linux-kernel - Re: [PATCH 05/10] added media specific (MS) TCP drivers

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:	Tue, 4 Nov 2014 10:02:10 -0800
From:	Greg KH <gregkh@...uxfoundation.org>
To:	Tobias Klauser <tklauser@...tanz.ch>
Cc:	Stephanie Wallick <stephanie.s.wallick@...el.com>,
	devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org,
	"Sean O. Stalley" <sean.stalley@...el.com>
Subject: Re: [PATCH 05/10] added media specific (MS) TCP drivers

On Tue, Nov 04, 2014 at 09:48:33AM +0100, Tobias Klauser wrote:
> On 2014-11-03 at 21:42:52 +0100, Stephanie Wallick <stephanie.s.wallick@...el.com> wrote:
> > This is where we handle media specific packets and transport. The MS driver
> > interfaces with a media agnostic (MA) driver via a series of transfer pairs.
> > Transfer pairs consist of a set of functions to pass MA USB packets back
> > and forth between MA and MS drivers. There is one transfer pair per device
> > endpoint and one transfer pair for control/management traffic. When the MA
> > driver needs to send an MA USB packet, it hands the packet off to the MS
> > layer where the packet is converted into an MS form and sent via TCP over
> > the underlying ethernet or wireless medium. When the MS driver receives a
> > packet, it converts it into an MA USB packet and hands it off the the MA
> > driver for handling.
> > 
> > In addition, the MS driver provides an interface to inititate connection events.
> > Because there are no physical MA USB ports in an MA USB host, the host must be
> > notified via software when a device is connected.
> > 
> > Lastly, the MS driver contains a number of ioctl functions that are used by a
> > utility to adjust medium-related driver parameters and connect or disconnect the
> > MA USB host and device drivers.
> > 
> > Signed-off-by: Sean O. Stalley <sean.stalley@...el.com>
> > Signed-off-by: Stephanie Wallick <stephanie.s.wallick@...el.com>
> > ---
> >  drivers/staging/mausb/drivers/mausb_ioctl.c      | 373 +++++++++++++++++++
> >  drivers/staging/mausb/drivers/mausb_ioctl.h      |  99 +++++
> >  drivers/staging/mausb/drivers/mausb_msapi.c      | 110 ++++++
> >  drivers/staging/mausb/drivers/mausb_msapi.h      | 232 ++++++++++++
> >  drivers/staging/mausb/drivers/mausb_tcp-device.c | 147 ++++++++
> >  drivers/staging/mausb/drivers/mausb_tcp-host.c   | 144 ++++++++
> >  drivers/staging/mausb/drivers/mausb_tcp.c        | 446 +++++++++++++++++++++++
> >  drivers/staging/mausb/drivers/mausb_tcp.h        | 129 +++++++
> >  8 files changed, 1680 insertions(+)
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_ioctl.c
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_ioctl.h
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_msapi.c
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_msapi.h
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_tcp-device.c
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_tcp-host.c
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_tcp.c
> >  create mode 100644 drivers/staging/mausb/drivers/mausb_tcp.h
> > 
> > diff --git a/drivers/staging/mausb/drivers/mausb_ioctl.c b/drivers/staging/mausb/drivers/mausb_ioctl.c
> > new file mode 100644
> > index 0000000..0c6c6bd
> > --- /dev/null
> > +++ b/drivers/staging/mausb/drivers/mausb_ioctl.c
> 
> [...]
> 
> > +/**
> > + * This function is used to send a message to the user, in other words, the
> > + * calling process. It basically copies the message one byte at a time.
> > + *
> > + * @msg:	The message to be sent to the user.
> > + * @buffer:	The buffer in which to put the message. This buffer was given to
> > + *	    	us to fill.
> > + */
> > +void to_user(char *msg, long unsigned int buffer)
> > +{
> > +	int length = (int)strlen(msg);
> > +	int bytes = 0;
> > +
> > +	while (length && *msg) {
> > +		put_user(*(msg++), (char *)buffer++);
> > +		length--;
> > +		bytes++;
> > +	}
> 
> Any reason not to use copy_to_user here? That way, access_ok would only
> need to be executed once for the whole range.
> 
> In any case, the return value of put_user/copy_to_user will need to be
> checked.

Never use put_user if you can help it, this whole function should go
away, and copy_to_user() should be used at the caller sites instead as
you point out.

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/