Comments on: Reading Notebook: 18-March-10 http://www.softwaregeneralist.com/2010/03/19/reading-notebook-18-march-10/ Connecting Software with Engineering, Science, Philosophy and Religion Thu, 09 Sep 2010 09:41:24 +0000 http://wordpress.org/?v=2.5 By: Sol_Ksacap http://www.softwaregeneralist.com/2010/03/19/reading-notebook-18-march-10/#comment-718 Sol_Ksacap Sun, 21 Mar 2010 18:56:01 +0000 http://www.softwaregeneralist.com/?p=101#comment-718 >PsPrioritySeperation Indeed, this variable's name was fixed in w7 – no more 'PsPrioritySeperation', but 'PsPrioritySeparation' present instead. That's slightly surprising, since misprints in some other places in windows code tend to persist from build to build. For example, 'bHasMeun' member of window structure: lkd> dt win32k!tagWND +0x000 head : _THRDESKHEAD +0x028 state : Uint4B +0x028 bHasMeun : Pos 0, 1 Bit +0x028 bHasVerticalScrollbar : Pos 1, 1 Bit Or misprint in address space layout table (\ntos\mm\amd64\miamd.h): FFFF080000000000 | Start of System space | MM_SYSTEM_RANGE_START Here, author of original layout probably meant "FFFF800000000000", since for virtual address to be regarded as Canonical, its bits 63..48 shall be sign-extension of bit 47 – but back in the day that erratum slipped through and its results are still there, in w7 kernel: lkd> dp MmSystemRangeStart l1 fffff800`02aa90e0 ffff0800`00000000 (Althought there might actually be some valid reason for this value – we don't really know). >PsPrioritySeperation
Indeed, this variable’s name was fixed in w7 – no more ‘PsPrioritySeperation’, but ‘PsPrioritySeparation’ present instead.
That’s slightly surprising, since misprints in some other places in windows code tend to persist from build to build. For example, ‘bHasMeun’ member of window structure:

lkd> dt win32k!tagWND
+0×000 head : _THRDESKHEAD
+0×028 state : Uint4B
+0×028 bHasMeun : Pos 0, 1 Bit
+0×028 bHasVerticalScrollbar : Pos 1, 1 Bit

Or misprint in address space layout table (\ntos\mm\amd64\miamd.h):
FFFF080000000000 | Start of System space | MM_SYSTEM_RANGE_START

Here, author of original layout probably meant “FFFF800000000000″, since for virtual address to be regarded as Canonical, its bits 63..48 shall be sign-extension of bit 47 – but back in the day that erratum slipped through and its results are still there, in w7 kernel:

lkd> dp MmSystemRangeStart l1
fffff800`02aa90e0 ffff0800`00000000
(Althought there might actually be some valid reason for this value – we don’t really know).

]]>