Re: [PATCH] hw/nvme: clear masked events from the aer queue

[Klaus Jensen]

On Sep  6 05:28, Arun Kumar wrote:
> clear masked events from the aer queue when get log page is issued with
> rae=0 without checking for the presence of outstanding aer requests
> 
> Signed-off-by: Arun Kumar <arun.kka@samsung.com>
> ---

Hi Arun,

Thanks, LGTM. One small nit below.

>  hw/nvme/ctrl.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
> index 127c3d2383..85039779da 100644
> --- a/hw/nvme/ctrl.c
> +++ b/hw/nvme/ctrl.c
> @@ -1649,9 +1649,16 @@ static void nvme_smart_event(NvmeCtrl *n, uint8_t 
> event)
>  
>  static void nvme_clear_events(NvmeCtrl *n, uint8_t event_type)
>  {
> +    NvmeAsyncEvent *event, *next;
>      n->aer_mask &= ~(1 << event_type);
>      if (!QTAILQ_EMPTY(&n->aer_queue)) {

It's safe to remove the QTAILQ_EMTPY check as well.

> -        nvme_process_aers(n);
> +        QTAILQ_FOREACH_SAFE(event, &n->aer_queue, entry, next) {
> +            if (event->result.event_type == event_type) {
> +                QTAILQ_REMOVE(&n->aer_queue, event, entry);
> +                n->aer_queued--;
> +                g_free(event);
> +            }
> +        }
>      }
>  }

We may want to check the mask when inserting as well? Currently
nvme_enqueue_event does not check the aer_mask and always inserts an
event, even if we have posted an AEN for a particular event type (but
event type has not been cleared yet).