[Bug 200871] F2FS experiences data loss (entry is completely lost) when an I/O failure occurs.

[bugzilla-daemon@bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=200871

--- Comment #17 from Chao Yu (chao@kernel.org) ---
(In reply to Stathis Maneas from comment #16)
> I am sorry for the delay! I can definitely help by adding the aforementioned
> statements!
> 
> After adding them and observing the output of dmesg, I can observe that only
> the second statement gets actually invoked:
> [ 1026.438986] f2fs_wait_on_node_pages_writeback, eio:0
> 
> Therefore, this indicates that the following condition is not true under the
> specified workload:
> if (unlikely(bio->bi_status)) { ... }
> 
> I tried moving the first print statement at the beginning of the following
> for loop,
> bio_for_each_segment_all(bvec, bio, i) { ... }
> 
> but still, I am not observing any errors for the segment (block) of interest:
> [   76.603891] f2fs_write_end_io, ino:1, idx:4, eio:0, type:7
> [   76.603893] f2fs_write_end_io, is_dummy:0 bio_status:0

Sorry for the delay.

Oh, so I guess that you trigger the error injection but didn't set
bio->bi_status to nonzero value, result in fs can not be aware of that.

I think we'd better to double check all your test steps and injection codes.

> 
> Is there something else we can try?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.