Re: [PATCH RFC] docs: document file-posix locking protocol

[Daniel P . Berrangé]

On Mon, Mar 22, 2021 at 09:27:38PM +0300, Vladimir Sementsov-Ogievskiy wrote:
> Let's document how we use file locks in file-posix driver, to allow
> external programs to "communicate" in this way with Qemu.
> 
> Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
> ---
> 
> Hi all!
> 
> We need to access disk images from non-Qemu code and coordinate with
> Qemu utilities which may use same image. So, we want to support Qemu
> file locking in the external code.
> 
> So, here is a patch to document how Qemu locking works, and make this
> thing "public".
> 
> This is an RFC, because I'm unsure how should we actually document
> different operations we have.
> 
> For example greaph-mod is a strange thing, I think we should get rid
> of it at all.. And at least, no sense in locking corresponding byte in a
> raw file.
> 
> The other thing is write-unchanged.. What it means when we consider a
> raw file opened in several processes? Probably we don't need it too..
> 
>  docs/system/qemu-block-drivers.rst.inc | 55 ++++++++++++++++++++++++++
>  1 file changed, 55 insertions(+)
> 
> diff --git a/docs/system/qemu-block-drivers.rst.inc 
> b/docs/system/qemu-block-drivers.rst.inc
> index b052a6d14e..3cd708b3dc 100644
> --- a/docs/system/qemu-block-drivers.rst.inc
> +++ b/docs/system/qemu-block-drivers.rst.inc
> @@ -952,3 +952,58 @@ on host and see if there are locks held by the QEMU 
> process on the image file.
>  More than one byte could be locked by the QEMU instance, each byte of which
>  reflects a particular permission that is acquired or protected by the running
>  block driver.
> +
> +Image locking protocol
> +~~~~~~~~~~~~~~~~~~~~~~
> +
> +QEMU holds rd locks and never rw locks. Instead, GETLK fcntl is used with 
> F_WRLCK
> +to handle permissions as described below.
> +QEMU process may rd-lock the following bytes of the image with corresponding
> +meaning:
> +
> +Permission bytes. If permission byte is rd-locked, it means that some process
> +uses corresponding permission on that file.
> +
> +Byte    Operation
> +100     read
> +          Lock holder can read
> +101     write
> +          Lock holder can write
> +102     write-unchanged
> +          Lock holder can write same data
> +103     resize
> +          Lock holder can resize the file
> +104     graph-mod
> +          Undefined. QEMU sometimes locks this byte, but external programs
> +          should not. QEMU will stop locking this byte in future
> +
> +Unshare bytes. If permission byte is rd-locked, it means that some process
> +does not allow the others use corresponding options on that file.
> +
> +Byte    Operation
> +200     read
> +          Lock holder don't allow read operation to other processes.
> +201     write
> +          Lock holder don't allow write operation to other processes.
> +202     write-unchanged
> +          Lock holder don't allow write-unchanged operation to other 
> processes.
> +203     resize
> +          Lock holder don't allow resizing the file by other processes.
> +204     graph-mod
> +          Undefined. QEMU sometimes locks this byte, but external programs
> +          should not. QEMU will stop locking this byte in future
> +
> +Handling the permissions works as follows: assume we want to open the file 
> to do
> +some operations and in the same time want to disallow some operation to other
> +processes. So, we want to lock some of the bytes described above. We operate 
> as
> +follows:
> +
> +1. rd-lock all needed bytes, both "permission" bytes and "unshare" bytes.
> +
> +2. For each "unshare" byte we rd-locked, do GETLK that "tries" to wr-lock
> +corresponding "permission" byte. So, we check is there any other process that
> +uses the permission we want to unshare. If it exists we fail.
> +
> +3. For each "permission" byte we rd-locked, do GETLK that "tries" to wr-lock
> +corresponding "unshare" byte. So, we check is there any other process that
> +unshares the permission we want to have. If it exists we fail.

I'd perhaps illustrate the few common scenarios - read-only, shared access,
read-write, exclusive and read-write share access, using C-psuedo-code, from
opening image, locking, doing operations, unlocking and closing image.

This would make it explicit that these locks need to be held open
for the duration of the i/o operations.

I'd also probably warn about the dangers of traditional fcntl locks  in
threaded programs, compared to the saner OFD locks.

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|