On 5/28/20 11:12 AM, Sedat Dilek wrote:
> On Thu, May 28, 2020 at 7:06 PM Jens Axboe <axboe@xxxxxxxxx> wrote:
>>
>> On 5/28/20 11:02 AM, Sedat Dilek wrote:
>>> On Tue, May 26, 2020 at 10:59 PM Jens Axboe <axboe@xxxxxxxxx> wrote:
>>>>
>>>> We technically support this already through io_uring, but it's
>>>> implemented with a thread backend to support cases where we would
>>>> block. This isn't ideal.
>>>>
>>>> After a few prep patches, the core of this patchset is adding support
>>>> for async callbacks on page unlock. With this primitive, we can simply
>>>> retry the IO operation. With io_uring, this works a lot like poll based
>>>> retry for files that support it. If a page is currently locked and
>>>> needed, -EIOCBQUEUED is returned with a callback armed. The callers
>>>> callback is responsible for restarting the operation.
>>>>
>>>> With this callback primitive, we can add support for
>>>> generic_file_buffered_read(), which is what most file systems end up
>>>> using for buffered reads. XFS/ext4/btrfs/bdev is wired up, but probably
>>>> trivial to add more.
>>>>
>>>> The file flags support for this by setting FMODE_BUF_RASYNC, similar
>>>> to what we do for FMODE_NOWAIT. Open to suggestions here if this is
>>>> the preferred method or not.
>>>>
>>>> In terms of results, I wrote a small test app that randomly reads 4G
>>>> of data in 4K chunks from a file hosted by ext4. The app uses a queue
>>>> depth of 32. If you want to test yourself, you can just use buffered=1
>>>> with ioengine=io_uring with fio. No application changes are needed to
>>>> use the more optimized buffered async read.
>>>>
>>>> preadv for comparison:
>>>> real 1m13.821s
>>>> user 0m0.558s
>>>> sys 0m11.125s
>>>> CPU ~13%
>>>>
>>>> Mainline:
>>>> real 0m12.054s
>>>> user 0m0.111s
>>>> sys 0m5.659s
>>>> CPU ~32% + ~50% == ~82%
>>>>
>>>> This patchset:
>>>> real 0m9.283s
>>>> user 0m0.147s
>>>> sys 0m4.619s
>>>> CPU ~52%
>>>>
>>>> The CPU numbers are just a rough estimate. For the mainline io_uring
>>>> run, this includes the app itself and all the threads doing IO on its
>>>> behalf (32% for the app, ~1.6% per worker and 32 of them). Context
>>>> switch rate is much smaller with the patchset, since we only have the
>>>> one task performing IO.
>>>>
>>>> Also ran a simple fio based test case, varying the queue depth from 1
>>>> to 16, doubling every time:
>>>>
>>>> [buf-test]
>>>> filename=/data/file
>>>> direct=0
>>>> ioengine=io_uring
>>>> norandommap
>>>> rw=randread
>>>> bs=4k
>>>> iodepth=${QD}
>>>> randseed=89
>>>> runtime=10s
>>>>
>>>> QD/Test Patchset IOPS Mainline IOPS
>>>> 1 9046 8294
>>>> 2 19.8k 18.9k
>>>> 4 39.2k 28.5k
>>>> 8 64.4k 31.4k
>>>> 16 65.7k 37.8k
>>>>
>>>> Outside of my usual environment, so this is just running on a virtualized
>>>> NVMe device in qemu, using ext4 as the file system. NVMe isn't very
>>>> efficient virtualized, so we run out of steam at ~65K which is why we
>>>> flatline on the patched side (nvme_submit_cmd() eats ~75% of the test app
>>>> CPU). Before that happens, it's a linear increase. Not shown is context
>>>> switch rate, which is massively lower with the new code. The old thread
>>>> offload adds a blocking thread per pending IO, so context rate quickly
>>>> goes through the roof.
>>>>
>>>> The goal here is efficiency. Async thread offload adds latency, and
>>>> it also adds noticable overhead on items such as adding pages to the
>>>> page cache. By allowing proper async buffered read support, we don't
>>>> have X threads hammering on the same inode page cache, we have just
>>>> the single app actually doing IO.
>>>>
>>>> Been beating on this and it's solid for me, and I'm now pretty happy
>>>> with how it all turned out. Not aware of any missing bits/pieces or
>>>> code cleanups that need doing.
>>>>
>>>> Series can also be found here:
>>>>
>>>> https://git.kernel.dk/cgit/linux-block/log/?h=async-buffered.5
>>>>
>>>> or pull from:
>>>>
>>>> git://git.kernel.dk/linux-block async-buffered.5
>>>>
>>>
>>> Hi Jens,
>>>
>>> I have pulled linux-block.git#async-buffered.5 on top of Linux v5.7-rc7.
>>>
>>> From first feelings:
>>> The booting into the system (until sddm display-login-manager) took a
>>> bit longer.
>>> The same after login and booting into KDE/Plasma.
>>
>> There is no difference for "regular" use cases, only io_uring with
>> buffered reads will behave differently. So I don't think you have longer
>> boot times due to this.
>>
>>> I am building/linking with LLVM/Clang/LLD v10.0.1-rc1 on Debian/testing AMD64.
>>>
>>> Here I have an internal HDD (SATA) and my Debian-system is on an
>>> external HDD connected via USB-3.0.
>>> Primarily, I use Ext4-FS.
>>>
>>> As said above is the "emotional" side, but I need some technical instructions.
>>>
>>> How can I see Async Buffer Reads is active on a Ext4-FS-formatted partition?
>>
>> You can't see that. It'll always be available on ext4 with this series,
>> and you can watch io_uring instances to see if anyone is using it.
>>
>
> Thanks for answering my questions.
>
> How can I "watch io_uring instances"?
You can enable io_uring tracing:
# echo 1 > /sys/kernel/debug/tracing/events/io_uring/io_uring_create/enable
# tail /sys/kernel/debug/tracing/trace
and see if you get any events for setup. Generally you can also look for
the existence of io_wq_manager processes, these will exist for an
io_uring instance.
> FIO?
> Debian has fio version 3.19-2 in its apt repositories.
> Version OK?
Yeah that should work.
--
Jens Axboe
