[ros-bugs] [Bug 1270] Crash in CoordDPtoLP

[ReactOS.Bugzilla at reactos.org]

http://www.reactos.org/bugzilla/show_bug.cgi?id=1270





------- Additional Comments From alex.ionescu at reactos.com  2006-01-14 09:08 CET -------
The crash in this instance happens because CurrentThread (fs:124h) seeminlgly
has the value 0xff000120 (this, for the kernel uninitiated, is &fs:120h). After
adding some dprints to the exp.c file right before calling the handler, thsi
value is a normal thread pointer. After adding a dprint right before the inlined
assembly in fpu.c, it's still a valid pointer. Adding anoter dprint after the
inlined assembly now caused a crash, because supposedly the IRQL was set to 0x80
(this, for the kernel uinitiated, is impossible. maximum irql is 1f). Removing
this dprint and setting it a bit later magically made the firefox installer and
firefox itself work, and I've been pounding the system trying to get it to crash
again, without success. Due to the recent enabling of optimizations and the fact
that this routine is very dirtyly written (I don't know who came up with the
ingenius idea of writing low-level trap handlers in C) and contains assembly
code and other inlined hacks, I believe this is gcc's optimizatizer messing up
the stack while optimizing, due to missing volatile statements or a bug. I will
send my binary to Waxdragon tomorrow for further testing.

-- 
Configure bugmail: http://www.reactos.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
You are the QA contact for the bug, or are watching the QA contact.