> From: Hartmut Birr
>
> Log message:
> - Rewrote (copied from ntdll) parts of the relocation
> and import fixup code.
> - Don't load no-load sections at the end of the driver.
> - Don't relocate the relocation section itself.
> - Set the correct page protection for section which are
> executable.
With current CVS, the VMware SVGA driver crashes. It happens in a
HwFindAdapter() call. While tracing through the driver, I got this eery
"I've been here before" feeling. And sure enough, I fixed this same crash
about a year and a half ago.
The problem is that vmx_svga.sys has a section alignment of 0x20 instead of
the normal 0x1000. So, sections are much smaller than the page size. When
setting the page protections you need to take into account all the sections
which might be mapped to that page. This was implemented in
ntoskrnl/ldr/loader.c rev 1.126.
Gé van Geldorp.
I started putting in ASSERT_IRQL in ntoskrnl, hoping that only my own code
would break from it. The results are ... interesting.
I've put a metabug (426) on bugzilla which is about reactos not honoring
documented irql requirements. I fixed a place in win32k.sys, but there
are probably more. I think we should clean these up and get the
assert_irqls into HEAD, so that people like me won't get bitten over and
over by assuming we're in one irql when we're in another.
Unrelated, please take a look at the following patch, anyone who's cozy
with the mm. The reason why this code is needed is that MmNotPresentFault
and MmAccessFault will try to lock the address space as well as
MmProbeAndLockPages. Since this is a spin lock, reactos will always
bugcheck when MmProbeAndLockPages faults in a page, even if that would
be legal otherwise. I'm not sure if this patch fixes it right but it does
seem to work this way. If i don't hear too many complaints i'll think
about committing something similar.
There are a couple of other places where MmNotPresentFault is called
(section.c) within a MmLockAddressSpace. Another idea would be to give
MmNotPresentFault and MmAccessFault an extra argument so they can ignore
locking if the address space is already locked.
--
Hey, Adam Smith, keep your invisible hands to yourself!
Hello!
I thought I would try running Cygwin on a ReactOS CVS build, and
after talking with some of you on IRC, I went ahead.
After setting PATH=c:\cygwin\bin;%PATH%, dd.exe, uniq.exe,
bash.exe, and ls.exe all drop ros to the kernel debugger with an
exception 14. This is on the now dead Command Prompt:
41 [Main] ? 10 cygheap_user::init: SetTokenInformation
(TokenDefaultDacl): Win32 error 1
Here is the debug output leading up to the exection of uniq.exe:
(objects/gdiobj.c:401) Can't delete hObj:0x00042d77, type:0x007f0000, flag:0
(objects/gdiobj.c:401) Can't delete hObj:0x00042f72, type:0x007f0000, flag:0
(ntuser/desktop.c:808) RtlQueryRegistryValues failed for
PaintDesktopVersion: status=c0000005
(NTDLL:ldr/utils.c:2060) Relocating (10000000 -> 631000)
C:\cygwin\bin\cygiconv-2.dll
(NTDLL:ldr/utils.c:2060) Relocating (77f00000 -> 941000)
C:\ReactOS\system32\KERNEL32.dll
(NTDLL:ldr/utils.c:2060) Relocating (77dc0000 -> 9ba000)
C:\ReactO(NTDLL:ldr/utils.c:2060) Relocating (61000000 -> 9d0000)
C:\cygwin\bin\cygwin1.dll
S\system32\advapi32.dll
(sec/misc.c:489) GetUserNameA: stub
Entered debugger on exception number 14.
Entered kernel debugger (type "help" for a list of commands)
kdb:> bt
Frames: <ntoskrnl.exe: e326c (ob/security.c:311 (NtSetSecurityObject))>
<ntoskrnl.exe: 3e5b (:148 (KeUserModeCallback))>
<ntdll.dll: 3379 (napi.c:404 (NtSetSecurityObject))>
All of the above commands result in the "GetUserNameA: stub" error and
the message on the command prompt.
Andy
--
The cheese stands alone.
(io/arcname.c:375) Booting from CD-ROM!
(io/arcname.c:319) Could not find ntoskrnl.exe
IoCreateSystemRootLink FAILED: (0xc0000001) - %KERNEL-WRN-UNSUCCESSFUL
UNSUCCESSFUL
KeBugCheck at ke/main.c:687
Bug detected (code 7b param 0 0 0 0)
INACCESSIBLE_BOOT_DEVICE
Frames: <ntoskrnl.exe: a859> <ntoskrnl.exe: a879> <C0082E12>
<ntoskrnl.exe: c0a9> <C0083681> <ntoskrnl.exe: 115c>
this is using the livecd, haven't figured out how to mod the bootcd to
display that output.
Hi,
can you disable line #32 (#define NDEBUG)in ntoskrnl\io\arcname.c and try it
again?
- Hartmut
> -----Original Message-----
> From: ros-kernel-bounces(a)reactos.com
> [mailto:ros-kernel-bounces@reactos.com] On Behalf Of Richard Campbell
> Sent: Tuesday, September 28, 2004 7:34 PM
> To: ros-kernel(a)reactos.com
> Subject: [ros-kernel] BootCD doesn't boot...
>
>
> Hello all
>
> I'm using a fresh CVS checkout and ROS fails to boot in
> either VMWare or
> QEMU via the iso built by mingw32-make bootcd.
>
> Attached is a screen.
>
Hi!
I e-mailed Brian Palmer, he said it comes down to what I want to do (I asked
him if I could help with ReactOS.
Anything need done?
my email address is either pattersonmpt(a)aol.com or xtreme101(a)btinternet.com
Darren,
Hello,
The Wine guys are interested in our implementation of userenv.dll. I
know you guys dont normally care about releasing the libs as LGPL but I
want to check with the four of you before I change the license
header(s) and merge attempt to get a patch ready for Winehq. Is this
cool with the four of you?
Thanks
Steven
__________________________________
Do you Yahoo!?
Take Yahoo! Mail with you! Get it on your mobile phone.
http://mobile.yahoo.com/maildemo
Hello,
I have been trying to test Firefox and OpenOffice under ReactOS and it
seems to be having trouble looking up mangled names in msvcrt. Firefox
fails like this
(NTDLL:ldr/utils.c:1205) LdrGetExportByName(): failed to find
??3@YAXPAX@Z
and OpenOffice fails to find this name ??1type_info@@UAE@XZ
Both of them are implemented and point to code I imported from Wine a
long time ago. Anyone have any idea whats up?
Thanks
Steven
__________________________________
Do you Yahoo!?
Yahoo! Mail - You care about security. So do we.
http://promotions.yahoo.com/new_mail
Attached is a test program that demonstrates a problem with memory mapped
files. Basically, if you try to MapViewOfFile() a part of the file at
non-zero file offset, ReactOS will still map the beginning of the file.
I think I found the cause of this, the attached .diff fixes it. My fix
uncovered another (small) problem, ViewOffset is relative to the start of
the segment, not the start of the section. So instead of passing
SectionSegments[i].VirtualAddress we need to pass 0.
I'd appreciate it if someone who actually understands the Mm stuff can take
a look at this.
Gé van Geldorp.
99% of the windows users consider networking as
browsing internet. Give suport for TCP/IP , for a
basic browser (wish it were graphical, and most common
ethernet cards (kind of realtek or others widely used)
and you can say. ¡We have networking!.
An extra would be suport for some widely used modems.
Noone expect apache or emule working in an alpha OS,
even FTP is prescindible in 0.3.
Regards,
Lucio Diaz.
______________________________________________
Renovamos el Correo Yahoo!: ¡100 MB GRATIS!
Nuevos servicios, más seguridad
http://correo.yahoo.es
Hi all
Ros-dev is now functional. Please refrain from posting to ros-kernel
and use this new list instead - I've moved all subscriptions over to
this list. Ros-kernel will be discontinued in a few days.
Cheers
Jason
