Re: [RFC PATCH v2 0/8] fallocate: introduce FALLOC_FL_WRITE_ZEROES flag
From: Zhang Yi
Date: Wed Jan 15 2025 - 22:05:01 EST
On 2025/1/16 5:07, Chaitanya Kulkarni wrote:
> On 1/15/25 03:46, Zhang Yi wrote:
>> Currently, we can use the fallocate command to quickly create a
>> pre-allocated file. However, on most filesystems, such as ext4 and XFS,
>> fallocate create pre-allocation blocks in an unwritten state, and the
>> FALLOC_FL_ZERO_RANGE flag also behaves similarly. The extent state must
>> be converted to a written state when the user writes data into this
>> range later, which can trigger numerous metadata changes and consequent
>> journal I/O. This may leads to significant write amplification and
>> performance degradation in synchronous write mode. Therefore, we need a
>> method to create a pre-allocated file with written extents that can be
>> used for pure overwriting. At the monent, the only method available is
>> to create an empty file and write zero data into it (for example, using
>> 'dd' with a large block size). However, this method is slow and consumes
>> a considerable amount of disk bandwidth, we must pre-allocate files in
>> advance but cannot add pre-allocated files while user business services
>> are running.
>
> it will be very useful if we can get some blktests for scsi/nvme/dm.
> Please note that this not a blocker to get this path series to be merged,
> but this will help everyone including regular tests runs we do to ensure
> the stability of new interface.
Hello, Chaitanya,
Thanks for your feedback! Yeah, the proposal for this series is still under
discussion, I will add counterpart tests to both blktests and fstests once
the solution is determined.
>
> if you do please CC and Shinichiro (added to CC list) to we can help those
> tests review and potentially also can provide tested by tag tht can help
> this work to move forward.
>
Sure, this will be very helpful.
Thanks,
Yi.