Re: [PULL] bcache: multiple updates

From: Jens Axboe
Date: Sun Oct 30 2016 - 11:01:18 EST


On 10/30/2016 08:00 AM, Kent Overstreet wrote:
On Sat, Oct 29, 2016 at 06:32:38PM -0700, Eric Wheeler wrote:
On Thu, 27 Oct 2016, Jens Axboe wrote:

On 10/27/2016 05:27 PM, Eric Wheeler wrote:
Hi Jens,

Please pull this v4.9-rc2 based series of bcache updates for v4.9-rc3:
(You may disregard the previous -rc1-based request.)

git pull https://bitbucket.org/ewheelerinc/linux.git
v4.9-rc2-bcache-updates

Thank you!

--
Eric Wheeler

]# git log --oneline v4.9-rc2..HEAD
bd532a6 bcache: partition support: add 16 minors per bcacheN device
3312845 bcache: Make gc wakeup sane, remove set_task_state()
6bb7f1e bcache: update bio->bi_opf bypass/writeback REQ_ flag hints
3d58a09 bcache: documentation for ioprio cache hinting
2e8884b bcache: introduce per-process ioprio-based bypass/writeback hints

How many of these are applicable to 4.9-rc3? I took a quick look at
them, and some of them look like they should go into the 4.10 branch
instead. We're after the merge window, so only strict fixes. Cleanups
and no features, no go.

3312845 might need to be in 4.9. Not sure, kent asked me to pick it up in
my next pull request. 4.10 is fine for the rest.

Kent, Davidlohr, does 3312845 need to land in 4.9 for some reason?

No, that can wait until 4.10.

Great, that's what it looked like.

Eric, three things. The first is that you based this branch as if it was
going into 4.9, which means if I pull it into my 4.10 branch, I get a
ton of extra stuff that has been added to master since. Not a huge
problem, I can just generate the patches and apply them.

Secondly, you are depending on REQ_THROTTLED, which in the 4.10 series,
has been modified:

commit 8d2bbd4c8236e9e38e6b36ac9e2c54fdcfe5b335
Author: Christoph Hellwig <hch@xxxxxx>
Date: Thu Oct 20 15:12:12 2016 +0200

block: replace REQ_THROTTLED with a bio flag

And lastly, would you mind using the regular git request-pull to
generate your pull requests? It's what everybody else is using, and
honestly I think it's a lot more readable than the oneline thing you are
doing.

Please resend the series against for-4.10/block and we can get it
applied, thanks.

--
Jens Axboe