Re: [PATCH net-next 1/2] vsock: Linger on unsent data

[Michal Luczaj]

On 4/11/25 12:32, Stefano Garzarella wrote:
> On Thu, Apr 10, 2025 at 12:51:48PM +0200, Paolo Abeni wrote:
>> On 4/7/25 8:41 PM, Michal Luczaj wrote:
>>> Change the behaviour of a lingering close(): instead of waiting for all
>>> data to be consumed, block until data is considered sent, i.e. until worker
>>> picks the packets and decrements virtio_vsock_sock::bytes_unsent down to 0.
>>
>> I think it should be better to expand the commit message explaining the
>> rationale.

Sure, will do.

>>> Do linger on shutdown() just as well.
>>
>> Why? Generally speaking shutdown() is not supposed to block. I think you
>> should omit this part.
> 
> I thought the same, but discussing with Michal we discovered this on
> socket(7) man page:
> 
>    SO_LINGER
>           Sets or gets the SO_LINGER option.  The argument is a
>           linger structure.
> 
>               struct linger {
>                   int l_onoff;    /* linger active */
>                   int l_linger;   /* how many seconds to linger for */
>               };
> 
>           When enabled, a close(2) or shutdown(2) will not return
>           until all queued messages for the socket have been
>           successfully sent or the linger timeout has been reached.
>           Otherwise, the call returns immediately and the closing is
>           done in the background.  When the socket is closed as part
>           of exit(2), it always lingers in the background.
> 
> In AF_VSOCK we supported SO_LINGER only on close(), but it seems that 
> shutdown must also do it from the manpage.

Even though shutdown() lingering isn't universally implemented :/

If I'm reading the code correctly, TCP lingers only on close(). So,
following the man page on the one hand, mimicking TCP on the other?