ion(a)svn.reactos.com wrote:
>- Small semaphore code simplification/optimization.
>
>
>Updated files:
>trunk/reactos/ntoskrnl/ke/kthread.c
>trunk/reactos/ntoskrnl/ke/sem.c
>
>_______________________________________________
>Ros-svn mailing list
>Ros-svn(a)reactos.org
>http://www.reactos.org/mailman/listinfo/ros-svn
>
>
>
Have you tried to compile sem.c before you did commit?
- Hartmut
Hello,
Since some weeks/months, I've done some work in setupapi.dll.
I'm now able to install devices using provided .inf files, ie copy needed
files and create registry entries.
Syntax:
rundll32.exe newdev.dll,DevInstall [InstanceID] where InstanceID is
something like ACPI\PNP0501\1 or
PCI\VEN_8086&DEV_7112&SUBSYS_00000000&REV_01\3&61aaa01&0&3A
However, I've still some bugs and some missing features:
- OpenSCManagerW freezes ReactOS [1]
- CreateServiceW is unimplemented [2]
- Service image path provided to CreateServiceW looks like
"C:\ReactOS\System32\drivers\driver.sys", instead of
"System32\drivers\driver.sys" [3]
- You have no way to use a .inf file which is not in ReactOS\Inf directory
[4]
- Files to be copied are always searched in ReactOS\Inf directory (so they
can't be found) [5]
- Files to be copied need to be already decompressed (ie not in a .cab file)
[6]
- A reboot is needed before being able to use the newly installed device [7]
- Having multiple times a same device is not supported, if the bus driver
can't provide a unique DeviceID [8]
- Last, but not least, device installation is not started automatically when
a new device is found/plugged [9]
[1] See bug #826: http://www.reactos.org/bugzilla/show_bug.cgi?id=826
[2] In lib/advapi32/service/scm.c. It needs to call umpnpmgr.exe by RPC, as
in OpenServiceW implementation
[3] In lib/setupapi/dirid.c: the value returned by create_system_dirid() is
always a full path
[4] It's due to missing functionnality in newdev.dll (lib/newdev/*), which
should ask the user when no driver is found
[5] In lib/setupapi/queue.c. If a file can't be found, SetupCommitFileQueueW
should ask the user where to find the file
[6] We need to support the [SourceDisksNames] section in .inf files
[7] Maybe calling IoInvalidateDeviceRelations() at the end of
SetupDiInstallDevice (lib/setupapi/devinst.c) would be enough
[8] We need to generate a ParentIdPrefix entry in registry in some(?) cases
[9] I suppose it should be initiated by umpnpmgr.exe, when it sees a device
for the first time
Hervé
PS: If someone is interested to work on one of this parts, please answer
this email
ion(a)svn.reactos.com wrote:
>- Add RtlSetProcessCritical
>- Some fixes.
>
>
>Updated files:
>trunk/reactos/include/ndk/ketypes.h
>trunk/reactos/include/ndk/pstypes.h
>trunk/reactos/include/ndk/rtlfuncs.h
>trunk/reactos/include/ndk/rtltypes.h
>trunk/reactos/include/ndk/umtypes.h
>
>_______________________________________________
>Ros-svn mailing list
>Ros-svn(a)reactos.org
>http://www.reactos.org/mailman/listinfo/ros-svn
>
>
>
Hi,
the header changes breaks compiling for me:
gcc -c boot\freeldr\freeldr\cache\blocklist.c -o
.\up\obj-i386\boot\freeldr\freeldr\cache\blocklist.o -nostdlib
-ffreestanding -fno-builtin -fno-inline -fno-zero-initialized-in-bss -Os
-Iboot\freeldr\freeldr\include -Iboot\freeldr\freeldr\cache
-Intoskrnl\include -D__USE_W32API -DDEBUG -Iinclude -Iinclude\reactos
-Iinclude\libs -Iinclude\drivers -Iinclude\subsys -Iw32api\include
-Iw32api\include\ddk -D_M_IX86 -DDBG -Wall -Os -Wno-strict-aliasing
-ftracer -momit-leaf-frame-pointer -mpreferred-stack-boundary=2
-Wpointer-arith -g -pipe
In file included from include/ndk/psfuncs.h:13,
from include/ndk/ntndk.h:34,
from boot/freeldr/freeldr/include/freeldr.h:24,
from boot\freeldr\freeldr\cache\blocklist.c:21:
include/ndk/pstypes.h:627: error: syntax error before 'PHANDLE_TABLE'
include/ndk/pstypes.h:627: warning: no semicolon at end of struct or union
include/ndk/pstypes.h:628: warning: type defaults to 'int' in
declaration of 'Token'
include/ndk/pstypes.h:628: warning: data definition has no type or
storage class
include/ndk/pstypes.h:655: warning: unnamed struct/union that defines no
instances
include/ndk/pstypes.h:668: error: syntax error before 'PrefetchTrace'
include/ndk/pstypes.h:668: warning: type defaults to 'int' in
declaration of 'PrefetchTrace'
include/ndk/pstypes.h:668: warning: data definition has no type or
storage class
include/ndk/pstypes.h:718: warning: unnamed struct/union that defines no
instances
include/ndk/pstypes.h:730: warning: unnamed struct/union that defines no
instances
include/ndk/pstypes.h:745: error: syntax error before '}' token
include/ndk/pstypes.h:745: warning: type defaults to 'int' in
declaration of 'EPROCESS'
include/ndk/pstypes.h:745: warning: data definition has no type or
storage class
_make: *** [.\up\obj-i386\boot\freeldr\freeldr\cache\blocklist.o] Error 1
- Hartmut
ion(a)svn.reactos.com wrote:
>Deleted files:
>trunk/reactos/include/subsys/csr/protocol.h
>trunk/reactos/subsys/csr/args.c
>trunk/reactos/subsys/csr/csr.h
>trunk/reactos/subsys/csr/csrsrv/debug.c
>trunk/reactos/subsys/csr/csrsrv/dllmain.c
>
>
Well, a bit unfair deleting header files etc. and leaving the subtree
not compilable.
My working copy depends on those very files.
Disabling compilation of csr also forces me to disable the vms module:
not a great loss, but we had only two left: os2 and posix are broken.
Keeping the whole modules compilable should be a goal, even if they are
minimal subsystems (if ROS is just Win32, well, let's rename it back
"FreeWin95").
--
:Emanuele Aliberti
Hi,
I found the OPTHDROFFSET not very useful.
I would rather use something lighter like:
(PIMAGE_NT_HEADERS)((PBYTE)a + ((PIMAGE_DOS_HEADER)a)->e_lfanew))
, same as RtlImageNtHeader()
May i ask someone with write access to remove it?
Hey everyone,
The website looks great. Props to the creators and website admin...
Rick Langschultz
rlangschultz(a)cox.net (Home)
rlangschultz(a)ellemaespa.com (Work)
rlangschultz(a)email.uophx.edu (School)
I want to make a spanish translation of the web
page...once i finish, how does i submit it?.
Second question:
I can not LOG!, i find no sense to the code validator,
what the hell imput is it waiting? The code validator
is NOT human readable unless you explain how does it
work!.
Best regards,
Lucio Diaz-Flores.
______________________________________________
Renovamos el Correo Yahoo!
Nuevos servicios, más seguridad
http://correo.yahoo.es