1 ide.txt -- Information regarding the Enhanced IDE drive in Linux 2.1/2.2
2 ===============================================================================
4 +-----------------------------------------------------------------+
5 | The hdparm utility for controlling various IDE features is |
6 | packaged separately. Look for it on popular linux FTP sites. |
7 +-----------------------------------------------------------------+
9 See description later on below for handling BIG IDE drives with >1024 cyls.
11 Major features of the 2.1/2.2 IDE driver ("NEW!" marks changes since 2.0.xx):
13 NEW! - support for IDE ATAPI *floppy* drives
14 - support for IDE ATAPI *tape* drives, courtesy of Gadi Oxman
15 (re-run MAKEDEV.ide to create the tape device entries in /dev/)
16 - support for up to *four* IDE interfaces on one or more IRQs
17 - support for any mix of up to *eight* IDE drives
18 - support for reading IDE ATAPI cdrom drives (NEC,MITSUMI,VERTOS,SONY)
19 - support for audio functions
20 - auto-detection of interfaces, drives, IRQs, and disk geometries
21 - "single" drives should be jumpered as "master", not "slave"
22 (both are now probed for)
23 - support for BIOSs which report "more than 16 heads" on disk drives
24 - uses LBA (slightly faster) on disk drives which support it
25 - support for lots of fancy (E)IDE drive functions with hdparm utility
26 - optional (compile time) support for 32-bit VLB data transfers
27 - support for IDE multiple (block) mode (same as hd.c)
28 - support for interrupt unmasking during I/O (better than hd.c)
29 - improved handshaking and error detection/recovery
30 - can co-exist with hd.c controlling the first interface
31 - run-time selectable 32bit interface support (using hdparm-2.3)
32 - support for reliable operation of buggy RZ1000 interfaces
33 - PCI support is automatic when rz1000 support is configured
34 - support for reliable operation of buggy CMD-640 interfaces
35 - PCI support is automatic when cmd640 support is configured
36 - for VLB, use kernel command line option: ide0=cmd640_vlb
37 - this support also enables the secondary i/f when needed
38 - interface PIO timing & prefetch parameter support
39 - experimental support for UMC 8672 interfaces
40 - support for secondary interface on the FGI/Holtek HT-6560B VLB i/f
41 - use kernel command line option: ide0=ht6560
42 - experimental support for various IDE chipsets
43 - use appropriate kernel command line option from list below
44 - support for drives with a stuck WRERR_STAT bit
45 - support for removable devices, including door lock/unlock
46 - transparent support for DiskManager 6.0x and "Dynamic Disk Overlay"
47 - works with Linux fdisk, LILO, loadlin, bootln, etc..
48 - mostly transparent support for EZ-Drive disk translation software
49 - to use LILO with EZ, install LILO on the linux partition
50 rather than on the master boot record, and then mark the
51 linux partition as "bootable" or "active" using fdisk.
52 (courtesy of Juha Laiho <jlaiho@ichaos.nullnet.fi>).
53 - auto-detect of disk translations by examining partition table
54 - ide-cd.c now compiles separate from ide.c
55 - ide-cd.c now supports door locking and auto-loading.
56 - Also preliminary support for multisession
57 and direct reads of audio data.
58 - experimental support for Promise DC4030VL caching interface card
59 - email thanks/problems to: peterd@pnd-pc.demon.co.uk
60 - the hdparm-3.1 package can be used to set PIO modes for some chipsets.
61 NEW! - support for setting PIO modes with the OPTi 82C621, courtesy of Jaromir Koutek.
62 NEW! - support for loadable modules
63 NEW! - optional SCSI host adapter emulation for ATAPI devices
64 NEW! - generic PCI Bus-Master DMA support
65 NEW! - works with most Pentium PCI systems, chipsets, add-on cards
66 NEW! - works with regular DMA as well as Ultra DMA
67 NEW! - automatically probes for all PCI IDE interfaces
68 NEW! - generic support for using BIOS-configured Ultra-DMA (UDMA) transfers
71 *** IMPORTANT NOTICES: BUGGY IDE CHIPSETS CAN CORRUPT DATA!!
73 *** PCI versions of the CMD640 and RZ1000 interfaces are now detected
74 *** automatically at startup when PCI BIOS support is configured.
76 *** Linux disables the "prefetch" ("readahead") mode of the RZ1000
77 *** to prevent data corruption possible due to hardware design flaws.
79 *** For the CMD640, linux disables "IRQ unmasking" (hdparm -u1) on any
80 *** drive for which the "prefetch" mode of the CMD640 is turned on.
81 *** If "prefetch" is disabled (hdparm -p8), then "IRQ unmasking" can be
84 *** For the CMD640, linux disables "32bit I/O" (hdparm -c1) on any drive
85 *** for which the "prefetch" mode of the CMD640 is turned off.
86 *** If "prefetch" is enabled (hdparm -p9), then "32bit I/O" can be
89 *** The CMD640 is also used on some Vesa Local Bus (VLB) cards, and is *NOT*
90 *** automatically detected by Linux. For safe, reliable operation with such
91 *** interfaces, one *MUST* use the "ide0=cmd640_vlb" kernel option.
93 *** Use of the "serialize" option is no longer necessary.
95 This is the multiple IDE interface driver, as evolved from hd.c.
96 It supports up to six IDE interfaces, on one or more IRQs (usually 14 & 15).
97 There can be up to two drives per interface, as per the ATA-2 spec.
99 Primary: ide0, port 0x1f0; major=3; hda is minor=0; hdb is minor=64
100 Secondary: ide1, port 0x170; major=22; hdc is minor=0; hdd is minor=64
101 Tertiary: ide2, port 0x1e8; major=33; hde is minor=0; hdf is minor=64
102 Quaternary: ide3, port 0x168; major=34; hdg is minor=0; hdh is minor=64
103 fifth.. ide4, usually PCI, probed
104 sixth.. ide5, usually PCI, probed
106 To access devices on interfaces > ide0, device entries must first be
107 created in /dev for them. To create such entries, simply run the included
108 shell script: /usr/src/linux/scripts/MAKEDEV.ide
110 Apparently many older releases of Slackware had incorrect entries
111 in /dev for hdc* and hdd* -- this can also be corrected by running MAKEDEV.ide
113 ide.c automatically probes for most IDE interfaces (including all PCI ones),
114 for the drives/geometries attached to those interfaces, and for the
115 IRQ numbers being used by the interfaces (normally 14, 15 for ide0/ide1).
117 For special cases, interfaces may be specified using kernel "command line"
118 options. For example,
120 ide3=0x168,0x36e,10 /* ioports 0x168-0x16f,0x36e, irq 10 */
122 Normally the irq number need not be specified, as ide.c will probe for it:
124 ide3=0x168,0x36e /* ioports 0x168-0x16f,0x36e */
126 The standard port, and irq values are these:
133 Note that the first parameter reserves 8 contiguous ioports, whereas the
134 second value denotes a single ioport. If in doubt, do a 'cat /proc/ioports'.
136 In all probability the device uses these ports and IRQs if it is attached
137 to the appropriate ide channel. Pass the parameter for the correct ide
138 channel to the kernel, as explained above.
140 Any number of interfaces may share a single IRQ if necessary, at a slight
141 performance penalty, whether on separate cards or a single VLB card.
142 The IDE driver automatically detects and handles this. However, this may
143 or may not be harmful to your hardware.. two or more cards driving the same IRQ
144 can potentially burn each other's bus driver, though in practice this
145 seldom occurs. Be careful, and if in doubt, don't do it!
147 Drives are normally found by auto-probing and/or examining the CMOS/BIOS data.
148 For really weird situations, the apparent (fdisk) geometry can also be specified
149 on the kernel "command line" using LILO. The format of such lines is:
151 hdx=cyls,heads,sects,wpcom,irq
154 where hdx can be any of hda through hdh, Three values are required
155 (cyls,heads,sects). For example:
157 hdc=1050,32,64 hdd=cdrom
159 either {hda,hdb} or {hdc,hdd}. The results of successful auto-probing may
160 override the physical geometry/irq specified, though the "original" geometry
161 may be retained as the "logical" geometry for partitioning purposes (fdisk).
163 If the auto-probing during boot time confuses a drive (ie. the drive works
164 with hd.c but not with ide.c), then an command line option may be specified
165 for each drive for which you'd like the drive to skip the hardware
166 probe/identification sequence. For example:
173 Note that when only one IDE device is attached to an interface,
174 it should be jumpered as "single" or "master", *not* "slave".
175 Many folks have had "trouble" with cdroms because of this requirement,
176 so ide.c now probes for both units, though success is more likely
177 when the drive is jumpered correctly.
179 Courtesy of Scott Snyder and others, the driver supports ATAPI cdrom drives
180 such as the NEC-260 and the new MITSUMI triple/quad speed drives.
181 Such drives will be identified at boot time, just like a hard disk.
183 If for some reason your cdrom drive is *not* found at boot time, you can force
184 the probe to look harder by supplying a kernel command line parameter
187 hdc=cdrom /* hdc = "master" on second interface */
189 hdd=cdrom /* hdd = "slave" on second interface */
191 For example, a GW2000 system might have a hard drive on the primary
192 interface (/dev/hda) and an IDE cdrom drive on the secondary interface
193 (/dev/hdc). To mount a CD in the cdrom drive, one would use something like:
195 ln -sf /dev/hdc /dev/cdrom
197 mount /dev/cdrom /cd -t iso9660 -o ro
199 If, after doing all of the above, mount doesn't work and you see
200 errors from the driver (with dmesg) complaining about `status=0xff',
201 this means that the hardware is not responding to the driver's attempts
202 to read it. One of the following is probably the problem:
204 - Your hardware is broken.
206 - You are using the wrong address for the device, or you have the
207 drive jumpered wrong. Review the configuration instructions above.
209 - Your IDE controller requires some nonstandard initialization sequence
210 before it will work properly. If this is the case, there will often
211 be a separate MS-DOS driver just for the controller. IDE interfaces
212 on sound cards usually fall into this category. Such configurations
213 can often be made to work by first booting MS-DOS, loading the
214 appropriate drivers, and then warm-booting linux (without powering
215 off). This can be automated using loadlin in the MS-DOS autoexec.
217 If you always get timeout errors, interrupts from the drive are probably
218 not making it to the host. Check how you have the hardware jumpered
219 and make sure it matches what the driver expects (see the configuration
220 instructions above). If you have a PCI system, also check the BIOS
221 setup; I've had one report of a system which was shipped with IRQ 15
222 disabled by the BIOS.
224 The kernel is able to execute binaries directly off of the cdrom,
225 provided it is mounted with the default block size of 1024 (as above).
227 Please pass on any feedback on any of this stuff to the maintainer,
228 whose address can be found in linux/MAINTAINERS.
230 Note that if BOTH hd.c and ide.c are configured into the kernel,
231 hd.c will normally be allowed to control the primary IDE interface.
232 This is useful for older hardware that may be incompatible with ide.c,
233 and still allows newer hardware to run on the 2nd/3rd/4th IDE ports
234 under control of ide.c. To have ide.c also "take over" the primary
235 IDE port in this situation, use the "command line" parameter: ide0=0x1f0
237 The IDE driver is partly modularized. The high level disk/cdrom/tape/floppy
238 drivers can always be compiled as loadable modules, the chipset drivers
239 can only be compiled into the kernel, and the core code (ide.c) can be
240 compiled as a loadable module provided no chipset support and no special
241 partition table translations are needed.
243 When using ide.c/ide-tape.c as modules in combination with kerneld, add:
245 alias block-major-3 ide-probe
246 alias char-major-37 ide-tape
248 respectively to /etc/conf.modules.
250 When ide.c is used as a module, you can pass command line parameters to the
251 driver using the "options=" keyword to insmod, while replacing any ',' with
254 insmod ide.o options="ide0=serialize ide2=0x1e8;0x3ee;11"
257 ================================================================================
259 Summary of ide driver parameters for kernel "command line":
260 ----------------------------------------------------------
261 "hdx=" is recognized for all "x" from "a" to "h", such as "hdc".
262 "idex=" is recognized for all "x" from "0" to "3", such as "ide1".
264 "hdx=noprobe" : drive may be present, but do not probe for it
265 "hdx=none" : drive is NOT present, ignore cmos and do not probe
266 "hdx=nowerr" : ignore the WRERR_STAT bit on this drive
267 "hdx=cdrom" : drive is present, and is a cdrom drive
268 "hdx=cyl,head,sect" : disk drive is present, with specified geometry
269 "hdx=autotune" : driver will attempt to tune interface speed
270 to the fastest PIO mode supported,
271 if possible for this drive only.
272 Not fully supported by all chipset types,
273 and quite likely to cause trouble with
274 older/odd IDE drives.
275 "hdx=slow" : insert a huge pause after each access to the data
276 port. Should be used only as a last resort.
277 "hdx=swapdata" : when the drive is a disk, byte swap all data
279 "idebus=xx" : inform IDE driver of VESA/PCI bus speed in MHz,
280 where "xx" is between 20 and 66 inclusive,
281 used when tuning chipset PIO modes.
282 For PCI bus, 25 is correct for a P75 system,
283 30 is correct for P90,P120,P180 systems,
284 and 33 is used for P100,P133,P166 systems.
285 If in doubt, use idebus=33 for PCI.
286 As for VLB, it is safest to not specify it.
287 Bigger values are safer than smaller ones.
289 "idex=noprobe" : do not attempt to access/use this interface
290 "idex=base" : probe for an interface at the addr specified,
291 where "base" is usually 0x1f0 or 0x170
292 and "ctl" is assumed to be "base"+0x206
293 "idex=base,ctl" : specify both base and ctl
294 "idex=base,ctl,irq" : specify base, ctl, and irq number
295 "idex=autotune" : driver will attempt to tune interface speed
296 to the fastest PIO mode supported,
297 for all drives on this interface.
298 Not fully supported by all chipset types,
299 and quite likely to cause trouble with
300 older/odd IDE drives.
301 "idex=noautotune" : driver will NOT attempt to tune interface speed
302 This is the default for most chipsets,
304 "idex=serialize" : do not overlap operations on idex and ide(x^1)
305 "idex=reset" : reset interface after probe
306 "idex=dma" : automatically configure/use DMA if possible.
308 The following are valid ONLY on ide0,
309 and the defaults for the base,ctl ports must not be altered.
311 "ide0=dtc2278" : probe/support DTC2278 interface
312 "ide0=ht6560b" : probe/support HT6560B interface
313 "ide0=cmd640_vlb" : *REQUIRED* for VLB cards with the CMD640 chip
314 (not for PCI -- automatically detected)
315 "ide0=qd6580" : probe/support qd6580 interface
316 "ide0=ali14xx" : probe/support ali14xx chipsets (ALI M1439/M1445)
317 "ide0=umc8672" : probe/support umc8672 chipsets
319 There may be more options than shown -- use the source, Luke!
321 Everything else is rejected with a "BAD OPTION" message.
323 ================================================================================
327 IDE = Integrated Drive Electronics, meaning that each drive has a built-in
328 controller, which is why an "IDE interface card" is not a "controller card".
330 IDE drives are designed to attach almost directly to the ISA bus of an AT-style
331 computer. The typical IDE interface card merely provides I/O port address
332 decoding and tri-state buffers, although several newer localbus cards go much
333 beyond the basics. When purchasing a localbus IDE interface, avoid cards with
334 an onboard BIOS and those which require special drivers. Instead, look for a
335 card which uses hardware switches/jumpers to select the interface timing speed,
336 to allow much faster data transfers than the original 8MHz ISA bus allows.
338 ATA = AT (the old IBM 286 computer) Attachment Interface, a draft American
339 National Standard for connecting hard drives to PCs. This is the official
342 The latest standards define some enhancements, known as the ATA-2 spec,
343 which grew out of vendor-specific "Enhanced IDE" (EIDE) implementations.
345 ATAPI = ATA Packet Interface, a new protocol for controlling the drives,
346 similar to SCSI protocols, created at the same time as the ATA2 standard.
347 ATAPI is currently used for controlling CDROM and TAPE devices, and will
348 likely also soon be used for Floppy drives, removable R/W cartridges,
349 and for high capacity hard disk drives.
351 How To Use *Big* ATA/IDE drives with Linux
352 ------------------------------------------
353 The ATA Interface spec for IDE disk drives allows a total of 28 bits
354 (8 bits for sector, 16 bits for cylinder, and 4 bits for head) for addressing
355 individual disk sectors of 512 bytes each (in "Linear Block Address" (LBA)
356 mode, there is still only a total of 28 bits available in the hardware).
357 This "limits" the capacity of an IDE drive to no more than 128GB (Giga-bytes).
358 All current day IDE drives are somewhat smaller than this upper limit, and
359 within a few years, ATAPI disk drives will raise the limit considerably.
361 All IDE disk drives "suffer" from a "16-heads" limitation: the hardware has
362 only a four bit field for head selection, restricting the number of "physical"
363 heads to 16 or less. Since the BIOS usually has a 63 sectors/track limit,
364 this means that all IDE drivers larger than 504MB (528Meg) must use a "physical"
365 geometry with more than 1024 cylinders.
367 (1024cyls * 16heads * 63sects * 512bytes/sector) / (1024 * 1024) == 504MB
369 (Some BIOSs (and controllers with onboard BIOS) pretend to allow "32" or "64"
370 heads per drive (discussed below), but can only do so by playing games with
371 the real (hidden) geometry, which is always limited to 16 or fewer heads).
373 This presents two problems to most systems:
375 1. The INT13 interface to the BIOS only allows 10-bits for cylinder
376 addresses, giving a limit of 1024cyls for programs which use it.
378 2. The physical geometry fields of the disk partition table only
379 allow 10-bits for cylinder addresses, giving a similar limit of 1024
380 cyls for operating systems that do not use the "sector count" fields
381 instead of the physical Cyl/Head/Sect (CHS) geometry fields.
383 Neither of these limitations affects Linux itself, as it (1) does not use the
384 BIOS for disk access, and it (2) is clever enough to use the "sector count"
385 fields of the partition table instead of the physical CHS geometry fields.
387 a) Most folks use LILO to load linux. LILO uses the INT13 interface
388 to the BIOS to load the kernel at boot time. Therefore, LILO can only
389 load linux if the files it needs (usually just the kernel images) are
390 located below the magic 1024 cylinder "boundary" (more on this later).
392 b) Many folks also like to have bootable DOS partitions on their
393 drive(s). DOS also uses the INT13 interface to the BIOS, not only
394 for booting, but also for operation after booting. Therefore, DOS
395 can normally only access partitions which are contained entirely below
396 the magic 1024 cylinder "boundary".
398 There are at least seven commonly used schemes for kludging DOS to work
399 around this "limitation". In the long term, the problem is being solved
400 by introduction of an alternative BIOS interface that does not have the
401 same limitations as the INT13 interface. New versions of DOS are expected
402 to detect and use this interface in systems whose BIOS provides it.
404 But in the present day, alternative solutions are necessary.
406 The most popular solution in newer systems is to have the BIOS shift bits
407 between the cylinder and head number fields. This is activated by entering
408 a translated logical geometry into the BIOS/CMOS setup for the drive.
409 Thus, if the drive has a geometry of 2100/16/63 (CHS), then the BIOS could
410 present a "logical" geometry of 525/64/63 by "shifting" two bits from the
411 cylinder number into the head number field for purposes of the partition table,
412 CMOS setup, and INT13 interfaces. Linux kernels 1.1.39 and higher detect and
413 "handle" this translation automatically, making this a rather painless solution
414 for the 1024 cyls problem. If for some reason Linux gets confused (unlikely),
415 then use the kernel command line parameters to pass the *logical* geometry,
418 If the BIOS does not support this form of drive translation, then several
419 options remain, listed below in order of popularity:
421 - use a partition below the 1024 cyl boundary to hold the linux
422 boot files (kernel images and /boot directory), and place the rest
423 of linux anywhere else on the drive. These files can reside in a DOS
424 partition, or in a tailor-made linux boot partition.
425 - use DiskManager software from OnTrack, supplied free with
426 many new hard drive purchases.
427 - use EZ-Drive software (similar to DiskManager). Note though,
428 that LILO must *not* use the MBR when EZ-Drive is present.
429 Instead, install LILO on the first sector of your linux partition,
430 and mark it as "active" or "bootable" with fdisk.
431 - boot from a floppy disk instead of the hard drive (takes 10 seconds).
433 If you cannot use drive translation, *and* your BIOS also restricts you to
434 entering no more than 1024 cylinders in the geometry field in the CMOS setup,
435 then just set it to 1024. As of v3.5 of this driver, Linux automatically
436 determines the *real* number of cylinders for fdisk to use, allowing easy
437 access to the full disk capacity without having to fiddle around.
439 Regardless of what you do, all DOS partitions *must* be contained entirely
440 within the first 1024 logical cylinders. For a 1Gig WD disk drive, here's
441 a good "half and half" partitioning scheme to start with:
443 geometry = 2100/16/63
444 /dev/hda1 from cyl 1 to 992 dos
445 /dev/hda2 from cyl 993 to 1023 swap
446 /dev/hda3 from cyl 1024 to 2100 linux
448 To ensure that LILO can boot linux, the boot files (kernel and /boot/*)
449 must reside within the first 1024 cylinders of the drive. If your linux
450 root partition is *not* completely within the first 1024 cyls (quite common),
451 then you can use LILO to boot linux from files on your DOS partition
452 by doing the following after installing Slackware (or whatever):
454 0. Boot from the "boot floppy" created during the installation
455 1. Mount your DOS partition as /dos (and stick it in /etc/fstab)
456 2. Move /boot to /dos/boot with: cp -a /boot /dos ; rm -r /boot
457 3. Create a symlink for LILO to use with: ln -s /dos/boot /boot
458 4. Move your kernel (/vmlinuz) to /boot/vmlinuz: mv /vmlinuz /boot
459 5. Edit /etc/lilo.conf to change /vmlinuz to /boot/vmlinuz
460 6. Re-run LILO with: lilo
462 A danger with this approach is that whenever an MS-DOS "defragmentation"
463 program is run (like Norton "speeddisk"), it may move the Linux boot
464 files around, confusing LILO and making the (Linux) system unbootable.
465 Be sure to keep a kernel "boot floppy" at hand for such circumstances.
466 A possible workaround is to mark the Linux files as S+H+R (System,
467 Hidden, Readonly), to prevent most defragmentation programs from
468 moving the files around.
470 If you "don't do DOS", then partition as you please, but remember to create
471 a small partition to hold the /boot directory (and vmlinuz) as described above
472 such that they stay within the first 1024 cylinders.
474 Note that when creating partitions that span beyond cylinder 1024,
475 Linux fdisk will complain about "Partition X has different physical/logical
476 endings" and emit messages such as "This is larger than 1024, and may cause
477 problems with some software". Ignore this for linux partitions. The "some
478 software" refers to DOS, the BIOS, and LILO, as described previously.
480 Western Digital ships a "DiskManager 6.03" diskette with all of their big
481 hard drives. Use BIOS translation instead of this if possible, as it is a
482 more generally compatible method of achieving the same results (DOS access
483 to the entire disk). However, if you must use DiskManager, it now works
484 with Linux 1.3.x in most cases. Let me know if you still have trouble.
486 My recommendations to anyone who asks about NEW systems are:
488 - buy a motherboard that uses the Intel Triton chipset -- very common.
489 - use IDE for the first two drives, placing them on separate interfaces.
490 - very fast 7200rpm drives are now available
491 (though many problems have been reported with Seagate ones).
492 - place the IDE cdrom drive as slave on either interface.
493 - if additional disks are to be connected, consider your needs:
494 - fileserver? Buy a SC200 SCSI adaptor for the next few drives.
495 - personal system? Use IDE for the next two drives.
496 - still not enough? Keep adding SC200 SCSI cards as needed.
498 Most manufacturers make both IDE and SCSI versions of each of their drives.
499 The IDE ones are usually as fast and cheaper, due to lower command overhead
500 and the higher data transfer speed of UDMA2. But fast/ultrawide/superlative
501 SCSI is still king of the heap, especially for servers, if you've got the bucks.
505 For current maintainers of this stuff, see the linux/MAINTAINERS file.