IDEINFO.EXE Page 1 Contents Introduction.......................... 1 Output................................ 2 How to interpret the output........... 3 Updating CMOS parameters...............5 Making an autodetect floppy............6 Speedtest..............................7 Some notes............................ 8 Credits............................... 8 Me.................................... 8 Changes, additions.................... 9 Introduction ------------- Try ? or H as parameter and you will see a short description of possible parameters. This is what the output looks like: ________________________________________________________________ IDEINFO [H || ?] || [R] || [CMOS] || [F [P xxx] [P xxx].. ] Version 1.1g. Tries to tell all about all IDE drives, including ATA-2 stuff. Options may be preceded with -, / or none -H or -? prints this help. -R Reads idedrive.* files from current directory. -CMOS allows to update CMOS with new drive parameters, works with AMI BIOS if it has user definable types -F Writes drive identify info to idedrive.* files to current directory. -P xxx looks only adapter at port xxx, xxx in hex format File extension is next available number when writing. File size is 512 bytes. You may redirect output. Example: Look only for adapters at 1F0h (primary) and 170h (normal secondary) and write info to files. IDEINFO P 1f0 P 170 F ________________________________________________________________ Default action is to scan all four addresses mentioned in ATA-2 draft spec for IDE adapters(1f0h, 0170h, 0f0h, 070h), if found, report characteristics of master drive and search also for slave. File format will remain the same even when new versions of ideinfo appear. That is because ideinfo uses the same raw format that the drive sends. Better way than redirecting output is therefore to store drive info to files. New version of ideinfo may find new features or be more accurate. If you have a file stored with old version you just run new version with -R parameter. Page 2 Output -------- Output for a normal ATA (not ATA-2) drive looks like this: Note: Identify Device information words used are in brackets. _______________________________________________________________ DRIVE 0 Adapter 1 at base address 170h Device is not removable. Disk Reports BIOS Reports Default Current # of Cylinders______: 665 [1] 665 [54] 0 # of Heads__________: 16 [3] 16 [55] 0 # of Sectors/Track__: 63 [6] 63 [56] 0 Model Number________: Conner Peripherals 340MB - CFA340A [27-46] Serial Number_______: BQBA27A [10-19] Controller Rev. #___: 6FT1.70 [23-26] Double Word Transfer: No [48] Buffer size (kB)____: 64 [21] # of ECC bytes______: 4 [22] DMA support Yes [49] IORDY supported No [49] IORDY can be disabled No [49] Standby timer support: No [49] Secure Mode Supported: No [50] # of secs/interrupt_: 64[47]Current setting 16 [59] LBA support No [49],[60-61],[53],[57-58] ATA compliant max. PIO transfer mode: 1 [51] SW DMA txfer cycle timing modes: [62] Active [62] MW DMA txfer cycle timing modes: 1, 0 Active 0 [63] Sorry, no ATA-2 features implemented Adapter 1 at 170h Drive 1 not found Last status 0h ______________________________________________________________ The output above was generated running under OS/2 2.11. For some reason OS/2 lets user programs access secondary adapters and correctly inhibits access to primary adapter. Note the effect of /SMS (Set Multiple Support) parameter of OS/2 IDE driver IBM1S506.ADD (# of secs/interrupt_:, Current setting 16 ). For newer drives you might get also something like this: _______________________________________________________________ Congratulations, your drive supports ATA-2 [53] Advanced PIO transfer modes supported: 4, 3 [64] Min MW DMA transfer cycle time/word: 120 ns 16.7MB/s [65] Mfg Recommended MW DMA transfer Cycle Time 120 ns 16.7MB/s [66] Min PIO transfer Cycle Time w/o Flow Control 200 ns 10.0MB/s [67] Min PIO transfer Cycle Time w IORDY Flow Control 120 ns 16.7MB/s [68] ________________________________________________________________ For ATA-2 reference I used Revision 2f july 26, 1994. I may have messed up some interpretations of spec, hope not too many. Page 3 How to interpret the output --------------------------- For a more complete reference you should get ATA-2 spec (draft) from: ATA Anonymous FTP Site fission.dt.wdc.com ATA directory is: "/pub/ata" SCSI Bulletin Board 719-574-0424 Document Distribution Global Engineering 15 Inverness Way East Englewood, CO 80112-5704 Tel: 303 -792-2181 or 800-854-7179 Fax: 303 -792-2192 Translation: ------------ Disk Reports Default Current # of Cylinders______: 665 665 # of Heads__________: 16 16 # of Sectors/Track__: 63 63 First column is default translation and next is current. Current may be all -1 if validity bit is not set. Some drives accept any translation that doesn't exceed capacity and some accept only one translation. Multiple sector setting: ------------------------ Normally the drive generates an interrupt for each sector read/written but enabling multiple sector setting and using read/write multiple commands enables the drive controller to work more efficiently and saves CPU cycles because interrupt is generated and processed only once for a given number of sectors. You must have BIOS support or a software driver to utilize this. Some drives get 30% data transfer speed boost, others only %5. Buffer size (kB)____: 64 # of secs/interrupt_: 64 Current setting 16 If Buffer size is half of secs/interrupt value and "Current setting" is missing the drive doesn't most propably support multiple sector per interrupt feature. If "Current setting" text (even with no number) is shown, the drive should work with multiple sector per interrupt. LBA support ----------- Found in most drives with capacity over 528MB. Normal BIOS and IDE translation parameters together have a limit of 528MB. BIOS alone has a 8GB limit. IDE alone has a 137GB limit. LBA (Logical Block Address) uses no translation and blocks or sectors are addressed with 0 to a device specific maximum LBA number. The maximum LBA number cannot exceed 268 435 455 (137GB). Using LBA should save some CPU cycles. Page 4 DMA support ----------- DMA (Direct Memory Access) means that the device writes directly to main memory thus freeing CPU from I/O handling. The other transfer mode is PIO (Programmable I/O). ISA bus has a limit of 16MB DMA addressable memory. IORDY supported --------------- IORDY is needed with PIO mode 3 and 4. Transfer modes -------------- +--------------------------------------------------+------+ | PIO |Mode 0|Mode 1|Mode 2|Mode 3|Mode 4| | timing parameters | nsec | nsec | nsec | nsec | nsec | +----+----------------------+------+------+------+------+------+ | t0 | Cycle time (1) (min) | 600 | 383 | 240 | 180 | 120 | Transfer rate MB/s | 3.3 | 5.2 | 8.3 | 11.1 | 16.6 | Modes 0 .. 2 are included in ATA spec, PIO mode 3 and 4 in ATA-2. +---------------------------------------------------------+ | Single word DMA | Mode 0| Mode 1| Mode 2| | timing parameters | nsec | nsec | nsec | +----+---------------------------------+-------+-------+-------| | t0 | Cycle time (min) | 960 | 480 | 240 | | Transfer rate (calculated) MB/s | 2.1 | 4.2 | 8.3 | +------------------------------------+--------+--------+ | Multiword DMA | Mode 0 | Mode 1 | Mode 2 | | timing parameters | nsec | nsec | nsec | +-----+---------------------------+--------|--------|--------| | t0 | Cycle time | 480 | 150 | 120 | | Transfer rate (calculated) MB/s | 4.2 | 13.3 | 16.6 | DMA mode 0 is included in ATA spec, mode 1 and 2 in ATA-2. Page 5 Updating CMOS parameters New with version 1.1 is similar feature that is found in new BIOSes. Ideinfo does more. If the drive has already been partitioned the partition table is analyzed and suitable parameters are given as one possible set to feed in to CMOS. If the drive has FAT partition at the begining of the drive then also BPB (BIOS Parameter Block) from DOS boot sector is analyzed and suitable parameters are given as one possible set to feed in to CMOS. The other two suggestions come directly from the drive. If CMOS parameter is given to Ideinfo the output might look like this: Disk 0 or C: Disk Reports Part. table BPB Default Current Cylinders 1024 1024 1024 0 Heads 16 16 16 0 Sectors/Track 63 63 63 0 Now you may update CMOS with new parameters for drive 0. Choose from above. Press D for Default, C for Current, P for Part. table, B for BPB or S for skip this drive Disk 1 or D: Disk Reports Part. table BPB Default Current Cylinders 1062 1062 1024 0 Heads 16 16 16 0 Sectors/Track 63 63 63 0 Now you may update CMOS with new parameters for drive 1. Choose fom above. Press D for Default, C for Current, P for Part. table, B for BPB or S for skip this drive Only primary adapter is searched for drives. Limitations ----------- Ideinfo can update CMOS succesfully only if the BIOS is by AMI and it uses drive type 47 as user definable drive parameters. CMOS has two checksum bytes and some BIOS versions might calculate them differently. Storage place for drive type 47 parameters varies but AMI should place them at the same bytes. New four drive BIOSes might differ in that. Ideinfo does no checking at all. It does not check that the BIOS is by AMI. It does not check that drive type 47 is user definable. It does not check that the bytes to store type 47 are correct. But you may skip update CMOS and write down the parameters that seem correct. Later you just start BIOS setup and update the parameters manually. Page 6 When partition table or BPB parameters should be used instead of what the drive suggests ---------------------------------- Some IDE drives work with any combination of parameters that don't exceed total capacity of the drive. Others work with two, possibly three set of parameters. If the drive has already been partitioned and the CMOS has lost its contents, the parameters that the drive suggests may not be the ones that were used when the drive was partitioned. If you use normal autodetect feature, the machine doesn't boot from hard disk, but hard disks are found when the machine is booted from floppy. BPB is DOS specific and cylinder count is calculated. If the first partition doesn't fill the entire drive then the cylinder count from BPB is too small. The other two parameters from BPB are reliable. Partition table has similar limitation. If the drive hasn't been all used up by partitions the cylinder count is too small. However, all existing partitions can be accessed if partition table is trusted. Making an autodetect floppy --------------------------- Ideinfo does not need BIOS to read IDE drives. To load Ideinfo into memory you need DOS. To load DOS you need at least a floppy. You may have to invoke BIOS setup to define your floppy drive. To make autodetect safe you should use CMOS utility to save current settings before autodetect. Make first a bootable floppy with "format a: /s" command, copy CMOS utility and Ideinfo to it and make following autoexec.bat ----------------------------- @ECHO OFF ECHO Press Ctrl-C if you want to restore previously saved ECHO CMOS settings. PAUSE CMOS IDEINFO F IDEINFO CMOS ------------------------------- CMOS saves first the current settings and if something goes wrong you may reboot with the same floppy, press Ctrl-C and after looking with "dir" what CMOS.* file is the latest, run CMOS nnn, where nnn is the same number that you saw as extension in the settings file you wish to restore to CMOS. Ideinfo F line stores drive identify data for later examination. It is not neccessary but convinient. Page 7 Speedtest --------- Speedtest is removed from Ideinfo, use separate Drvspeed instead. Page 8 Some notes ---------- IDEINFO was compiled with BC++ 3.1 using 286 intructions. XT or 8-bit interfaces were not even considered, so you should have a 286 or better to use this program. Currently IDEINFO tries to read idedrive.0 .. idedrive.99 files with R parameter. I hope that 100 files is enough for a while. Increasing it would slow it down or extensions would have to be in sequence. The line # of secs/interrupt_: 64[47]Current setting 16 [59] apparently tells also if OS/2 (2.1, 2.11) IBM1S506.ADD will support /SMS parameter. Word 59 has a validity bit and if it is not set "Current setting" will not be displayed and IBM1S506.ADD ignores /SMS. There are some drives that have validity bit set but still IBM1S506.ADD ignores /SMS but the only such drive I have encountered is Seagate ST3391A. I am not sure about which last status values (displayed when adapter is not found) mean that there is no device responding. I have seen 7fh on EISA bus machine and OS/2, ffh on others. If you see some other value there is propably something using that address. Credits -------- Great help in getting started was the source code of: DUG_IDE : Provides IDE disk info on both drives in the system Version 1.0 Copyright (c) 12 Jan 1993 by Doug Merrett - All rights reserved Linux 1.1 kernel sources were also consulted. Me --- Copyright (c) July 1994 by Raimo Koski - All rights reserved Postal Address: Uudenmaantie 23 K 3 20720 Finland I can receive Internet mail with address: raimo.koski@mikropc.fi You can also try Fidonet with address: 2:222/70 Page 9 Changes, additions ------- 8/29/94 Checked that ATA-2 rev. F caused no changes. The original version of ideinfo was based on ATA-2 rev. c. Added PIO mode 4 and MW DMA mode 2 to this document. Source remains the same. 11/15/94 Found a bug! PIO 4 and 5 were not reported correctly. Yes, there are PIO 4 drives commonly available. Some cosmetic changes. Raised timeout value. Only one person has complained about ideinfo not finding a drive. Hope the problems have not been common. The worst case seems to be fast CPU, slow drive and two drive IDE bus. If you encounter this problem, simply run ideinfo again. You can also read the drive that is not found with ordinary means (say DIR or something) and then try ideinfo again. Seems that ideinfo is not resetting drives properly. I should look at this sometime but it is not so important. Use your operating system (actually it usually calls BIOS to do this) to do resetting. Say DIR or similar. Had a possibility to test a drive with 108 sector/interrupt support. OS/2 WARP II used 16. Seems that OS/2 never uses more than 16 and if the drive can't accept that then OS/2 can use 8 also. I haven't seen other values with OS/2. 11/31/94 Checked ATA-2 r.2i. Words 20 and 21 are now obsolete. Removed word 20 reporting and added removable/non removable reporting. Added delay(50) lines after writing to ports, that might be better than raising the time out value. On subject of time out, it seems to be the root of all trouble of drive light staying lit and drives not found errors. This version raises time out value once again. Changed max_cntr value to 3. Now only three adapters are searched. I have wondered about base address 0x70 mentioned in ATA-2 informative annex. I think that it can't be used for secondary IDE adapter in normal AT class machines because that address belongs to CMOS. Anyway few people have more than two IDE adapters in their machines. I was one of the very few people that used secondary IDE adapter back in spring -93 when OS/2 2.1 beta IDE driver was the first good driver to support two adapters. Now they are more common. Base address 120h is used by some Plus Hardcards. 12/11/94 Added CMOS update and smart autodetect. Now Ideinfo reads IDE drives without any help from BIOS or DOS. Identify device command output is generated usually by drive controller and not read from disks. 06/08/95 Added speedtest, check for AMI BIOS. A bug in second disk partition table cyls seek found and corrected. Updated to ATA-3 R1 level. Added logic for avoiding divisions by zero (happened occasionally with older drives, which were reporting incorrect values). Page 10 08/03/96 Moved speedtest to separate utility. Updated to ATA-4 level. Added ATAPI device detection. Added following fields for ATAPI devices: Device types [0] Direct-access device Sequential-access device Printer device Processor device Write-once device CD-ROM device Scanner device Optical memory device Medium changer device Communications device Array controller device Reserved device type! Reserved for ACS IT8 Unknown or no device type Command packet length [0] with 16 byte command packet with 12 byte command packet with reserved length command packet Typical time from receipt of PACKET command to release: %6u ns [71] Typical time from receipt of SERVICE command to BSY cleared to zero: %6u ns [72] ATA devices: ATA version major, minor [80], [81] Power management features supported [82] Removable features supported [82] Security features supported [82] Security state [128] Security enabled Security locked Security frozen Security count expired Security level maximum / high SMART features supported [82] Most of these fields are untested. One reason is that I have not seen devices that support them, the other is that there propably are yet none that support them all :-) Third is that I have not seen yet other programs that report them, so I have no comparison.