Re: [PATCH v2 3/4] input: evdev: Indicate buffer overrun with SYN_DROPPED.

[Jeffrey Brown]

Hi Dmitry,

On Mon, Apr 4, 2011 at 2:33 PM, Dmitry Torokhov
<dmitry.torokhov@xxxxxxxxx> wrote:
> On Fri, Apr 01, 2011 at 11:54:18PM -0700, Jeff Brown wrote:
>> +      * When the client buffer is full, drain the buffer and enqueue a
>> +      * SYN_DROPPED event to let the client know that events were dropped
>> +      * and the last packet was incomplete.  We then consume all remaining
>> +      * events from the dropped packet until the next packet begins.
>
> I do not think we (kernel) should be doing this. Clients should take
> care and not allow overruns but if they happen the pain should be on the
> client to restore the context. It has to query the device to get its
> current state and it can drop the events till next sync as well.

That's fine.  We should document SYN_DROPPED and mention that the
client should wait until after the next SYN_REPORT to get back in
sync.

> Therefore I intend to apply the patch in the form below.

Looks good to me.  It's more like my original patch.

Jeff.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/