Re: [RFC PATCH v9 19/19] af_vsock: serialize writes to shared socket

[Stefano Garzarella]

On Thu, May 13, 2021 at 05:48:19PM +0300, Arseny Krasnov wrote:
> On 13.05.2021 17:46, Stefano Garzarella wrote:
> > On Thu, May 13, 2021 at 04:01:50PM +0200, Stefano Garzarella wrote:
> > > On Sat, May 08, 2021 at 07:37:35PM +0300, Arseny Krasnov wrote:
> > > > This add logic, that serializes write access to single socket
> > > > by multiple threads. It is implemented be adding field with TID
> > > > of current writer. When writer tries to send something, it checks
> > > > that field is -1(free), else it sleep in the same way as waiting
> > > > for free space at peers' side.
> > > >
> > > > Signed-off-by: Arseny Krasnov <arseny.krasnov@xxxxxxxxxxxxx>
> > > > ---
> > > > include/net/af_vsock.h   |  1 +
> > > > net/vmw_vsock/af_vsock.c | 10 +++++++++-
> > > > 2 files changed, 10 insertions(+), 1 deletion(-)
> > > I think you forgot to move this patch at the beginning of the series.
> > > It's important because in this way we can backport to stable branches
> > > easily.
> > >
> > > About the implementation, can't we just add a mutex that we hold until
> > > we have sent all the payload?
> > Re-thinking, I guess we can't because we have the timeout to deal
> > with...
> Yes, i forgot about why i've implemented it using 'tid_owner' :)

It is not clear to me if we need to do this also for stream.

I think will be better to follow af_inet/af_unix, but I need to check 
their implementation.

Thanks,
Stefano