20 months ago[SHELLEXT][ZIPFLDR] Performance: Initialize m_Filename quickly
Katayama Hirofumi MZ [Tue, 22 Oct 2019 00:55:06 +0000 (09:55 +0900)]
[SHELLEXT][ZIPFLDR] Performance: Initialize m_Filename quickly

20 months ago[CPL][DESK] hShell32 is always non-NULL in epilogue
Katayama Hirofumi MZ [Tue, 22 Oct 2019 00:43:34 +0000 (09:43 +0900)]
[CPL][DESK] hShell32 is always non-NULL in epilogue

20 months ago[CPL][HDWWIZ] Move 'Index' and 'Item' variables to prologue
Katayama Hirofumi MZ [Tue, 22 Oct 2019 00:41:09 +0000 (09:41 +0900)]
[CPL][HDWWIZ] Move 'Index' and 'Item' variables to prologue

20 months ago[WIN32SS][NTGDI] Unite duplicate of condition 'UsesSource'
Katayama Hirofumi MZ [Tue, 22 Oct 2019 00:33:32 +0000 (09:33 +0900)]
[WIN32SS][NTGDI] Unite duplicate of condition 'UsesSource'

20 months ago[WIN32SS][NTGDI] 'otm' is always non-NULL in epilogue
Katayama Hirofumi MZ [Tue, 22 Oct 2019 00:29:15 +0000 (09:29 +0900)]
[WIN32SS][NTGDI] 'otm' is always non-NULL in epilogue

20 months ago[COMCTL32] Address regression CORE-14622
Joachim Henze [Mon, 21 Oct 2019 23:10:28 +0000 (01:10 +0200)]
[COMCTL32] Address regression CORE-14622

"Avira Antivir Personal" had issues painting its
setup wizard dialog with correct background color.
That part was a regression of SVN r74136 == git

and also did not draw its umbrella bmp, which did never work
in ros up to now afaik.,69478 VBOX,69479 KVM

Thanks to patches author Doug Lyons and HBelusca.

Wine has been reported to be affected by that bug as well:
Not sure whether our patch will be accepted by Wine,
to my naive eye this looks like an interim solution.

20 months ago[NTOSKRNL] Rewrite IoAssignDriveLetters to make NT5 compliant
Pierre Schweitzer [Mon, 21 Oct 2019 16:28:40 +0000 (18:28 +0200)]
[NTOSKRNL] Rewrite IoAssignDriveLetters to make NT5 compliant

The major change with this rewrite is the support for the mount
manager. Fstub will now assume that most of the devices are PnP
and that they are already registered to the mount manager.
It will thus ask the mount manager to assign the drive letter.
Fstub will keep assigning drive letters non mission critical devices
such as CDs, floppies and other removable devices.

See MountMgr:QueryPoints API test that will now return mount points :-).

20 months ago[DISK] Forcibly declare our partitions to the MountMgr
Pierre Schweitzer [Mon, 21 Oct 2019 14:50:36 +0000 (16:50 +0200)]
[DISK] Forcibly declare our partitions to the MountMgr

Because our disk.sys doesn't do anything related to PnP
(compared to disk_new.sys), forcibly declare our partitions
to the MountMgr so that it can references them and assign
them a DOS drive letter on demand later on.

20 months ago[CLASS2] Hackplement support for IOCTL_MOUNTDEV_QUERY_UNIQUE_ID
Pierre Schweitzer [Mon, 21 Oct 2019 14:45:28 +0000 (16:45 +0200)]

This is required so that MountMgr can handle devices that are still
using class2 instead of classpnp.
Given we have no unique ID to return, we'll return device path, which
is far from perfect but which is enough for now to have everything

20 months ago[NTOSKRNL] Define TAG_FSTUB in the dedicated internal header
Pierre Schweitzer [Mon, 21 Oct 2019 12:26:00 +0000 (14:26 +0200)]
[NTOSKRNL] Define TAG_FSTUB in the dedicated internal header

20 months ago[NTOSKRNL] Declare IoRemoteBootClient in internal headers
Pierre Schweitzer [Mon, 21 Oct 2019 12:16:58 +0000 (14:16 +0200)]
[NTOSKRNL] Declare IoRemoteBootClient in internal headers

20 months ago[MOUNTMGR] Fix setting up reparse index file name
Pierre Schweitzer [Mon, 21 Oct 2019 10:58:30 +0000 (12:58 +0200)]
[MOUNTMGR] Fix setting up reparse index file name

This fixes memory smashing while attempting to volume
reparse index (we were previously trying to copy the
name on itself, in the middle of itself...).
This code won't go farther on FAT, it requires NTFS.

Now, with this, ReactOS can properly boot with MountMgr
handling DOS devices without any crash or code disabled.
Cf: what was written in 7608ac9.

Modifications in class2, disk, and ntoskrnl are still to
be committed to enable all this.

20 months ago[MOUNTMGR] That was not meant to be committed
Pierre Schweitzer [Mon, 21 Oct 2019 09:03:43 +0000 (11:03 +0200)]
[MOUNTMGR] That was not meant to be committed

Even though it shows there might be a bug in the
code handling remote databases in the MountMgr ;-)

Addendum to 7608ac9

20 months ago[MOUNTMGR] Misc. fixes for WorkerThread()
Pierre Schweitzer [Mon, 21 Oct 2019 08:54:51 +0000 (10:54 +0200)]
[MOUNTMGR] Misc. fixes for WorkerThread()

- Properly quit the active loop when we're out of work items;
- Fix timeout duration (setting it to 1s);
- Fix handling the "Unloading" variable in case of a shutdown
  so that waiting loop is properly stopped;
- Documented why we're waiting on VolumesSafeForWriteAccess.

This fixes shutting down ReactOS with work items queued.
This is needed here because no one ever sets that event (properly)
created by SMSS though. A. Ionescu was explaining in 2018 that it's
autochk responsibility, but it doesn't seem to be the case in W2K3.
To be investigated.

This fix with all the previous ones and more uncommitted stuff (yet ;-))
allows reaching the first steps towards a NT5 storage stack:

20 months ago[IPCONFIG] Implement the /flushdns option and most of the /displaydns option
Eric Kohl [Sun, 20 Oct 2019 21:24:28 +0000 (23:24 +0200)]
[IPCONFIG] Implement the /flushdns option and most of the /displaydns option

20 months ago[DNSAPI][REACTOS] Add a header for undocumented dnsapi functions and use it in dnsapi.dll
Eric Kohl [Sun, 20 Oct 2019 21:13:46 +0000 (23:13 +0200)]
[DNSAPI][REACTOS] Add a header for undocumented dnsapi functions and use it in dnsapi.dll

20 months ago[HNETCFG_WINETEST] Sync with Wine Staging 4.18. CORE-16441
Amine Khaldi [Sun, 20 Oct 2019 19:04:04 +0000 (20:04 +0100)]
[HNETCFG_WINETEST] Sync with Wine Staging 4.18. CORE-16441

20 months ago[HNETCFG] Sync with Wine Staging 4.18. CORE-16441
Amine Khaldi [Sun, 20 Oct 2019 19:03:33 +0000 (20:03 +0100)]
[HNETCFG] Sync with Wine Staging 4.18. CORE-16441

20 months ago[PSDK][UUID] Import natupnp.idl from Wine Staging 4.18. CORE-16441
Amine Khaldi [Sun, 20 Oct 2019 19:01:58 +0000 (20:01 +0100)]
[PSDK][UUID] Import natupnp.idl from Wine Staging 4.18. CORE-16441

20 months ago[DEVENUM_WINETEST] Sync with Wine Staging 4.18. CORE-16441
Amine Khaldi [Sun, 20 Oct 2019 17:56:45 +0000 (18:56 +0100)]
[DEVENUM_WINETEST] Sync with Wine Staging 4.18. CORE-16441

20 months ago[DEVENUM] Sync with Wine Staging 4.18. CORE-16441
Amine Khaldi [Sun, 20 Oct 2019 17:56:05 +0000 (18:56 +0100)]
[DEVENUM] Sync with Wine Staging 4.18. CORE-16441

20 months ago[PSDK] Add missing OVERLAPPED_ENTRY. CORE-16441
Amine Khaldi [Sun, 20 Oct 2019 16:58:55 +0000 (17:58 +0100)]

20 months ago[SHELL32] CDefaultContextMenu: Don't render a medium just to query if it is present 1977/head
Mark Jansen [Sun, 20 Oct 2019 12:21:28 +0000 (14:21 +0200)]
[SHELL32] CDefaultContextMenu: Don't render a medium just to query if it is present

20 months ago[SHELL32] CDefaultContextMenu: Always add the wanted operation to the IDataObject
Mark Jansen [Sat, 19 Oct 2019 22:54:11 +0000 (00:54 +0200)]
[SHELL32] CDefaultContextMenu: Always add the wanted operation to the IDataObject

20 months ago[SDK][SHELL32] Augment the internally used IDataObject with some extra formats
Mark Jansen [Sat, 19 Oct 2019 22:44:03 +0000 (00:44 +0200)]
[SDK][SHELL32] Augment the internally used IDataObject with some extra formats
This is needed because our code seems to use CF_HDROP a lot, instead of HIDA...

20 months ago[SHELL32] Mark an unexpected failure as unexpected
Mark Jansen [Sat, 19 Oct 2019 22:42:16 +0000 (00:42 +0200)]
[SHELL32] Mark an unexpected failure as unexpected

20 months ago[SHELL32_APITEST] Show that SHCreateDataObject behaves exactly like CIDLData_CreateFr...
Mark Jansen [Sat, 19 Oct 2019 22:37:05 +0000 (00:37 +0200)]
[SHELL32_APITEST] Show that SHCreateDataObject behaves exactly like CIDLData_CreateFromIDArray

20 months ago[SHELL32] Properly release an StgMedium
Mark Jansen [Sat, 19 Oct 2019 22:27:58 +0000 (00:27 +0200)]
[SHELL32] Properly release an StgMedium

20 months ago[SHELL32_APITEST] Show that the default data object does not have a preferred drop...
Mark Jansen [Sat, 19 Oct 2019 21:56:19 +0000 (23:56 +0200)]
[SHELL32_APITEST] Show that the default data object does not have a preferred drop effect

20 months ago[FONTEXT] Return the correct error in GetDisplayNameOf
Mark Jansen [Sat, 19 Oct 2019 21:55:39 +0000 (23:55 +0200)]
[FONTEXT] Return the correct error in GetDisplayNameOf

20 months ago[NETCFGX] Free all component data on INetCfg_fnUninitialize()
Eric Kohl [Sun, 20 Oct 2019 08:04:08 +0000 (10:04 +0200)]
[NETCFGX] Free all component data on INetCfg_fnUninitialize()

20 months ago[MBEDTLS] Update to version 2.7.12. CORE-16440
Thomas Faber [Sun, 20 Oct 2019 10:44:34 +0000 (12:44 +0200)]
[MBEDTLS] Update to version 2.7.12. CORE-16440

20 months ago[MPG123] Update to version 1.25.12. CORE-16440
Thomas Faber [Sun, 20 Oct 2019 10:19:51 +0000 (12:19 +0200)]
[MPG123] Update to version 1.25.12. CORE-16440

20 months ago[BZIP2] Remove unused library. CORE-16440
Thomas Faber [Sun, 20 Oct 2019 10:04:41 +0000 (12:04 +0200)]
[BZIP2] Remove unused library. CORE-16440

This was once used by the "ramdrv" driver, removed in r37739 (c47ab26e825).

20 months ago[MSVCRT] Export __acrt_iob_func to fix GCC build with latest RosBE 2.2 (#1835)
Victor Perevertkin [Sun, 20 Oct 2019 10:44:17 +0000 (13:44 +0300)]
[MSVCRT] Export __acrt_iob_func to fix GCC build with latest RosBE 2.2 (#1835)

This adds an evil hack to persuade libstdc++, which tries to import __acrt_iob_func from a DLL.
This can only be solved cleanly by adding a GCC-compatible C++ standard library to our tree later.

20 months ago[FONTEXT] Initial implementation 1907/head
Mark Jansen [Sat, 3 Aug 2019 13:14:20 +0000 (15:14 +0200)]
[FONTEXT] Initial implementation
Create Fonts\desktop.ini when registering the shell ext
Also list the shell extension as needing to be registered at install

20 months ago[SHELL32] Update CDataObject,
Mark Jansen [Sat, 12 Oct 2019 00:11:25 +0000 (02:11 +0200)]
[SHELL32] Update CDataObject,
allowing it to accept new formats and behave a bit more like windows

20 months ago[SHELL32_APITEST] Add test for CIDLData_CreateFromIDArray
Mark Jansen [Thu, 10 Oct 2019 22:15:41 +0000 (00:15 +0200)]
[SHELL32_APITEST] Add test for CIDLData_CreateFromIDArray

20 months ago[FONTVIEW] Add /d option
Mark Jansen [Sat, 14 Sep 2019 21:06:10 +0000 (23:06 +0200)]
[FONTVIEW] Add /d option

20 months ago[SHELL32] Add support for custom View objects defined with UICLSID
Mark Jansen [Sun, 4 Aug 2019 00:42:09 +0000 (02:42 +0200)]
[SHELL32] Add support for custom View objects defined with UICLSID

20 months ago[ZIPFLDR] Fix number of menu items returned for DFM_MERGECONTEXTMENU
Mark Jansen [Fri, 13 Sep 2019 19:37:07 +0000 (21:37 +0200)]
[ZIPFLDR] Fix number of menu items returned for DFM_MERGECONTEXTMENU

20 months ago[PSDK] Add some missing defines to shlobj.h
Mark Jansen [Fri, 13 Sep 2019 19:35:51 +0000 (21:35 +0200)]
[PSDK] Add some missing defines to shlobj.h

20 months ago[FONTEXT_APITEST] Add simple api test showing interfaces
Mark Jansen [Sat, 3 Aug 2019 13:13:21 +0000 (15:13 +0200)]
[FONTEXT_APITEST] Add simple api test showing interfaces

20 months ago[SHELL32] Demote an ERR to WARN
Mark Jansen [Sun, 4 Aug 2019 00:42:31 +0000 (02:42 +0200)]
[SHELL32] Demote an ERR to WARN

20 months ago[MOUNTMGR] Don't kill Mm when a device has several symlinks
Pierre Schweitzer [Sat, 19 Oct 2019 16:04:30 +0000 (18:04 +0200)]
[MOUNTMGR] Don't kill Mm when a device has several symlinks

On preflight to compute output size, device name & unique ID were
counted only once per device.
Then, on copy, these two were copied on every MOUNTMGR_MOUNT_POINT
structure. This is counter efficient (data duplication) but also,
it was overruning the output buffer, since the preflight was not
expecting these extra copies.

20 months ago[MOUNTMGR_APITEST] Dump offsets
Pierre Schweitzer [Sat, 19 Oct 2019 16:00:36 +0000 (18:00 +0200)]

That shows that some offset are shared (everything but the symlinks)

20 months ago[MOUNTMGR] Fix enumerating drive letter for creating new mountpoint
Pierre Schweitzer [Sat, 19 Oct 2019 14:56:21 +0000 (16:56 +0200)]
[MOUNTMGR] Fix enumerating drive letter for creating new mountpoint

This is purely a copypasta error fix, which was causing MountMgrNextDriveLetterWorker
to fail as no drive letters were enumerated previously.

With that set of patches, MountMgr now properly assigns drive letters to
new devices!

20 months ago[MOUNTMGR] Fix interpretation of QueryDeviceInformation GptDriveLetter
Pierre Schweitzer [Sat, 19 Oct 2019 14:53:45 +0000 (16:53 +0200)]
[MOUNTMGR] Fix interpretation of QueryDeviceInformation GptDriveLetter

Being TRUE doesn't mean the device is GPT and has a drive letter. It just
means that it's not a GPT device with GPT_BASIC_DATA_ATTRIBUTE_NO_DRIVE_LETTER
attribute. In short, if TRUE, it means that the device can receive a drive
letter mount point.

This fixes MountMgrNextDriveLetterWorker bailing out for any attempt to
assign a drive letter to a device.

20 months ago[MOUNTMGR] Fix IsFtVolume so that it returns TRUE only for FT volumes
Pierre Schweitzer [Sat, 19 Oct 2019 14:52:29 +0000 (16:52 +0200)]
[MOUNTMGR] Fix IsFtVolume so that it returns TRUE only for FT volumes

20 months ago[MOUNTMGR] Query proper device when creating mount point
Pierre Schweitzer [Sat, 19 Oct 2019 14:34:38 +0000 (16:34 +0200)]
[MOUNTMGR] Query proper device when creating mount point

We must query the target device, and not the symbolic link
we attempt to create. The later will always fail as it doesn't
exist yet.

This fixes MountMgrCreatePointWorker not working.

20 months ago[MOUNTMGR] Remove cancel routine before completing pending IRP
Pierre Schweitzer [Sat, 19 Oct 2019 14:29:37 +0000 (16:29 +0200)]
[MOUNTMGR] Remove cancel routine before completing pending IRP

20 months ago[BROWSEUI][SHELLFIND] Update de-DE.rc CORE-16427
Joachim Henze [Sat, 19 Oct 2019 14:36:02 +0000 (16:36 +0200)]
[BROWSEUI][SHELLFIND] Update de-DE.rc CORE-16427

Translate the new checkboxes text to german

Addendum to 0.4.14-dev-60-g

20 months ago[SHELLFIND] Add Hidden Files & Folders Selection. CORE-16427 1968/head
Doug Lyons [Sun, 13 Oct 2019 23:48:08 +0000 (18:48 -0500)]
[SHELLFIND] Add Hidden Files & Folders Selection. CORE-16427

20 months ago[WSHTCPIP] Print IOCTL as hexa value
Pierre Schweitzer [Sat, 19 Oct 2019 09:46:24 +0000 (11:46 +0200)]
[WSHTCPIP] Print IOCTL as hexa value

20 months ago[MSAFD] SizeOf*Buffer: Type is ULONG, not INT/DWORD
Serge Gautherie [Thu, 17 Oct 2019 10:33:29 +0000 (12:33 +0200)]
[MSAFD] SizeOf*Buffer: Type is ULONG, not INT/DWORD

20 months ago[GDI32] Update Gdi Driver Header.
jimtabor [Sat, 19 Oct 2019 04:49:13 +0000 (23:49 -0500)]
[GDI32] Update Gdi Driver Header.

Patch by : Akihiro Sagawa : Implement GetCharWidthInfo().

20 months agoRevert "[NTOS:KDBG] Use CONTEXT instead of KTRAP_FRAME"
Hervé Poussineau [Fri, 18 Oct 2019 22:12:41 +0000 (00:12 +0200)]
Revert "[NTOS:KDBG] Use CONTEXT instead of KTRAP_FRAME"

This reverts commit e5bffe49da21f5faf2a88a0983ad4876d388e0ea.


20 months ago[MOUNTMGR] QueryPointsFromMemory: take into account the multiple MOUNTMGR_MOUNT_POINT
Pierre Schweitzer [Thu, 17 Oct 2019 20:40:23 +0000 (22:40 +0200)]
[MOUNTMGR] QueryPointsFromMemory: take into account the multiple MOUNTMGR_MOUNT_POINT

This fixes returning too small structure on an IOCTL_MOUNTMGR_QUERY_POINTS call.
The multiple MOUNTMGR_MOUNT_POINT structures were ignored and thus the data of the
first one were erased by the multiple structures.
MountMgr now returns consistent output on this IOCTL call.

20 months ago[CODEOWNERS] Add /drivers/filters/*/ entries
Serge Gautherie [Sat, 5 Oct 2019 21:53:48 +0000 (23:53 +0200)]
[CODEOWNERS] Add /drivers/filters/*/ entries

Also add Pierre's Jira username.

20 months ago[MSAFD] WSPSetSockOpt(): Fix a copypasta
Serge Gautherie [Thu, 17 Oct 2019 00:39:30 +0000 (02:39 +0200)]
[MSAFD] WSPSetSockOpt(): Fix a copypasta

Addendum to 6bc61f63f1e0ad20ab09f881d6968558eba9a8e6.


20 months ago[MOUNTMGR_APITEST] Dump returned mount points
Pierre Schweitzer [Wed, 16 Oct 2019 20:22:30 +0000 (22:22 +0200)]
[MOUNTMGR_APITEST] Dump returned mount points

20 months ago[MOUNTMGR] QueryPointsFromSymbolicLinkName(): Fix a copypasta
Serge Gautherie [Mon, 14 Oct 2019 12:01:44 +0000 (14:01 +0200)]
[MOUNTMGR] QueryPointsFromSymbolicLinkName(): Fix a copypasta

Addendum to r55156.


20 months ago[umpnpmgr][ndk] Implement PNP_RequestDeviceEject() and update unknown type in PLUGPLA...
Eric Kohl [Tue, 15 Oct 2019 20:40:06 +0000 (22:40 +0200)]
[umpnpmgr][ndk] Implement PNP_RequestDeviceEject() and update unknown type in PLUGPLAY_CONTROL_QUERY_REMOVE_DATA

20 months ago[SDK] Fix a typo, caught by Stas'M :)
Hermès Bélusca-Maïto [Sun, 13 Oct 2019 23:36:42 +0000 (01:36 +0200)]
[SDK] Fix a typo, caught by Stas'M :)

20 months ago[FREELDR] Set LDRP_IMAGE_INTEGRITY_FORCED if necessary. (Part of PR #1905)
Mark Harmstone [Wed, 9 Oct 2019 20:28:33 +0000 (21:28 +0100)]
[FREELDR] Set LDRP_IMAGE_INTEGRITY_FORCED if necessary. (Part of PR #1905)

+ Add extra information.
Co-authored-by: Hermès Bélusca-Maïto <>
20 months ago[NDK] Fix coding style and indentation. No code changes!
Eric Kohl [Sun, 13 Oct 2019 21:49:10 +0000 (23:49 +0200)]
[NDK] Fix coding style and indentation. No code changes!

20 months ago[FREELDR] Fix build with the 'Grossly Crying Compiler' (aka. GCC).
Hermès Bélusca-Maïto [Sun, 13 Oct 2019 19:07:23 +0000 (21:07 +0200)]
[FREELDR] Fix build with the 'Grossly Crying Compiler' (aka. GCC).

20 months ago[FREELDR] Improvements for GDT descriptors initialization.
Hermès Bélusca-Maïto [Sun, 13 Oct 2019 17:33:00 +0000 (19:33 +0200)]
[FREELDR] Improvements for GDT descriptors initialization.
Triggered by a fix attempt by M.Harmstone in PR #1905.

- Introduce #defines and helper functions so as to initialize the
  descriptors in a human-readable and tractable way, without magic values.

- Rework some comments.

- Document what the changes are in (Longhorn and) Vista+.

Data has been obtained through using the WinDbg command: dg 0x00 0xFF
(for example), that lists all the selectors from 0x00 to 0xFF included.
for more details.

20 months ago[SDK][NDK] Add two missing documented GDT selector names.
Hermès Bélusca-Maïto [Sun, 13 Oct 2019 17:21:34 +0000 (19:21 +0200)]
[SDK][NDK] Add two missing documented GDT selector names.

Add KGDT_NULL and KGDT_VDM_TILE to the list of known GDT name defines.

20 months ago[UMPNPMGR] Implement PNP_QueryRemove()
Eric Kohl [Sun, 13 Oct 2019 13:49:13 +0000 (15:49 +0200)]
[UMPNPMGR] Implement PNP_QueryRemove()

Eric Kohl [Sun, 13 Oct 2019 13:24:29 +0000 (15:24 +0200)]

20 months ago[SDK][INCLUDE] Update msgdump.h
Katayama Hirofumi MZ [Sun, 13 Oct 2019 01:45:17 +0000 (10:45 +0900)]
[SDK][INCLUDE] Update msgdump.h

20 months ago[WIN32K:menu] Use signed types for negative values... 1943/head
Mark Jansen [Mon, 30 Sep 2019 21:15:20 +0000 (23:15 +0200)]
[WIN32K:menu] Use signed types for negative values...

20 months ago[BROWSEUI][SHELLFIND] Allow to 'Open Containing Folder'
Joachim Henze [Sat, 12 Oct 2019 13:44:58 +0000 (15:44 +0200)]
[BROWSEUI][SHELLFIND] Allow to 'Open Containing Folder'

of a search result again.

The feature regressed by 0.4.13-dev-1133-g

I chose the shorter solution of Brock Mammen instead
of the patch of Doug Lyons from PR1960.

Still many thanks to Doug Lyons for his encouragement
in creating a fix as well and performing

I intend to merge that back into 0.4.13RCs as well.

20 months ago[WIN32SS] Fix regression CORE-16393
Joachim Henze [Fri, 11 Oct 2019 21:21:39 +0000 (23:21 +0200)]
[WIN32SS] Fix regression CORE-16393

Symptom "cmd did not hide the taskbar anymore when switching to fullscreen"

It regressed by SVN r75407 == git

VBox LGTM,69297
 KVM LGTM,69298

Many thanks to the patches author James Tabor.

20 months ago[FREELDR][XBOXVMP] Retrieve screen resolution directly from NV2A GPU (#1962)
Stanislav Motylkov [Wed, 9 Oct 2019 21:27:22 +0000 (00:27 +0300)]
[FREELDR][XBOXVMP] Retrieve screen resolution directly from NV2A GPU (#1962)


20 months ago[FREELDR] Advance the file pointers every time a read operation is performed, in...
Hermès Bélusca-Maïto [Wed, 9 Oct 2019 00:26:29 +0000 (02:26 +0200)]
[FREELDR] Advance the file pointers every time a read operation is performed, in accordance with the ARC specification.

20 months ago[FREELDR] Improvements for the RamDisk support.
Hermès Bélusca-Maïto [Fri, 13 Sep 2019 17:04:06 +0000 (19:04 +0200)]
[FREELDR] Improvements for the RamDisk support.

- Implement support for the "RDIMAGELENGTH" and "RDIMAGEOFFSET" boot
  options. Fixes CORE-15432.

- Separate the initialization of the global gInitRamDiskBase /
  gInitRamDiskSize variables from the FreeLdr command-line, and the
  actual initialization of the internal variables of the RamDisk.
  The latter are initialized via calls to RamDiskInitialize().

- Implement 'SeekRelative' mode in RamDiskSeek().

- Make RamDiskLoadVirtualFile() internal function that gets called by
  RamDiskInitialize(), and we use the latter in the NT loader instead.

20 months ago[FREELDR] Xbox memory management improvements (#1961)
Stanislav Motylkov [Wed, 9 Oct 2019 00:35:07 +0000 (03:35 +0300)]
[FREELDR] Xbox memory management improvements (#1961)

- Reuse the framebuffer address that was set up by the firmware.
- Get rid of XboxMemReserveMemory() and use ReserveMemory() instead.


20 months ago[EXPLORER] Fix right-click menu for taskbar clock (#1944)
Brock Mammen [Tue, 8 Oct 2019 23:18:36 +0000 (18:18 -0500)]
[EXPLORER] Fix right-click menu for taskbar clock (#1944)


20 months ago[XBOXVMP] Fix broken pixels and general refactoring (#1896)
Stanislav Motylkov [Tue, 8 Oct 2019 15:08:44 +0000 (18:08 +0300)]
[XBOXVMP] Fix broken pixels and general refactoring (#1896)

- Remove old hacky code based on MmHighestPhysicalPage.
- Split I2C SMBus code into a separate source file.

CORE-16216 CORE-16357

20 months ago[TRANSLATION] Add/Update Indonesian translations (#1904)
Mas4hmad [Tue, 8 Oct 2019 15:03:14 +0000 (22:03 +0700)]
[TRANSLATION] Add/Update Indonesian translations (#1904)


20 months ago[CMDUTILS] Add italian translation for 'comp' utility. (#1931)
Carlo Bramini [Tue, 8 Oct 2019 14:52:30 +0000 (16:52 +0200)]
[CMDUTILS] Add italian translation for 'comp' utility. (#1931)

20 months ago[FREELDR] Handle Btrfs sparse extents (#1959)
maharmstone [Tue, 8 Oct 2019 14:50:58 +0000 (15:50 +0100)]
[FREELDR] Handle Btrfs sparse extents (#1959)

20 months ago[FREELDR] Add support for loading Linux in x64 FreeLdr. Part 2/2: C code.
Hermès Bélusca-Maïto [Sun, 6 Oct 2019 17:22:18 +0000 (19:22 +0200)]
[FREELDR] Add support for loading Linux in x64 FreeLdr. Part 2/2: C code.

- Re-enable custom and Linux boot menu entries.
- Fix compilation and add address limit checks for x64 build of linux loader.

20 months ago[FREELDR] Fix an inconsistency bug when switching the CPU state back to 16-bit real...
Hermès Bélusca-Maïto [Wed, 2 Oct 2019 01:20:10 +0000 (03:20 +0200)]
[FREELDR] Fix an inconsistency bug when switching the CPU state back to 16-bit real mode.

Basically it appeared that without this fix, the CPU was somehow residing
in an inconsistent state, that made it crash when a full 16-bit real-mode
to 32-bit protected mode transition occurred. (Encountered when trying
to load Linux, see problem description below.)
In that situation, Bochs reports that the CPU is in "compatibility mode".

The fix is based from information from Fig.1-6 "Operating Modes of the
AMD64 Architecture" (page 12) and Chapter 14 (pages 429-446) of the
"AMD64 Architecture Programmer’s Manual Volume 2: System Programming"


When booting Linux using x86 FreeLdr everything goes well.

When trying to do the same using x64 FreeLdr, the Linux code (both the
boot sector, the setup sector and the main kernel contents) is all
correctly loaded and relocated in memory as in the x86 case. We then pass
control to the decompressing code that appears to succeed. However, once
it has finished and "Parsing ELF file..." step has been run, the next
step "Booting the kernel..." crashes with a CPU Triple-Fault.

This problem **DOES NOT HAPPEN** when booting from GRUB.

Log excerpt from Bochs:


00089459736i[BIOS  ] Booting from 07c0:0000
00089782775i[SER   ] com1: FIFO enabled
00095994535i[BIOS  ] int13_harddisk: function 41, unmapped device for ELDL=81
00095998517i[BIOS  ] int13_harddisk: function 08, unmapped device for ELDL=81
00397139785i[BIOS  ] KBD: unsupported int 16h function 03
00397143625i[BIOS  ] *** int 15h function AX=e980, BX=0000 not yet supported!
00523008104e[CPU0  ] interrupt(long mode): vector must be within IDT table limits, IDT.limit = 0x0
00523008104e[CPU0  ] interrupt(long mode): vector must be within IDT table limits, IDT.limit = 0x0
00523008104i[CPU0  ] CPU is in compatibility mode (active)
00523008104i[CPU0  ] CS.mode = 32 bit
00523008104i[CPU0  ] SS.mode = 32 bit
00523008104i[CPU0  ] EFER   = 0x00000500
00523008104i[CPU0  ] | RAX=00000000e0000011  RBX=0000000000000000
00523008104i[CPU0  ] | RCX=0000000000000000  RDX=0000000000000000
00523008104i[CPU0  ] | RSP=00000000004f8000  RBP=000000000082e003
00523008104i[CPU0  ] | RSI=0000000000099800  RDI=00000000c0611000
00523008104i[CPU0  ] |  R8=0000000000109000   R9=0000000000009020
00523008104i[CPU0  ] | R10=00000000000007e3  R11=000000000000e958
00523008104i[CPU0  ] | R12=0000000000000000  R13=0000000000000000
00523008104i[CPU0  ] | R14=0000000000000000  R15=0000000000000000
00523008104i[CPU0  ] | IOPL=0 id vip vif ac vm RF nt of df if tf SF zf af PF cf
00523008104i[CPU0  ] | SEG sltr(index|ti|rpl)     base    limit G D
00523008104i[CPU0  ] |  CS:0010( 0002| 0|  0) 00000000 ffffffff 1 1
00523008104i[CPU0  ] |  DS:0018( 0003| 0|  0) 00000000 ffffffff 1 1
00523008104i[CPU0  ] |  SS:0018( 0003| 0|  0) 00000000 ffffffff 1 1
00523008104i[CPU0  ] |  ES:0018( 0003| 0|  0) 00000000 ffffffff 1 1
00523008104i[CPU0  ] |  FS:0018( 0003| 0|  0) 00000000 ffffffff 1 1
00523008104i[CPU0  ] |  GS:0018( 0003| 0|  0) 00000000 ffffffff 1 1
00523008104i[CPU0  ] |  MSR_FS_BASE:0000000000000000
00523008104i[CPU0  ] |  MSR_GS_BASE:0000000000000000
00523008104i[CPU0  ] | RIP=0000000000409327 (0000000000409327)
00523008104i[CPU0  ] | CR0=0xe0000011 CR2=0x0000000000409327
00523008104i[CPU0  ] | CR3=0x005b5000 CR4=0x000000a0
00523008104i[CPU0  ] 0x0000000000409327: (instruction unavailable) page not present
00523008104p[CPU0  ] >>PANIC<< exception(): 3rd (13) exception with no resolution


20 months ago[FREELDR] Add support for loading Linux in x64 FreeLdr. Part 1/2: ASM code.
Hermès Bélusca-Maïto [Tue, 1 Oct 2019 01:50:29 +0000 (03:50 +0200)]
[FREELDR] Add support for loading Linux in x64 FreeLdr. Part 1/2: ASM code.

Add also ASM implementation for intrinsics that may not be always
present on MSVC (e.g. MSVC 2010).

20 months ago[FREELDR] Several changes regarding chainloading and Linux boot.
Hermès Bélusca-Maïto [Sun, 29 Sep 2019 17:20:15 +0000 (19:20 +0200)]
[FREELDR] Several changes regarding chainloading and Linux boot.

- Introduce "Relocator16Boot()". So far its aim is just to correctly set
  the CPU state (segments, registers, flags) to what is expected by a
  given boot image before running it.
  This function can be seen as the embryonic state of a future boot relocator
  (see e.g. GRUB or SYSLINUX) that would also relocate the boot image at
  the correct places. (Such feature is needed when boot images have to
  be loaded in memory areas that cover where the boot loader is in memory.)

- Implement ChainLoadBiosBootSectorCode() around it.

- Replace BootOldLinuxKernel() and BootNewLinuxKernel() by a new
  BootLinuxKernel() function (in assembly) that relocates the kernel
  to a given position and then boot it, using Relocator16Boot().
  Ideally the relocation should be done by a future boot relocator...

Implementation notes for Relocator16Boot():

For setting the CPU state the function is based on a similar code as the
Int386() helper, namely it takes a pointer to REGS structure and pass
this information through the 32->16 bits call before setting the CPU state
in accordance.
New stack segment/pointer and code segment/pointer are also specified.
For passing these values through the 32->16 bits call the 16-bit BSS
memory offsets "BSS_CallbackReturn" and "BSS_RealModeEntry" (respectively)
are reused.

20 months ago[SDK:CRT] For intrinsics that don't exist on MSVC 2010 we cannot use an implementatio...
Hermès Bélusca-Maïto [Sun, 6 Oct 2019 13:41:30 +0000 (15:41 +0200)]
[SDK:CRT] For intrinsics that don't exist on MSVC 2010 we cannot use an implementation in inlined ASM for x64 builds.

20 months ago[SYSSETUP][NETCFGX][INF] Install the TCPIP protocol driver properly
Eric Kohl [Sun, 6 Oct 2019 20:36:19 +0000 (22:36 +0200)]
[SYSSETUP][NETCFGX][INF] Install the TCPIP protocol driver properly

Get rid of the special RandomProtocolGUID_TCPIP :-)
Works with the Setup CD and Live CD.

20 months ago[FREELDR] NT loader: Allocate the Loader Block Extension much earlier in the process.
Hermès Bélusca-Maïto [Sun, 6 Oct 2019 16:24:48 +0000 (18:24 +0200)]
[FREELDR] NT loader: Allocate the Loader Block Extension much earlier in the process.

20 months ago[FREELDR] INFFILE Parser: Fix while-loop conditions, this bug existed for 16+ years...
Hermès Bélusca-Maïto [Sun, 6 Oct 2019 16:18:34 +0000 (18:18 +0200)]
[FREELDR] INFFILE Parser: Fix while-loop conditions, this bug existed for 16+ years (commit 939d5316, revision 4758).

20 months ago[MEDIA][FONTS] Delete many fonts CORE-16407 (#1955)
Katayama Hirofumi MZ [Sun, 6 Oct 2019 13:09:19 +0000 (22:09 +0900)]
[MEDIA][FONTS] Delete many fonts CORE-16407 (#1955)

Delete many fonts (about a hundred) to reduce memory consumption. CORE-16407

20 months ago[MOUNTMGR] Properly look for device and break when found
Serge Gautherie [Sat, 5 Oct 2019 20:44:25 +0000 (22:44 +0200)]
[MOUNTMGR] Properly look for device and break when found

Also use same pattern in MountMgrCreatePointWorker().

Addendum to a7f97cc70f62f6805e71b28eba51f52b942d22b4.

20 months ago[MOUNTMGR] QueryPointsFromMemory(): Do not leak DeviceName.Buffer
Serge Gautherie [Sat, 5 Oct 2019 23:01:31 +0000 (01:01 +0200)]
[MOUNTMGR] QueryPointsFromMemory(): Do not leak DeviceName.Buffer

Follow-up to 7601011f4ed578ae4a15a55298ebbe96fe3a1129.

20 months ago[FREELDR] Retrieve the correct ACPI RSDP table address. (#1951) 1951/head
Mark Harmstone [Wed, 2 Oct 2019 11:43:30 +0000 (12:43 +0100)]
[FREELDR] Retrieve the correct ACPI RSDP table address. (#1951)

Use either the RSDT (ACPI 1.0) or the XSDT (ACPI >1.0).
This fixes ACPI issues with Vista on VirtualBox.

20 months ago[FREELDR] Load the Firmware Errata file specified in the registry. (#1951)
Mark Harmstone [Mon, 9 Sep 2019 19:09:57 +0000 (20:09 +0100)]
[FREELDR] Load the Firmware Errata file specified in the registry. (#1951)

On Windows <= 2003 (and current ReactOS), this is the "InfName" value
in the "HKEY_LOCAL_MACHINE\CurrentControlSet\Control\BiosInfo\" key.
(The default file name found there is 'biosinfo.inf'.)

On Vista+, this is the "InfName" value in the
"HKEY_LOCAL_MACHINE\CurrentControlSet\Control\Errata\" key.
(The default file name found there is 'errata.inf'.)

For 1st-stage setup, the file is specified in the "InfName" value of
the "BiosInfo" section in the TXTSETUP.SIF file.

Co-authored-by: Hermès Bélusca-Maïto <>
20 months ago[FREELDR] Insert DLLs into load order before what references them. (#1951)
Mark Harmstone [Thu, 3 Oct 2019 18:54:49 +0000 (19:54 +0100)]
[FREELDR] Insert DLLs into load order before what references them. (#1951)

20 months ago[FREELDR] Set LDRP_DRIVER_DEPENDENT_DLL where necessary. (#1951)
Mark Harmstone [Thu, 3 Oct 2019 18:36:43 +0000 (19:36 +0100)]
[FREELDR] Set LDRP_DRIVER_DEPENDENT_DLL where necessary. (#1951)

20 months ago[FREELDR] Fix bug in Btrfs filename lookups. (#1951)
Mark Harmstone [Mon, 9 Sep 2019 18:39:58 +0000 (19:39 +0100)]
[FREELDR] Fix bug in Btrfs filename lookups. (#1951)