linux-kernel - RE: [Resend PATCH v2] testusb: add path /dev/bus/usb to default search paths of usbfs

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]

Message-ID: <alpine.DEB.2.00.1205212055140.2666@chino.kir.corp.google.com>
Date:	Mon, 21 May 2012 20:58:28 -0700 (PDT)
From:	David Rientjes <rientjes@...gle.com>
To:	"Du, ChangbinX" <changbinx.du@...el.com>
cc:	"gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
	Sergei Shtylyov <sshtylyov@...sta.com>,
	"mina86@...a86.com" <mina86@...a86.com>,
	"Fleming, Matt" <matt.fleming@...el.com>,
	"balbi@...com" <balbi@...com>, "hpa@...or.com" <hpa@...or.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>
Subject: RE: [Resend PATCH v2] testusb: add path /dev/bus/usb to default
 search paths of usbfs

On Tue, 22 May 2012, Du, ChangbinX wrote:

> > diff --git a/tools/usb/testusb.c b/tools/usb/testusb.c index 
> > 6e0f567..82d7c59 100644
> > --- a/tools/usb/testusb.c
> > +++ b/tools/usb/testusb.c
> > @@ -358,6 +358,7 @@ static const char *usbfs_dir_find(void)  {
> >  	static char usbfs_path_0[] = "/dev/usb/devices";
> >  	static char usbfs_path_1[] = "/proc/bus/usb/devices";
> > +	static char udev_usb_path[] = "/dev/bus/usb";
> >  
> >  	static char *const usbfs_paths[] = {
> >  		usbfs_path_0, usbfs_path_1
> > @@ -376,6 +377,10 @@ static const char *usbfs_dir_find(void)
> >  		}
> >  	} while (++it != end);
> >  
> > +	/* real device-nodes managed by udev */
> > +	if (access(udev_usb_path, F_OK) == 0)
> > +		return udev_usb_path;
> > +
> >  	return NULL;
> >  }
> >  
> 
> Two issues with this: F_OK only guarantees that the path exists, it 
> does not guarantee that it is readable like this function guarantees for 
> usbfs_paths, and access() shouldn't be used because of its security 
> implications, you're better off using open() and testing for fd.
> 
> Hello, David. I think this function doesn't need check the permission. 
> What this function need do is to find the usbfs mount point. If the path 
> exists but cannot read, we cannot report as it's not exits. And when we 
> read files, open() will return an error and user can check if it's need 
> to upgrade his access right.
> 

Your email client doesn't seem to be quoting messages correctly :)

Anyway, this patch is inconsistent with the rest of the function.  The 
other two paths are checked with open(O_RDONLY) followed by a close() to 
address my second comment and we certainly wouldn't want to return a path 
that exists by is unreadable: we'd want to fallback to one of the other 
possibilities.  So if anybody is going to extend this in the future like 
you have, it would be possible to return /dev/bus/usb even though we can't 
read it.  That's a bad result.

Please consider doing it the proper way: by doing open(O_RDONLY), close() 
instead of access() -- if you don't understand why, read access(2) -- and 
in an extendable way.
--
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/