- `MountMgrVolumeMountPointChanged()`: Set status code in all failure
paths and correct the check for remote or removable media.
- `DeleteRemoteDatabaseEntry()`: Do not allow the database with size of zero.
- `QueryVolumeName()`: Use reparse point file reference when no file name is supplied.
`DeviceInformation->DeviceName.Buffer` was already freed via the previous
`FreePool(TargetDeviceName.Buffer);` call,
since `DeviceInformation->DeviceName` was set to `TargetDeviceName` above
in the code. This resulted in a pool double-free, triggering a corruption
of the pool, and a BSoD.
What had to be freed instead, is `DeviceInformation->SymbolicName.Buffer`
that is allocated at the beginning of the function.
- When trying to find a device, don't fail with STATUS_NOT_FOUND if no
associated symbolic links are present. Instead, that test was wrong,
and should go out of the search loop to do the drive-letter processing.
Addendum to commit f9f5a78715.
- In addition, when using the associated-device method, fix the list
used to retrieve the corresponding device.
- In the TryWithVolumeName: block, reset DeviceString to NULL after
freeing, so that we can correctly fail with STATUS_NOT_FOUND if no
suitable device was found, before initializing the output data.
- Use FIELD_OFFSET to correct structure and member instead of hardcoding
sizeof-s of fields until the member of interest.
- Fix a bug in MountMgrQueryDosVolumePath() where the FIELD_OFFSET used
in the entry structure size validation was incorrect.
FIELD_OFFSET(MOUNTMGR_TARGET_NAME, DeviceNameLength) is == 0 since
DeviceNameLength is the first member of the MOUNTMGR_TARGET_NAME
structure. The intended member was DeviceName.
Addendum to commit f9f5a78715.
1. MountMgrMountedDeviceArrival():
Fix the conditions under which the device's online notifications
are skipped (SkipNotifications == TRUE) and fix the code comments.
Now, things make much more sense:
online notifications are skipped when the device is already offline
or is a legacy (NT <= 4) fault-tolerant volume (see point 2 below),
or is NOT mounted (doesn't have a drive letter).
Previously, we were sending an online notification if the device
was NOT mounted (why?!...) or if it was deemed as "valid" (wrongly
determined, see point 2 below).
2. QueryDeviceInformation():
* The usage of the "Valid" parameter didn't make much sense. Indeed,
when a partition/volume device is reported to the Mount Manager,
it's already valid. (Also, setting "Valid" to TRUE only in the case
of an MBR partition while ignoring GPT ones, and resetting it to
FALSE if IOCTL_STORAGE_GET_DEVICE_NUMBER returned success, pointed
to something incorrect was going on.)
Instead, what we are checking here is whether the device is a
legacy fault-tolerant volume: such volume can only reside on an
MBR disk, have the expected partition type, and does not really
reside on a specific storage device (hence the check for
IOCTL_STORAGE_GET_DEVICE_NUMBER returning failure).
* Take also the opportunity to SAL2-ify the function.
This "NoAutoMount" member was not consistently used. Sometimes it was
used correctly, some other times it was used as "not NoAutoMount" i.e.
"AutoMount" enabled.
Fix this consistently throughout the source, and fix also some comments.
Fix IOCTL_MOUNTDEV_LINK_CREATED, IOCTL_MOUNTDEV_LINK_DELETED,
IOCTL_MOUNTDEV_UNIQUE_ID_CHANGE_NOTIFY, to Win2k3+ compatible definitions.
The previous versions of these IOCTLs (in Win2000 and XP) were defined without any access protection.
This was noticed here:
https://community.osr.com/t/ioctl-mountdev-link-created-definition-changed/29428
Do not treat target device change notification as
DEVICE_INTERFACE_CHANGE_NOTIFICATION. The notification have to be
unregistered while handling GUID_DEVICE_INTERFACE_REMOVAL, so
GUID_TARGET_DEVICE_REMOVE_COMPLETE should never be sent to mountmgr in a
normal case.
CORE-16106
- Do not access uninitialized SymlinkInformation on failure case
and just return
- Do not make an assumption that MOUNTMGR_TARGET_NAME has a zero-string
- ReparseFile was concatenated with itself, instead of ReparseIndex
- Meanwhile, use RtlAppendUnicodeStringToString for concatenating
strings instead of raw memory operations
- Change INIT_FUNCTION and INIT_SECTION to CODE_SEG("INIT") and DATA_SEG("INIT") respectively
- Remove INIT_FUNCTION from function prototypes
- Remove alloc_text pragma calls as they are not needed anymore
Adam Słaboń