[lkml]   [2011]   [Mar]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [Regression] Please revert a91a2785b20
>>>>> "Thomas" == Thomas Gleixner <> writes:


Thomas> But the changelog does not give the courtesy of explaining these
Thomas> changes. Also there is no fcking reason why the kernel cannot
Thomas> deal with the missing integrity capabilities of a drive just by
Thomas> emitting a warning msg and dealing gracefully with the outcome.

My mistake. I was made aware of it earlier today and I'm working on a
patch. Surprised we didn't see any reports of this in -next. It's been
in there for a while.

Thomas> All my RAID setups have been working perfectly fine until now,
Thomas> so what's the rationale to break this?

People were complaining about excessive mempool usage with the block
integrity bits enabled (thanks to MD and DM allocating a bioset per
device to prevent deadlocks).

Making allocation conditional meant we had to deal with memory
allocation errors in the setup path. I tested various combinations of on
and off but apparently not all off. Sorry about that.

Martin K. Petersen Oracle Linux Engineering

 \ /
  Last update: 2011-03-29 00:49    [W:0.106 / U:2.416 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site