I agree.
The patch "dm: ensure bio submission follows a depth-first tree walk" doesn't really fix any bug (and it may introduce new bugs), so it isn't applicable to the stable kernel branches.
The kernel 4.15 and below have a "dm_offload" mechanism that avoids the possible deadlocks in dm, there is no reason why this should be converted to the "depth-first tree walk".
Mikulas
On Thu, 22 Mar 2018, Mike Snitzer wrote:
On Thu, Mar 22 2018 at 9:06am -0400, gregkh@linuxfoundation.org gregkh@linuxfoundation.org wrote:
This is a note to let you know that I've just added the patch titled
dm: ensure bio submission follows a depth-first tree walkto the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git%3Ba=su...
The filename of the patch is: dm-ensure-bio-submission-follows-a-depth-first-tree-walk.patch and it can be found in the queue-4.15 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree, please let stable@vger.kernel.org know about it.
The following applies to 4.14 stable@ too.
I think it very questionable to pull this into stable trees. How'd it even elevate to be considered for stable@?
But if you do, you definitely need this additional upstream commit: 8dd601fa8317 ("dm: correctly handle chained bios in dec_pending()")
Even with that I'm not excited about supporting some partial backport of these changes because there were a lot more related changes -- only taking a subset makes these stable@ kernels unicorns.. I deal with enough unicorn vendor kernels (but on my terms, with my control over what is "supportable"):
Anyway, other related commits are: 80cd17578310 dm crypt: remove BIOSET_NEED_RESCUER flag c110a4b6e603 dm io: remove BIOSET_NEED_RESCUER flag from bios bioset f31c21e4365c dm: remove unused 'num_write_bios' target interface 318716ddea08 dm: safely allocate multiple bioset bios 4a3f54d94d5c dm: remove BIOSET_NEED_RESCUER based dm_offload infrastructure 0776aa0e30aa dm: ensure bio-based DM's bioset and io_pool support targets' maximum IOs 3d7f45625a84 dm: fix __send_changing_extent_only() to send first bio and chain remainder
So NAK from me. If others have a compelling argument I'm open to considering.
Mike