| 1 |
.TH HDPARM 8 "Sep 2022" "Version 9.65" |
| 2 |
|
| 3 |
.SH NAME |
| 4 |
hdparm \- get/set SATA/IDE device parameters |
| 5 |
.SH SYNOPSIS |
| 6 |
.B hdparm |
| 7 |
[options] [device ...] |
| 8 |
.SH DESCRIPTION |
| 9 |
.B hdparm |
| 10 |
provides a command line interface to various kernel interfaces |
| 11 |
supported by the Linux SATA/PATA/SAS "libata" subsystem |
| 12 |
and the older IDE driver subsystem. Many newer (2008 and later) |
| 13 |
USB drive enclosures now also support "SAT" (SCSI-ATA Command Translation) |
| 14 |
and therefore may also work with hdparm. E.g., recent WD "Passport" models |
| 15 |
and recent NexStar-3 enclosures. |
| 16 |
Some options may work correctly only with the latest kernels. |
| 17 |
.SH OPTIONS |
| 18 |
When no options are given, |
| 19 |
.B \-acdgkmur |
| 20 |
is assumed. |
| 21 |
For "Get/set" options, a query without the optional parameter (e.g., |
| 22 |
\-d) will query (get) the device state, and with a parameter (e.g., |
| 23 |
\-d0) will set the device state. |
| 24 |
.TP |
| 25 |
.I \-a |
| 26 |
Get/set sector count for filesystem (software) read-ahead. |
| 27 |
This is used to improve performance in sequential reads of large files, |
| 28 |
by prefetching additional |
| 29 |
blocks in anticipation of them being needed by the running task. |
| 30 |
Many IDE drives also have a separate built-in read-ahead function, |
| 31 |
which augments this filesystem (software) read-ahead function. |
| 32 |
.TP |
| 33 |
.I \-A |
| 34 |
Get/set the IDE drive's read-lookahead feature (usually ON by default). |
| 35 |
Usage: |
| 36 |
.B \-A0 |
| 37 |
(disable) or |
| 38 |
.B \-A1 |
| 39 |
(enable). |
| 40 |
.TP |
| 41 |
.I \-b |
| 42 |
Get/set bus state. |
| 43 |
.TP |
| 44 |
.I \-B |
| 45 |
Get/set Advanced Power Management feature, if the drive supports it. A low value |
| 46 |
means aggressive power management and a high value means better performance. |
| 47 |
Possible settings range from values 1 through 127 (which permit spin-down), |
| 48 |
and values 128 through 254 (which do not permit spin-down). |
| 49 |
The highest degree of power management is attained with a setting of 1, |
| 50 |
and the highest I/O performance with a setting of 254. |
| 51 |
A value of 255 tells hdparm to disable Advanced Power Management altogether |
| 52 |
on the drive (not all drives support disabling it, but most do). |
| 53 |
.TP |
| 54 |
.I \-c |
| 55 |
Get/set (E)IDE 32-bit I/O support. A numeric parameter can be |
| 56 |
used to enable/disable 32-bit I/O support. |
| 57 |
Currently supported values include |
| 58 |
.B 0 |
| 59 |
to disable 32-bit I/O support, |
| 60 |
.B 1 |
| 61 |
to enable 32-bit data transfers, and |
| 62 |
.B 3 |
| 63 |
to enable 32-bit data transfers with a special |
| 64 |
.B sync |
| 65 |
sequence required by many chipsets. The value |
| 66 |
.B 3 |
| 67 |
works with nearly all |
| 68 |
32-bit IDE chipsets, but incurs slightly more overhead. |
| 69 |
Note that "32-bit" refers to data transfers across a PCI or VLB bus to the |
| 70 |
interface card only; all (E)IDE drives still have only a 16-bit connection |
| 71 |
over the ribbon cable from the interface card. |
| 72 |
.TP |
| 73 |
.I \-C |
| 74 |
Check the current IDE power mode status, which will always be one of |
| 75 |
.B unknown |
| 76 |
(drive does not support this command), |
| 77 |
.B active/idle |
| 78 |
(normal operation), |
| 79 |
.B standby |
| 80 |
(low power mode, drive has spun down), |
| 81 |
or |
| 82 |
.B sleeping |
| 83 |
(lowest power mode, drive is completely shut down). |
| 84 |
The |
| 85 |
.B \-S, \-y, \-Y, |
| 86 |
and |
| 87 |
.B \-Z |
| 88 |
options can be used to manipulate the IDE power modes. |
| 89 |
.TP |
| 90 |
.I \-d |
| 91 |
Get/set the "using_dma" flag for this drive. This option now works |
| 92 |
with most combinations of drives and PCI interfaces which support DMA |
| 93 |
and which are known to the kernel IDE driver. |
| 94 |
It is also a good idea to use the appropriate |
| 95 |
.B \-X |
| 96 |
option in combination with |
| 97 |
.B \-d1 |
| 98 |
to ensure that the drive itself is programmed for the correct DMA mode, |
| 99 |
although most BIOSs should do this for you at boot time. |
| 100 |
Using DMA nearly always gives the best performance, |
| 101 |
with fast I/O throughput and low CPU usage. |
| 102 |
But there are at least a few configurations of chipsets and drives |
| 103 |
for which DMA does not make much of a difference, or may even slow |
| 104 |
things down (on really messed up hardware!). Your mileage may vary. |
| 105 |
.TP |
| 106 |
.I \-\-dco-freeze |
| 107 |
DCO stands for Device Configuration Overlay, a way for vendors to selectively |
| 108 |
disable certain features of a drive. The |
| 109 |
.B \-\-dco-freeze |
| 110 |
option will freeze/lock the current drive configuration, |
| 111 |
thereby preventing software (or malware) |
| 112 |
from changing any DCO settings until after the next power-on reset. |
| 113 |
.TP |
| 114 |
.I \-\-dco-identify |
| 115 |
Query and dump information regarding drive configuration settings |
| 116 |
which can be disabled by the vendor or OEM installer. |
| 117 |
These settings show capabilities of the drive which might be disabled |
| 118 |
by the vendor for "enhanced compatibility". |
| 119 |
When disabled, they are otherwise hidden and will not show in the |
| 120 |
.B \-I |
| 121 |
identify output. For example, system vendors sometimes disable 48_bit |
| 122 |
addressing on large drives, for compatibility (and loss of capacity) |
| 123 |
with a specific BIOS. In such cases, |
| 124 |
.B \-\-dco-identify |
| 125 |
will show that the drive is 48_bit capable, but |
| 126 |
.B \-I |
| 127 |
will not show it, and nor will the drive accept 48_bit commands. |
| 128 |
.TP |
| 129 |
.I \-\-dco-restore |
| 130 |
Reset all drive settings, features, and accessible capacities back to |
| 131 |
factory defaults and full capabilities. |
| 132 |
This command will fail if DCO is frozen/locked, |
| 133 |
or if a |
| 134 |
.B \-Np |
| 135 |
maximum size restriction has also been set. |
| 136 |
This is |
| 137 |
.B EXTREMELY DANGEROUS |
| 138 |
and will very likely cause massive loss of data. |
| 139 |
.B DO NOT USE THIS COMMAND. |
| 140 |
.TP |
| 141 |
.I \-\-direct |
| 142 |
Use the kernel's "O_DIRECT" flag when performing a |
| 143 |
.B \-t |
| 144 |
timing test. This bypasses the page cache, causing the reads |
| 145 |
to go directly from the drive into hdparm's buffers, using so-called |
| 146 |
"raw" I/O. In many cases, this can produce results that appear |
| 147 |
much faster than the usual page cache method, giving a better indication |
| 148 |
of raw device and driver performance. |
| 149 |
.TP |
| 150 |
.I \-\-drq-hsm-error |
| 151 |
.B VERY DANGEROUS, DON'T EVEN THINK ABOUT USING IT. |
| 152 |
This option causes hdparm to issue an IDENTIFY command |
| 153 |
to the kernel, but incorrectly marked as a "non-data" command. |
| 154 |
This results in the drive being left with its DataReQust(DRQ) line |
| 155 |
"stuck" high. This confuses the kernel drivers, and may crash the system |
| 156 |
immediately with massive data loss. The option exists to help in testing |
| 157 |
and fortifying the kernel against similar real-world drive malfunctions. |
| 158 |
.B VERY DANGEROUS, DO NOT USE!! |
| 159 |
.TP |
| 160 |
.I \-D |
| 161 |
Enable/disable the on-drive defect management feature, |
| 162 |
whereby the drive firmware tries to automatically manage |
| 163 |
defective sectors by relocating them to "spare" sectors |
| 164 |
reserved by the factory for such. Control of this feature |
| 165 |
via the |
| 166 |
.B \-D |
| 167 |
option is not supported for most modern drives |
| 168 |
since ATA-4; thus this command may fail. |
| 169 |
.TP |
| 170 |
.I \-E |
| 171 |
Set cd/dvd drive speed. This is NOT necessary for regular operation, |
| 172 |
as the drive will automatically switch speeds on its own. |
| 173 |
But if you want to play with it, just supply a speed number |
| 174 |
after the option, usually a number like 2 or 4. |
| 175 |
This can be useful in some cases, though, to smooth out DVD video playback. |
| 176 |
.TP |
| 177 |
.I \-f |
| 178 |
Sync and flush the buffer cache for the device on exit. |
| 179 |
This operation is also performed internally as part of the |
| 180 |
.B \-t |
| 181 |
and |
| 182 |
.B \-T |
| 183 |
timings and other options. |
| 184 |
.TP |
| 185 |
.I \-\-fallocate |
| 186 |
This option currently works only on ext4 and xfs filesystem types. |
| 187 |
When used, this must be the only option given. |
| 188 |
It requires two parameters: the desired file size in kibibytes |
| 189 |
(byte count divided by 1024), followed by the pathname for the new file. |
| 190 |
It will create a new file of the specified size, |
| 191 |
but without actually having to write any data to the file. |
| 192 |
This will normally complete very quickly, and without thrashing the |
| 193 |
storage device. |
| 194 |
.IP |
| 195 |
E.g., create a 10\ kibibyte file: |
| 196 |
.B hdparm \-\-fallocate 10 temp_file |
| 197 |
.TP |
| 198 |
.I \-\-fibmap |
| 199 |
When used, this must be the only option given. |
| 200 |
It requires a file path as a parameter, and will print |
| 201 |
out a list of the block extents (sector ranges) |
| 202 |
occupied by that file on disk. |
| 203 |
Sector numbers are given as absolute LBA numbers, |
| 204 |
referenced from sector 0 of the physical device rather |
| 205 |
than from the partition or filesystem. |
| 206 |
This information can then be used for a variety of purposes, |
| 207 |
such as examining the degree of fragmentation of larger files, or |
| 208 |
determining appropriate sectors to deliberately corrupt |
| 209 |
during fault-injection testing procedures. |
| 210 |
.IP |
| 211 |
This option uses the new FIEMAP (file extent map) ioctl() when available, |
| 212 |
and falls back to the older FIBMAP (file block map) ioctl() otherwise. |
| 213 |
Note that FIBMAP suffers from a 32-bit block-number interface, |
| 214 |
and thus not work beyond 8\ TB or 16\ TB. FIBMAP is also very slow, |
| 215 |
and does not deal well with preallocated uncommitted extents |
| 216 |
in ext4/xfs filesystems, unless a sync() is done before using this option. |
| 217 |
.TP |
| 218 |
.I \-\-fwdownload |
| 219 |
When used, this should be the only option given. |
| 220 |
It requires a file path immediately after the |
| 221 |
option, indicating where the new drive firmware should be read from. |
| 222 |
The contents of this file will be sent to the drive using the |
| 223 |
(S)ATA |
| 224 |
.B DOWNLOAD MICROCODE |
| 225 |
command, using either transfer protocol 7 (entire file at once), |
| 226 |
or, if the drive supports it, transfer protocol 3 (segmented download). |
| 227 |
This command is |
| 228 |
.B EXTREMELY DANGEROUS |
| 229 |
and could destroy both the drive and all data on it. |
| 230 |
.B DO NOT USE THIS COMMAND. |
| 231 |
The |
| 232 |
.B \-\-fwdownload-mode3 |
| 233 |
, |
| 234 |
.B \-\-fwdownload-mode3-max |
| 235 |
, and |
| 236 |
.B \-\-fwdownload-mode7 |
| 237 |
variations on basic |
| 238 |
.B \-\-fwdownload |
| 239 |
allow overriding automatic protocol detection in favour of |
| 240 |
forcing hdparm to use a specific transfer protocol, for testing purposes only. |
| 241 |
.TP |
| 242 |
.I \-F |
| 243 |
Flush the on-drive write cache buffer (older drives may not implement this). |
| 244 |
.TP |
| 245 |
.I \-g |
| 246 |
Display the drive geometry (cylinders, heads, sectors), |
| 247 |
the size (in sectors) of the device, |
| 248 |
and the starting offset (in sectors) of the device from |
| 249 |
the beginning of the drive. |
| 250 |
.TP |
| 251 |
.I \-h |
| 252 |
Display terse usage information (help). |
| 253 |
.TP |
| 254 |
.I \-H |
| 255 |
Read the temperature from some (mostly Hitachi) drives. |
| 256 |
Also reports if the temperature is within operating condition range |
| 257 |
(this may not be reliable). Does not cause the drive to spin up if idle. |
| 258 |
.TP |
| 259 |
.I \-i |
| 260 |
Display the identification info which the kernel drivers (IDE, libata) |
| 261 |
have stored from boot/configuration time. This may differ from the |
| 262 |
current information obtainable directly from the drive itself |
| 263 |
with the |
| 264 |
.B \-I |
| 265 |
option. |
| 266 |
The data returned may or may not be current, depending on activity |
| 267 |
since booting the system. |
| 268 |
For a more detailed interpretation of the identification info, |
| 269 |
refer to |
| 270 |
.I AT Attachment Interface for Disk Drives, |
| 271 |
ANSI ASC X3T9.2 working draft, revision 4a, April 19/93, and later editions. |
| 272 |
.TP |
| 273 |
.I \-\-idle-immediate |
| 274 |
Issue an ATA IDLE_IMMEDIATE command, to put the drive into a lower power state. |
| 275 |
Usually the device remains spun-up. |
| 276 |
.TP |
| 277 |
.I \-\-idle-unload |
| 278 |
Issue an ATA IDLE_IMMEDIATE_WITH_UNLOAD command, to unload or park the heads |
| 279 |
and put the drive into a lower power state. Usually the device remains spun-up. |
| 280 |
.TP |
| 281 |
.I \-I |
| 282 |
Request identification info directly from the drive, |
| 283 |
which is displayed in a new expanded format with considerably |
| 284 |
more detail than with the older |
| 285 |
.B \-i |
| 286 |
option. |
| 287 |
.TP |
| 288 |
.I \-\-Iraw <pathname> |
| 289 |
This option dumps the drive's identify data in raw binary to the specified file. |
| 290 |
.TP |
| 291 |
.I \-\-Istdin |
| 292 |
This is a special variation on the |
| 293 |
.B \-I |
| 294 |
option, |
| 295 |
which accepts a drive identification block as standard input |
| 296 |
instead of using a /dev/hd* parameter. |
| 297 |
The format of this block must be |
| 298 |
.B exactly |
| 299 |
the same as that found in the /proc/ide/*/hd*/identify "files", |
| 300 |
or that produced by the |
| 301 |
.B \-\-Istdout |
| 302 |
option described below. |
| 303 |
This variation is designed for use with collected "libraries" of drive |
| 304 |
identification information, and can also be used on ATAPI |
| 305 |
drives which may give media errors with the standard mechanism. |
| 306 |
When |
| 307 |
.B \-\-Istdin |
| 308 |
is used, it must be the *only* parameter given. |
| 309 |
Note that some information will be incomplete, as the drive LOG PAGES are unavailable when |
| 310 |
.B \-\-Istdin |
| 311 |
is used. |
| 312 |
.TP |
| 313 |
.I \-\-Istdout |
| 314 |
This option dumps the drive's identify data in hex to stdout, |
| 315 |
in a format similar to that from /proc/ide/*/identify, and suitable for |
| 316 |
later use with the |
| 317 |
.B \-\-Istdin |
| 318 |
option. |
| 319 |
.TP |
| 320 |
.I \-J |
| 321 |
Get/set the Western Digital (WD) Green Drive's "idle3" timeout value. |
| 322 |
This timeout controls how often the drive parks its heads and enters |
| 323 |
a low power consumption state. The factory default is eight (8) seconds, |
| 324 |
which is a very poor choice for use with Linux. Leaving it at the default |
| 325 |
will result in hundreds of thousands of head load/unload cycles in a very |
| 326 |
short period of time. The drive mechanism is only rated for 300,000 to 1,000,000 |
| 327 |
cycles, so leaving it at the default could result in premature failure, |
| 328 |
not to mention the performance impact of the drive often having to wake-up |
| 329 |
before doing routine I/O. |
| 330 |
.IP |
| 331 |
WD supply a WDIDLE3.EXE DOS utility for tweaking this setting, |
| 332 |
and you should use that program instead of hdparm |
| 333 |
if at all possible. The reverse-engineered implementation in hdparm |
| 334 |
is not as complete as the original official program, even though it does |
| 335 |
seem to work on at a least a few drives. A full power cycle is required |
| 336 |
for any change in setting to take effect, regardless of which program is |
| 337 |
used to tweak things. |
| 338 |
.IP |
| 339 |
A setting of 30 seconds is recommended for Linux use. |
| 340 |
Permitted values are from 8 to 12 seconds, and from 30 to 300 seconds |
| 341 |
in 30-second increments. |
| 342 |
Specify a value of zero (0) to disable the WD idle3 timer completely |
| 343 |
(NOT RECOMMENDED!). |
| 344 |
.TP |
| 345 |
.I \-k |
| 346 |
Get/set the "keep_settings_over_reset" flag for the drive. |
| 347 |
When this flag is set, the drive will preserve the |
| 348 |
.B \-dmu |
| 349 |
settings over a soft reset, (as done during the error recovery sequence). |
| 350 |
This option defaults to off, |
| 351 |
to prevent drive reset loops which could be caused by combinations of |
| 352 |
.B \-dmu |
| 353 |
settings. The |
| 354 |
.B \-k |
| 355 |
option should therefore only be set after one has achieved confidence in |
| 356 |
correct system operation with a chosen set of configuration settings. |
| 357 |
In practice, all that is typically necessary to test a configuration |
| 358 |
(prior to using \-k) is to verify that the drive can be read/written, |
| 359 |
and that no error logs (kernel messages) are generated in the process |
| 360 |
(look in /var/log/messages on most systems). |
| 361 |
.TP |
| 362 |
.I \-K |
| 363 |
Set the drive's "keep_features_over_reset" flag. Setting this enables |
| 364 |
the drive to retain the settings for |
| 365 |
.B \-APSWXZ |
| 366 |
over a soft reset (as done during the error recovery sequence). |
| 367 |
Not all drives support this feature. |
| 368 |
.TP |
| 369 |
.I \-L |
| 370 |
Set the drive's doorlock flag. Setting this to |
| 371 |
.B 1 |
| 372 |
will lock the door mechanism of some removable hard drives |
| 373 |
(e.g., Syquest, ZIP, Jazz..), and setting it to |
| 374 |
.B 0 |
| 375 |
will unlock the door mechanism. Normally, Linux |
| 376 |
maintains the door locking mechanism automatically, depending on drive usage |
| 377 |
(locked whenever a filesystem is mounted). But on system shutdown, this can |
| 378 |
be a nuisance if the root partition is on a removable disk, since the root |
| 379 |
partition is left mounted (read-only) after shutdown. So, by using this |
| 380 |
command to unlock the door |
| 381 |
.B after |
| 382 |
the root filesystem is remounted read-only, one can then remove the cartridge |
| 383 |
from the drive after shutdown. |
| 384 |
.TP |
| 385 |
.I \-m |
| 386 |
Get/set sector count for multiple sector I/O on the drive. A setting of |
| 387 |
.B 0 |
| 388 |
disables this feature. Multiple sector mode (aka IDE Block Mode), is a feature |
| 389 |
of most modern IDE hard drives, permitting the transfer of multiple sectors per |
| 390 |
I/O interrupt, rather than the usual one sector per interrupt. When this |
| 391 |
feature is enabled, it typically reduces operating system overhead for disk |
| 392 |
I/O by 30\(en50%. On many systems, it also provides increased data throughput |
| 393 |
of anywhere from 5% to 50%. Some drives, however |
| 394 |
(most notably the WD Caviar series), |
| 395 |
seem to run slower with multiple mode enabled. Your mileage may vary. |
| 396 |
Most drives support the minimum settings of |
| 397 |
2, 4, 8, or 16 (sectors). Larger settings may also be possible, depending on |
| 398 |
the drive. A setting of 16 or 32 seems optimal on many systems. |
| 399 |
Western Digital recommends lower settings of 4 to 8 on many of their drives, |
| 400 |
due tiny (32 KiB) drive buffers and non-optimized buffering algorithms. |
| 401 |
The |
| 402 |
.B \-i |
| 403 |
option can be used to find the maximum setting supported by an installed drive |
| 404 |
(look for MaxMultSect in the output). |
| 405 |
Some drives claim to support multiple mode, but lose data at some settings. |
| 406 |
Under rare circumstances, such failures can result in |
| 407 |
.B massive filesystem corruption. |
| 408 |
.TP |
| 409 |
.I \-\-make-bad-sector |
| 410 |
Deliberately create a bad sector (aka. "media error") on the disk. |
| 411 |
.B EXCEPTIONALLY DANGEROUS. DO NOT USE THIS OPTION!! |
| 412 |
This can be useful for testing of device/RAID error recovery mechanisms. |
| 413 |
The sector number is given as a (base10) parameter after the option. |
| 414 |
Depending on the device, hdparm will choose one of two possible ATA |
| 415 |
commands for corrupting the sector. |
| 416 |
The WRITE_LONG works on most drives, but only up to the 28-bit |
| 417 |
sector boundary. |
| 418 |
Some very recent drives (2008) may support the new WRITE_UNCORRECTABLE_EXT |
| 419 |
command, which works for any LBA48 sector. |
| 420 |
If available, hdparm will use that in |
| 421 |
preference to WRITE_LONG. |
| 422 |
The WRITE_UNCORRECTABLE_EXT command itself presents a |
| 423 |
choice of how the new bad sector should behave. |
| 424 |
By default, it will look like any other bad sector, and the drive may |
| 425 |
take some time to retry and fail on subsequent READs of the sector. |
| 426 |
However, if a single letter |
| 427 |
.B f |
| 428 |
is prepended immediately in front of the first digit of the sector |
| 429 |
number parameter, |
| 430 |
then hdparm will issue a "flagged" WRITE_UNCORRECTABLE_EXT, |
| 431 |
which causes the drive to merely flag the sector as bad (rather than |
| 432 |
genuinely corrupt it), |
| 433 |
and subsequent READs of the sector will fail immediately (rather than |
| 434 |
after several retries). |
| 435 |
Note also that the |
| 436 |
.B \-\-repair-sector |
| 437 |
option can be used to restore (any) bad sectors when they are no |
| 438 |
longer needed, |
| 439 |
including sectors that were genuinely bad (the drive will likely remap |
| 440 |
those to a fresh area on the media). |
| 441 |
.TP |
| 442 |
.I \-M |
| 443 |
Get/set Automatic Acoustic Management (AAM) setting. Most modern harddisk drives |
| 444 |
have the ability to speed down the head movements to reduce their noise output. |
| 445 |
The possible values are between 0 and 254. 128 is the most quiet (and therefore |
| 446 |
slowest) setting and 254 the fastest (and loudest). Some drives have only two |
| 447 |
levels (quiet / fast), while others may have different levels between 128 and 254. |
| 448 |
At the moment, most drives only support 3 options, off, quiet, and fast. |
| 449 |
These have been assigned the values 0, 128, and 254 at present, respectively, |
| 450 |
but integer space has been incorporated for future expansion, should this change. |
| 451 |
.TP |
| 452 |
.I \-n |
| 453 |
Get or set the "ignore_write_errors" flag in the driver. |
| 454 |
Do NOT play with this without grokking the driver source code first. |
| 455 |
.TP |
| 456 |
.I \-N |
| 457 |
Get/set max visible number of sectors, also known as the |
| 458 |
.B Host Protected Area |
| 459 |
setting. Without a parameter, |
| 460 |
.B \-N |
| 461 |
displays the current setting, which is reported as two values: the first |
| 462 |
gives the current max sectors setting, and the second shows |
| 463 |
the native (real) hardware limit for the disk. |
| 464 |
The difference between these two values indicates how many sectors of the disk |
| 465 |
are currently hidden from the operating system, in the form of a |
| 466 |
.B Host Protected Area (HPA). |
| 467 |
This area is often used by computer makers to hold diagnostic software, |
| 468 |
and/or a copy of the originally provided operating system for recovery purposes. |
| 469 |
Another possible use is to hide the true capacity of a very large disk |
| 470 |
from a BIOS/system that cannot normally cope with drives of that size |
| 471 |
(e.g., most current {2010} BIOSs cannot deal with drives larger than 2\ TB, |
| 472 |
so an HPA could be used to cause a 3\ TB drive to report itself as a 2\ TB drive). |
| 473 |
To change the current max (VERY DANGEROUS, DATA LOSS IS EXTREMELY LIKELY), |
| 474 |
a new value should be provided (in base10) immediately |
| 475 |
following the |
| 476 |
.B \-N |
| 477 |
option. |
| 478 |
This value is specified as a count of sectors, rather than |
| 479 |
the "max sector address" of the drive. |
| 480 |
Drives have the concept of a temporary (volatile) setting which is |
| 481 |
lost on the next hardware reset, as well as a more permanent (non-volatile) |
| 482 |
value which survives resets and power cycles. By default, |
| 483 |
.B \-N |
| 484 |
affects only the temporary (volatile) setting. To change the permanent |
| 485 |
(non-volatile) value, prepend a leading |
| 486 |
.B p |
| 487 |
character immediately before the first digit of the value. |
| 488 |
Drives are supposed to allow only a single permanent |
| 489 |
change per session. A hardware reset (or power cycle) is required |
| 490 |
before another permanent |
| 491 |
.B \-N |
| 492 |
operation can succeed. Note that any attempt to set this value |
| 493 |
may fail if the disk is being accessed by other software at the same time. |
| 494 |
This is because setting the value requires a pair of back-to-back drive commands, |
| 495 |
but there is no way to prevent some other command from being inserted between |
| 496 |
them by the kernel. So if it fails initially, just try again. |
| 497 |
Kernel support for |
| 498 |
.B \-N |
| 499 |
is buggy for many adapter types across many kernel versions, |
| 500 |
in that an incorrect (too small) max size value is sometimes reported. |
| 501 |
As of the 2.6.27 kernel, this does finally seem to be working on most hardware. |
| 502 |
.TP |
| 503 |
.I \-\-offset |
| 504 |
Offsets to given number of GiB (1024*1024*1024) when performing |
| 505 |
.B \-t |
| 506 |
timings of device reads. |
| 507 |
Speed changes (about twice) along many mechanical drives. |
| 508 |
Usually the maximum is at the beginning, but not always. |
| 509 |
Solid-state drives (SSDs) should show similar timings regardless of offset. |
| 510 |
.TP |
| 511 |
.I \-p |
| 512 |
Attempt to reprogram the IDE interface chipset for the specified PIO mode, |
| 513 |
or attempt to auto-tune for the "best" PIO mode supported by the drive. |
| 514 |
This feature is supported in the kernel for only a few "known" chipsets, |
| 515 |
and even then the support is iffy at best. Some IDE chipsets are unable |
| 516 |
to alter the PIO mode for a single drive, in which case this option may cause |
| 517 |
the PIO mode for |
| 518 |
.I both |
| 519 |
drives to be set. Many IDE chipsets support either fewer or more than the |
| 520 |
standard six (0 to 5) PIO modes, so the exact speed setting that is actually |
| 521 |
implemented will vary by chipset/driver sophistication. |
| 522 |
.I Use with extreme caution! |
| 523 |
This feature includes zero protection for the unwary, |
| 524 |
and an unsuccessful outcome may result in |
| 525 |
.I severe filesystem corruption! |
| 526 |
.TP |
| 527 |
.I \-P |
| 528 |
Set the maximum sector count for the drive's internal prefetch mechanism. |
| 529 |
Not all drives support this feature, and it was dropped from the official spec |
| 530 |
as of ATA-4. |
| 531 |
.TP |
| 532 |
.I \-\-prefer-ata12 |
| 533 |
When using the SAT (SCSI ATA Translation) protocol, hdparm normally prefers |
| 534 |
to use the 16-byte command format whenever possible. |
| 535 |
But some USB drive enclosures don't work correctly with 16-byte commands. |
| 536 |
This option can be used to force use of the smaller 12-byte command format |
| 537 |
with such drives. hdparm will still revert to 16-byte commands for things |
| 538 |
that cannot be done with the 12-byte format (e.g., sector accesses beyond 28-bits). |
| 539 |
.TP |
| 540 |
.I \-q |
| 541 |
Handle the next option quietly, suppressing normal output (but not error messages). |
| 542 |
This is useful for reducing screen clutter when running from system startup scripts. |
| 543 |
Not applicable to the |
| 544 |
.B \-i |
| 545 |
or |
| 546 |
.B \-v |
| 547 |
or |
| 548 |
.B \-t |
| 549 |
or |
| 550 |
.B \-T |
| 551 |
options. |
| 552 |
.TP |
| 553 |
.I \-Q |
| 554 |
Get or set the device's command queue_depth, if supported by the hardware. |
| 555 |
This only works with 2.6.xx (or later) kernels, and only with |
| 556 |
device and driver combinations which support changing the queue_depth. |
| 557 |
For SATA disks, this is the Native Command Queuing (NCQ) queue depth. |
| 558 |
.TP |
| 559 |
.I \-r |
| 560 |
Get/set read-only flag for the device. |
| 561 |
When set, Linux disallows write operations on the device. |
| 562 |
.TP |
| 563 |
.I \-R |
| 564 |
Get/set Write-Read-Verify feature, if the drive supports it. |
| 565 |
Usage: |
| 566 |
.B \-R0 |
| 567 |
(disable) or |
| 568 |
.B \-R1 |
| 569 |
(enable). This feature is intended to have the drive firmware automatically |
| 570 |
read-back any data that is written by software, to verify that the data was |
| 571 |
successfully written. This is generally overkill, and can slow down disk |
| 572 |
writes by as much as a factor of two (or more). |
| 573 |
.TP |
| 574 |
.I \-\-read-sector |
| 575 |
Reads from the specified sector number, |
| 576 |
and dumps the contents in hex to standard output. |
| 577 |
The sector number must be given (base10) after this option. |
| 578 |
hdparm will issue a low-level read (completely bypassing the usual |
| 579 |
block layer read/write mechanisms) for the specified sector. |
| 580 |
This can be used to definitively check whether a given sector is bad |
| 581 |
(media error) or not (doing so through the usual mechanisms can |
| 582 |
sometimes give false positives). |
| 583 |
.TP |
| 584 |
.I \-\-repair-sector |
| 585 |
This is an alias for the |
| 586 |
.B \-\-write-sector |
| 587 |
option. VERY DANGEROUS. |
| 588 |
.TP |
| 589 |
.I \-s |
| 590 |
Enable/disable the power-on in standby feature, if supported by |
| 591 |
the drive. |
| 592 |
.B VERY DANGEROUS. |
| 593 |
Do not use unless you are absolutely certain |
| 594 |
that both the system BIOS (or firmware) and the operating system kernel |
| 595 |
(Linux >= 2.6.22) support probing for drives that use this feature. |
| 596 |
When enabled, the drive is powered-up in the |
| 597 |
.B standby |
| 598 |
mode to allow the controller to sequence the spin-up of devices, |
| 599 |
reducing the instantaneous current draw burden when many drives |
| 600 |
share a power supply. Primarily for use in large RAID setups. |
| 601 |
This feature is usually disabled and the drive is powered-up in the |
| 602 |
.B active |
| 603 |
mode (see \-C above). |
| 604 |
Note that a drive may also allow enabling this feature by a jumper. |
| 605 |
Some SATA drives support the control of this feature by pin 11 of |
| 606 |
the SATA power connector. In these cases, this command may be |
| 607 |
unsupported or may have no effect. |
| 608 |
.TP |
| 609 |
.I \-S |
| 610 |
Put the drive into idle (low-power) mode, and also set the standby |
| 611 |
(spindown) timeout for the drive. This timeout value is used |
| 612 |
by the drive to determine how long to wait (with no disk activity) |
| 613 |
before turning off the spindle motor to save power. Under such |
| 614 |
circumstances, the drive may take as long as 30 seconds to respond to |
| 615 |
a subsequent disk access, though most drives are much quicker. The |
| 616 |
encoding of the timeout value is somewhat peculiar. A value of zero |
| 617 |
means "timeouts are disabled": the device will not automatically enter |
| 618 |
standby mode. Values from 1 to 240 specify multiples of 5 seconds, |
| 619 |
yielding timeouts from 5 seconds to 20 minutes. Values from 241 to |
| 620 |
251 specify from 1 to 11 units of 30 minutes, yielding timeouts from |
| 621 |
30 minutes to 5.5 hours. A value of 252 signifies a timeout of 21 |
| 622 |
minutes. A value of 253 sets a vendor-defined timeout period between 8 |
| 623 |
and 12 hours, and the value 254 is reserved. 255 is interpreted as 21 |
| 624 |
minutes plus 15 seconds. Note that some older drives may have very |
| 625 |
different interpretations of these values. |
| 626 |
.TP |
| 627 |
.I \-\-set-sector-size |
| 628 |
For drives which support reconfiguring of the Logical Sector Size, |
| 629 |
this flag can be used to specify the new desired sector size in bytes. |
| 630 |
.B VERY DANGEROUS. This most likely will scramble all data on the drive. |
| 631 |
The specified size must be one of 512, 520, 528, 4096, 4160, or 4224. |
| 632 |
Very few drives support values other than 512 and 4096. Eg. |
| 633 |
.B hdparm --set-sector-size 4096 /dev/sdb |
| 634 |
.TP |
| 635 |
.I \-t |
| 636 |
Perform timings of device reads for benchmark and comparison purposes. |
| 637 |
For meaningful results, this operation should be repeated 2\(en3 times on |
| 638 |
an otherwise inactive system (no other active processes) with at least a |
| 639 |
couple of megabytes of free memory. This displays the speed of reading |
| 640 |
through the buffer cache to the disk without any prior caching of data. |
| 641 |
This measurement is an indication of how fast the drive can sustain |
| 642 |
sequential data reads under Linux, without any filesystem overhead. To |
| 643 |
ensure accurate measurements, the buffer cache is flushed during the |
| 644 |
processing of |
| 645 |
.I \-t |
| 646 |
using the BLKFLSBUF ioctl. |
| 647 |
.TP |
| 648 |
.I \-T |
| 649 |
Perform timings of cache reads for benchmark and comparison purposes. |
| 650 |
For meaningful results, this operation should be repeated 2\(en3 times |
| 651 |
on an otherwise inactive system (no other active processes) with at |
| 652 |
least a couple of megabytes of free memory. This displays the speed |
| 653 |
of reading directly from the Linux buffer cache without disk access. |
| 654 |
This measurement is essentially an indication of the throughput of the |
| 655 |
processor, cache, and memory of the system under test. |
| 656 |
.TP |
| 657 |
.I \-\-trim-sector-ranges |
| 658 |
For Solid State Drives (SSDs). |
| 659 |
.B EXCEPTIONALLY DANGEROUS. DO NOT USE THIS OPTION!! |
| 660 |
Tells the drive firmware |
| 661 |
to discard unneeded data sectors, destroying any data that may have |
| 662 |
been present within them. This makes those sectors available for |
| 663 |
immediate use by the firmware's garbage collection mechanism, to |
| 664 |
improve scheduling for wear-leveling of the flash media. |
| 665 |
This option expects one or more sector range pairs immediately after |
| 666 |
the option: |
| 667 |
an LBA starting address, a colon, and a sector count (max 65535), |
| 668 |
with no intervening spaces. |
| 669 |
.B EXCEPTIONALLY DANGEROUS. DO NOT USE THIS OPTION!! |
| 670 |
.IP |
| 671 |
E.g., |
| 672 |
.B hdparm \-\-trim-sector-ranges 1000:4 7894:16 /dev/sdz |
| 673 |
.TP |
| 674 |
.I \-\-trim-sector-ranges-stdin |
| 675 |
Identical to |
| 676 |
.B \-\-trim-sector-ranges |
| 677 |
above, except the list of lba:count pairs is read from stdin |
| 678 |
rather than being specified on the command line. This can be used |
| 679 |
to avoid problems with excessively long command lines. It also permits |
| 680 |
batching of many more sector ranges into single commands to the drive, |
| 681 |
up to the currently configured transfer limit (max_sectors_kb). |
| 682 |
.TP |
| 683 |
.I \-u |
| 684 |
Get/set the interrupt-unmask flag for the drive. A setting of |
| 685 |
.B 1 |
| 686 |
permits the |
| 687 |
driver to unmask other interrupts during processing of a disk interrupt, |
| 688 |
which greatly improves Linux's responsiveness and eliminates "serial port |
| 689 |
overrun" errors. |
| 690 |
.B Use this feature with caution: |
| 691 |
some drive/controller combinations do |
| 692 |
not tolerate the increased I/O latencies possible when this feature is enabled, |
| 693 |
resulting in |
| 694 |
.B massive filesystem corruption. |
| 695 |
In particular, |
| 696 |
.B CMD-640B |
| 697 |
and |
| 698 |
.B RZ1000 |
| 699 |
(E)IDE interfaces can be |
| 700 |
.B unreliable |
| 701 |
(due to a hardware flaw) when this option is used with kernel versions earlier |
| 702 |
than 2.0.13. Disabling the |
| 703 |
.B IDE prefetch |
| 704 |
feature of these interfaces (usually a BIOS/CMOS setting) |
| 705 |
provides a safe fix for the problem for use with earlier kernels. |
| 706 |
.TP |
| 707 |
.I \-v |
| 708 |
Display some basic settings, similar to \-acdgkmur for IDE. |
| 709 |
This is also the default behaviour when no options are specified. |
| 710 |
.TP |
| 711 |
.I \-V |
| 712 |
Display program version and exit immediately. |
| 713 |
.TP |
| 714 |
.I \-\-verbose |
| 715 |
Display extra diagnostics from some commands. |
| 716 |
.TP |
| 717 |
.I \-w |
| 718 |
Perform a device reset |
| 719 |
.B (DANGEROUS). |
| 720 |
Do NOT use this option. |
| 721 |
It exists for unlikely situations where a reboot might otherwise be |
| 722 |
required to get a confused drive back into a useable state. |
| 723 |
.TP |
| 724 |
.I \-\-write-sector |
| 725 |
Writes zeros to the specified sector number. VERY DANGEROUS. |
| 726 |
The sector number must be given (base10) after this option. |
| 727 |
hdparm will issue a low-level write (completely bypassing the usual |
| 728 |
block layer read/write mechanisms) to the specified sector. |
| 729 |
This can be used to force a drive to repair a bad sector (media error). |
| 730 |
.TP |
| 731 |
.I \-W |
| 732 |
Get/set the IDE/SATA drive's write-caching feature. |
| 733 |
.TP |
| 734 |
.I \-X |
| 735 |
Set the IDE transfer mode for (E)IDE/ATA drives. |
| 736 |
This is typically used in combination with |
| 737 |
.B \-d1 |
| 738 |
when enabling DMA to/from a drive on a supported interface chipset, where |
| 739 |
.B \-X mdma2 |
| 740 |
is used to select multiword DMA mode2 transfers and |
| 741 |
.B \-X sdma1 |
| 742 |
is used to select simple mode 1 DMA transfers. |
| 743 |
With systems which support UltraDMA burst timings, |
| 744 |
.B \-X udma2 |
| 745 |
is used to select UltraDMA mode2 transfers (you'll need to prepare |
| 746 |
the chipset for UltraDMA beforehand). |
| 747 |
Apart from that, use of this option is |
| 748 |
.B seldom necessary |
| 749 |
since most/all modern IDE drives default to their fastest PIO transfer mode |
| 750 |
at power-on. Fiddling with this can be both needless and risky. |
| 751 |
On drives which support alternate transfer modes, |
| 752 |
.B \-X |
| 753 |
can be used to switch the mode of the drive |
| 754 |
.B only. |
| 755 |
Prior to changing the transfer mode, the IDE interface should be jumpered |
| 756 |
or programmed (see |
| 757 |
.B \-p |
| 758 |
option) |
| 759 |
for the new mode setting to prevent loss and/or corruption of data. |
| 760 |
.I Use this with extreme caution! |
| 761 |
For the PIO (Programmed Input/Output) |
| 762 |
transfer modes used by Linux, this value is simply the desired |
| 763 |
PIO mode number plus 8. |
| 764 |
Thus, a value of 09 sets PIO mode1, 10 enables PIO mode2, |
| 765 |
and 11 selects PIO mode3. |
| 766 |
Setting 00 restores the drive's "default" PIO mode, and 01 disables IORDY. |
| 767 |
For multiword DMA, the value used is the desired DMA mode number |
| 768 |
plus 32. for UltraDMA, the value is the desired UltraDMA mode number |
| 769 |
plus 64. |
| 770 |
.TP |
| 771 |
.I \-y |
| 772 |
Force an IDE drive to immediately enter the low power consumption |
| 773 |
.B standby |
| 774 |
mode, usually causing it to spin down. |
| 775 |
The current power mode status can be checked using the |
| 776 |
.B \-C |
| 777 |
option. |
| 778 |
.TP |
| 779 |
.I \-Y |
| 780 |
Force an IDE drive to immediately enter the lowest power consumption |
| 781 |
.B sleep |
| 782 |
mode, causing it to shut down completely. A hard or soft reset |
| 783 |
is required before the drive can be accessed again |
| 784 |
(the Linux IDE driver will automatically handle issuing a reset if/when needed). |
| 785 |
The current power mode status can be checked using the |
| 786 |
.B \-C |
| 787 |
option. |
| 788 |
.TP |
| 789 |
.I \-z |
| 790 |
Force a kernel re-read of the partition table of the specified device(s). |
| 791 |
.TP |
| 792 |
.I \-Z |
| 793 |
Disable the automatic power-saving function of certain Seagate drives |
| 794 |
(ST3xxx models?), to prevent them from idling/spinning-down |
| 795 |
at inconvenient times. |
| 796 |
. |
| 797 |
.SH ATA Security Feature Set |
| 798 |
. |
| 799 |
These switches are |
| 800 |
.B DANGEROUS |
| 801 |
to experiment with, and might not work with some kernels. |
| 802 |
.B USE AT YOUR OWN RISK. |
| 803 |
.TP |
| 804 |
.I \-\-security-help |
| 805 |
Display terse usage info for all of the \-\-security-* options. |
| 806 |
.TP |
| 807 |
.I \-\-security-freeze |
| 808 |
Freeze the drive's security settings. |
| 809 |
The drive does not accept any security commands until next power-on reset. |
| 810 |
Use this function in combination with \-\-security-unlock to protect drive |
| 811 |
from any attempt to set a new password. Can be used standalone, too. |
| 812 |
No other options are permitted on the command line with this one. |
| 813 |
.TP |
| 814 |
.I \-\-security-prompt-for-password |
| 815 |
Prompt for the \-\-security PWD rather than getting from the command line args. |
| 816 |
This avoids having passwords show up in shell history or in |
| 817 |
/proc/self/cmdline during execution. |
| 818 |
.TP |
| 819 |
.I \-\-security-unlock PWD |
| 820 |
Unlock the drive, using password PWD. |
| 821 |
Password is given as an ASCII string and is padded with NULs to reach 32 bytes. |
| 822 |
The applicable drive password is selected with the \-\-user-master switch |
| 823 |
(default is "user" password). |
| 824 |
No other options are permitted on the command line with this one. |
| 825 |
.TP |
| 826 |
.I \-\-security-set-pass PWD |
| 827 |
Lock the drive, using password PWD (Set Password) |
| 828 |
.B (DANGEROUS). |
| 829 |
Password is given as an ASCII string and is padded with NULs to reach 32 bytes. |
| 830 |
Use the special password |
| 831 |
.B NULL |
| 832 |
to set an empty password. |
| 833 |
The applicable drive password is selected with the \-\-user-master switch |
| 834 |
(default is "user" password) |
| 835 |
and the applicable security mode with the \-\-security-mode switch. |
| 836 |
No other options are permitted on the command line with this one. |
| 837 |
.TP |
| 838 |
.I \-\-security-disable PWD |
| 839 |
Disable drive locking, using password PWD. |
| 840 |
Password is given as an ASCII string and is padded with NULs to reach 32 bytes. |
| 841 |
The applicable drive password is selected with the \-\-user-master switch |
| 842 |
(default is "user" password). |
| 843 |
No other options are permitted on the command line with this one. |
| 844 |
.TP |
| 845 |
.I \-\-security-erase PWD |
| 846 |
Erase (locked) drive, using password PWD |
| 847 |
.B (DANGEROUS). |
| 848 |
Password is given as an ASCII string and is padded with NULs to reach 32 bytes. |
| 849 |
Use the special password |
| 850 |
.B NULL |
| 851 |
to represent an empty password. |
| 852 |
The applicable drive password is selected with the \-\-user-master switch |
| 853 |
(default is "user" password). |
| 854 |
No other options are permitted on the command line with this one. |
| 855 |
.TP |
| 856 |
.I \-\-security-erase-enhanced PWD |
| 857 |
Enhanced erase (locked) drive, using password PWD |
| 858 |
.B (DANGEROUS). |
| 859 |
Password is given as an ASCII string and is padded with NULs to reach 32 bytes. |
| 860 |
The applicable drive password is selected with the \-\-user-master switch |
| 861 |
(default is "user" password). |
| 862 |
No other options are permitted on the command line with this one. |
| 863 |
.TP |
| 864 |
.I \-\-user-master USER |
| 865 |
Specifies which password (user/master) to select. |
| 866 |
.B Defaults to "user" password. |
| 867 |
Only useful in combination with \-\-security-unlock, \-\-security-set-pass, |
| 868 |
\-\-security-disable, \-\-security-erase or \-\-security-erase-enhanced. |
| 869 |
u user password |
| 870 |
m master password |
| 871 |
|
| 872 |
.TP |
| 873 |
.I \-\-security-mode MODE |
| 874 |
Specifies which security mode (high/maximum) to set. |
| 875 |
.B Defaults to high. |
| 876 |
Only useful in combination with \-\-security-set-pass. |
| 877 |
h high security |
| 878 |
m maximum security |
| 879 |
|
| 880 |
.B THIS FEATURE IS EXPERIMENTAL AND NOT WELL TESTED. USE AT YOUR OWN RISK. |
| 881 |
.SH FILES |
| 882 |
/etc/hdparm.conf |
| 883 |
.SH BUGS |
| 884 |
As noted above, the |
| 885 |
.B \-m sectcount |
| 886 |
and |
| 887 |
.B \-u 1 |
| 888 |
options should be used with caution at first, preferably on a |
| 889 |
read-only filesystem. Most drives work well with these features, but |
| 890 |
a few drive/controller combinations are not 100% compatible. Filesystem |
| 891 |
corruption may result. Backup everything before experimenting! |
| 892 |
.PP |
| 893 |
Some options (e.g., \-r for SCSI) may not work with old kernels as |
| 894 |
necessary ioctl()'s were not supported. |
| 895 |
.PP |
| 896 |
Although this utility is intended primarily for use with SATA/IDE hard disk |
| 897 |
devices, several of the options are also valid (and permitted) for use with |
| 898 |
SCSI hard disk devices and MFM/RLL hard disks with XT interfaces. |
| 899 |
.PP |
| 900 |
The Linux kernel up until 2.6.12 (and probably later) doesn't handle the |
| 901 |
security unlock and disable commands gracefully and will segfault and in some |
| 902 |
cases even panic. The security commands however might indeed have been executed |
| 903 |
by the drive. This poor kernel behaviour makes the PIO data security commands |
| 904 |
rather useless at the moment. |
| 905 |
.PP |
| 906 |
Note that the "security erase" and "security disable" commands have been |
| 907 |
implemented as two consecutive PIO data commands and will not succeed on a |
| 908 |
locked drive because the second command will not be issued after the segfault. |
| 909 |
See the code for hints how patch it to work around this problem. Despite the |
| 910 |
segfault it is often still possible to run two instances of hdparm |
| 911 |
consecutively and issue the two necessary commands that way. |
| 912 |
.SH AUTHOR |
| 913 |
.B hdparm |
| 914 |
has been written by Mark Lord <[email protected]>, the original primary |
| 915 |
developer and maintainer of the (E)IDE driver for Linux, and current contributor |
| 916 |
to the libata subsystem, along with suggestions and patches from many netfolk. |
| 917 |
.PP |
| 918 |
The disable Seagate auto-powersaving code |
| 919 |
is courtesy of Tomi Leppikangas([email protected]). |
| 920 |
.PP |
| 921 |
Security freeze command by Benjamin Benz, 2005. |
| 922 |
.PP |
| 923 |
PIO data out security commands by Leonard den Ottolander, 2005. |
| 924 |
Some other parts by Benjamin Benz and others. |
| 925 |
.SH SEE ALSO |
| 926 |
.B http://www.t13.org/ |
| 927 |
Technical Committee T13 AT Attachment (ATA/ATAPI) Interface. |
| 928 |
.PP |
| 929 |
.B http://www.serialata.org/ |
| 930 |
Serial ATA International Organization. |
| 931 |
.PP |
| 932 |
.B http://www.compactflash.org/ |
| 933 |
CompactFlash Association. |