Re: [PATCH] usb: host: xhci: Handle the right timeout command

[Baolin Wang]

On 2 December 2016 at 19:21, Baolin Wang <baolin.wang@xxxxxxxxxx> wrote:
> If a command event is found on the event ring during an interrupt,
> we need to stop the command timer with del_timer(). Since del_timer()
> can fail if the timer is running and waiting on the xHCI lock, then
> it maybe get the wrong timeout command in xhci_handle_command_timeout()
> if host fetched a new command and updated the xhci->current_cmd in
> handle_cmd_completion(). For this situation, we need a way to signal
> to the command timer that everything is fine and it should exit.
>
> We should introduce a counter (xhci->current_cmd_pending) for the number
> of pending commands. If we need to cancel the command timer and del_timer()
> succeeds, we decrement the number of pending commands. If del_timer() fails,
> we leave the number of pending commands alone.
>
> For handling timeout command, in xhci_handle_command_timeout() we will check
> the counter after decrementing it, if the counter (xhci->current_cmd_pending)
> is 0, which means xhci->current_cmd is the right timeout command. If the
> counter (xhci->current_cmd_pending) is greater than 0, which means current
> timeout command has been handled by host and host has fetched new command
> as xhci->current_cmd, then just return and wait for new current command.
>
> Signed-off-by: Baolin Wang <baolin.wang@xxxxxxxxxx>
> ---
> This patch is based on Lu Baolu's new fix patch:
> usb: xhci: fix possible wild pointer
> ---
>  drivers/usb/host/xhci-ring.c |   26 +++++++++++++++++++++++++-
>  drivers/usb/host/xhci.h      |    1 +
>  2 files changed, 26 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
> index 62dd1c7..a62904e 100644
> --- a/drivers/usb/host/xhci-ring.c
> +++ b/drivers/usb/host/xhci-ring.c
> @@ -1253,6 +1253,7 @@ static void xhci_handle_stopped_cmd_ring(struct xhci_hcd *xhci,
>         if ((xhci->cmd_ring->dequeue != xhci->cmd_ring->enqueue) &&
>             !(xhci->xhc_state & XHCI_STATE_DYING)) {
>                 xhci->current_cmd = cur_cmd;
> +               xhci->current_cmd_pending++;
>                 mod_timer(&xhci->cmd_timer, jiffies + XHCI_CMD_DEFAULT_TIMEOUT);
>                 xhci_ring_cmd_db(xhci);
>         }
> @@ -1274,6 +1275,21 @@ void xhci_handle_command_timeout(unsigned long data)
>                 return;
>         }
>
> +       xhci->current_cmd_pending--;

I realized I should move the decrement before 'xhci->current_cmd'
checking, and I will wait for a while for new comments before sending
out v2 patch.

> +       /*
> +        * If the current_cmd_pending is 0, which means current command is
> +        * timeout.
> +        *
> +        * If the current_cmd_pending is greater than 0, which means current
> +        * timeout command has been handled by host and host has fetched new
> +        * command as xhci->current_cmd, then just return and wait for new
> +        * current command.
> +        */
> +       if (xhci->current_cmd_pending > 0) {
> +               spin_unlock_irqrestore(&xhci->lock, flags);
> +               return;
> +       }
> +
>         if (xhci->current_cmd->status == COMP_CMD_ABORT)
>                 second_timeout = true;
>         xhci->current_cmd->status = COMP_CMD_ABORT;
> @@ -1336,7 +1352,13 @@ static void handle_cmd_completion(struct xhci_hcd *xhci,
>
>         cmd = list_entry(xhci->cmd_list.next, struct xhci_command, cmd_list);
>
> -       del_timer(&xhci->cmd_timer);
> +       /*
> +        * If the command timer is running on another CPU, we don't decrement
> +        * current_cmd_pending, since we didn't successfully stop the command
> +        * timer.
> +        */
> +       if (del_timer(&xhci->cmd_timer))
> +               xhci->current_cmd_pending--;
>
>         trace_xhci_cmd_completion(cmd_trb, (struct xhci_generic_trb *) event);
>
> @@ -1424,6 +1446,7 @@ static void handle_cmd_completion(struct xhci_hcd *xhci,
>         if (cmd->cmd_list.next != &xhci->cmd_list) {
>                 xhci->current_cmd = list_entry(cmd->cmd_list.next,
>                                                struct xhci_command, cmd_list);
> +               xhci->current_cmd_pending++;
>                 mod_timer(&xhci->cmd_timer, jiffies + XHCI_CMD_DEFAULT_TIMEOUT);
>         } else if (xhci->current_cmd == cmd) {
>                 xhci->current_cmd = NULL;
> @@ -3924,6 +3947,7 @@ static int queue_command(struct xhci_hcd *xhci, struct xhci_command *cmd,
>         if (xhci->cmd_list.next == &cmd->cmd_list &&
>             !timer_pending(&xhci->cmd_timer)) {
>                 xhci->current_cmd = cmd;
> +               xhci->current_cmd_pending++;
>                 mod_timer(&xhci->cmd_timer, jiffies + XHCI_CMD_DEFAULT_TIMEOUT);
>         }
>
> diff --git a/drivers/usb/host/xhci.h b/drivers/usb/host/xhci.h
> index 9dbaacf..5d81257 100644
> --- a/drivers/usb/host/xhci.h
> +++ b/drivers/usb/host/xhci.h
> @@ -1567,6 +1567,7 @@ struct xhci_hcd {
>         unsigned int            cmd_ring_reserved_trbs;
>         struct timer_list       cmd_timer;
>         struct xhci_command     *current_cmd;
> +       u32                     current_cmd_pending;
>         struct xhci_ring        *event_ring;
>         struct xhci_erst        erst;
>         /* Scratchpad */
> --
> 1.7.9.5
>



-- 
Baolin.wang
Best Regards