From: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
To: AKASH KUMAR <quic_akakum@quicinc.com>
Cc: Thinh Nguyen <Thinh.Nguyen@synopsys.com>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Jing Leng <jleng@ambarella.com>, Felipe Balbi <balbi@kernel.org>,
Jack Pham <quic_jackp@quicinc.com>,
"kernel@quicinc.com" <kernel@quicinc.com>,
Wesley Cheng <quic_wcheng@quicinc.com>,
Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
Daniel Scally <dan.scally@ideasonboard.com>,
Vijayavardhan Vennapusa <quic_vvreddy@quicinc.com>,
Krishna Kurapati <quic_kriskura@quicinc.com>,
"linux-usb@vger.kernel.org" <linux-usb@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] usb: dwc3: gadget: Fix TX FIFO size for HS ISOC endpoints
Date: Thu, 29 Aug 2024 23:08:34 +0000 [thread overview]
Message-ID: <20240829230825.7tzgz44f3it2fdmk@synopsys.com> (raw)
In-Reply-To: <aebb87ab-d57e-4cb3-8cdb-835678f31154@quicinc.com>
On Thu, Aug 29, 2024, AKASH KUMAR wrote:
> Hi Thinh,
>
> On 8/29/2024 5:15 AM, Thinh Nguyen wrote:
> > On Wed, Aug 28, 2024, AKASH KUMAR wrote:
> > > Hi Thinh,
> > >
> > > On 8/28/2024 4:45 AM, Thinh Nguyen wrote:
> > > > On Tue, Aug 27, 2024, Akash Kumar wrote:
> > > > > Use 2K TX FIFO size for low-resolution UVC cameras to support the
> > > > > maximum possible UVC instances. Restrict 2K TX FIFO size based on
> > > > > the minimum maxburst required to run low-resolution UVC cameras.
> > > > >
> > > > > Signed-off-by: Akash Kumar <quic_akakum@quicinc.com>
> > > > > ---
> > > > > drivers/usb/dwc3/gadget.c | 4 ++++
> > > > > 1 file changed, 4 insertions(+)
> > > > >
> > > > > diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> > > > > index 89fc690fdf34..f342ccda6705 100644
> > > > > --- a/drivers/usb/dwc3/gadget.c
> > > > > +++ b/drivers/usb/dwc3/gadget.c
> > > > > @@ -788,6 +788,10 @@ static int dwc3_gadget_resize_tx_fifos(struct dwc3_ep *dep)
> > > > > usb_endpoint_xfer_isoc(dep->endpoint.desc)) && DWC3_IP_IS(DWC31))
> > > > > num_fifos = dwc->tx_fifo_resize_max_num;
> > > > > + if (dep->endpoint.maxburst <= 1 &&
> > > > > + usb_endpoint_xfer_isoc(dep->endpoint.desc))
> > > > > + num_fifos = 2;
> > > > > +
> > > > > /* FIFO size for a single buffer */
> > > > > fifo = dwc3_gadget_calc_tx_fifo_size(dwc, 1);
> > > > > --
> > > > > 2.17.1
> > > > >
> > > > These settings are starting to get too specific for each application.
> > > > Can we find a better calculation?
> > > >
> > > > Perhaps something like this? (code not tested)
> > > >
> > > > diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> > > > index 9a18973ebc05..d54b08f92aea 100644
> > > > --- a/drivers/usb/dwc3/gadget.c
> > > > +++ b/drivers/usb/dwc3/gadget.c
> > > > @@ -908,15 +908,10 @@ static int dwc3_gadget_resize_tx_fifos(struct dwc3_ep *dep)
> > > > ram1_depth = DWC3_RAM1_DEPTH(dwc->hwparams.hwparams7);
> > > > - if ((dep->endpoint.maxburst > 1 &&
> > > > - usb_endpoint_xfer_bulk(dep->endpoint.desc)) ||
> > > > + if (usb_endpoint_xfer_bulk(dep->endpoint.desc) ||
> > > > usb_endpoint_xfer_isoc(dep->endpoint.desc))
> > > > - num_fifos = 3;
> > > > -
> > > > - if (dep->endpoint.maxburst > 6 &&
> > > > - (usb_endpoint_xfer_bulk(dep->endpoint.desc) ||
> > > > - usb_endpoint_xfer_isoc(dep->endpoint.desc)) && DWC3_IP_IS(DWC31))
> > > > - num_fifos = dwc->tx_fifo_resize_max_num;
> > > > + num_fifos = min_t(unsigned int, dep->endpoint.maxburst + 1,
> > > > + dwc->tx_fifo_resize_max_num);
> > > > /* FIFO size for a single buffer */
> > > > fifo = dwc3_gadget_calc_tx_fifo_size(dwc, 1);
> > > >
> > > should be fine for me, as earlier there was no case handling for maxburst <=
> > > 1, by allocating fifo based on maxburst itself
> > >
> > > should be a good solution and will work for all as they customize based on
> > > maxburst through init scripts, let me test and update.
> > >
> > As I noted in the follow up response, you need to also account for
> > additional transaction opportunities for isoc if operating in highspeed.
> > (ie. use usb_endpoint_maxp_mult()).
> yeah i think that we can control with maxpacket size for high speed cams and
> we can
> work with maxburst requested only to allocate fifo as both are configurable
> entity and one
> should be free to adjust fifo as per his environment and not add any
> additional limitation.
> Let me know if we aligned and go ahead with above suggestion only.
>
I'm not quite clear here. Perhaps you can provide the change so we can
review.
BR,
Thinh
prev parent reply other threads:[~2024-08-29 23:09 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-27 5:49 Akash Kumar
2024-08-27 23:15 ` Thinh Nguyen
2024-08-28 2:06 ` Thinh Nguyen
2024-08-28 4:11 ` AKASH KUMAR
2024-08-28 23:45 ` Thinh Nguyen
2024-08-29 7:37 ` AKASH KUMAR
2024-08-29 23:08 ` Thinh Nguyen [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240829230825.7tzgz44f3it2fdmk@synopsys.com \
--to=thinh.nguyen@synopsys.com \
--cc=balbi@kernel.org \
--cc=dan.scally@ideasonboard.com \
--cc=gregkh@linuxfoundation.org \
--cc=jleng@ambarella.com \
--cc=kernel@quicinc.com \
--cc=laurent.pinchart@ideasonboard.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-usb@vger.kernel.org \
--cc=quic_akakum@quicinc.com \
--cc=quic_jackp@quicinc.com \
--cc=quic_kriskura@quicinc.com \
--cc=quic_vvreddy@quicinc.com \
--cc=quic_wcheng@quicinc.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®