Linux 5.1 Hit By A Data Loss Bug Due To Overly Aggressive FSTRIM

[Guest]

Ouch, my desktop satisfies all of those conditions. I haven't noticed any problems so far, but I did spend a few days on 5.1.3 .

[Melcar]

Sounds to me more like yet another Samsung SSD firmware bug.

I noticed there is a new firmaware being pushed for my drive (860 EVO). Just updated and moved to 5.1.5 kernel. I don't use any encryption however.

[shmerl]

I noticed there is a new firmaware being pushed for my drive (860 EVO). Just updated and moved to 5.1.5 kernel. I don't use any encryption however.

How do you update firmware for Samsung drives on Linux?

[shmerl]

5.1.5 seems to include the fix:
* https://www.redhat.com/archives/dm-d.../msg00102.html
* https://cdn.kernel.org/pub/linux/ker...hangeLog-5.1.5

commit 871e122d55e8d1cd7c0d5dec9bdba1fe45406196
Author: Michael Lass <[email protected]>
Date:   Tue May 21 21:58:07 2019 +0200

    dm: make sure to obey max_io_len_target_boundary
    
    commit 51b86f9a8d1c4bb4e3862ee4b4c5f46072f7520d upstream.
    
    Commit 61697a6abd24 ("dm: eliminate 'split_discard_bios' flag from DM
    target interface") incorrectly removed code from
    __send_changing_extent_only() that is required to impose a per-target IO
    boundary on IO that exceeds max_io_len_target_boundary().  Otherwise
    "special" IO (e.g. DISCARD, WRITE SAME, WRITE ZEROES) can write beyond
    where allowed.
    
    Fix this by restoring the max_io_len_target_boundary() limit in
    __send_changing_extent_only()
    
    Fixes: 61697a6abd24 ("dm: eliminate 'split_discard_bios' flag from DM target interface")
    Cc: [email protected] # 5.1+
    Signed-off-by: Michael Lass <[email protected]>
    Signed-off-by: Mike Snitzer <[email protected]>
    Signed-off-by: Greg Kroah-Hartman <[email protected]>

[Melcar]

How do you update firmware for Samsung drives on Linux?

I cheat. I still have a small SSD with a W10 install. However, I think you can make your own bootable USB key with the firmware and flash manually.

[shmerl]

I cheat. I still have a small SSD with a W10 install. However, I think you can make your own bootable USB key with the firmware and flash manually.

How did you notice the new firmware being pushed though, you have some monitor for that?

[Melcar]

How did you notice the new firmware being pushed though, you have some monitor for that?

I have Samsung Magician loaded on my Windows install, which notifies you of firmware updates. I boot into my W10 SSD every two weeks or so, mostly to update hardware that is not that Linux friendly software wise. Yesterday I noticed the notification in the Magician software.

[DanglingPointer]

Kernel 5.1.5 fixes the problem. For those tracking mainline upgrade to 5.1.5 ASAP if you have the conditions which may cause the data loss.
Available with the Ubuntu Mainline Kernels as well if you're using those.

Michael Lass did the fix in the kernel...

Detailed Summary : https://github.com/zen-kernel/zen-ke...c5f46072f7520d...

Summary line...
"Michael Lass (1):
dm: make sure to obey max_io_len_target_boundary"

[Wojcian]

Is it the third data corruption bug for the past several months? The Linux kernel is so darn stable.

If you were little smarter pathetic troll, you will realize to use stable distributions when you want stability. Show me winblos.kernel.org or shut up moron.

[DrYak]

@Schmerl it happens that their updater USB boot key images actually run on linux.

So you either update on the linux (...that happens to be on the update boot key).

Or you can extract the software and run it from your normal linux installation, if you are updating a non-mounted drive