In-depth technology research: finding new ways to recover data, accessing firmware, writing programs, reading bits off the platter, recovering data from dust.
Forum rules
Please
do not post questions about data recovery cases here (use
this forum instead). This forum is for topics on finding new ways to recover data. Accessing firmware, writing programs, reading bits off the platter, recovering data from dust...
February 8th, 2011, 20:22
Did hard drives always have firmware information on platters?
Where is the majority of firmware info stored on a modern hdd - the
chip on the PCB or on the SA?
February 9th, 2011, 3:45
AFAIK there has always been a certain amount of fw on the platters.
The vast majority of FW (at least on modern drives) is on the platters, with the ROM info on the PCB just a "kick-start" to read the actual FW from the platters.
December 12th, 2011, 16:23
pcimage wrote:The vast majority of FW (at least on modern drives) is on the platters, with the ROM info on the PCB just a "kick-start" to read the actual FW from the platters.
So, the operational code necessary to the bare workings of our disk drives is in large part stored on the disk platters themselves! Calling data stored on magnetic disks "firmware" is an abuse in itself BTW.
Isn't such a scheme a shame in your opinion (even if said code is stored in more than one place for redundance) ? Shouldn't the firmware be entirely in (reP)ROM ? Else, as soon as part the "firmware" is damaged, -all- data become inaccessible to an ordinary user who can't afford professional recovery services.
At least if they were honest, disk drive vendors IMO should provide ways for *end users* to update/rewrite a damaged firmware for no additional (or minimal) added cost.
Questions : How *big* is a typical disk "firmware" (that part which is stored on the disk instead of ROM) ? What additional cost would it make on a new disk if enough PROM were provided for the whole operative firmware ? Isn't that industry simply ripping off us customers for its exclusive benefit ?
TIA
PS: I had a disk (Maxtor) trashed and all data lost in practice for (probably) the cause outlined above. Vendor greed
December 12th, 2011, 17:23
It's always a good idea to keep a separate backup of your data
December 12th, 2011, 17:46
guru wrote:It's always a good idea to keep a separate backup of your data
Sure thing. Care to answer my questions ? How large is the piece of code that is so carelessly stored on the disk platters, and is there a good reason why it isn't kept in ROM instead ?
That code has to be read from the disk - presumably upon power-on - into semiconductor based memory before it can be executed, there has to be memory for it. Why don't they use ROM for the program itself, plus of course RAM for transient data and buffers, is a serious question.
December 12th, 2011, 17:56
http://files.hddguru.com/index.phpYou can find lots of examples here
Czerno wrote:is there a good reason why it isn't kept in ROM instead ?
That would be like trying to solder Windows onto the motherboard
December 12th, 2011, 18:34
Czerno
Unfortunately computer storage is not incapable of error or failure. SSD drives store all firmware in flash and these devices still fail.
I understand that you are not happy with the way firmware is stored on the platters. Before it completely eats out your heart just learn from your experience and always backup your data and then move on.
December 13th, 2011, 3:19
drc wrote:Czerno wrote:is there a good reason why it isn't kept in ROM instead ?
That would be like trying to solder Windows onto the motherboard
My GPS has an embedded Windows CE OS ...
December 13th, 2011, 3:25
pcimage wrote:AFAIK there has always been a certain amount of fw on the platters.
You don't go back as far as me, then.
What about the old MFM drives?
December 13th, 2011, 6:15
drc wrote:http://files.hddguru.com/index.php
You can find lots of examples here
Czerno wrote:is there a good reason why it isn't kept in ROM instead ?
That would be like trying to solder Windows onto the motherboard
From a cursory peek into one or two of the FW files, it appears disk FW is a few tens of kilobytes at most. How's that comparable to Windows, or even DOS ?
December 13th, 2011, 6:26
guru wrote:Czerno
I understand that you are not happy with the way firmware is stored on the platters. Before it completely eats out your heart just learn from your experience and always backup your data and then move on.
Unhappy is an understatement, disgusted and horrified by what might amount to conspiration.
Before I "move on", as you invite me to, aren't there /any/ brand & models (consumer) hard disk drives having the firmware stored in permanent semiconductor memory ? Failing that, a brand/model which retains enough functionality, in case the operative software stored on disk cannot be loaded at power on, that
the end user can (try to) regenerate it using some manufacturer provided software but without special tools ?
December 13th, 2011, 6:37
Czerno wrote:drc wrote:http://files.hddguru.com/index.php
You can find lots of examples here
Czerno wrote:is there a good reason why it isn't kept in ROM instead ?
That would be like trying to solder Windows onto the motherboard
From a cursory peek into one or two of the FW files, it appears disk FW is a few tens of kilobytes at most. How's that comparable to Windows, or even DOS ?
ISTM that it's more like megabytes than kilobytes. Even so, it shouldn't add more than a dollar to the cost.
In fact this 8Mbyte SPI serial flash memory is a drop-in replacement for existing 4Mbit chips:
http://www.winbond.com/NR/rdonlyres/092 ... 25X64V.pdfThat said, hard drive margins have been wafer thin for a long time ...
December 13th, 2011, 6:53
fzabkar wrote:What about the old MFM drives?
I still have a 80286 mobo from the 80s in working (more or less) order with two of those, 20
megabytes each, and the associated 8-bit ISA controller card.
Though hard disks must fail eventually in a way or another, and I understand the notion of economic compromises, I have yet to accept that they be craftfully
designed to fail, as the thread subject demonstrates.
fzabkar wrote:ISTM that it's more like megabytes than kilobytes. Even so, it shouldn't add more than a dollar to the cost.
In fact this 8Mbyte SPI serial flash memory is a drop-in replacement for existing 4Mbit chips:
http://www.winbond.com/NR/rdonlyres/092 ... 25X64V.pdfThat said, hard drive margins have been wafer thin for a long time ...
I, and probably most customers if they were properly informed, would pay a couple dollars extra to have that nonsense suppressed. In fact I would not mind trading half the storage capacity for quality and reliabiilty. There should be obligatory regulations, not just pseudo norms made by the industry - but I know the US hate the very idea.
December 13th, 2011, 6:58
Czerno wrote:guru wrote:Czerno
I understand that you are not happy with the way firmware is stored on the platters. Before it completely eats out your heart just learn from your experience and always backup your data and then move on.
:)
Unhappy is an understatement, disgusted and horrified by what might amount to conspiration.
I feel the same way about softmodems, Win printers, and any other device where software has been substituted for hardware. This includes NAS boxes where the embedded OS is booted from the HDD, and DVD recorders where part of the "firmware" resides on the HDD, thereby preventing you from replacing a faulty drive, or upgrading to a higher capacity without being gouged by the manufacturer.
At the very least I would like HDD manufacturers to maintain a database of adaptive data for each serial number so that the user could purchase a replacement board that comes preprogrammed for his particular drive.
December 13th, 2011, 11:48
Regarding disasters which surely will happen:
Every HDD (incl USB attached HDDs) with embedded encryption should be clearly labelled "DO NOT USE FOR STORAGE OF DATA!".
In other words: "Should only be used for transporting copies of data".
December 13th, 2011, 11:57
CT512M4SSD2
December 14th, 2011, 5:36
Guru :
CT512M4SSD2
Praytell what you meant ? Would that be a drive with firmware stored in additional ROM rather than on the magnetic surfaces ?
If drives of the kind exist at all, where is a list to be found please gentlemen ?
December 14th, 2011, 20:24
Hi,
Your only choice is SSD HDD.
March 15th, 2012, 1:42
Eleg wrote:Regarding disasters which surely will happen:
Every HDD (incl USB attached HDDs) with embedded encryption should be clearly labelled "DO NOT USE FOR STORAGE OF DATA!".
In other words: "Should only be used for transporting copies of data".
And drives that don't have encryption should be marked "Should not be used for transporting copies of data." (Unless you like it getting stolen!)
Heck, even put TrueCrypt on it instead of internal encryption. It's not like you're needing the performance of built-in crypto hardware when moving files via sneaker net.
Yeah, there should be some jumper or firmware version that prevents passwords being set or 'fixes' them to a default as well. For security it should not work unless the drive is already unlocked, obviously. Glitches causing corrupted passwords are very much possible (and not exactly unheard of).
March 16th, 2012, 17:02
Hmm, didn't even MFM and RLL drives have sectors reserved for mapping info? I know that some early hard drives were directly controlled by the controller card. And yes, they were dog slow.
BTW: You can tell how old there are, because they're named after the technology and not the protocol/interface! IDE is an interface and ATA is a protocol? PATA is revisionist version of IDE, in practice. It was technically always ATA? I never heard that term until I got into ATAPI programming. SATA is the name of the interface and ATA is the protocol for SATA drives?
Powered by phpBB © phpBB Group.