NSA injected spyware into your hard drive

[Valerion]

Probably not. A greater deterrent is that America wants the world to play nice, honour contracts, copyright, intellectual property, etc. This kind of blatant hypocrisy would undermine that.

Note: I concede that this is theoretical and the NSA may consider themselves above laws that apply to mortals. Benefit of the doubt?

Protect the intellectual property of their own companies perhaps.

 

[zamicro]

So what is next? Open source firmware for HDDs?

 

[Jet-Fighter7700]

So what is next? Open source firmware for HDDs?

Impossible to do, like creating printer driver open source,

Honestly the nsa can easily get into anything they want,
Probably have quite a few tricks we don't yet know about, sure they can screengrab from certain monitors directly, trace printouts to specific printers,

Batman like sonar from speakers/tv set to identify people in a room, by the sound bouncing off them

Only foolproof way is analog or caveman tech, pen and paper..

 

[zamicro]

Impossible to do, like creating printer driver open source

Why?

 

[Valerion]

Why?

There are proprietary secrets in the firmware, some of them giving the manufacturer some advantage. And it shows how the internals work. And there are very few people that work directly with hardware at that level. It's not like writing a device driver through an API, it's working directly with the raw IC's in the device to provide the higher-level API.

Same reason there are few assembler developers these days, more C/C++ and a lot more JAVA/C# developers.

 

[Sonic2k]

There are proprietary secrets in the firmware, some of them giving the manufacturer some advantage. And it shows how the internals work. And there are very few people that work directly with hardware at that level. It's not like writing a device driver through an API, it's working directly with the raw IC's in the device to provide the higher-level API.

Same reason there are few assembler developers these days, more C/C++ and a lot more JAVA/C# developers.

Uh, most of the embedded stuff these days is programmed in C, even the smallest, cheapest of chips.
Assembler is too time-consuming.

The firmware is what implements the ATA register set you see when talking to the drive directly via a device driver.
So when you send a READ command for example, this firmware is what kicks into action, reads the track, decodes the information and checks it for errors, and presents that, as data in the buffer, to the host.

Most drive manufacturers implement proprietary commands to upgrade/write to the FLASH on the hard disk controller. These commands are not defined in the SFF ATA spec and vary by manufacturer most likely.
I used to patch my DVD-ROM drive firmware using this method, to get rid of region coding, about 10 years ago.

 

[Valerion]

Uh, most of the embedded stuff these days is programmed in C, even the smallest, cheapest of chips.
Assembler is too time-consuming.

The firmware is what implements the ATA register set you see when talking to the drive directly via a device driver.
So when you send a READ command for example, this firmware is what kicks into action, reads the track, decodes the information and checks it for errors, and presents that, as data in the buffer, to the host.

Most drive manufacturers implement proprietary commands to upgrade/write to the FLASH on the hard disk controller. These commands are not defined in the SFF ATA spec and vary by manufacturer most likely.
I used to patch my DVD-ROM drive firmware using this method, to get rid of region coding, about 10 years ago.

The question was around OSS-type firmware. My argument is that there are few people capable of writing firmware drivers (because it's much more low-level, and because of proprietary code, patents, etc.). These people are also paid well. So that will likely never become a reality, and we will have to keep relying on the HDD manufacturers. And they have no real reasons to release that, unless there's a general firmware upgrade for a specific model.

 

[Sonic2k]

My argument is that there are few people capable of writing firmware drivers (because it's much more low-level, and because of proprietary code, patents, etc.). These people are also paid well.

Partly fact, partly fiction. Very few can do this yes, however all the gurus I know are certainly not paid well, and I certainly wasn't paid well when I used to this sort of work. The people who get paid well are always the sales/marketing people and the bean counters.