lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Sun, 14 Aug 2022 11:32:07 -0400
From:   Sasha Levin <sashal@...nel.org>
To:     linux-kernel@...r.kernel.org, stable@...r.kernel.org
Cc:     Michael Grzeschik <m.grzeschik@...gutronix.de>,
        Dan Vacura <w36195@...orola.com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Sasha Levin <sashal@...nel.org>,
        laurent.pinchart@...asonboard.com, balbi@...nel.org,
        linux-usb@...r.kernel.org
Subject: [PATCH AUTOSEL 5.15 06/46] usb: gadget: uvc: calculate the number of request depending on framesize

From: Michael Grzeschik <m.grzeschik@...gutronix.de>

[ Upstream commit 87d76b5f1d8eeb49efa16e2018e188864cbb9401 ]

The current limitation of possible number of requests being handled is
dependent on the gadget speed. It makes more sense to depend on the
typical frame size when calculating the number of requests. This patch
is changing this and is using the previous limits as boundaries for
reasonable minimum and maximum number of requests.

For a 1080p jpeg encoded video stream with a maximum imagesize of
e.g. 800kB with a maxburst of 8 and an multiplier of 1 the resulting
number of requests is calculated to 49.

        800768         1
nreqs = ------ * -------------- ~= 49
          2      (1024 * 8 * 1)

Tested-by: Dan Vacura <w36195@...orola.com>
Signed-off-by: Michael Grzeschik <m.grzeschik@...gutronix.de>
Link: https://lore.kernel.org/r/20220529223848.105914-2-m.grzeschik@pengutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 drivers/usb/gadget/function/uvc_queue.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/drivers/usb/gadget/function/uvc_queue.c b/drivers/usb/gadget/function/uvc_queue.c
index 99dc9adf56ef..a64b842665b9 100644
--- a/drivers/usb/gadget/function/uvc_queue.c
+++ b/drivers/usb/gadget/function/uvc_queue.c
@@ -44,7 +44,8 @@ static int uvc_queue_setup(struct vb2_queue *vq,
 {
 	struct uvc_video_queue *queue = vb2_get_drv_priv(vq);
 	struct uvc_video *video = container_of(queue, struct uvc_video, queue);
-	struct usb_composite_dev *cdev = video->uvc->func.config->cdev;
+	unsigned int req_size;
+	unsigned int nreq;
 
 	if (*nbuffers > UVC_MAX_VIDEO_BUFFERS)
 		*nbuffers = UVC_MAX_VIDEO_BUFFERS;
@@ -53,10 +54,16 @@ static int uvc_queue_setup(struct vb2_queue *vq,
 
 	sizes[0] = video->imagesize;
 
-	if (cdev->gadget->speed < USB_SPEED_SUPER)
-		video->uvc_num_requests = 4;
-	else
-		video->uvc_num_requests = 64;
+	req_size = video->ep->maxpacket
+		 * max_t(unsigned int, video->ep->maxburst, 1)
+		 * (video->ep->mult);
+
+	/* We divide by two, to increase the chance to run
+	 * into fewer requests for smaller framesizes.
+	 */
+	nreq = DIV_ROUND_UP(DIV_ROUND_UP(sizes[0], 2), req_size);
+	nreq = clamp(nreq, 4U, 64U);
+	video->uvc_num_requests = nreq;
 
 	return 0;
 }
-- 
2.35.1

Powered by blists - more mailing lists