Menu
▾
▴
reactos-kernel
You can subscribe to this list here.
| 2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(2) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(64) |
Oct
(438) |
Nov
(183) |
Dec
|
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
(132) |
May
(466) |
Jun
(366) |
Jul
(392) |
Aug
(31) |
Sep
(18) |
Oct
|
Nov
|
Dec
|
|
From: <kr...@mb...> - 2002-06-18 19:08:07
|
make -C drivers/lib/zlib install make[1]: *** No rule to make target `zlib.sym', needed by `../../../reactos/system32/zlib.a'. Stop. make[1]: Entering directory `C:/tmp/reactos/drivers/lib/zlib' make[1]: Leaving directory `C:/tmp/reactos/drivers/lib/zlib' make: *** [zlib_install] Error 2 BTW Source tree can be compiled with DBG=1 and KDBG=1 now. Thanks. D. |
|
From: Eric K. <ek...@rz...> - 2002-06-18 18:21:31
|
"Casper Hornstrup" <ch...@us...> wrote: > Great. IIRC symbolic registry links are volatile and must be recreated > after each boot. Does your tests confirm this? I found out that the registry link can be persistent (non volatile). The "SymbolicLinkValue" value can be deleted when the key is opened using REG_OPTION_OPEN_LINK. The key seems to be unaccessible from regedit or regedt32 once it is closed but I guess there is a way to delete it too. Eric |
|
From: Steven E. <ste...@ya...> - 2002-06-18 14:49:15
|
Note: forwarded message attached. __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com |
|
From: Casper H. <ch...@us...> - 2002-06-18 14:06:35
|
tir, 2002-06-18 kl. 15:42 skrev Eric Kohl: > Hi! > > > HKEY hKeyHandle; > DWORD dwDisposition; > DWORD dwLength; > > PWCHAR ValueName = L"SymbolicLinkValue"; > PWCHAR Buffer= L"\\Registry\\Machine\\SOFTWARE\\SMT"; > > /* create the key */ > RegCreateKeyExW(HKEY_LOCAL_MACHINE, > L"SOFTWARE\\Test\\Link", > 0, > NULL, > REG_OPTION_VOLATILE | REG_OPTION_CREATE_LINK, > KEY_ALL_ACCESS | KEY_CREATE_LINK, > NULL, > &hKeyHandle, > &dwDisposition); > > /* Note: length WITHOUT the terminating zero */ > dwLength = wcslen(Buffer) * sizeof(WCHAR); > > /* set the link value */ > RegSetValueExW(hKeyHandle, > ValueName, > 0, > REG_LINK, > (const BYTE *)Buffer, > dwLength); > > RegCloseKey(hKeyHandle); > > > I still have to find out whether registry links can be removed or changed. > > > Regards, > Eric Great. IIRC symbolic registry links are volatile and must be recreated after each boot. Does your tests confirm this? Casper |
|
From: Eric K. <ek...@rz...> - 2002-06-18 13:40:25
|
Hi!
One of the most mysterious details of the registry has been uncovered, the
registry link. Didn't you ever wonder what this REG_LINK value type is used
for? Or what about the key access right KEY_CREATE_LINK?
At least one registry link is used in Windows NT/2K/XP.
"HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet" is a link to one of the
"HKEY_LOCAL_MACHINE\SYSTEM\ControlSetXXX" keys (XXX can be any three digit
number, starting at 001).
The available documentation does not provide any useful information about
registry links other than trivial statements, like: 'A registry link is a
link to link registry keys.'
On the net, I found a posting about someones unsuccessful attempt to create
a registry link. This was the only useful information I found.
After some extensive research I got it working:
Let's assume you have a existing key "HKEY_LOCAL_MACHINE\SOFTWARE\SMT" and
the new key "HKEY_LOCAL_MACHINE\SOFTWARE\Test\Link" should point to it. Then
the following code will do the trick:
HKEY hKeyHandle;
DWORD dwDisposition;
DWORD dwLength;
PWCHAR ValueName = L"SymbolicLinkValue";
PWCHAR Buffer= L"\\Registry\\Machine\\SOFTWARE\\SMT";
/* create the key */
RegCreateKeyExW(HKEY_LOCAL_MACHINE,
L"SOFTWARE\\Test\\Link",
0,
NULL,
REG_OPTION_VOLATILE | REG_OPTION_CREATE_LINK,
KEY_ALL_ACCESS | KEY_CREATE_LINK,
NULL,
&hKeyHandle,
&dwDisposition);
/* Note: length WITHOUT the terminating zero */
dwLength = wcslen(Buffer) * sizeof(WCHAR);
/* set the link value */
RegSetValueExW(hKeyHandle,
ValueName,
0,
REG_LINK,
(const BYTE *)Buffer,
dwLength);
RegCloseKey(hKeyHandle);
I still have to find out whether registry links can be removed or changed.
Regards,
Eric
|
|
From: Steven E. <ste...@ya...> - 2002-06-18 09:31:05
|
http://www.k9wk.com/alinfo.html didnt know if this would be of any use to anyone but in case you want to see sources/examples of another, here it is. He has also started doing a little work on SEH. Thanks Steven __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com |
|
From: Eric K. <ek...@rz...> - 2002-06-18 07:08:49
|
ntoskrnl/cm/rtlfunc.c:
RtlQueryRegistryValues() must treat REG_MULTI_SZ and REG_EXPAND_SZ
values like REG_SZ values upon direct queries.
ntoskrnl/io/driver.c
IoDestroyDriverList(): Free memory of group and service entries.
IopCreateServiceListEntry(): simplified function and reduced memory
consumption.
Eric
|
|
From: David W. <we...@cw...> - 2002-06-17 23:19:40
|
On Mon, Jun 17, 2002 at 07:35:07PM +0200, KJK::Hyperion wrote: > whoopsie. Must be my old half-baked attempt at kernel-mode programming. I > was investigating why our null.sys is so big, and discovered that > Microsoft's uses a tighter alignment, and started experimenting linker > flags. Apologies > I think your changes should work; it is most likely a bug in some versions of gcc or binutils. In any case the space saving is marginal. |
|
From: KJK::Hyperion <no...@li...> - 2002-06-17 23:12:02
|
At 01.00 17/06/2002, you wrote: >2002-06-16 David Welch <we...@wh...> > > * drivers/dd/null/makefile: Commented out local LDFLAGS as > these cause bad relocations in the stripped image. whoopsie. Must be my old half-baked attempt at kernel-mode programming. I was investigating why our null.sys is so big, and discovered that Microsoft's uses a tighter alignment, and started experimenting linker flags. Apologies |
|
From: Joseph G. <gal...@va...> - 2002-06-17 23:09:21
|
apps/tests/tokentest.c * Fix stupidity that prevented tokentest from compiling Joseph |
|
From: Casper H. <ch...@us...> - 2002-06-17 23:08:12
|
man, 2002-06-17 kl. 17:51 skrev Eric Kohl: > PS.: Where, in the CVS tree, are we going to put user-mode services like > rpcss and eventlog? I think we should put them into the old 'services' > directory. Comments? > > > Eric I think so too. That seems like a good place to put them. Casper |
|
From: <kr...@mb...> - 2002-06-17 22:52:38
|
Errgh. With w98 boot floppy reactos kernel boots fine ! I am sorry for that. Could you update FAQ about that ? David V 06/17/2002 - 21:32, Eric Kohl napsal: > > <kr...@mb...> wrote: > > > > Is it possible, that PCDOS causes my problem ? > > > > I use PCDOS bootable diskette to switch my NT/Linux station into real DOS. > > That's not impossible! > > The current FreeDOS works as well as FreeLoader. Btw, using FreeLoader to > boot ReactOS is a lot faster than using DOS. > > Eric > > > > ---------------------------------------------------------------------------------------------------- > Sponsor's Message > ---------------------------------------------------------------------------------------------------- > Bringing you mounds of caffeinated joy > >>> http://thinkgeek.com/sf <<< > > _______________________________________________ > reactos-kernel mailing list > rea...@li... > https://lists.sourceforge.net/lists/listinfo/reactos-kernel > |
|
From: Joseph G. <gal...@va...> - 2002-06-17 22:50:08
|
ntoskrnl/ke/main.c * Move SeInit2() to before we initialize process management so we can create an initial system token to go with the initial system process. ntoskrnl/ps/process.c * Implement NtOpenProcessToken() by calling _NtOpenProcessToken() -- why do we have both of these? * Fix bug in _NtOpenProcessToken() that caused us to crash creating the token handle. * Return STATUS_INVALID_INFO_CLASS from NtQueryInformationProcess() and NtSetInformationProcess() for information classes documented as unqueryable or unsettable by Gary Nebbet in his book "Native API Reference". * Fix bug in NtSetInformationProcess() where it assigns the primary token -- it derefeneced the old token if SeExchangePrimaryToken() failed; it should have derferenced it if it succeeded. * Call SepCreateSystemProcessToken() when creating the initial system process. Cleanup the token during in PiDeleteProcess(). * Call SepInitializeNewProcess() during NtCreateProcess to copy the parent processes token to the new process. ntoskrnl/include/internal/se.h ntoskrnl/se/token.c * Change name of SepDuplicationToken() [which was unimplemented] to SepDuplicateToken(). Implement it. * Implement new functions (which are used by process.c) SepInitializeNewProcess() and SepCreateSystemToken(). * Correctly (I think) set Token->ImpersonationLevel in NtCreateToken(). apps/tests/tokentest/tokentest.c * Dump the current processes token. - Joseph |
|
From: <kr...@mb...> - 2002-06-17 22:16:01
|
(ke/main.c: 198) : NtOpenSymbolicLinkObject() failed (Status c0000001) David > Then have a look at InitSystemSharedUserPage() in ntoskrnl/ke/main.c. There > are several calls to KeBugCheck(). If the function above the KeBugCheck() > call is a CPRINT macro, change "CPRINT" to "DPRINT1" and rebuild ntoskrnl > (run "make ntoskrnl"), install it and boot it. Now you should get more > detailed information about where ntoskrnl actually crashes. > > Eric > > > > ---------------------------------------------------------------------------------------------------- > Sponsor's Message > ---------------------------------------------------------------------------------------------------- > Bringing you mounds of caffeinated joy > >>> http://thinkgeek.com/sf <<< > > _______________________________________________ > reactos-kernel mailing list > rea...@li... > https://lists.sourceforge.net/lists/listinfo/reactos-kernel > |
|
From: Eric K. <ek...@rz...> - 2002-06-17 21:21:26
|
<kr...@mb...> wrote: > Is it possible, that PCDOS causes my problem ? > > I use PCDOS bootable diskette to switch my NT/Linux station into real DOS. That's not impossible! The current FreeDOS works as well as FreeLoader. Btw, using FreeLoader to boot ReactOS is a lot faster than using DOS. Eric |
|
From: Eric K. <ek...@rz...> - 2002-06-17 21:17:36
|
<kr...@mb...> wrote: > I'm so paranoid that I do full rebuild from clear updated source every time. Then have a look at InitSystemSharedUserPage() in ntoskrnl/ke/main.c. There are several calls to KeBugCheck(). If the function above the KeBugCheck() call is a CPRINT macro, change "CPRINT" to "DPRINT1" and rebuild ntoskrnl (run "make ntoskrnl"), install it and boot it. Now you should get more detailed information about where ntoskrnl actually crashes. Eric |
|
From: <kr...@mb...> - 2002-06-17 21:11:24
|
Is it possible, that PCDOS causes my problem ? I use PCDOS bootable diskette to switch my NT/Linux station into real DOS. David |
|
From: <kr...@mb...> - 2002-06-17 21:04:49
|
I'm so paranoid that I do full rebuild from clear updated source every time. David V 06/17/2002 - 20:53, Eric Kohl napsal: > > "David Kredba" <kr...@mb...> wrote: > > Hello David! > > > It told : > > > > (ke/main.c:710) CommandLine: > > multi(0)disk(0)rdisk(0)partition(1)reactos /DEBUGPORT=SCREEN > > > > and crash in the same funtctions. > > Do you have a clean source tree or did you update your source tree when the > crashes started? > If you updated your source tree, make a full rebuild. Run "make clean" and > then "make". > > Since you seem to be the only one who experiences this crash, I assume you > didn't run a full rebuild after the update. > > > Eric > > > > _______________________________________________________________ > > Sponsored by: > ThinkGeek at http://www.ThinkGeek.com/ > _______________________________________________ > reactos-kernel mailing list > rea...@li... > https://lists.sourceforge.net/lists/listinfo/reactos-kernel > |
|
From: Eric K. <ek...@rz...> - 2002-06-17 20:45:14
|
"David Kredba" <kr...@mb...> wrote: Hello David! > It told : > > (ke/main.c:710) CommandLine: > multi(0)disk(0)rdisk(0)partition(1)\reactos /DEBUGPORT=SCREEN > > and crash in the same funtctions. Do you have a clean source tree or did you update your source tree when the crashes started? If you updated your source tree, make a full rebuild. Run "make clean" and then "make". Since you seem to be the only one who experiences this crash, I assume you didn't run a full rebuild after the update. Eric |
|
From: David K. <kr...@mb...> - 2002-06-17 19:47:39
|
It told : (ke/main.c:710) CommandLine: multi(0)disk(0)rdisk(0)partition(1)\reactos /DEBUGPORT=SCREEN and crash in the same funtctions. David > > Is should say something like > "multi(0)disk(0)rdisk(0)partition(1)\reactos...". > > > Eric > |
|
From: Eric K. <ek...@rz...> - 2002-06-17 15:48:20
|
lib/ntdll/rtl/registry.c
RtlQueryRegistryValues(): Implemented direct query of REG_EXPAND_SZ
values.
subsys/system/services/database.c
Improved driver loading sequence with respect to the group order.
Added ability to run user-mode services.
PS.: Where, in the CVS tree, are we going to put user-mode services like
rpcss and eventlog? I think we should put them into the old 'services'
directory. Comments?
Eric
|
|
From: Steven E. <ste...@ya...> - 2002-06-17 05:56:51
|
Changelog: reactos/apps/tcpsvr * tcpsvr.c * makefile Added test tcp server. __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com |
|
From: David W. <we...@cw...> - 2002-06-16 21:58:14
|
2002-06-16 David Welch <we...@wh...> * tools/helper.mk: Make the install target depend on all the files to be installed. 2002-06-16 David Welch <we...@wh...> * ntoskrnl/ps/thread.c (NtCallbackReturn): Set TSS.Esp0 to the top of the old stack. * ntoskrnl/ps/thread.c (NtW32Call): Set TSS.Esp0 to the top of the new stack. Free the callback stack correctly. Don't copy portion of the trap frame that doesn't exist in non-v86-mode interrupts. * ntoskrnl/ps/thread.c (PsFreeCallbackStack): New function to free a stack allocated with PsAllocateCallbackStack. 2002-06-16 David Welch <we...@wh...> * drivers/dd/null/makefile: Commented out local LDFLAGS as these cause bad relocations in the stripped image. 2002-06-16 David Welch <we...@wh...> * config: Corrected spelling error. |
|
From: James M. <jid...@sa...> - 2002-06-16 20:57:03
|
Thanks for pointing this out. James Marjie GnuPG KeyID: 0x7C837497 "Take your life in your own hands, and what happens? A terrible thing: no one to blame." -Erica Jong |
|
From: Jason F. <jas...@ya...> - 2002-06-16 20:48:55
|
reactos/ntoskrnl/ps/thread.c NtW32Call: replaced sizeof(KTRAP_FRAME) with 124 bytes (thanks to Eugene Ingerman for pointing this out) NtW32Call now crashes when calling KePushAndStackSwitchAndSysRet. Hartmut you worked on that function recently - does the change to 124 bytes affect this function? Thanks - Jason __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com |
8 messages has been excluded from this view by a project administrator.
