Good morning!
I would like to comment on this issue again, because to my surprise there seem
to have been some significant changes about the problems I was having.
I downloaded another trunk build yesterday to retry my stuff, and this time
several things worked, that would previously cause bluescreens. The test was
to download a ~690MB file from the internet using either ftp.exe on the
commandline or FileZilla2 (both with and without Implicit AND Explicit SSL),
bug was always triggered.
Additionally, I also ran the x264 video transcoder, statically linked against
libav libraries. This thing runs CPU- and memory-intensive computation using
several CPU instruction extensions when present (SSE, SSE2, SSSE3, SSE4.1,
SSE4.2, AVX, FMA4 etc.).
Tested builds:
* 0.3.14 -> NTOSKRNL.EXE BSOD on data transfer and crunching.
* r57415 -> NTOSKRNL.EXE BSOD on data transfer and crunching.
* r57452 -> tcpip.sys BSOD on data transfer, NTOSKRNL.EXE BSOD on crunching.
* r57481 -> everything works!
For those builds which would cause BSOD when tested with x264, the BSOD would
usually arrive around 200-400 completed video frames (out of 15691 which are
processed twice). This time, I completed an entire run of encoding pass 1 + 2,
which took around 8 hours (i7 980X), and I am now running another one, which
also does not seem to cause any problems!!
See here (german version, shouldn't matter):
http://www.xin.at/x264/pics/guide_screenshot_reactos.png
I studied the changelogs since the last version, but I didn't see anything
big that would have said "fix tcpip.sys" or "fix LOTS of bluescreens",
but...
WHATEVER you guys (developers) did, it really, really helped in my case!
Thanks for that!
--
Michael Lackner
Lehrstuhl für Informationstechnologie (CiT)
Montanuniversität Leoben
Tel.: +43 (0)3842/402-1505 | Mail: michael.lackner(a)unileoben.ac.at
Fax.: +43 (0)3842/402-1502 | Web:
http://institute.unileoben.ac.at/infotech