James, the bug is that with iscsi running, if ipv6 is provisioned on system, the kernel would oops on "scheduling while atomic". So it could be hit easily if ipv6 is used.
The changeset does look big, but it really just split a list original protected by mutex lock into two list, The 2nd list uses rcu lock. Karen ________________________________________ From: James Bottomley [[email protected]] Sent: Wednesday, October 01, 2014 4:36 AM To: Anish Bhatt Cc: Christoph Hellwig; [email protected]; Karen Xie; Manoj Malviya Subject: Re: [PATCH scsi] cxgb4i : Avoid holding mutex in interrupt context On Wed, 2014-10-01 at 08:08 +0000, Anish Bhatt wrote: > James, > This is bugfix and prevents a "scheduling while atomic" bug that can be > hit when using an ipv6 address with cxgb4i, and can possibly be seen without > Chelsio hardware. Could this be cherry picked for 3.17 please ? Can you describe the bug it's fixing (what are the consequences) and how often is it seen? If it's not something everyone hits, it may be better to go via stable, because the code change is quite large, so validating that nothing else breaks with it at this stage (a week away from release) is quite important. > Christoph, > My bad, this is a bugfix for code that was added for 3.17, and I assumed > the > bugfixes would automatically be picked for the next rc. If James doesn't pick > this up, I'll send it to stable when appropriate. You said in your change log: Applies on top of core-for-3.18 So the natural assumption would be this is for the merge window queue. James -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
