From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH] add icswx support From: Michael Ellerman To: "Tseng-Hui (Frank) Lin" In-Reply-To: <1272405410.6329.16.camel@flin.austin.ibm.com> References: <1272060275.6329.13.camel@flin.austin.ibm.com> <1272070551.2330.213.camel@pasglop> <1272405410.6329.16.camel@flin.austin.ibm.com> Content-Type: multipart/signed; micalg="pgp-sha1"; protocol="application/pgp-signature"; boundary="=-3a0/Zc9r0AEVl6unjoFy" Date: Fri, 30 Apr 2010 23:35:22 +1000 Message-ID: <1272634522.9367.0.camel@concordia> Mime-Version: 1.0 Cc: linuxppc-dev@ozlabs.org Reply-To: michael@ellerman.id.au List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , --=-3a0/Zc9r0AEVl6unjoFy Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On Tue, 2010-04-27 at 16:56 -0500, Tseng-Hui (Frank) Lin wrote: > On Sat, 2010-04-24 at 10:55 +1000, Benjamin Herrenschmidt wrote: > > On Fri, 2010-04-23 at 17:04 -0500, Tseng-Hui (Frank) Lin wrote: > > > Add Power7 icswx co-processor instruction support. > >=20 > > Please provide a -much- more detailed explanation of what it is, what i= t > > does and why it requires hooking into the MMU context switch code. _I_ > > know these things but nobody else on the list does which limits the > > ability of people to review your patch. > > >=20 > icswx is a PowerPC co-processor instruction to send data to a=20 > co-processor. On Book-S processors the LPAR_ID and process ID (PID) of=20 > the owning process are registered in the window context of the > co-processor at initial time. When the icswx instruction is executed, > the L2 generates a cop-reg transaction on PowerBus. The transaction has > no address and the processor does not perform an MMU access to=20 > authenticate the transaction. The coprocessor compares the LPAR_ID and > the PID included in the transaction and the LPAR_ID and PID held in the > window context to determine if the process is authorized to generate the > transaction. How does userspace discover that there are coprocessors to send requests to? And how does the coprocessor send results back to the process? cheers --=-3a0/Zc9r0AEVl6unjoFy Content-Type: application/pgp-signature; name="signature.asc" Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEABECAAYFAkva3JYACgkQdSjSd0sB4dKrbACglS1/4p4gJG0+flRwII/3Kt5S MdMAoI0PocSRyBCVvejqvw3Lhwh+SsWi =hTea -----END PGP SIGNATURE----- --=-3a0/Zc9r0AEVl6unjoFy--