1.\" 2.\" Copyright (c) 2003 Robert N. M. Watson 3.\" All rights reserved. 4.\" 5.\" Redistribution and use in source and binary forms, with or without 6.\" modification, are permitted provided that the following conditions 7.\" are met: 8.\" 1. Redistributions of source code must retain the above copyright 9.\" notice(s), this list of conditions and the following disclaimer as 10.\" the first lines of this file unmodified other than the possible 11.\" addition of one or more copyright notices. 12.\" 2. Redistributions in binary form must reproduce the above copyright 13.\" notice(s), this list of conditions and the following disclaimer in the 14.\" documentation and/or other materials provided with the distribution. 15.\" 16.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY 17.\" EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED 18.\" WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 19.\" DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY 20.\" DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES 21.\" (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR 22.\" SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER 23.\" CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 24.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 25.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH 26.\" DAMAGE. 27.\" 28.Dd April 30, 2020 29.Dt DISK 9 30.Os 31.Sh NAME 32.Nm disk 33.Nd kernel disk storage API 34.Sh SYNOPSIS 35.In geom/geom_disk.h 36.Ft struct disk * 37.Fn disk_alloc void 38.Ft void 39.Fn disk_create "struct disk *disk" "int version" 40.Ft void 41.Fn disk_gone "struct disk *disk" 42.Ft void 43.Fn disk_destroy "struct disk *disk" 44.Ft int 45.Fn disk_resize "struct disk *disk" "int flags" 46.Ft void 47.Fn disk_add_alias "struct disk *disk" "const char *alias" 48.Sh DESCRIPTION 49The disk storage API permits kernel device drivers providing access to 50disk-like storage devices to advertise the device to other kernel 51components, including 52.Xr GEOM 4 53and 54.Xr devfs 5 . 55.Pp 56Each disk device is described by a 57.Vt "struct disk" 58structure, which contains a variety of parameters for the disk device, 59function pointers for various methods that may be performed on the device, 60as well as private data storage for the device driver. 61In addition, some fields are reserved for use by GEOM in managing access 62to the device and its statistics. 63.Pp 64GEOM has the ownership of 65.Vt "struct disk" , 66and drivers must allocate storage for it with the 67.Fn disk_alloc 68function, 69fill in the fields and call 70.Fn disk_create 71when the device is ready to service requests. 72.Fn disk_add_alias 73adds an alias for the disk and must be called before 74.Fn disk_create , 75but may be called multiple times. 76For each alias added, a device node will be created with 77.Xr make_dev_alias 9 78in the same way primary device nodes are created with 79.Xr make_dev 9 80for 81.Va d_name 82and 83.Va d_unit . 84Care should be taken to ensure that only one driver creates aliases 85for any given name. 86.Fn disk_resize 87can be called by the driver after modifying 88.Va d_mediasize 89to notify GEOM about the disk capacity change. 90The 91.Fa flags 92field should be set to either M_WAITOK, or M_NOWAIT. 93.Fn disk_gone 94orphans all of the providers associated with the drive, setting an error 95condition of ENXIO in each one. 96In addition, it prevents a re-taste on last close for writing if an error 97condition has been set in the provider. 98After calling 99.Fn disk_destroy , 100the device driver is not allowed to access the contents of 101.Vt "struct disk" 102anymore. 103.Pp 104The 105.Fn disk_create 106function 107takes a second parameter, 108.Fa version , 109which must always be passed 110.Dv DISK_VERSION . 111If GEOM detects that the driver is compiled against an unsupported version, 112it will ignore the device and print a warning on the console. 113.Ss Descriptive Fields 114The following fields identify the disk device described by the structure 115instance, and must be filled in prior to submitting the structure to 116.Fn disk_create 117and may not be subsequently changed: 118.Bl -tag -width indent 119.It Vt u_int Va d_flags 120Optional flags indicating to the storage framework what optional features 121or descriptions the storage device driver supports. 122Currently supported flags are 123.Dv DISKFLAG_OPEN 124(maintained by storage framework), 125.Dv DISKFLAG_CANDELETE 126(maintained by device driver), 127and 128.Dv DISKFLAG_CANFLUSHCACHE 129(maintained by device driver). 130.It Vt "const char *" Va d_name 131Holds the name of the storage device class, e.g., 132.Dq Li ahd . 133This value typically uniquely identifies a particular driver device, 134and must not conflict with devices serviced by other device drivers. 135.It Vt u_int Va d_unit 136Holds the instance of the storage device class, e.g., 137.Dq Li 4 . 138This namespace is managed by the device driver, and assignment of unit 139numbers might be a property of probe order, or in some cases topology. 140Together, the 141.Va d_name 142and 143.Va d_unit 144values will uniquely identify a disk storage device. 145.El 146.Ss Disk Device Methods 147The following fields identify various disk device methods, if implemented: 148.Bl -tag -width indent 149.It Vt "disk_open_t *" Va d_open 150Optional: invoked when the disk device is opened. 151If no method is provided, open will always succeed. 152.It Vt "disk_close_t *" Va d_close 153Optional: invoked when the disk device is closed. 154Although an error code may be returned, the call should always terminate 155any state setup by the corresponding open method call. 156.It Vt "disk_strategy_t *" Va d_strategy 157Mandatory: invoked when a new 158.Vt "struct bio" 159is to be initiated on the disk device. 160.It Vt "disk_ioctl_t *" Va d_ioctl 161Optional: invoked when an I/O control operation is initiated on the disk device. 162Please note that for security reasons these operations should not 163be able to affect other devices than the one on which they are performed. 164.It Vt "dumper_t *" Va d_dump 165Optional: if configured with 166.Xr dumpon 8 , 167this function is invoked from a very restricted system state after a 168kernel panic to record a copy of the system RAM to the disk. 169.It Vt "disk_getattr_t *" Va d_getattr 170Optional: if this method is provided, it gives the disk driver the 171opportunity to override the default GEOM response to BIO_GETATTR requests. 172This function should return -1 if the attribute is not handled, 0 if the 173attribute is handled, or an errno to be passed to 174.Fn g_io_deliver . 175.It Vt "disk_gone_t *" Va d_gone 176Optional: if this method is provided, it will be called after 177.Fn disk_gone 178is called, once GEOM has finished its cleanup process. 179Once this callback is called, it is safe for the disk driver to free all of 180its resources, as it will not be receiving further calls from GEOM. 181.El 182.Ss Mandatory Media Properties 183The following fields identify the size and granularity of the disk device. 184These fields must stay stable from return of the drivers open method until 185the close method is called, but it is perfectly legal to modify them in 186the open method before returning. 187.Bl -tag -width indent 188.It Vt u_int Va d_sectorsize 189The sector size of the disk device in bytes. 190.It Vt off_t Va d_mediasize 191The size of the disk device in bytes. 192.It Vt u_int Va d_maxsize 193The maximum supported size in bytes of an I/O request. 194Requests larger than this size will be chopped up by GEOM. 195.El 196.Ss Optional Media Properties 197These optional fields can provide extra information about the disk 198device. 199Do not initialize these fields if the field/concept does not apply. 200These fields must stay stable from return of the drivers open method until 201the close method is called, but it is perfectly legal to modify them in 202the open method before returning. 203.Bl -tag -width indent 204.It Vt u_int Va d_fwsectors , Vt u_int Va d_fwheads 205The number of sectors and heads advertised on the disk device by the 206firmware or BIOS. 207These values are almost universally bogus, but on some architectures 208necessary for the correct calculation of disk partitioning. 209.It Vt u_int Va d_stripeoffset , Vt u_int Va d_stripesize 210These two fields can be used to describe the width and location of 211natural performance boundaries for most disk technologies. 212Please see 213.Pa src/sys/geom/notes 214for details. 215.It Vt char Va d_ident[DISK_IDENT_SIZE] 216This field can and should be used to store disk's serial number if the 217d_getattr method described above isn't implemented, or if it does not 218support the GEOM::ident attribute. 219.It Vt char Va d_descr[DISK_IDENT_SIZE] 220This field can be used to store the disk vendor and product description. 221.It Vt uint16_t Va d_hba_vendor 222This field can be used to store the PCI vendor ID for the HBA connected to 223the disk. 224.It Vt uint16_t Va d_hba_device 225This field can be used to store the PCI device ID for the HBA connected to 226the disk. 227.It Vt uint16_t Va d_hba_subvendor 228This field can be used to store the PCI subvendor ID for the HBA connected to 229the disk. 230.It Vt uint16_t Va d_hba_subdevice 231This field can be used to store the PCI subdevice ID for the HBA connected to 232the disk. 233.El 234.Ss Driver Private Data 235This field may be used by the device driver to store a pointer to 236private data to implement the disk service. 237.Bl -tag -width indent 238.It Vt "void *" Va d_drv1 239Private data pointer. 240Typically used to store a pointer to the drivers 241.Vt softc 242structure for this disk device. 243.El 244.Sh SEE ALSO 245.Xr GEOM 4 , 246.Xr devfs 5 247.Sh HISTORY 248The 249.Nm kernel disk storage API 250first appeared in 251.Fx 4.9 . 252.Sh AUTHORS 253This manual page was written by 254.An Robert Watson . 255.Sh BUGS 256Disk aliases are not a general purpose aliasing mechanism, but are 257intended only to ease the transition from one name to another. 258They can be used to ensure that nvd0 and nda0 are the same thing. 259They cannot be used to implement the diskX concept from macOS. 260