2022-07-19 02:27:12

by Thinh Nguyen

[permalink] [raw]
Subject: [PATCH v2 17/25] usb: gadget: f_tcm: Handle abort command

Implemented aborted_task to cancel outstanding request.

Signed-off-by: Thinh Nguyen <[email protected]>
---
Changes in v2:
- None

drivers/usb/gadget/function/f_tcm.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)

diff --git a/drivers/usb/gadget/function/f_tcm.c b/drivers/usb/gadget/function/f_tcm.c
index d7318c84af98..8f7eb7045e64 100644
--- a/drivers/usb/gadget/function/f_tcm.c
+++ b/drivers/usb/gadget/function/f_tcm.c
@@ -1278,6 +1278,22 @@ static void usbg_queue_tm_rsp(struct se_cmd *se_cmd)

static void usbg_aborted_task(struct se_cmd *se_cmd)
{
+ struct usbg_cmd *cmd = container_of(se_cmd, struct usbg_cmd, se_cmd);
+ struct f_uas *fu = cmd->fu;
+ struct uas_stream *stream = cmd->stream;
+ int ret = 0;
+
+ if (stream->req_out->status == -EINPROGRESS)
+ ret = usb_ep_dequeue(fu->ep_out, stream->req_out);
+ else if (stream->req_in->status == -EINPROGRESS)
+ ret = usb_ep_dequeue(fu->ep_in, stream->req_in);
+ else if (stream->req_status->status == -EINPROGRESS)
+ ret = usb_ep_dequeue(fu->ep_status, stream->req_status);
+
+ if (ret)
+ pr_err("Unable to dequeue se_cmd out %p\n", se_cmd);
+
+ cmd->state = UASP_QUEUE_COMMAND;
}

static const char *usbg_check_wwn(const char *name)
--
2.28.0


Subject: Re: [PATCH v2 17/25] usb: gadget: f_tcm: Handle abort command

On 2022-07-18 18:27:51 [-0700], Thinh Nguyen wrote:
> diff --git a/drivers/usb/gadget/function/f_tcm.c b/drivers/usb/gadget/function/f_tcm.c
> --- a/drivers/usb/gadget/function/f_tcm.c
> +++ b/drivers/usb/gadget/function/f_tcm.c
> @@ -1278,6 +1278,22 @@ static void usbg_queue_tm_rsp(struct se_cmd *se_cmd)
>
> static void usbg_aborted_task(struct se_cmd *se_cmd)
> {
> + struct usbg_cmd *cmd = container_of(se_cmd, struct usbg_cmd, se_cmd);
> + struct f_uas *fu = cmd->fu;
> + struct uas_stream *stream = cmd->stream;
> + int ret = 0;
> +
> + if (stream->req_out->status == -EINPROGRESS)
> + ret = usb_ep_dequeue(fu->ep_out, stream->req_out);
> + else if (stream->req_in->status == -EINPROGRESS)
> + ret = usb_ep_dequeue(fu->ep_in, stream->req_in);
> + else if (stream->req_status->status == -EINPROGRESS)
> + ret = usb_ep_dequeue(fu->ep_status, stream->req_status);
> +
> + if (ret)
> + pr_err("Unable to dequeue se_cmd out %p\n", se_cmd);

I know I wasn't the best example here. But if you continue to work on
that one, if would be nice if you could replace all the pr_err() here
with a dev_err() (or so) so it is bound to an actual device and the
reader can actually pin point the message to the actually device/
subsystem. I'm not sure if we lack a device or I was extremely lazy.
Again, not something you need change now.

> + cmd->state = UASP_QUEUE_COMMAND;
> }
>
> static const char *usbg_check_wwn(const char *name)

Sebastian

2022-08-26 19:14:19

by Thinh Nguyen

[permalink] [raw]
Subject: Re: [PATCH v2 17/25] usb: gadget: f_tcm: Handle abort command

On Fri, Aug 26, 2022, Sebastian Andrzej Siewior wrote:
> On 2022-07-18 18:27:51 [-0700], Thinh Nguyen wrote:
> > diff --git a/drivers/usb/gadget/function/f_tcm.c b/drivers/usb/gadget/function/f_tcm.c
> > --- a/drivers/usb/gadget/function/f_tcm.c
> > +++ b/drivers/usb/gadget/function/f_tcm.c
> > @@ -1278,6 +1278,22 @@ static void usbg_queue_tm_rsp(struct se_cmd *se_cmd)
> >
> > static void usbg_aborted_task(struct se_cmd *se_cmd)
> > {
> > + struct usbg_cmd *cmd = container_of(se_cmd, struct usbg_cmd, se_cmd);
> > + struct f_uas *fu = cmd->fu;
> > + struct uas_stream *stream = cmd->stream;
> > + int ret = 0;
> > +
> > + if (stream->req_out->status == -EINPROGRESS)
> > + ret = usb_ep_dequeue(fu->ep_out, stream->req_out);
> > + else if (stream->req_in->status == -EINPROGRESS)
> > + ret = usb_ep_dequeue(fu->ep_in, stream->req_in);
> > + else if (stream->req_status->status == -EINPROGRESS)
> > + ret = usb_ep_dequeue(fu->ep_status, stream->req_status);
> > +
> > + if (ret)
> > + pr_err("Unable to dequeue se_cmd out %p\n", se_cmd);
>
> I know I wasn't the best example here. But if you continue to work on
> that one, if would be nice if you could replace all the pr_err() here
> with a dev_err() (or so) so it is bound to an actual device and the
> reader can actually pin point the message to the actually device/
> subsystem. I'm not sure if we lack a device or I was extremely lazy.
> Again, not something you need change now.

Yes. I noticed that too. However, I think that can be done later as it
doesn't seem critical. I tried to keep the code consistent for now.

Thanks,
Thinh

>
> > + cmd->state = UASP_QUEUE_COMMAND;
> > }
> >
> > static const char *usbg_check_wwn(const char *name)
>
> Sebastian