Would
@Cindori or anybody else with knowledge of OS X's inner workings with HFS+ and TRIM please clear up the confusion with Samsung 8xx series and some Crucial SSDs?
From what I understand, the implementation of NCQ (queued) TRIM commands is buggy on some of these SSDs and may result in blocks being zeroed when it's occupied by a file, resulting in data corruption.
Some drives apparently never had the problem fixed (Crucial M500), som have had firmware updates that say they fixed the problem, but apparently didn't or did, depending on who you ask (M550).
With Samsung drives, some people are saying the bug isn't related to TRIM at all, which is strange.
Windows apparently doesn't issue queued TRIM commands and Linux has worked around it by blacklisting the affected drives so that only non-queued TRIM commands are issued.
Now, my question is, does this issue affect OS X's handling of TRIM? The Linux pages describing it suggest it affects both Linux and OS X, but some people point out that OS X's implementation of TRIM on HFS+ doesn't use queued TRIM commands either.
Which one of these is true?
If the queued TRIM bug does occur on OS X, the best solution would be for
@Cindori to release updated Disk Sensei and TRIM Enabler with custom signed drivers as mentioned previously that do the same sort of queued TRIM blacklisting on affected drives as Linux is doing.
Thanks!
For the record, I have Crucial M500 and MX100 drives on various Macs and I found out about this issue around the same time people noticed the new trimforce command in El Capitan and have since turned off TRIM on all of them until I can be reasonably sure it's safe.