Method to virtualize PCIe controllers to support boot/hibernation/crash-dump from a spanned virtual disk

1. A method, comprising:

associating each physical storage controller coupled to a bus with a protocol function driver;

performing a first initialization comprising initializing each physical storage controller with its protocol function driver;

virtualizing all physical storage controllers of a particular protocol with a storage controller bus driver and exposing a single virtual storage controller for each particular protocol;

associating a virtual miniport driver with each of the virtual storage controllers;

performing a second initialization comprising initializing each of the virtual storage controllers with its associated virtual miniport driver;

identifying redundant array of independent drives (RAID) virtual disks associated with the virtual storage controllers;

responsive to receiving a request for logical unit (LUN) information, exposing the RAID virtual disks;

associating a disk driver with each of the RAID virtual disks; and

exposing the RAID virtual disks to users.

2. The method of claim 1 , further comprising:

sending, to a port driver, an input/output (I/O) request for a particular LUN associated with a virtual RAID controller;

routing, by the port driver, the I/O request to a particular virtual miniport driver associated with the virtual RAID controller;

generating, by the particular virtual miniport driver, a plurality of I/O request packets in accordance with a RAID level of the virtual RAID controller;

sending, by an I/O manager, the plurality of information request packets (IRPs) to the storage controller bus driver;

identifying, by the storage controller bus driver, storage protocol function driver objects corresponding to the plurality of IRP's; and

sending, by the I/O manager, the plurality of IRPs to the storage protocol function drivers.

3. The method of claim 1 , further comprising:

responsive to receiving an indication of command completion from hardware associated with the storage protocol function driver, calling an interrupt service routine registered by the storage protocol function driver.

4. The method of claim 3 , further comprising:

responsive to receiving an indication of command completion from the interrupt service routine registered by the storage protocol function driver, calling an interrupt service routine registered by the storage controller bus driver.

5. The method of claim 4 , further comprising:

responsive to receiving an indication of command completion from the interrupt service routine registered by the storage controller bus driver, calling an interrupt service routine associated with the virtual miniport driver.

6. The method of claim 5 , wherein the interrupt service routine associated with the virtual miniport driver is configured to:

process command completions for individual IRPs received; and

perform any error handling required.

7. The method of claim 1 , further comprising:

loading a crash dump driver to obtain crash dump stack information;

querying a disk driver to obtain boot disk information indicative of a boot disk;

identifying a dump port driver associated with the boot disk;

sending an I/O control code to the dump port driver to obtain miniport dump driver information and other hibernation crash dump drivers;

loading a storage controller bus driver and obtaining crash hibernation context information;

storing information for the virtual storage controller associated with the boot LUN into the dump context;

loading storage protocol function driver information to obtain crash dump hibernation context information; and

storing peripheral component interconnect (PCI) configuration space of storage controllers associated with a boot LUN into a dump context.

8. The method of claim 7 , further comprising:

responsive to detecting a crash dump or hibernation trigger, invoking the crash dump driver;

obtaining boot device information from the stored dump context;

loading, by the crash dump driver, a dump storage controller bus (SCB) driver and a dump storage protocol function (SPF) driver;

initializing, by the dump SPF driver, storage controllers associated with the boot LUN;

registering, by the dump SPF driver, callbacks for communicating with the dump SCB driver;

identifying, by the dump SCB driver, a virtual storage controller associated with the boot LUN;

creating a single virtual protocol controller for each particular protocol;

loading a dump port driver and a dump miniport driver;

registering, by the dump virtual miniport driver, callbacks for communication with the dump SCB driver;

requesting, by the dump port driver, initialization of the virtual storage controller;

initializing, by the dump virtual miniport driver, the virtual storage controller and a RAID core;

obtaining, from the dump SCB driver and the dump SPF driver, metadata indicative of virtual RAID disks associated with the virtual storage controllers;

obtaining, by the dump port driver, INQUIRY data for the boot LUN; and

processing a WRITE request for a dump file.

9. The method of claim 8 , wherein processing the dump file includes:

generating, by the dump port driver, WRITE SCSI request blocks (SRBs) for the RAID virtual disk;

sending the WRITE SRBs to the dump virtual miniport driver;

generating WRITE I/O requests in accordance with the RAID level for the boot LUN;

sending the WRITE I/O requests to the dump SCB driver;

identifying, by the dump SCB driver, virtual storage controllers associated with the WRITE I/O requests;

forwarding, via the callbacks, the write I/O requests to the dump SPF driver; and

sending, via callbacks, the WRITE I/O requests to disks associated with the dump SPF driver.

10. An information handling system, comprising:

a processor; and

a computer readable medium, including processor executable instructions for performing device stack operations comprising:

establishing a driver stack including:

a physical device object for each physical storage controller coupled to a PCI bus;

a protocol functional driver object (FDO) corresponding to each physical device object;

a storage controller bus driver FDO coupled to each of the protocol FDOs and configured to expose a single virtual controller physical device object (PDO) for each of a plurality of storage protocols;

a virtual miniport driver FDO coupled to each virtual controller PDO

a storport FDO coupled to each of the virtual miniport FDOs and further configured to expose a virtual disk PDO corresponding to each of the virtual miniport driver FDOs; and

a disk driver FDO associated with each of the virtual disk PDOs.

11. The information handling system of claim 10 , wherein the physical storage controllers include first storage controllers associated with a first storage protocol and second storage controllers associated with a second storage protocol.

12. The information handling system of claim 11 , wherein the first storage controllers include non-volatile memory express (NVMe) storage controllers.

13. The information handling system of claim 12 , wherein the virtual disk PDO corresponds to a redundant array of independent drives (RAID) virtual disk spanning a plurality of physical NVMe storage controllers.

14. The information handling system of claim 13 , wherein the RAID virtual disk includes a boot logical unit (LUN).

15. The information handling system of claim 12 , wherein the second storage controllers comprise serial AT attachment (SATA) storage controllers.

16. The information handling system of claim 10 , wherein the operations include:

establishing a crash dump/hibernation stack including a dump port driver, a dump virtual miniport driver, a dump storage controller bus driver, and a dump storage protocol function driver; and

registering function call backs enabling the dump virtual miniport driver, the dump storage protocol function driver and the dump storage controller bus driver to communicate during crash or hibernation.

17. A computer readable storage medium including processor executable instructions that, when executed, cause the processor to perform operations including:

discovering and initializing, by a storage protocol function driver, physical storage controllers coupled to a bus;

exposing, by a storage controller bus driver, a protocol-specific virtual controller, corresponding to all physical storage controllers of a particular protocol, to a virtual miniport driver;

initializing, by the virtual miniport driver, the protocol-specific virtual controller and discovering a redundant array of independent drives (RAID) virtual disk associated with the protocol-specific virtual controller, wherein the RAID virtual disk spans a plurality of the physical storage controllers of the particular protocol; and

responsive to receiving a request for logical units (LUNs) from a storport driver, exposing, by the virtual miniport driver, the RAID virtual disk to the storport driver.

18. The computer readable medium of claim 17 , wherein the RAID virtual disk corresponds to a boot device.

19. The computer readable medium of claim 18 , wherein the RAID virtual disk spans a plurality of non-volatile memory express (NVMe) storage controllers.

20. The computer readable medium of claim 17 , wherein the operations include:

establishing a crash dump/hibernation stack including a dump port driver, a dump virtual miniport driver, a dump storage controller bus driver, and a dump storage protocol function driver; and

registering function call backs enabling the dump virtual miniport driver, the dump storage protocol function drier and the dump storage controller bus driver to communicate during crash or hibernation.