Imported: 13 Feb '17 | Published: 18 Jan '11
USPTO - Utility Patents
A portable computing device, in which an internal non-volatile memory drive that is used to boot to a functional device GUI is automatically swapped with a temporary RAM drive if the internal non-volatile memory drive is found to be corrupted. The non-volatile memory is typically Flash memory, but the principle of the present invention can be applied to any kind of non-volatile memory that could become corrupted. Default configuration files may be automatically copied to the RAM drive. These may, for example, allow at least emergency voice calls to be made.
This application claims the priority of PCT/GB2004/002725 filed on 23 Jun. 2004 and GB 0314623.0 filed on 23 Jun. 2003, the entire contents of which are hereby incorporated in total by reference.
This invention relates to portable computing devices with non-volatile (e.g. flash) memory drives that are used to boot the device to a functional GUI (graphical user interface)—i.e. load the necessary files for the device to display the device GUI so that the device is operational by a user.
Many portable computing devices, such as Smartphones, include some form of Flash non-volatile memory for internal data storage that is used to boot up to a functional device GUI (i.e. to a functional state that includes the graphical user interface). In many cases, this is the C: drive and the term “C: drive” will be used hereafter to mean the primary internal storage drive or medium for user and application data.
There are two possible types of Flash memory: NOR and NAND. Each has its own limitations and quirks. But the problems are essentially the same:
Yet it is necessary to have a working C: drive in order to start enough code for the device to do anything useful. There are two conventional approaches to solving this:
In a first aspect of the invention, there is a portable computing device controlled by an operating system, in which, if the operating system is intact but an internal non-volatile memory drive that is used to boot the device to a functional GUI is found to be corrupted, then the non-volatile memory is automatically swapped with a temporary RAM drive to enable the operating system to boot.
A ‘RAM drive’ is a portion of device RAM that provides the same data storage capabilities as the non-volatile memory (typically Flash memory, as explained above) but is volatile. In the present invention, it uses part of the device RAM to store the files and data needed to boot the device to a GUI that is at least partly functional—i.e. so that the device can display a graphical user interface that enables the end-user to use some device function(s). At a minimum, this might be to allow the device to make an emergency voice call.
Default configuration files and data may be automatically copied to the RAM drive from the ROM (read-only memory). These may, for example, allow at least emergency voice calls to be made. Higher levels of functionality are also possible. The ROM is used to hold the binary executable code and factory-programmed data and is never modified in normal use so can be assumed to always hold valid versions of these configuration files.
This is much easier and quicker to implement than redesigning all the code to be able to run without an internal non-volatile memory on the C: drive, and also avoids the need to have a lot of extra error handling and fallback conditions within applications. Typically, it also means that applications can continue to offer much of the normal functionality that requires a C: drive. It is also much more acceptable to end-users than automatically reformatting a corrupt C: drive, since that process would lead to the loss of potentially recoverable user data. The non-volatile memory is typically Flash memory, but the principle of the present invention can be applied to any kind of non-volatile memory that could become corrupted, such as EEPROM, hard drives or any future non-volatile memory products.
Further, the corrupt drive may be automatically moved to a different drive letter to allow subsequent reformatting. A user notification asking if reformatting should take place may then be displayed.
Other display messages may include the following:
A second aspect of the invention is a method of enabling a portable computing device to boot up to a functional GUI when its resident operating system is intact but an internal non-volatile memory drive that is normally used to boot up from is found to be corrupt, comprising the step of automatically swapping the corrupt non-volatile memory drive with a temporary RAM drive to enable the operating system to boot.
A third aspect is operating system software, for a portable computing device, comprising an internal non-volatile memory drive that is normally used to boot up the device to a functional GUI, in which the operating system software automatically swaps the non-volatile memory drive with a temporary RAM drive if the non-volatile memory drive is found to be corrupt to enable the operating system to boot.
The invention will be described with reference to an implementation using SymbianOS from Symbian Limited of London, United Kingdom. A SymbianOS based device, as in FIG. 1, 101, in which the SymbianOS operating system is intact, 103, always has a C: drive, which is the internal read/write drive, and the system cannot boot to a functional GUI without a usable C: drive. The C: drive is normally Flash-based, which means that there is always the possibility of it becoming corrupt. This invention envisages a method of replacing a corrupt C: drive 105 with a temporary RAM drive 107 so that a SymbianOS device remains usable 109 in the event of a failure of the C: drive.
The procedure is this:
What happens next can be configured to the device manufacturer's requirements. The user may have the option to either reformat the drive or return the device to a service centre. A more user-friendly approach would be to allow the user to at least call emergency numbers. Depending on the functionality provided by the device, it might be possible for it to remain largely usable as normal, though of course any data saved will be lost when the device is turned off, so either save options should be disabled from the GUI or the user should be warned that any newly created data will be lost.
When is a Drive “Corrupt”?
We will consider here what is meant by “corrupt” in the context of the present invention. First, the data in the file system can be divided into two parts—user data and metadata. User data is the content of files. Metadata is the information used internally by the file system to record its state. If the user data is corrupt but the metadata is not, then technically the file system is not corrupt because it still knows what state it is in. Realistically though, it is unlikely to be of any use in this condition. It is also possible that the file system is not totally corrupt—as an example Symbian's LFFS file system will enter read-only mode if it is unable to satisfactorily determine its state but existing data is still likely to be valid.
Therefore the likely failure cases are:
In practice, it is unlikely that the device will be able to operate usefully in any of these situations. It will want to read existing files, which may be corrupt, and very likely it will also want to write new data. Whatever condition the file system is in, we can call them all “corrupt” with reference to the present invention.
How to Detect Corruption
There are two methods—if the file system performs some sort of integrity scan when it is mounted then if the scan fails the mount should also fail with an error indicating drive corruption. Alternatively some sort of scan of the file system could be invoked to determine its integrity. A spot-check on the validity of files known to be essential for the correct operation of the device is another method to determine whether the drive is corrupt. How corruption should be detected is really outside the scope of this document since it depends on the type of file system, the degree of thoroughness required, the time available, and what files are critical to the system.
The swapping of the corrupt drive with a RAM drive must happen early in boot before any code that is dependant on the C: drive. The ideal place in SymbianOS is ESTART, which is the first program run after the kernel and file server have been started. ESTART will do something like this:
At the end of this either the C: drive will be the normal Flash drive, or it will be a temporary RAM drive and the Flash will have been moved to a different drive letter.
The GUI must be able to detect that the temporary RAM drive is in use because it has to at least inform the user that the Flash is corrupt. This can be done conveniently in SymbianOS by checking the media type of the C: drive (RFs::Drive function) which will be EMediaRam if the drives have been swapped.
The first thing the GUI will want to do is to tell the user that the drive is corrupt and also offer the opportunity to reformat it—important data may already have been backed up and the user is happy to just reformat.
Further GUI handling will be partly dependent on the device capabilities and partly on the device manufacturer's requirements. The device could be limited to just emergency calls. Or it could be mostly functional but the detection of the RAM drive could be used to disable save options in programs (e.g. calendars and address books) or warn the user that any data saved will be lost when the device is turned off There could even be the option to attempt to extract information from the corrupt drive—for the cases where the drive is still partly accessible—attempting to find address books, calendars etc on the corrupt drive.
The following technical terms and abbreviations are used within this specification.