I leave any scenario uncovered?



But wait, there is more.[J]

MS02-032 was just an example. There are 4 more mentioned in the global
list as applicable to
W2K4: MS03-026, MS03-023, MS03-022 & MS02-050 - ALL within the period of
time that SP4 is
suppose to cover. The first 3 (MS03-.) though are NOT listed as included
in the SP4 release (see
my link above) although they belong there by date according release nots
for SP4. MS02-050 (like
MS02-032) IS explicitly incl. in the SP4 release, and unlike MS02-032, it
doesn't pertain to
some add-on such as the WMP but purely to the core OS. And there is even
more to this MS02-050
patch. It even shows up in the list of patches included with Rollup_1(V2)
for SP4 (hereafter
R1V2). Now the release notes for R1V2 say clearly that SP4 must be
installed first.



My hunch is that I should not install patches such as MS02-050 on top of
SP4 (let alone on top
of R1V2) but what say you in this case?



And what should I make about the other 3? I have no clue whether I should
go ahead and add them
on top of SP4 or not (and again, let alone on top of R1V2). They are
listed in the R1V2's list
of patches.

See: http://support.microsoft.com/kb/891861



Luckily, nothing in the global list references the R1V2.



All in all, I am in square one of my original thread. I would obviously
start with manual
installation of individual patches after installing R1V2 and again the
question is whether it is
OK to go by date (in order of course), that is, to pickup from the global
list all patches dated
after the last patch included in R1V2's list and ignore anything else.


In general, I agree with you that there could be, for historical, legal,
and other reasons some
peculiar scenarios but this doesn't mean that the "Affected Software
Service Packs" column, like
any other piece of technical info should not be precise and unequivocal
stated.
You are attempting to make thing too clear-cut; that way does not match
the realities encountered in world.


W2k4 ?
Yes.

Look. At a fixed point in time MS took away from the web all downloads
that included an install of the MS Java VM. This was required by the Sun
lawsuite settlement. That does not mean that people could not use things
obtained before that time to cause it to install.
Hence, it is known that the VM might be installed on certain versions,
that it certainly is installed on some or via specific built histories,
etc..
But, of those, only some can have the statement "the VM is installed"
made with certainty. Hence, how can you expect them to say the patch
is needed with certainty?

You will not find the VM on a fresh W2k3 install nor after SP1.
Yet W2k3 is called out in this bulletin. Why? Because some
histories leading to that instance of a W2k3 build could have
caused the VM to become installed.


Take the case of the VM patch.
To meet the court mandate MS removed all downloadables the
would install the VM. This included purging the MSDN downloads
for subscribers of W2k with SP4 inlined CD image, etc.
They also released IIRC version 3810 of the VM, which would
install only if the VM was already present on the machine.
Prior to 3810 the VM distributions would update an existing
VM if present or install the VM if not.
Many people wanted to not use the Sun JRE, so you will find
even today downloads of the VM that will install when the VM
is not present, provided only that it is a support OS. So, people
can build a XP SP2, which today will be without the VM, and
then apply one of these older VM distributions (obtained from
third-party archives) and it will install and it will leave the
machine at a vulnerable VM version if they omit downloading
and applying the 3810 VM update.
So, can XP SP2 be in a state that needs the patch? Yes.
Would a XP SP2 installed from a XP SP2 CD need it? No.
Would an XP installed from the original gold XP CD need it? Yes.
Would a gold XP install with SP 2 applied need it? No
Could a gold XP plus SP2 be caused to need it? Perhaps, likely not.
(one could uninstall 3810 and then use outdated download . . .)

Things are just not as clear-cut as you would like.



If you cannot see the simplicity of mbsacli usage I do not know
how you could be convinced, nor do I wish to try.


I believe you are correct, that only the old MBSA will detect for O2k
http://support.microsoft.com/kb/895660/en-us
But, IIRC MS will stop providing scan files in the old form as of March,
whence only ver 2.x releases of MBSA will be supported.
Also, is not O2k beyond its service lifetime? so one would not expect
to see further patches soon (like W2k which will soon no longer have
even security patches released to the public).

That is OK. I never referred to future expectations but only to what has
been
released in the past. And actually, MS promissed to make sure that
released
security patches for legacy products will remain available. But even w/o
such
a promiss it is OK because one can download and keep them by himself.

I don't jump on every new version MS releases. I do it only on "have to"
basis either because of HW or SW applications requirements. I am fairly
happy with W2K for development (and Office 2K). Others, such as WXP
I use only for testing.
Originally the MSSecure.xml file used by 1.2.1 was to be discontinued
6 months after 2.0 released, but I believe they extended that (I do not
recall until when)



You are correct that info on where to get the offline file is very,
very thouroughly buried, and since v2 it has no longer been as
simple as grabbing the file from disk on a system attached to the
network after running mbsacli

The download link for the offline WsusScn2.cab file is mentioned in
http://support.microsoft.com/kb/926464
For the more on the change from WsusScan.cab to the new format see
http://support.microsoft.com/kb/927745
 
D

David F

Roger Abell said:
I really do no know how best to respond to you expect to
say that if it is called out as affected (note that this does not
always mean needed, just that it can be needed) then there
is/are case(s) in which it can be needed.

I have every reason to believe the folks at MSRC know what
they are doing this regard, having been in on email exhanges
previously (nearly monthly) about making sure the released
bulletins are clear and accurate.

Also, keep in mind that it was only somewhere around the
end of 2002 when the patch updater technology started to
see the results from the push for improved strategy and for
unification of the multiple updaters in use. Prior to that time
it was possible to have a patch not take just because it was
applied in the wrong order while applying a specific group
of reboot requiring patches at the same time without reboots.

Anyway, the scanning tools are the best way to determine
what is needed by the current state of a system. Yes, this
does get more difficult for a non-connected system, and yes,
for older, end-of-life products one may have to use a variety
of tools to scan as much as possible. Nevertheless, trying
to wind one's way through the fine details would really
mean looking at the installed binaries on the system and
comparing their versions with the versions called out as
supplied by the patches (in the bulletin or associated KB).
One cannot just say, this is a W2k at SP 4, and since the

So if we stick to the WMP7.1 case, and lets say that it is only
applicable under the W2K+SP3 for example, it should simply
say so: "It affects WMP7.1, dated xx-xx-xxxx or earlier, when
installed under W2K+SP3. If SP4 or highier was installed
after WMP7.1, this patch has been taken care of."

It is a little bit of work, to think first and then to type it in
and maybe someone else to verify but:
1. It is done once. 2. Used million of times. 3. It is after all
their bug, not the customer's.

And remember that they have to go through this exact process,
with even much more precision and more meticulously when
they implement the automatic tools (MBSA, etc.).

To me it looks like they were skimpy in constructing that
global list. It does look like there is a lot of "copy & paste"
work there plus simple mistakes (such as in the case of
MS02-050, I think). And we know what it leads to.
high-level statement of affected versions includes that OS
therefore this one needs this. If you really want to do the
decision process you might need to read into the cab files
used in the scans and look at just what detection info they
are keying in on, which you will find in the xml called out
per OS version.

For the VM example you say, well yes that particular install
history would result in a vulnerable system, but it is that
person's fault that they did that. That may be true enough,
but you know where that person will place the blame, right?

There used to be a number of pages on the web that tracked
the patches, trying to organize the info much as you seem to
be attempting. As far as I know, those people have mostly
all stopped keeping their listings updated subsequent to the
release by MS of the scanners. There is a reason, it is just
hard to do better than one gets by running a scan.

Roger

So my conclusion for now is that I am going to experiment
on a best guess basis of my understanding the global list
vis-à-vis the individual list of each major update such as
SP4 and R1V2 (in the case of W2K), etc.

I have no reason to think that the automated tools and their
companion definition files which are basically another form
of those lists, are in a better shape than those lists. If any,
I would think that they are likely to have ADDITIONAL
and or other mistakes. Only if I will bump into a major
road block I will have to resort to the automated tools,
still striving to do as much as possible off line.

Anyway, many thanks for all the insights.

David
David F said:
Roger Abell said:
The global detailed Security Bulleting is given in:
http://www.microsoft.com/technet/security/current.aspx

I did find the following problems there.
I normally use Win 2K + SP4.
According the Security Bulletin for the release of SP4,
(http://www.microsoft.com/technet/security/prodtech/windows2000/w2ksp4.mspx)
it contains all new (new - with respect to SP3) patches in the
range:
MS01-022 (4.18.2001) till MS03-030 (7.23.2003)
and of course also all prior patches from SP1 to SP3.

But when I looked in that detailed global list, I find for example
that
as
old
MS00--077 (10.13.2000 !) and many other older than MS03-030
should be applied to W2K/SP4. This doesn't make sense.


I do not see statement that ms00-077 applies to W2k Sp4
Where do you see this? In what you term the Global Sec Bulletin it
states this applies to W2k at Sp1 and Sp2, and if you read into it you
will see that the Sp2 part was added when the specific patch was
reissued to address a further, similar exploit. If you look at the KB
http://support.microsoft.com/?kbid=299796
you will see that the updated patch was first included in SP3


You are right - I copy and paste the wrong patch #. I meant to specify
MS02-032 dated: 6.26.2002, which is listed specifically as already
included
in the SP4.


That is a Windows Media Player (WMP) patch.
Tell me, what happens in the following scenario . . .
Someone has W2k with WMP 6.4 (which I think is what came with W2k),
and they install SP 4 when it come out.
Then they elect to upgrade to WMP 7.1, but they use a download of the
WMP 7.1 installable obtained before the download was updated to have
the patch of MS02-032 included within it.
So, they have a W2k at SP4 with a vulnerable version of WMP 7.1, no?

I think the answer is NO - if one behaves reasonably. And here is my
rational.



If the downloaded WMP 7.1 is from before the release of MS02-032, then
when SP4 would come out
and installed, the WMP would be updated.

If the WMP 7.1 was downloaded after the release of MS02-032, whether
before the release of SP4
OR EVEN AFTER the release of SP4, then the WMP should incl. it and the
release and installation
of SP4 should have no impact (as far as MS02-032 is concerned).

If there was no updated version of WMP7.1 after MS02-032 release (and
actually even if there
was), it is OK for the global list to say that it is applicable to such
and such version or
release date or file date of WMP7.1.


Either way, there is no point to say in the global list that MS02-032
applies to W2K4 because it
is not.

Of course, if one downloaded WMP7.1 before MS02-032 released, kept it in
the drawer, then
downloads and installs SP4, AND ONLY THEN decided to install that OLD
WMP7.1 then it should be
HIS problem - stupidity is indeed costly. It is reasonable to expect that
when one installs
something, s/he would use the most recently available version and/or in
proper order.

I thought that this is an ABC and elementary logic. Isn't it?

You tell me, am I missing something? Did I leave any scenario uncovered?



But wait, there is more.[J]

MS02-032 was just an example. There are 4 more mentioned in the global
list as applicable to
W2K4: MS03-026, MS03-023, MS03-022 & MS02-050 - ALL within the period of
time that SP4 is
suppose to cover. The first 3 (MS03-.) though are NOT listed as included
in the SP4 release (see
my link above) although they belong there by date according release nots
for SP4. MS02-050 (like
MS02-032) IS explicitly incl. in the SP4 release, and unlike MS02-032, it
doesn't pertain to
some add-on such as the WMP but purely to the core OS. And there is even
more to this MS02-050
patch. It even shows up in the list of patches included with Rollup_1(V2)
for SP4 (hereafter
R1V2). Now the release notes for R1V2 say clearly that SP4 must be
installed first.



My hunch is that I should not install patches such as MS02-050 on top of
SP4 (let alone on top
of R1V2) but what say you in this case?



And what should I make about the other 3? I have no clue whether I should
go ahead and add them
on top of SP4 or not (and again, let alone on top of R1V2). They are
listed in the R1V2's list
of patches.

See: http://support.microsoft.com/kb/891861



Luckily, nothing in the global list references the R1V2.



All in all, I am in square one of my original thread. I would obviously
start with manual
installation of individual patches after installing R1V2 and again the
question is whether it is
OK to go by date (in order of course), that is, to pickup from the global
list all patches dated
after the last patch included in R1V2's list and ignore anything else.


In general, I agree with you that there could be, for historical, legal,
and other reasons some
peculiar scenarios but this doesn't mean that the "Affected Software
Service Packs" column, like
any other piece of technical info should not be precise and unequivocal
stated.
You are attempting to make thing too clear-cut; that way does not match
the realities encountered in world.


I found the exact same contradictions with regard to Win XP SP2.
In this case, I found even a worse situation, when patch # MS03-011
showing up in the detailed global list, which according to its date
and patch # should be included in the Security Bulletin list for
WXP/SP2 but it is not!. This list is given in:
http://www.microsoft.com/technet/security/prodtech/windowsXP/xpsp2.mspx


ms03-011 is something of a peculiar situation as it deals with the
MS distributed Java VM, which MS had to pull from availability
as part of a settlement with Sun. In fact, this was involved in the
reissue of XP Sp1 as XP Sp1a.
It is possible for even W2k3 to need this patch. It all depends on

I am confused by "...possible...". This one is ranked "critical" for
installing
explicitly on W2K3 (as a matter of fact, on W2K2/3/4). But then again,
like MS02-032, it should have already been included in W2K4. But again,

W2k4 ?
Yes.

Look. At a fixed point in time MS took away from the web all downloads
that included an install of the MS Java VM. This was required by the Sun
lawsuite settlement. That does not mean that people could not use things
obtained before that time to cause it to install.
Hence, it is known that the VM might be installed on certain versions,
that it certainly is installed on some or via specific built histories,
etc..
But, of those, only some can have the statement "the VM is installed"
made with certainty. Hence, how can you expect them to say the patch
is needed with certainty?

You will not find the VM on a fresh W2k3 install nor after SP1.
Yet W2k3 is called out in this bulletin. Why? Because some
histories leading to that instance of a W2k3 build could have
caused the VM to become installed.

what is worst, in spite of being released before SP4 was released, I
would
expect it to be included in SP4 and me NOT needed to install it on a
W2K4.
And like with WXP, it is missing from the Security Bulletin list for
SP4.

the history of the machine. One could install an XP SP2 from an
XP Sp2 CD and not get the VM installed. Later, install of an older
download of some product might install the VM at a level before
3810, causing need for the patch at that point.
Read into the bulletin and you will see there is quite a bit that is
abnormal about this specific patch. Behind the scenes there is a
bit of legal history. One must take the history of a specific machine
into account to determine need, and just being at an OS level
mentioned
in the bulletin does not mean one needs the patch (the VM might not be
present).

I am confused from the explanation.

My key point is that when I install a new system, I am obviously
expected
to
install the highest SPi (and for WXP I do have and use indeed the XP
SP2
CD
you mentioned to install XP) and then simply add (manually in my case)
each patch listed in that global list that was released after the
release
of that
SPi, and that that list explicitly says it applies to that SPi.
It is that simply and that straight forward.
Accordingly, I would not expect any patch listed as released before the
release
of a specific SPi, being mentioned as still applicable for that SPi.

And in the case of W2K, we have something called Rollup_1_ver_2-for SP4
which is even highier than SP4 (and I did not see an explicit list of
what
security patches it contains) and it is not mentioned in that global
list.
So the only way I can relate to it is by relaying to the date of the
last
patch
included as mentioned in its release notes. So I would use any patch
listed
in the global list as mentioned above but that its release date is not
just highier
than the last one to be included in SP4 but highier than that last date
of
that
Rollup.

BTW, I also would think that it should not matter, that for the case of
XP,
for example, that it would not matter whether one installs the very
original XP he has and add later the downloadable SP2 or use that CD
mentioned above.


Take the case of the VM patch.
To meet the court mandate MS removed all downloadables the
would install the VM. This included purging the MSDN downloads
for subscribers of W2k with SP4 inlined CD image, etc.
They also released IIRC version 3810 of the VM, which would
install only if the VM was already present on the machine.
Prior to 3810 the VM distributions would update an existing
VM if present or install the VM if not.
Many people wanted to not use the Sun JRE, so you will find
even today downloads of the VM that will install when the VM
is not present, provided only that it is a support OS. So, people
can build a XP SP2, which today will be without the VM, and
then apply one of these older VM distributions (obtained from
third-party archives) and it will install and it will leave the
machine at a vulnerable VM version if they omit downloading
and applying the 3810 VM update.
So, can XP SP2 be in a state that needs the patch? Yes.
Would a XP SP2 installed from a XP SP2 CD need it? No.
Would an XP installed from the original gold XP CD need it? Yes.
Would a gold XP install with SP 2 applied need it? No
Could a gold XP plus SP2 be caused to need it? Perhaps, likely not.
(one could uninstall 3810 and then use outdated download . . .)

Things are just not as clear-cut as you would like.



So I don't know what to make of it and how I can rely on that
detailed global list in terms of what to pickup from it and apply to
W2K/SP4
(and the same for WXP/SP2).


You seem to be going about this the hard way David.
You can either use the Microsoft Update site in Custom mode to
get a list of what a specific machine needs based on what is at
the moment installed; or one can download MBSA and after it
is installed open a cmd prompt at the install directory and issue
mbsacli /?
for syntax, and then use mbsacli with a switch
/n OS+IIS+SQL+Password
in order to only check patches

You are right about "the hard way" - to some degree. But - when I
looked
once
into that MBSA "thing", it looks to me overly cumbersome (like most
other
things from MS) and also things being changed all the time.
So I am not sure if in this case it is worth the time invested in it.
I might be wrong of course.


If you cannot see the simplicity of mbsacli usage I do not know
how you could be convinced, nor do I wish to try.

My understanding is that there is something called a "scan file" which
is
something
like a definition file (like for viruses detection for example), that
the
executable
MBSA program uses to figure out what my SW product(s) already have and
produces a list of MS0X-0YYs (or similar) I do need still to install.
And, if
the executable and "scan file" are installed, it is all dome offline.

It seems that for Office 2K I still have to use MBSA ver. 1.2.1. But I
could
NOT find an offline scan file for that ver. and my impression is that
this
version
works online only - exactly the kind of things I am trying to avoid

I believe you are correct, that only the old MBSA will detect for O2k
http://support.microsoft.com/kb/895660/en-us
But, IIRC MS will stop providing scan files in the old form as of March,
whence only ver 2.x releases of MBSA will be supported.
Also, is not O2k beyond its service lifetime? so one would not expect
to see further patches soon (like W2k which will soon no longer have
even security patches released to the public).

That is OK. I never referred to future expectations but only to what has
been
released in the past. And actually, MS promissed to make sure that
released
security patches for legacy products will remain available. But even w/o
such
a promiss it is OK because one can download and keep them by himself.

I don't jump on every new version MS releases. I do it only on "have to"
basis either because of HW or SW applications requirements. I am fairly
happy with W2K for development (and Office 2K). Others, such as WXP
I use only for testing.
Originally the MSSecure.xml file used by 1.2.1 was to be discontinued
6 months after 2.0 released, but I believe they extended that (I do not
recall until when)


Even for MBSA ver 2.0.1, I could not find the scan file, especially the
new one.


You are correct that info on where to get the offline file is very,
very thouroughly buried, and since v2 it has no longer been as
simple as grabbing the file from disk on a system attached to the
network after running mbsacli

The download link for the offline WsusScn2.cab file is mentioned in
http://support.microsoft.com/kb/926464
For the more on the change from WsusScan.cab to the new format see
http://support.microsoft.com/kb/927745
 
R

Roger Abell [MVP]

You said within
I have no reason to think that the automated tools and their
companion definition files which are basically another form
of those lists, are in a better shape than those lists. If any,
I would think that they are likely to have ADDITIONAL
and or other mistakes. Only if I will bump into a major
road block I will have to resort to the automated tools,
still striving to do as much as possible off line.

which I think is getting things backwards.
The scan files are generated by the experts, the web content
by content specialists albeit ones ultimately responsible to
revision directives from those experts.

David F said:
Roger Abell said:
I really do no know how best to respond to you expect to
say that if it is called out as affected (note that this does not
always mean needed, just that it can be needed) then there
is/are case(s) in which it can be needed.

I have every reason to believe the folks at MSRC know what
they are doing this regard, having been in on email exhanges
previously (nearly monthly) about making sure the released
bulletins are clear and accurate.

Also, keep in mind that it was only somewhere around the
end of 2002 when the patch updater technology started to
see the results from the push for improved strategy and for
unification of the multiple updaters in use. Prior to that time
it was possible to have a patch not take just because it was
applied in the wrong order while applying a specific group
of reboot requiring patches at the same time without reboots.

Anyway, the scanning tools are the best way to determine
what is needed by the current state of a system. Yes, this
does get more difficult for a non-connected system, and yes,
for older, end-of-life products one may have to use a variety
of tools to scan as much as possible. Nevertheless, trying
to wind one's way through the fine details would really
mean looking at the installed binaries on the system and
comparing their versions with the versions called out as
supplied by the patches (in the bulletin or associated KB).
One cannot just say, this is a W2k at SP 4, and since the

So if we stick to the WMP7.1 case, and lets say that it is only
applicable under the W2K+SP3 for example, it should simply
say so: "It affects WMP7.1, dated xx-xx-xxxx or earlier, when
installed under W2K+SP3. If SP4 or highier was installed
after WMP7.1, this patch has been taken care of."

It is a little bit of work, to think first and then to type it in
and maybe someone else to verify but:
1. It is done once. 2. Used million of times. 3. It is after all
their bug, not the customer's.

And remember that they have to go through this exact process,
with even much more precision and more meticulously when
they implement the automatic tools (MBSA, etc.).

To me it looks like they were skimpy in constructing that
global list. It does look like there is a lot of "copy & paste"
work there plus simple mistakes (such as in the case of
MS02-050, I think). And we know what it leads to.
high-level statement of affected versions includes that OS
therefore this one needs this. If you really want to do the
decision process you might need to read into the cab files
used in the scans and look at just what detection info they
are keying in on, which you will find in the xml called out
per OS version.

For the VM example you say, well yes that particular install
history would result in a vulnerable system, but it is that
person's fault that they did that. That may be true enough,
but you know where that person will place the blame, right?

There used to be a number of pages on the web that tracked
the patches, trying to organize the info much as you seem to
be attempting. As far as I know, those people have mostly
all stopped keeping their listings updated subsequent to the
release by MS of the scanners. There is a reason, it is just
hard to do better than one gets by running a scan.

Roger

So my conclusion for now is that I am going to experiment
on a best guess basis of my understanding the global list
vis-à-vis the individual list of each major update such as
SP4 and R1V2 (in the case of W2K), etc.

I have no reason to think that the automated tools and their
companion definition files which are basically another form
of those lists, are in a better shape than those lists. If any,
I would think that they are likely to have ADDITIONAL
and or other mistakes. Only if I will bump into a major
road block I will have to resort to the automated tools,
still striving to do as much as possible off line.

Anyway, many thanks for all the insights.

David
David F said:
The global detailed Security Bulleting is given in:
http://www.microsoft.com/technet/security/current.aspx

I did find the following problems there.
I normally use Win 2K + SP4.
According the Security Bulletin for the release of SP4,
(http://www.microsoft.com/technet/security/prodtech/windows2000/w2ksp4.mspx)
it contains all new (new - with respect to SP3) patches in the
range:
MS01-022 (4.18.2001) till MS03-030 (7.23.2003)
and of course also all prior patches from SP1 to SP3.

But when I looked in that detailed global list, I find for
example
that
as
old
MS00--077 (10.13.2000 !) and many other older than MS03-030
should be applied to W2K/SP4. This doesn't make sense.


I do not see statement that ms00-077 applies to W2k Sp4
Where do you see this? In what you term the Global Sec Bulletin it
states this applies to W2k at Sp1 and Sp2, and if you read into it
you
will see that the Sp2 part was added when the specific patch was
reissued to address a further, similar exploit. If you look at the
KB
http://support.microsoft.com/?kbid=299796
you will see that the updated patch was first included in SP3


You are right - I copy and paste the wrong patch #. I meant to
specify
MS02-032 dated: 6.26.2002, which is listed specifically as already
included
in the SP4.


That is a Windows Media Player (WMP) patch.
Tell me, what happens in the following scenario . . .
Someone has W2k with WMP 6.4 (which I think is what came with W2k),
and they install SP 4 when it come out.
Then they elect to upgrade to WMP 7.1, but they use a download of the
WMP 7.1 installable obtained before the download was updated to have
the patch of MS02-032 included within it.
So, they have a W2k at SP4 with a vulnerable version of WMP 7.1, no?


I think the answer is NO - if one behaves reasonably. And here is my
rational.



If the downloaded WMP 7.1 is from before the release of MS02-032, then
when SP4 would come out
and installed, the WMP would be updated.

If the WMP 7.1 was downloaded after the release of MS02-032, whether
before the release of SP4
OR EVEN AFTER the release of SP4, then the WMP should incl. it and the
release and installation
of SP4 should have no impact (as far as MS02-032 is concerned).

If there was no updated version of WMP7.1 after MS02-032 release (and
actually even if there
was), it is OK for the global list to say that it is applicable to such
and such version or
release date or file date of WMP7.1.


Either way, there is no point to say in the global list that MS02-032
applies to W2K4 because it
is not.

Of course, if one downloaded WMP7.1 before MS02-032 released, kept it
in
the drawer, then
downloads and installs SP4, AND ONLY THEN decided to install that OLD
WMP7.1 then it should be
HIS problem - stupidity is indeed costly. It is reasonable to expect
that
when one installs
something, s/he would use the most recently available version and/or in
proper order.

I thought that this is an ABC and elementary logic. Isn't it?

You tell me, am I missing something? Did I leave any scenario
uncovered?



But wait, there is more.[J]

MS02-032 was just an example. There are 4 more mentioned in the global
list as applicable to
W2K4: MS03-026, MS03-023, MS03-022 & MS02-050 - ALL within the period
of
time that SP4 is
suppose to cover. The first 3 (MS03-.) though are NOT listed as
included
in the SP4 release (see
my link above) although they belong there by date according release
nots
for SP4. MS02-050 (like
MS02-032) IS explicitly incl. in the SP4 release, and unlike MS02-032,
it
doesn't pertain to
some add-on such as the WMP but purely to the core OS. And there is
even
more to this MS02-050
patch. It even shows up in the list of patches included with
Rollup_1(V2)
for SP4 (hereafter
R1V2). Now the release notes for R1V2 say clearly that SP4 must be
installed first.



My hunch is that I should not install patches such as MS02-050 on top
of
SP4 (let alone on top
of R1V2) but what say you in this case?



And what should I make about the other 3? I have no clue whether I
should
go ahead and add them
on top of SP4 or not (and again, let alone on top of R1V2). They are
listed in the R1V2's list
of patches.

See: http://support.microsoft.com/kb/891861



Luckily, nothing in the global list references the R1V2.



All in all, I am in square one of my original thread. I would obviously
start with manual
installation of individual patches after installing R1V2 and again the
question is whether it is
OK to go by date (in order of course), that is, to pickup from the
global
list all patches dated
after the last patch included in R1V2's list and ignore anything else.


In general, I agree with you that there could be, for historical,
legal,
and other reasons some
peculiar scenarios but this doesn't mean that the "Affected Software
Service Packs" column, like
any other piece of technical info should not be precise and unequivocal
stated.

You are attempting to make thing too clear-cut; that way does not
match
the realities encountered in world.


I found the exact same contradictions with regard to Win XP SP2.
In this case, I found even a worse situation, when patch #
MS03-011
showing up in the detailed global list, which according to its
date
and patch # should be included in the Security Bulletin list for
WXP/SP2 but it is not!. This list is given in:
http://www.microsoft.com/technet/security/prodtech/windowsXP/xpsp2.mspx


ms03-011 is something of a peculiar situation as it deals with the
MS distributed Java VM, which MS had to pull from availability
as part of a settlement with Sun. In fact, this was involved in
the
reissue of XP Sp1 as XP Sp1a.
It is possible for even W2k3 to need this patch. It all depends on

I am confused by "...possible...". This one is ranked "critical" for
installing
explicitly on W2K3 (as a matter of fact, on W2K2/3/4). But then
again,
like MS02-032, it should have already been included in W2K4. But
again,

W2k4 ?

Yes.

Look. At a fixed point in time MS took away from the web all
downloads
that included an install of the MS Java VM. This was required by the
Sun
lawsuite settlement. That does not mean that people could not use
things
obtained before that time to cause it to install.
Hence, it is known that the VM might be installed on certain versions,
that it certainly is installed on some or via specific built
histories,
etc..
But, of those, only some can have the statement "the VM is installed"
made with certainty. Hence, how can you expect them to say the patch
is needed with certainty?

You will not find the VM on a fresh W2k3 install nor after SP1.
Yet W2k3 is called out in this bulletin. Why? Because some
histories leading to that instance of a W2k3 build could have
caused the VM to become installed.

what is worst, in spite of being released before SP4 was released, I
would
expect it to be included in SP4 and me NOT needed to install it on a
W2K4.
And like with WXP, it is missing from the Security Bulletin list for
SP4.

the history of the machine. One could install an XP SP2 from an
XP Sp2 CD and not get the VM installed. Later, install of an older
download of some product might install the VM at a level before
3810, causing need for the patch at that point.
Read into the bulletin and you will see there is quite a bit that
is
abnormal about this specific patch. Behind the scenes there is a
bit of legal history. One must take the history of a specific
machine
into account to determine need, and just being at an OS level
mentioned
in the bulletin does not mean one needs the patch (the VM might not
be
present).

I am confused from the explanation.

My key point is that when I install a new system, I am obviously
expected
to
install the highest SPi (and for WXP I do have and use indeed the XP
SP2
CD
you mentioned to install XP) and then simply add (manually in my
case)
each patch listed in that global list that was released after the
release
of that
SPi, and that that list explicitly says it applies to that SPi.
It is that simply and that straight forward.
Accordingly, I would not expect any patch listed as released before
the
release
of a specific SPi, being mentioned as still applicable for that SPi.

And in the case of W2K, we have something called Rollup_1_ver_2-for
SP4
which is even highier than SP4 (and I did not see an explicit list
of
what
security patches it contains) and it is not mentioned in that global
list.
So the only way I can relate to it is by relaying to the date of the
last
patch
included as mentioned in its release notes. So I would use any patch
listed
in the global list as mentioned above but that its release date is
not
just highier
than the last one to be included in SP4 but highier than that last
date
of
that
Rollup.

BTW, I also would think that it should not matter, that for the case
of
XP,
for example, that it would not matter whether one installs the very
original XP he has and add later the downloadable SP2 or use that CD
mentioned above.


Take the case of the VM patch.
To meet the court mandate MS removed all downloadables the
would install the VM. This included purging the MSDN downloads
for subscribers of W2k with SP4 inlined CD image, etc.
They also released IIRC version 3810 of the VM, which would
install only if the VM was already present on the machine.
Prior to 3810 the VM distributions would update an existing
VM if present or install the VM if not.
Many people wanted to not use the Sun JRE, so you will find
even today downloads of the VM that will install when the VM
is not present, provided only that it is a support OS. So, people
can build a XP SP2, which today will be without the VM, and
then apply one of these older VM distributions (obtained from
third-party archives) and it will install and it will leave the
machine at a vulnerable VM version if they omit downloading
and applying the 3810 VM update.
So, can XP SP2 be in a state that needs the patch? Yes.
Would a XP SP2 installed from a XP SP2 CD need it? No.
Would an XP installed from the original gold XP CD need it? Yes.
Would a gold XP install with SP 2 applied need it? No
Could a gold XP plus SP2 be caused to need it? Perhaps, likely not.
(one could uninstall 3810 and then use outdated download . . .)

Things are just not as clear-cut as you would like.



So I don't know what to make of it and how I can rely on that
detailed global list in terms of what to pickup from it and apply
to
W2K/SP4
(and the same for WXP/SP2).


You seem to be going about this the hard way David.
You can either use the Microsoft Update site in Custom mode to
get a list of what a specific machine needs based on what is at
the moment installed; or one can download MBSA and after it
is installed open a cmd prompt at the install directory and issue
mbsacli /?
for syntax, and then use mbsacli with a switch
/n OS+IIS+SQL+Password
in order to only check patches

You are right about "the hard way" - to some degree. But - when I
looked
once
into that MBSA "thing", it looks to me overly cumbersome (like most
other
things from MS) and also things being changed all the time.
So I am not sure if in this case it is worth the time invested in
it.
I might be wrong of course.


If you cannot see the simplicity of mbsacli usage I do not know
how you could be convinced, nor do I wish to try.

My understanding is that there is something called a "scan file"
which
is
something
like a definition file (like for viruses detection for example),
that
the
executable
MBSA program uses to figure out what my SW product(s) already have
and
produces a list of MS0X-0YYs (or similar) I do need still to
install.
And, if
the executable and "scan file" are installed, it is all dome
offline.

It seems that for Office 2K I still have to use MBSA ver. 1.2.1. But
I
could
NOT find an offline scan file for that ver. and my impression is
that
this
version
works online only - exactly the kind of things I am trying to avoid

I believe you are correct, that only the old MBSA will detect for O2k
http://support.microsoft.com/kb/895660/en-us
But, IIRC MS will stop providing scan files in the old form as of
March,
whence only ver 2.x releases of MBSA will be supported.
Also, is not O2k beyond its service lifetime? so one would not expect
to see further patches soon (like W2k which will soon no longer have
even security patches released to the public).

That is OK. I never referred to future expectations but only to what
has
been
released in the past. And actually, MS promissed to make sure that
released
security patches for legacy products will remain available. But even
w/o
such
a promiss it is OK because one can download and keep them by himself.

I don't jump on every new version MS releases. I do it only on "have
to"
basis either because of HW or SW applications requirements. I am fairly
happy with W2K for development (and Office 2K). Others, such as WXP
I use only for testing.

Originally the MSSecure.xml file used by 1.2.1 was to be discontinued
6 months after 2.0 released, but I believe they extended that (I do
not
recall until when)


Even for MBSA ver 2.0.1, I could not find the scan file, especially
the
new one.


You are correct that info on where to get the offline file is very,
very thouroughly buried, and since v2 it has no longer been as
simple as grabbing the file from disk on a system attached to the
network after running mbsacli

The download link for the offline WsusScn2.cab file is mentioned in
http://support.microsoft.com/kb/926464
For the more on the change from WsusScan.cab to the new format see
http://support.microsoft.com/kb/927745
 
D

David F

I forgot to ask you one more thing.
Two examples in a W2K4 system (with R1V2).
The MS02-032 is necessary only for WMP7.1 which is no more relevant
(assuming one uses WMP9).
The MS02-050 is explicitly listed as included in SP4 AND in Rollup 1
(which by itself is a redundancy).
Now these two might be too trivial - it suggests quite strongly that they
can be skipped but some others might be more fuzzy.

So the question is: What happen if one runs an unnecessary patch?
Do they know to do nothing or at least cause no harm?

David

Roger Abell said:
You said within
I have no reason to think that the automated tools and their
companion definition files which are basically another form
of those lists, are in a better shape than those lists. If any,
I would think that they are likely to have ADDITIONAL
and or other mistakes. Only if I will bump into a major
road block I will have to resort to the automated tools,
still striving to do as much as possible off line.

which I think is getting things backwards.
The scan files are generated by the experts, the web content
by content specialists albeit ones ultimately responsible to
revision directives from those experts.

David F said:
Roger Abell said:
I really do no know how best to respond to you expect to
say that if it is called out as affected (note that this does not
always mean needed, just that it can be needed) then there
is/are case(s) in which it can be needed.

I have every reason to believe the folks at MSRC know what
they are doing this regard, having been in on email exhanges
previously (nearly monthly) about making sure the released
bulletins are clear and accurate.

Also, keep in mind that it was only somewhere around the
end of 2002 when the patch updater technology started to
see the results from the push for improved strategy and for
unification of the multiple updaters in use. Prior to that time
it was possible to have a patch not take just because it was
applied in the wrong order while applying a specific group
of reboot requiring patches at the same time without reboots.

Anyway, the scanning tools are the best way to determine
what is needed by the current state of a system. Yes, this
does get more difficult for a non-connected system, and yes,
for older, end-of-life products one may have to use a variety
of tools to scan as much as possible. Nevertheless, trying
to wind one's way through the fine details would really
mean looking at the installed binaries on the system and
comparing their versions with the versions called out as
supplied by the patches (in the bulletin or associated KB).
One cannot just say, this is a W2k at SP 4, and since the

So if we stick to the WMP7.1 case, and lets say that it is only
applicable under the W2K+SP3 for example, it should simply
say so: "It affects WMP7.1, dated xx-xx-xxxx or earlier, when
installed under W2K+SP3. If SP4 or highier was installed
after WMP7.1, this patch has been taken care of."

It is a little bit of work, to think first and then to type it in
and maybe someone else to verify but:
1. It is done once. 2. Used million of times. 3. It is after all
their bug, not the customer's.

And remember that they have to go through this exact process,
with even much more precision and more meticulously when
they implement the automatic tools (MBSA, etc.).

To me it looks like they were skimpy in constructing that
global list. It does look like there is a lot of "copy & paste"
work there plus simple mistakes (such as in the case of
MS02-050, I think). And we know what it leads to.
high-level statement of affected versions includes that OS
therefore this one needs this. If you really want to do the
decision process you might need to read into the cab files
used in the scans and look at just what detection info they
are keying in on, which you will find in the xml called out
per OS version.

For the VM example you say, well yes that particular install
history would result in a vulnerable system, but it is that
person's fault that they did that. That may be true enough,
but you know where that person will place the blame, right?

There used to be a number of pages on the web that tracked
the patches, trying to organize the info much as you seem to
be attempting. As far as I know, those people have mostly
all stopped keeping their listings updated subsequent to the
release by MS of the scanners. There is a reason, it is just
hard to do better than one gets by running a scan.

Roger

So my conclusion for now is that I am going to experiment
on a best guess basis of my understanding the global list
vis-à-vis the individual list of each major update such as
SP4 and R1V2 (in the case of W2K), etc.

I have no reason to think that the automated tools and their
companion definition files which are basically another form
of those lists, are in a better shape than those lists. If any,
I would think that they are likely to have ADDITIONAL
and or other mistakes. Only if I will bump into a major
road block I will have to resort to the automated tools,
still striving to do as much as possible off line.

Anyway, many thanks for all the insights.

David
The global detailed Security Bulleting is given in:
http://www.microsoft.com/technet/security/current.aspx

I did find the following problems there.
I normally use Win 2K + SP4.
According the Security Bulletin for the release of SP4,
(http://www.microsoft.com/technet/security/prodtech/windows2000/w2ksp4.mspx)
it contains all new (new - with respect to SP3) patches in the
range:
MS01-022 (4.18.2001) till MS03-030 (7.23.2003)
and of course also all prior patches from SP1 to SP3.

But when I looked in that detailed global list, I find for
example
that
as
old
MS00--077 (10.13.2000 !) and many other older than MS03-030
should be applied to W2K/SP4. This doesn't make sense.


I do not see statement that ms00-077 applies to W2k Sp4
Where do you see this? In what you term the Global Sec Bulletin it
states this applies to W2k at Sp1 and Sp2, and if you read into it
you
will see that the Sp2 part was added when the specific patch was
reissued to address a further, similar exploit. If you look at the
KB
http://support.microsoft.com/?kbid=299796
you will see that the updated patch was first included in SP3


You are right - I copy and paste the wrong patch #. I meant to
specify
MS02-032 dated: 6.26.2002, which is listed specifically as already
included
in the SP4.


That is a Windows Media Player (WMP) patch.
Tell me, what happens in the following scenario . . .
Someone has W2k with WMP 6.4 (which I think is what came with W2k),
and they install SP 4 when it come out.
Then they elect to upgrade to WMP 7.1, but they use a download of the
WMP 7.1 installable obtained before the download was updated to have
the patch of MS02-032 included within it.
So, they have a W2k at SP4 with a vulnerable version of WMP 7.1, no?


I think the answer is NO - if one behaves reasonably. And here is my
rational.



If the downloaded WMP 7.1 is from before the release of MS02-032, then
when SP4 would come out
and installed, the WMP would be updated.

If the WMP 7.1 was downloaded after the release of MS02-032, whether
before the release of SP4
OR EVEN AFTER the release of SP4, then the WMP should incl. it and the
release and installation
of SP4 should have no impact (as far as MS02-032 is concerned).

If there was no updated version of WMP7.1 after MS02-032 release (and
actually even if there
was), it is OK for the global list to say that it is applicable to such
and such version or
release date or file date of WMP7.1.


Either way, there is no point to say in the global list that MS02-032
applies to W2K4 because it
is not.

Of course, if one downloaded WMP7.1 before MS02-032 released, kept it
in
the drawer, then
downloads and installs SP4, AND ONLY THEN decided to install that OLD
WMP7.1 then it should be
HIS problem - stupidity is indeed costly. It is reasonable to expect
that
when one installs
something, s/he would use the most recently available version and/or in
proper order.

I thought that this is an ABC and elementary logic. Isn't it?

You tell me, am I missing something? Did I leave any scenario
uncovered?



But wait, there is more.[J]

MS02-032 was just an example. There are 4 more mentioned in the global
list as applicable to
W2K4: MS03-026, MS03-023, MS03-022 & MS02-050 - ALL within the period
of
time that SP4 is
suppose to cover. The first 3 (MS03-.) though are NOT listed as
included
in the SP4 release (see
my link above) although they belong there by date according release
nots
for SP4. MS02-050 (like
MS02-032) IS explicitly incl. in the SP4 release, and unlike MS02-032,
it
doesn't pertain to
some add-on such as the WMP but purely to the core OS. And there is
even
more to this MS02-050
patch. It even shows up in the list of patches included with
Rollup_1(V2)
for SP4 (hereafter
R1V2). Now the release notes for R1V2 say clearly that SP4 must be
installed first.



My hunch is that I should not install patches such as MS02-050 on top
of
SP4 (let alone on top
of R1V2) but what say you in this case?



And what should I make about the other 3? I have no clue whether I
should
go ahead and add them
on top of SP4 or not (and again, let alone on top of R1V2). They are
listed in the R1V2's list
of patches.

See: http://support.microsoft.com/kb/891861



Luckily, nothing in the global list references the R1V2.



All in all, I am in square one of my original thread. I would obviously
start with manual
installation of individual patches after installing R1V2 and again the
question is whether it is
OK to go by date (in order of course), that is, to pickup from the
global
list all patches dated
after the last patch included in R1V2's list and ignore anything else.


In general, I agree with you that there could be, for historical,
legal,
and other reasons some
peculiar scenarios but this doesn't mean that the "Affected Software
Service Packs" column, like
any other piece of technical info should not be precise and unequivocal
stated.

You are attempting to make thing too clear-cut; that way does not
match
the realities encountered in world.


I found the exact same contradictions with regard to Win XP SP2.
In this case, I found even a worse situation, when patch #
MS03-011
showing up in the detailed global list, which according to its
date
and patch # should be included in the Security Bulletin list for
WXP/SP2 but it is not!. This list is given in:
http://www.microsoft.com/technet/security/prodtech/windowsXP/xpsp2.mspx


ms03-011 is something of a peculiar situation as it deals with the
MS distributed Java VM, which MS had to pull from availability
as part of a settlement with Sun. In fact, this was involved in
the
reissue of XP Sp1 as XP Sp1a.
It is possible for even W2k3 to need this patch. It all depends on

I am confused by "...possible...". This one is ranked "critical" for
installing
explicitly on W2K3 (as a matter of fact, on W2K2/3/4). But then
again,
like MS02-032, it should have already been included in W2K4. But
again,

W2k4 ?

Yes.

Look. At a fixed point in time MS took away from the web all
downloads
that included an install of the MS Java VM. This was required by the
Sun
lawsuite settlement. That does not mean that people could not use
things
obtained before that time to cause it to install.
Hence, it is known that the VM might be installed on certain versions,
that it certainly is installed on some or via specific built
histories,
etc..
But, of those, only some can have the statement "the VM is installed"
made with certainty. Hence, how can you expect them to say the patch
is needed with certainty?

You will not find the VM on a fresh W2k3 install nor after SP1.
Yet W2k3 is called out in this bulletin. Why? Because some
histories leading to that instance of a W2k3 build could have
caused the VM to become installed.

what is worst, in spite of being released before SP4 was released, I
would
expect it to be included in SP4 and me NOT needed to install it on a
W2K4.
And like with WXP, it is missing from the Security Bulletin list for
SP4.

the history of the machine. One could install an XP SP2 from an
XP Sp2 CD and not get the VM installed. Later, install of an older
download of some product might install the VM at a level before
3810, causing need for the patch at that point.
Read into the bulletin and you will see there is quite a bit that
is
abnormal about this specific patch. Behind the scenes there is a
bit of legal history. One must take the history of a specific
machine
into account to determine need, and just being at an OS level
mentioned
in the bulletin does not mean one needs the patch (the VM might not
be
present).

I am confused from the explanation.

My key point is that when I install a new system, I am obviously
expected
to
install the highest SPi (and for WXP I do have and use indeed the XP
SP2
CD
you mentioned to install XP) and then simply add (manually in my
case)
each patch listed in that global list that was released after the
release
of that
SPi, and that that list explicitly says it applies to that SPi.
It is that simply and that straight forward.
Accordingly, I would not expect any patch listed as released before
the
release
of a specific SPi, being mentioned as still applicable for that SPi.

And in the case of W2K, we have something called Rollup_1_ver_2-for
SP4
which is even highier than SP4 (and I did not see an explicit list
of
what
security patches it contains) and it is not mentioned in that global
list.
So the only way I can relate to it is by relaying to the date of the
last
patch
included as mentioned in its release notes. So I would use any patch
listed
in the global list as mentioned above but that its release date is
not
just highier
than the last one to be included in SP4 but highier than that last
date
of
that
Rollup.

BTW, I also would think that it should not matter, that for the case
of
XP,
for example, that it would not matter whether one installs the very
original XP he has and add later the downloadable SP2 or use that CD
mentioned above.


Take the case of the VM patch.
To meet the court mandate MS removed all downloadables the
would install the VM. This included purging the MSDN downloads
for subscribers of W2k with SP4 inlined CD image, etc.
They also released IIRC version 3810 of the VM, which would
install only if the VM was already present on the machine.
Prior to 3810 the VM distributions would update an existing
VM if present or install the VM if not.
Many people wanted to not use the Sun JRE, so you will find
even today downloads of the VM that will install when the VM
is not present, provided only that it is a support OS. So, people
can build a XP SP2, which today will be without the VM, and
then apply one of these older VM distributions (obtained from
third-party archives) and it will install and it will leave the
machine at a vulnerable VM version if they omit downloading
and applying the 3810 VM update.
So, can XP SP2 be in a state that needs the patch? Yes.
Would a XP SP2 installed from a XP SP2 CD need it? No.
Would an XP installed from the original gold XP CD need it? Yes.
Would a gold XP install with SP 2 applied need it? No
Could a gold XP plus SP2 be caused to need it? Perhaps, likely not.
(one could uninstall 3810 and then use outdated download . . .)

Things are just not as clear-cut as you would like.



So I don't know what to make of it and how I can rely on that
detailed global list in terms of what to pickup from it and apply
to
W2K/SP4
(and the same for WXP/SP2).


You seem to be going about this the hard way David.
You can either use the Microsoft Update site in Custom mode to
get a list of what a specific machine needs based on what is at
the moment installed; or one can download MBSA and after it
is installed open a cmd prompt at the install directory and issue
mbsacli /?
for syntax, and then use mbsacli with a switch
/n OS+IIS+SQL+Password
in order to only check patches

You are right about "the hard way" - to some degree. But - when I
looked
once
into that MBSA "thing", it looks to me overly cumbersome (like most
other
things from MS) and also things being changed all the time.
So I am not sure if in this case it is worth the time invested in
it.
I might be wrong of course.


If you cannot see the simplicity of mbsacli usage I do not know
how you could be convinced, nor do I wish to try.

My understanding is that there is something called a "scan file"
which
is
something
like a definition file (like for viruses detection for example),
that
the
executable
MBSA program uses to figure out what my SW product(s) already have
and
produces a list of MS0X-0YYs (or similar) I do need still to
install.
And, if
the executable and "scan file" are installed, it is all dome
offline.

It seems that for Office 2K I still have to use MBSA ver. 1.2.1. But
I
could
NOT find an offline scan file for that ver. and my impression is
that
this
version
works online only - exactly the kind of things I am trying to avoid

I believe you are correct, that only the old MBSA will detect for O2k
http://support.microsoft.com/kb/895660/en-us
But, IIRC MS will stop providing scan files in the old form as of
March,
whence only ver 2.x releases of MBSA will be supported.
Also, is not O2k beyond its service lifetime? so one would not expect
to see further patches soon (like W2k which will soon no longer have
even security patches released to the public).

That is OK. I never referred to future expectations but only to what
has
been
released in the past. And actually, MS promissed to make sure that
released
security patches for legacy products will remain available. But even
w/o
such
a promiss it is OK because one can download and keep them by himself.

I don't jump on every new version MS releases. I do it only on "have
to"
basis either because of HW or SW applications requirements. I am fairly
happy with W2K for development (and Office 2K). Others, such as WXP
I use only for testing.

Originally the MSSecure.xml file used by 1.2.1 was to be discontinued
6 months after 2.0 released, but I believe they extended that (I do
not
recall until when)


Even for MBSA ver 2.0.1, I could not find the scan file, especially
the
new one.


You are correct that info on where to get the offline file is very,
very thouroughly buried, and since v2 it has no longer been as
simple as grabbing the file from disk on a system attached to the
network after running mbsacli

The download link for the offline WsusScn2.cab file is mentioned in
http://support.microsoft.com/kb/926464
For the more on the change from WsusScan.cab to the new format see
http://support.microsoft.com/kb/927745
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top