On 14-01-2016 21:50, Gustavo De Nardin (spuk) wrote:

According to (updates on) that same blog post, that bug was found not to be specific to Samsung or any other drives, but in the kernel handling of TRIM on software RAID.

Quoting: """... UPDATE July 17: We have just finished a conference call with Samsung considering the failure analysis of this issue. Samsung engineering team has been able to successfully reproduce the issue with our latest provided binary.

Samsung had a concrete conclusion that the issue is not related to Samsung SSD or Algolia software but is related to the Linux kernel. ..."""

The patch you linked is from May, the last email in the thread is from 22 July 2015. This seems to be the real fix (commited also on 22 July): https://github.com/torvalds/linux/commit/f3f5da624e0a891c34d8cd513c57f1d9b0c...

Here is the kernel thread start (also linked from that post), with the discussion and patches: http://www.spinics.net/lists/raid/msg49440.html

So this is in the kernel since July 22, Arch kernel is newer.

Now, thanks for bringing this up, the last I knew was also about the blacklist mitigation. Feel safer about my own EVO 850 now. :]

t'

On 14 January 2016 at 17:19, Andrew Martin andrew.s.martin@gmail.com wrote:

...

Hello,

I installed Arch on an EVO 850 120GB SSD. I read about the linux EVO 800-series TRIM bug a few months ago: https://blog.algolia.com/when-solid-state-drives-are-not-that-solid/

However it appeared to have been fixed in the 4.1 kernel release:

https://github.com/torvalds/linux/commit/9a9324d3969678d44b330e1230ad2c8ae67...

However a few days ago, my root partition suddenly became corrupted and I was forced to reinstall. While I cannot be 100% certain it was this bug, it seems likely. Has it been confirmed that this bug has been fixed in Arch? If so, at which kernel version?

Thanks,

Andrew

There were two different bugs that came up at different times, one is drives that don't work properly with queued trim and another was the bug with raid.

The fix for the first problem is still to blacklist queued trim for the affected drives, it hasn't changed. Both problems should not affect either the "normal" or lts kernels provided by arch.

Thanks for the clarification on this!

Andrew

On 14 January 2016 at 19:43, Gustavo De Nardin (spuk) gustavodn@gmail.com wrote:

On 14 January 2016 at 21:10, Mauro Santos registo.mailling@gmail.com wrote:

...

There were two different bugs that came up at different times, one is drives that don't work properly with queued trim and another was the bug with raid.

Alright, thanks for clearing that up. I see there's still a big list of drives needing workarounds. (

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers... )

...

The fix for the first problem is still to blacklist queued trim for the affected drives, it hasn't changed. Both problems should not affect either the "normal" or lts kernels provided by arch.

-- (nil)