From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kenneth Graunke Subject: Re: [PATCH v2 3/3] drm/i915: Track OACONTROL register enable/disable during parsing Date: Fri, 28 Mar 2014 10:37:29 -0700 Message-ID: <5335B359.9070809@whitecape.org> References: <1395945820-20376-4-git-send-email-bradley.d.volkin@intel.com> <1396027310-16951-1-git-send-email-bradley.d.volkin@intel.com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1617904076==" Return-path: Received: from homiemail-a10.g.dreamhost.com (caiajhbdccbe.dreamhost.com [208.97.132.214]) by gabe.freedesktop.org (Postfix) with ESMTP id 0DF7A6ED09 for ; Fri, 28 Mar 2014 10:37:05 -0700 (PDT) In-Reply-To: <1396027310-16951-1-git-send-email-bradley.d.volkin@intel.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" To: bradley.d.volkin@intel.com, intel-gfx@lists.freedesktop.org List-Id: intel-gfx@lists.freedesktop.org This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --===============1617904076== Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="pe5ihs3jm7vsteRM1nr7meDkuxW71gUFL" This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --pe5ihs3jm7vsteRM1nr7meDkuxW71gUFL Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable On 03/28/2014 10:21 AM, bradley.d.volkin@intel.com wrote: > From: Brad Volkin >=20 > There is some thought that the data from the performance counters enabl= ed > via OACONTROL should only be available to the process that enabled coun= ting. > To limit snooping, require that any batch buffer which sets OACONTROL t= o a > non-zero value also sets it back to 0 before the end of the batch. >=20 > This requires limiting OACONTROL writes to happen via MI_LOAD_REGISTER_= IMM > so that we can access the value being written. This should be in line w= ith > the expected use case for writing OACONTROL. >=20 > v2: Drop an unnecessary '? true : false' >=20 > Cc: Kenneth Graunke > Signed-off-by: Brad Volkin I still don't see any rationale for prohibiting LRM - there's zero security benefit, and you can write to every other register with either LRI or LRM, so it's just extra inconsistency. But, Daniel. Reviewed-by: Kenneth Graunke --pe5ihs3jm7vsteRM1nr7meDkuxW71gUFL Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iQIcBAEBAgAGBQJTNbNdAAoJEFtb2gcdScw4oykP/1eAV6Src+h0Y3ndjiHDWLTl Fx8upGS8SvOILxIHJy+liTVMdYvKoGFCENcrG/i2RLNkSH4cH4pcyhxafoWrk6BE MXakUvMcqHbWmTmzIXW7/O2GkYJvrsxf4LFQvXnpdzYeuZPs9ofursOtYNANPCGo cdpcAWV19pAxWyGUKSQLeLxpTQqSTzfRFr4el2cHZvWIJTGt++JxqX1pLyC3vWYm wVXBKwRJyEO3JN+L++BwC+i0YXrs3/kLIh2CjRxh+GAeHluNqo6/dy/AoTnfeVy1 EonrcNS0o2ugs5Qb6j6zEF9ye0ihl96rCFeS3E/q0mv9ocgPU+xgW63fQCEQZbyP maMExL0oHjmLFjk02P3rZN+YVDkfgbNN4RJsbrMkI8gNlW2j6BjSgpT49mdMjXEX raG8rQeYMyjP8nYg7wE/txoztZ/vaZVNGPe3L8xOtBf4xzMzZyZiObHnZ1psXy4G sjmyCTovw9tt2vZn6VdV5R16P3FEc5yEEJ7OXYNiy2z2ccjsl+mWEDD8fjkf1OOI gdwnH9RcXqCNruXsxbjn0/Hqo2sIWCW5DILbyjLtWnWd0NwYNR9kQHLEm2+Cq1Ry jURYVYXQi4lxvwtZg/wpTMqtu54AKDGmePNIycRcOqrDb/vHmgDWUZmqwJpbpk6o Pn+aoy0z49d0tuOtetgD =GhIN -----END PGP SIGNATURE----- --pe5ihs3jm7vsteRM1nr7meDkuxW71gUFL-- --===============1617904076== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx --===============1617904076==--