Re: [PATCH 3.0.y] md: protect against crash upon fsync on ro array

2013-03-14 Thread Sebastian Riemer
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Exactly! Greg, please queue it that way! It only differs in requiring the return value. Cheers, Sebastian On 14.03.2013 01:54, Ben Hutchings wrote: Actually, Sebastian already sent an equivalent patch in 510ac77a.2000...@profitbricks.com. And

[PATCH 3.0.y] md: protect against crash upon fsync on ro array

2013-03-13 Thread Ben Hutchings
From: Sebastian Riemer sebastian.rie...@profitbricks.com commit bbfa57c0f2243a7c31fd248d22e9861a2802cad5 upstream. If an fsync occurs on a read-only array, we need to send a completion for the IO and may not increment the active IO count. Otherwise, we hit a bug trace and can't stop the MD array

Re: [PATCH 3.0.y] md: protect against crash upon fsync on ro array

2013-03-13 Thread Ben Hutchings
Actually, Sebastian already sent an equivalent patch in 510ac77a.2000...@profitbricks.com. And he did test it. Here's his patch with the upstream reference added. Ben. -- Ben Hutchings Humans are not rational beings; they are rationalising beings. From: Sebastian Riemer