mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Tobias Herzog <t-herzog@gmx.de>
To: Oliver Neukum <oneukum@suse.com>
Cc: gregkh@linuxfoundation.org, linux-kernel@vger.kernel.org,
	linux-usb@vger.kernel.org
Subject: Re: [PATCH 1/4] cdc-acm: reassemble fragmented notifications
Date: Sat, 18 Mar 2017 19:52:26 +0100	[thread overview]
Message-ID: <1489863146.3390.8.camel@gmx.de> (raw)
In-Reply-To: <1489569961.30434.6.camel@suse.com>

Hi,

thank you for your feedback. I tried to fix all issues with v2. I was
just unsure with one point (see comment).


/tobias


> Am Dienstag, den 14.03.2017, 21:14 +0100 schrieb Tobias Herzog:
> > 
> > USB devices may have very limitited endpoint packet sizes, so that
> > notifications can not be transferred within one single usb packet.
> > Reassembling of multiple packages may be necessary.
> Hi,
> 
> thank you for the patch. Unfortunately it has some issue.
> Please see the comments inside.
> 
> 	Regards
> 		Oliver
> 
> > 
> > 
> > Signed-off-by: Tobias Herzog <t-herzog@gmx.de>
> > ---
> >  drivers/usb/class/cdc-acm.c | 102 +++++++++++++++++++++++++++++++-
> > ------------
> >  drivers/usb/class/cdc-acm.h |   2 +
> >  2 files changed, 75 insertions(+), 29 deletions(-)
> > 
> > diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-
> > acm.c
> > index e35b150..40714fe 100644
> > --- a/drivers/usb/class/cdc-acm.c
> > +++ b/drivers/usb/class/cdc-acm.c
> > @@ -282,39 +282,13 @@ static DEVICE_ATTR(iCountryCodeRelDate,
> > S_IRUGO, show_country_rel_date, NULL);
> >   * Interrupt handlers for various ACM device responses
> >   */
> >  
> > -/* control interface reports status changes with "interrupt"
> > transfers */
> > -static void acm_ctrl_irq(struct urb *urb)
> > +static void acm_process_notification(struct acm *acm, unsigned
> > char *buf)
> >  {
> > -	struct acm *acm = urb->context;
> > -	struct usb_cdc_notification *dr = urb->transfer_buffer;
> > -	unsigned char *data;
> >  	int newctrl;
> >  	int difference;
> > -	int retval;
> > -	int status = urb->status;
> > +	struct usb_cdc_notification *dr = (struct
> > usb_cdc_notification *)buf;
> > +	unsigned char *data = (unsigned char *)(dr + 1);
> >  
> > -	switch (status) {
> > -	case 0:
> > -		/* success */
> > -		break;
> > -	case -ECONNRESET:
> > -	case -ENOENT:
> > -	case -ESHUTDOWN:
> > -		/* this urb is terminated, clean up */
> > -		dev_dbg(&acm->control->dev,
> > -			"%s - urb shutting down with status:
> > %d\n",
> > -			__func__, status);
> > -		return;
> > -	default:
> > -		dev_dbg(&acm->control->dev,
> > -			"%s - nonzero urb status received: %d\n",
> > -			__func__, status);
> > -		goto exit;
> > -	}
> > -
> > -	usb_mark_last_busy(acm->dev);
> > -
> > -	data = (unsigned char *)(dr + 1);
> >  	switch (dr->bNotificationType) {
> >  	case USB_CDC_NOTIFY_NETWORK_CONNECTION:
> >  		dev_dbg(&acm->control->dev,
> > @@ -363,8 +337,74 @@ static void acm_ctrl_irq(struct urb *urb)
> >  			__func__,
> >  			dr->bNotificationType, dr->wIndex,
> >  			dr->wLength, data[0], data[1]);
> > +	}
> > +}
> > +
> > +/* control interface reports status changes with "interrupt"
> > transfers */
> > +static void acm_ctrl_irq(struct urb *urb)
> > +{
> > +	struct acm *acm = urb->context;
> > +	struct usb_cdc_notification *dr = urb->transfer_buffer;
> > +	unsigned int current_size = urb->actual_length;
> > +	unsigned int expected_size, copy_size;
> > +	int retval;
> > +	int status = urb->status;
> > +
> > +	switch (status) {
> > +	case 0:
> > +		/* success */
> >  		break;
> > +	case -ECONNRESET:
> > +	case -ENOENT:
> > +	case -ESHUTDOWN:
> > +		/* this urb is terminated, clean up */
> > +		kfree(acm->notification_buffer);
> > +		acm->notification_buffer = NULL;
> Why? Disconnect() will free it anyway. It should be enough
> to discard the content.
> 
> > 
> > +		dev_dbg(&acm->control->dev,
> > +			"%s - urb shutting down with status:
> > %d\n",
> > +			__func__, status);
> > +		return;
> > +	default:
> > +		dev_dbg(&acm->control->dev,
> > +			"%s - nonzero urb status received: %d\n",
> > +			__func__, status);
> > +		goto exit;
> >  	}
> > +
> > +	usb_mark_last_busy(acm->dev);
> > +
> > +	if (acm->notification_buffer)
> > +		dr = (struct usb_cdc_notification *)acm-
> > >notification_buffer;
> > +
> > +	/* assume the first package contains at least two bytes */
> > +	expected_size = dr->wLength + 8;
> You need the explain where you got the 8 from. In fact a define would
> be best.
I feel better to use 'sizeof(struct usb_cdc_notification)' here, than
using an separate, hard-coded define. In fact this is really what we
want here, as the buffer needs to cover the notification header (i.e.
struct usb_cdc_notification) + the additional (optional) data.

> 
> > 
> > +
> > +	if (current_size < expected_size) {
> > +		/* notification is transmitted framented,
> > reassemble */
> Please fix the typo.
> 
> > 
> > +		if (!acm->notification_buffer) {
> > +			acm->notification_buffer =
> > +					kmalloc(expected_size,
> > GFP_ATOMIC);
> This can fail. You _must_ check for that.
> 
> > 
> > +			acm->nb_index = 0;
> > +		}
> > +
> > +		copy_size = min(current_size,
> > +				expected_size - acm->nb_index);
> > +
> > +		memcpy(&acm->notification_buffer[acm->nb_index],
> > +		       urb->transfer_buffer, copy_size);
> > +		acm->nb_index += copy_size;
> > +		current_size = acm->nb_index;
> > +	}
> > +
> > +	if (current_size < expected_size)
> > +		goto exit;
> This is an unneeded goto.
> 
> > 
> > +	/* notification complete */
> > +	acm_process_notification(acm, (unsigned char *)dr);
> > +
> > +	kfree(acm->notification_buffer);
> Why? If one message was fragmented, the next one will also likely be
> fragmented. Why release the buffer before you know whether it can be
> reused?
> 
> > 
> > +	acm->notification_buffer = NULL;
> > +
> >  exit:
> >  	retval = usb_submit_urb(urb, GFP_ATOMIC);
> >  	if (retval && retval != -EPERM)
> > @@ -1488,6 +1528,8 @@ static int acm_probe(struct usb_interface
> > *intf,
> >  			 epctrl->bInterval ? epctrl->bInterval :
> > 16);
> >  	acm->ctrlurb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP;
> >  	acm->ctrlurb->transfer_dma = acm->ctrl_dma;
> > +	acm->notification_buffer = NULL;
> > +	acm->nb_index = 0;
> >  
> >  	dev_info(&intf->dev, "ttyACM%d: USB ACM device\n", minor);
> >  
> > @@ -1580,6 +1622,8 @@ static void acm_disconnect(struct
> > usb_interface *intf)
> >  	usb_free_coherent(acm->dev, acm->ctrlsize, acm-
> > >ctrl_buffer, acm->ctrl_dma);
> >  	acm_read_buffers_free(acm);
> >  
> > +	kfree(acm->notification_buffer);
> > +
> >  	if (!acm->combined_interfaces)
> >  		usb_driver_release_interface(&acm_driver, intf ==
> > acm->control ?
> >  					acm->data : acm->control);
> > diff --git a/drivers/usb/class/cdc-acm.h b/drivers/usb/class/cdc-
> > acm.h
> > index c980f11..bc07fb2 100644
> > --- a/drivers/usb/class/cdc-acm.h
> > +++ b/drivers/usb/class/cdc-acm.h
> > @@ -98,6 +98,8 @@ struct acm {
> >  	struct acm_wb *putbuffer;			/* for
> > acm_tty_put_char() */
> >  	int rx_buflimit;
> >  	spinlock_t read_lock;
> > +	u8 *notification_buffer;			/* to
> > reassemble fragmented notifications */
> > +	unsigned int nb_index;
> >  	int write_used;					/*
> > number of non-empty write buffers */
> >  	int transmitting;
> >  	spinlock_t write_lock;

  reply	other threads:[~2017-03-18 18:52 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <1479118868.21146.4.camel@suse.com>
2017-03-14 20:14 ` Tobias Herzog
2017-03-14 20:14   ` [PATCH 2/4] cdc-acm: fix possible invalid access when processing notification Tobias Herzog
2017-03-15  9:26     ` Oliver Neukum
2017-03-14 20:14   ` [PATCH 3/4] cdc-acm: log message for serial state notification Tobias Herzog
2017-03-14 20:14   ` [PATCH 4/4] cdc-acm: remove unused element of struct acm Tobias Herzog
2017-03-15  9:26   ` [PATCH 1/4] cdc-acm: reassemble fragmented notifications Oliver Neukum
2017-03-18 18:52     ` Tobias Herzog [this message]
2017-03-18 18:52 ` [PATCH v2 0/4] " Tobias Herzog
2017-03-18 18:52   ` [PATCH v2 1/4] " Tobias Herzog
2017-03-20 15:02     ` Oliver Neukum
2017-03-24 21:50       ` Tobias Herzog
2017-03-28  8:04         ` Oliver Neukum
2017-03-18 18:52   ` [PATCH v2 2/4] cdc-acm: fix possible invalid access when processing notification Tobias Herzog
2017-03-19  9:50     ` Sergei Shtylyov
2017-03-20 15:04     ` Oliver Neukum
2017-03-18 18:52   ` [PATCH v2 3/4] cdc-acm: log message for serial state notification Tobias Herzog
2017-03-18 18:52   ` [PATCH v2 4/4] cdc-acm: remove unused element of struct acm Tobias Herzog
2017-03-30 20:15 ` [PATCH v3 0/4] cdc-acm: reassemble fragmented notifications Tobias Herzog
2017-03-30 20:15   ` [PATCH v3 1/4] cdc-acm: fix possible invalid access when processing notification Tobias Herzog
2017-03-31  9:31     ` Oliver Neukum
2017-03-30 20:15   ` [PATCH v3 2/4] cdc-acm: reassemble fragmented notifications Tobias Herzog
2017-03-31  9:33     ` Oliver Neukum
2017-03-30 20:15   ` [PATCH v3 3/4] cdc-acm: log message for serial state notification Tobias Herzog
2017-03-31  9:34     ` Oliver Neukum
2017-03-30 20:15   ` [PATCH v3 4/4] cdc-acm: remove unused element of struct acm Tobias Herzog
2017-03-31  9:35     ` Oliver Neukum

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=1489863146.3390.8.camel@gmx.de \
    --to=t-herzog@gmx.de \
    --cc=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-usb@vger.kernel.org \
    --cc=oneukum@suse.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

Powered by JetHome