[GIT PULL v2] Block fixes for 5.7-rc5

[GIT PULL v2] Block fixes for 5.7-rc5

[Jens Axboe]

Hi Linus,

Let's try this again... BFQ was missing a header, I fixed that up. To
avoid introducing a bisection breakage point, I rebased the branch.

A few fixes that should go into this series:

- Small series fixing a use-after-free of bdi name (Christoph,Yufen)

- NVMe fix for a regression with the smaller CQ update (Alexey)

- NVMe fix for a hang at namespace scanning error recovery (Sagi)

- Fix race with blk-iocost iocg->abs_vdebt updates (Tejun)

Please pull!


  git://git.kernel.dk/linux-block.git tags/block-5.7-2020-05-09


----------------------------------------------------------------
Alexey Dobriyan (1):
      nvme-pci: fix "slimmer CQ head update"

Christoph Hellwig (3):
      vboxsf: don't use the source name in the bdi name
      bdi: move bdi_dev_name out of line
      bdi: add a ->dev_name field to struct backing_dev_info

Sagi Grimberg (1):
      nvme: fix possible hang when ns scanning fails during error recovery

Tejun Heo (1):
      iocost: protect iocg->abs_vdebt with iocg->waitq.lock

Yufen Yu (1):
      bdi: use bdi_dev_name() to get device name

 block/bfq-iosched.c              |   6 +-
 block/blk-cgroup.c               |   2 +-
 block/blk-iocost.c               | 117 ++++++++++++++++++++++++---------------
 drivers/nvme/host/core.c         |   2 +-
 drivers/nvme/host/pci.c          |   6 +-
 fs/ceph/debugfs.c                |   2 +-
 fs/vboxsf/super.c                |   2 +-
 include/linux/backing-dev-defs.h |   1 +
 include/linux/backing-dev.h      |   9 +--
 include/trace/events/wbt.h       |   8 +--
 mm/backing-dev.c                 |  13 ++++-
 tools/cgroup/iocost_monitor.py   |   7 ++-
 12 files changed, 107 insertions(+), 68 deletions(-)

-- 
Jens Axboe

Re: [GIT PULL v2] Block fixes for 5.7-rc5

[Linus Torvalds]

On Sat, May 9, 2020 at 6:33 PM Jens Axboe <axboe@kernel.dk> wrote:
>
> Let's try this again... BFQ was missing a header, I fixed that up.

The fix looked trivial to me. That wasn't what worries me.

Why did you send me something that was clearly NOT TESTED AT ALL.

If it hadn't even gotten build-testing, what _did_ it get?

The fact that it now builds doesn't make me much happier.

Why should I believe that this clearly totally untested pull request
is now any good?

Why should I believe that your _future_ pull requests are any good,
when they clearly have absolutely _zero_ testing at all?

Jens, in case you didn't catch on, this is a BIG PROBLEM.

Sending me completely untested crap is a bigger deal than "let's just
polish the crap until it at least compiles".

What have you done to make sure this doesn't happen again?

             Linus

Re: [GIT PULL v2] Block fixes for 5.7-rc5

[pr-tracker-bot]

The pull request you sent on Sat, 9 May 2020 19:33:19 -0600:

> git://git.kernel.dk/linux-block.git tags/block-5.7-2020-05-09

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/0a85ed6e7fce8075bb3090f8eac05ca1000f5969

Thank you!

-- 
Deet-doot-dot, I am a bot.
https://korg.wiki.kernel.org/userdoc/prtracker

Re: [GIT PULL v2] Block fixes for 5.7-rc5

[Jens Axboe]

On 5/9/20 9:36 PM, Linus Torvalds wrote:
> On Sat, May 9, 2020 at 6:33 PM Jens Axboe <axboe@kernel.dk> wrote:
>>
>> Let's try this again... BFQ was missing a header, I fixed that up.
> 
> The fix looked trivial to me. That wasn't what worries me.
> 
> Why did you send me something that was clearly NOT TESTED AT ALL.
> 
> If it hadn't even gotten build-testing, what _did_ it get?
> 
> The fact that it now builds doesn't make me much happier.
> 
> Why should I believe that this clearly totally untested pull request
> is now any good?
> 
> Why should I believe that your _future_ pull requests are any good,
> when they clearly have absolutely _zero_ testing at all?
> 
> Jens, in case you didn't catch on, this is a BIG PROBLEM.
> 
> Sending me completely untested crap is a bigger deal than "let's just
> polish the crap until it at least compiles".
> 
> What have you done to make sure this doesn't happen again?

I was out all day yesterday, so didn't get a chance to follow up on this
one. I did notice that you pulled it, thanks for that.

My test box is currently out of commission. While that doesn't mean it
gets no testing, it does mean I'm using my laptop and qemu configs.
Neither one of those have BFQ set, which is why it was missed. While
that isn't ideal, it's not a core kernel file, or it would have never
been missed. If you look at the patches, it's also not like they weren't
reviewed, in fact they were quite heavily reviewed.

So, yes, I agree this was unfortunate, and I hate having sent something
that didn't compile for a pretty common option. That sucks.

-- 
Jens Axboe