Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

To: David Howells <dhowells@xxxxxxxxxx>
Subject: Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
From: Amir Goldstein <amir73il@xxxxxxxxx>
Date: Fri, 31 Jan 2020 10:56:19 +0200
Cc: "hch@xxxxxx" <hch@xxxxxx>, Dave Chinner <david@xxxxxxxxxxxxx>, Al Viro <viro@xxxxxxxxxxxxxxxxxx>, Omar Sandoval <osandov@xxxxxxxxxxx>, Trond Myklebust <trondmy@xxxxxxxxxxxxxxx>, "lsf-pc@xxxxxxxxxxxxxxxxxxxxxxxxxx" <lsf-pc@xxxxxxxxxxxxxxxxxxxxxxxxxx>, "miklos@xxxxxxxxxx" <miklos@xxxxxxxxxx>, "linux-fsdevel@xxxxxxxxxxxxxxx" <linux-fsdevel@xxxxxxxxxxxxxxx>, "Darrick J. Wong" <darrick.wong@xxxxxxxxxx>
In-reply-to: <3773161.1580457868@warthog.procyon.org.uk>

On Fri, Jan 31, 2020 at 10:04 AM David Howells <dhowells@xxxxxxxxxx> wrote:
>
> hch@xxxxxx <hch@xxxxxx> wrote:
>
> > > I'm using direct I/O, so I'm assuming I don't need to fsync().
> >
> > Direct I/O of course requires fsync.  What makes you think different?
>
> I guess that's fair.  Even if the data makes it directly to storage, that's
> not a guarantee that the metadata does.
>

It's not only that. DIO gets data to storage *layer*, but it doesn't
tell storage *layer* to make the data durable (usually with FLUSH/FUA).

Thanks,
Amir.

References:
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Al Viro
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Omar Sandoval
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Al Viro
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Omar Sandoval
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Al Viro
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Omar Sandoval
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Amir Goldstein
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Omar Sandoval
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: Al Viro
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: David Howells
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: hch@xxxxxx
- Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
  - From: David Howells

Prev by Date: Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
Next by Date: Re: [PATCH 12/12] iomap: Convert from readpages to readahead
Previous by thread: Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
Next by thread: Re: [LSF/MM/BPF TOPIC] Allowing linkat() to replace the destination
Index(es):
- Date
- Thread

[Index of Archives] [Linux Ext4 Filesystem] [Union Filesystem] [Filesystem Testing] [Ceph Users] [Ecryptfs] [AutoFS] [Kernel Newbies] [Share Photos] [Security] [Netfilter] [Bugtraq] [Yosemite News] [MIPS Linux] [ARM Linux] [Linux Security] [Linux Cachefs] [Reiser Filesystem] [Linux RAID] [Samba] [Device Mapper] [CEPH Development]