* [LSF/MM TOPIC] Atomic Writes
@ 2019-02-11 17:45 Bart Van Assche
0 siblings, 0 replies; 1+ messages in thread
From: Bart Van Assche @ 2019-02-11 17:45 UTC (permalink / raw)
To: lsf-pc; +Cc: linux-block, linux-fsdevel
Atomic writes are writes that either succeed in their entirety or that are
not executed if a power failure occurs. It is well known that using atomic
writes can improve database and filesystem performance significantly +AFs-1,2+AF0.
Although the NVMe and SCSI standards support atomic writes, neither the
block layer nor filesystems offer a standardized interface for submitting
atomic writes. Hence the proposal to add block device and filesystem
independent interfaces for atomic writes.
Block Layer Proposal
+ACo Block drivers (NVMe, SCSI, ...) that support atomic writes set a queue
flag that makes it clear to the block layer core that support for atomic
writes is present.
+ACo Atomic writes are submitted from kernel context by marking individual
requests as atomic. One possible approach is to introduce a new bio flag.
Another possible approach is to introduce a new request type, e.g.
+ACo Introduce new limits for atomic writes such that it is guaranteed that
atomic writes will respect the device atomic write alignment and size
restrictions. We will probably need limits that correspond to the NAWUN,
NAWUNPF, NABSN, NABO and NABSPF parameters from the NVMe Identify
+ACo Kernel code that submits atomic writes is responsible for ensuring that
the write request size does not exceed the maximum size advertised by the
request queue. Fail atomic writes that are too large, not aligned or do
not satisfy the atomic write limits in some other way.
+ACo Add support in blk+AF8-stack+AF8-limits() for the atomic write limits.
+ACo Allow merging of regular writes with other regular writes. Allow merging
of atomic writes with other atomic writes. Do not allow merging of regular
writes with atomic writes. Respect the device limits when merging atomic
+ACo Continue allowing splitting of regular write requests but do not allow
splitting of atomic writes.
+ACo Make it possible to submit atomic writes from user space. One possible
approach is to add an O+AF8-ATOMIC flag to the open() system call.
+ACo Applications that want to submit both atomic and non-atomic writes must
open the block device twice - once with and once without the O+AF8-ATOMIC
+ACo Another possible approach is to add a new flag to the flags arguments of
the pwritev2() system call and the asynchronous I/O iocb structure.
+ACo Make it possible to submit atomic writes from user space. Just like for
block devices, one possible approach is to add an O+AF8-ATOMIC flag to the
open() system call. Another possible approach is to add a new flag to the
flags arguments of the pwritev2() system call and the asynchronous I/O iocb
structure. Note: Chris Mason had already proposed in 2013 to introduce the
O+AF8-ATOMIC flag for filesystems +AFs-3+AF0.
+ACo Filesystems may but do not have to submit atomic writes to the block layer
to implement O+AF8-ATOMIC. Using a traditional transaction mechanism to
implement O+AF8-ATOMIC is also fine but will result in write amplification.
+ACo Introduce a standardized interface for querying the filesystem atomic write
limits, e.g. by adding attributes under /sys/fs/.
+AFs-1+AF0 Ouyang, Xiangyong+ADs Nellans, David+ADs Wipfel, Robert+ADs Flynn, David+ADs Panda,
Dhabaleswar K. (February 2011). +ACI-Beyond block I/O: Rethinking traditional
storage primitives+ACI. 2011 IEEE 17th International Symposium on High
Performance Computer Architecture: 301+IBM-311
+AFs-2+AF0 MariaDB Knowledgebase, Atomic Write Support
+AFs-3+AF0 Chris Mason, Support for atomic IOs, fsdevel mailing list, November 2013
+AFs-4+AF0 Jonathan Corbet, Atomic I/O Operations, March 2013
+AFs-5+AF0 Jonathan Corbet, Support for atomic block I/O operations, November 2013
^ permalink raw reply [flat|nested] 1+ messages in thread
only message in thread, back to index
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-02-11 17:45 [LSF/MM TOPIC] Atomic Writes Bart Van Assche
Linux-Fsdevel Archive on lore.kernel.org
Archives are clonable:
git clone --mirror https://lore.kernel.org/linux-fsdevel/0 linux-fsdevel/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 linux-fsdevel linux-fsdevel/ https://lore.kernel.org/linux-fsdevel \
Newsgroup available over NNTP:
AGPL code for this site: git clone https://public-inbox.org/ public-inbox