From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stefan Berger Subject: Re: [PATCH v6 11/11] A test program for vTPM devicecreation Date: Fri, 11 Mar 2016 08:16:28 -0500 Message-ID: <56E2C52C.20606@linux.vnet.ibm.com> References: <1457545170-30120-1-git-send-email-stefanb@linux.vnet.ibm.com> <1457545170-30120-12-git-send-email-stefanb@linux.vnet.ibm.com> <201603102058.u2AKw7Ie013400@d01av04.pok.ibm.com> <20160311103001.GA13368@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20160311103001.GA13368-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: tpmdd-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: Jarkko Sakkinen , Stefan Berger Cc: tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org List-Id: tpmdd-devel@lists.sourceforge.net On 03/11/2016 05:30 AM, Jarkko Sakkinen wrote: > On Thu, Mar 10, 2016 at 03:57:58PM -0500, Stefan Berger wrote: >> Stefan Berger wrote on 03/09/2016 12:39:30 >> PM: >> > >> > This patch provides a program that is for testing purposes only. >> > >> > Build it using the following commands: >> > >> > make headers_install ARCH=x86_64 INSTALL_HDR_PATH=/usr >> > >> > gcc vtpmctrl.c -o vtpmctrl >> > >> > To use it: >> > >> > To create a device pair and have vtpmctrl listen for commands, display >> > them and respond with TPM success messages do: >> > >> > Created TPM device /dev/tpm0; vTPM device has fd 4, major/minor = >> 10/224. >> > >> > In another shell do >> > >> > 00000000 00 c4 00 00 00 0a 00 00 00 00 >> > 00000012 >> >> This test program is now available as part of a test suite at: >> >> https://github.com/stefanberger/linux-vtpm-tests > I've ported TPM 2.0 simulator proxy here > > git://git.infradead.org/users/jjs/tpm2-scripts.git > > At the moment it gets blocked on write() at client side and on read() at > server side. Haven't yet applied the timeouts patch but I'd expect > things to work without that patch since it's a separate patch With this patch applied it works for me. I am using a file descriptor rather than the file. If using a file, it seems to want to read more bytes to fill up 4096 bytes or get an EOF. Also make sure that /dev/tpmX is not a normal file. I ran into that problem when using exec 100<>/dev/tpm1 that creates /dev/tpm1 as file if it isn't there... diff --git a/tpm2-simulator-vtpm b/tpm2-simulator-vtpm index f9f7d82..6a9a0d9 100755 --- a/tpm2-simulator-vtpm +++ b/tpm2-simulator-vtpm @@ -56,6 +56,7 @@ class TPM2Simulator(): """ Initializes the connection to the simulator and initializes the simulator device by sending some control commands """ + return self.ctrl = socket.socket(socket.AF_INET, socket.SOCK_STREAM) self.ctrl.connect((host, 2322)) self.data = socket.socket(socket.AF_INET, socket.SOCK_STREAM) @@ -109,16 +110,14 @@ class TPM2Simulator(): with open('/dev/vtpmx', 'w+b', buffering = 0) as vtpmx_fp: ioctl(vtpmx_fp, VTPM_IOC_NEW_DEV, addressof(new_dev)) - vtpm_fp = os.fdopen(new_dev.fd) - for n, t in new_dev._fields_: print(n, getattr(new_dev, n)) while True: - stream = vtpm_fp.read(4096) + stream = os.read(new_dev.fd, 4096) print("IN {}".format(binascii.hexlify(stream))) - resp = self.transmit_command(stream) - vtpm_fp.write(resp) + #resp = self.transmit_command(stream) + os.write(new_dev.fd, stream) class c_vtpm_new_dev(Structure): _fields_ = \ Stefan ------------------------------------------------------------------------------ Transform Data into Opportunity. Accelerate data analysis in your applications with Intel Data Analytics Acceleration Library. Click to learn more. http://pubads.g.doubleclick.net/gampad/clk?id=278785111&iu=/4140