where to send SSDs for development?

[FliesLikeABrick]

Hi,

I have a few fake SSDs which were not detected as such by f3. Are there any contributors/developers that would like me to send them their way? I believe I have 3 identical drives, they claim to be 1TB. I can't remember from my testing how big they actually are. Happy to mail these out within the US at no cost.

[LordRafa]

Hi, I just got a wd ssd from china and I was testing it with f3 to see if it is fake.

I was taking a look how f3 work and its rnd number generator and it looks pretty flaky perhaps due to the price of an ssd it is worth to invest in some anti-f3 firmware solution. Why not use a real rnd generator with a seed provided by the user, or at least make it available as an option, something like --extreme_paranoid_seed xxxx

Which test did you run to detect that they were fake?

[AltraMayor]

Hi @LordRafa,

Long ago, users asked me to keep F3 compatible with H2testw. They even persuaded the author of H2testw to provide me with the random number generator he used since H2testw is not open source. F3 and H2testw have been compatible since then. Nobody has found a drive that tries to trick the number generator. But, of course, it's possible, and F3 will respond accordingly if fake drives evolve as F3 has responded in the past.

[LordRafa]

And indeed the original random generator still valid....

I run f3 with the original generator and a modified version patched to use rand/srand on my SSD and both return the same result --- GOOD but the original generator was much faster (as expected). I run/modify both the f3read/write and the f3brew.

Unfortunately I didn't have time to understand the f3probe... I would say that it is testing only a chosen random upper block but with the other tests I think that I am happy enough even knowing that rand function is a weak generator (paranoia never stops I guess :P)....

I just wonder how the original user found that hims were fake

[FliesLikeABrick]

@LordRafa these are super shady SSDs that came in a box of random crap at a local auction, someone before me fell for the cheap price and then they ended up in a box with some other items I did want at the auction. They're completely generic chinesium and just say 1TB on the box. Similar in many ways to what is shown here https://linustechtips.com/topic/1373318-psa-counterfeit-1tb-ssds-being-sold-on-amazon/

I can't remember the exact test I did, I think I put one in a machine and did dd if=/dev/zero to it and monitored its rate of writing and status. I think towards the end of the storage it did have (32gbyte?) the write speed slowly but infinitely approached zero. It has been a while since I did that test and I do not remember for certain the failure mode -- just that they definitely took an ext4 format as if you had a 1tbyte raw device, but when you tried writing data to it it would not go past a certain point successfully.

I have 3 left here, happy to mail them (I'll cover postage) to anyone who wants to develop or experiment.

[mikaey]

Hi!

I recently bought a pair of 2TB microSD cards off of AliExpress. f3probe was able to identify them as fake flash, but it identified the real capacity as 1TB. But...f3write/f3read showed that the real capacity was closer to 8GB. Any ideas how they might have done it and/or can I provide any info to you to help you figure this out?

[AltraMayor]

Hi @mikaey, could you open a new issue for your case? Besides the information you already posted above, add the output of the F3 commands you've run, any related information, and insights you may have had during the tests and analysis.