From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Axtens Subject: Re: [PATCH] KVM: PPC: Exit guest upon fatal machine check exception Date: Fri, 13 Nov 2015 08:37:09 +1100 Message-ID: <87pozezypm.fsf@gamma.ozlabs.ibm.com> References: <20151111165845.3721.98296.stgit@aravindap> <876118ymy4.fsf@gamma.ozlabs.ibm.com> <20151112033816.GJ5852@voom.redhat.com> <87bnazizkp.fsf@gamma.ozlabs.ibm.com> <5644CABE.3010905@linux.vnet.ibm.com> Mime-Version: 1.0 Content-Type: multipart/signed; boundary="=-=-="; micalg=pgp-sha512; protocol="application/pgp-signature" Cc: David Gibson , kvm@vger.kernel.org, michaele@au1.ibm.com, mahesh@linux.vnet.ibm.com, agraf@suse.de, kvm-ppc@vger.kernel.org, linuxppc-dev@ozlabs.org To: Aravinda Prasad Return-path: Received: from mail-pa0-f41.google.com ([209.85.220.41]:35742 "EHLO mail-pa0-f41.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753918AbbKLVh0 (ORCPT ); Thu, 12 Nov 2015 16:37:26 -0500 Received: by pasz6 with SMTP id z6so79445348pas.2 for ; Thu, 12 Nov 2015 13:37:26 -0800 (PST) In-Reply-To: <5644CABE.3010905@linux.vnet.ibm.com> Sender: kvm-owner@vger.kernel.org List-ID: --=-=-= Content-Type: text/plain Aravinda Prasad writes: >> Yeah, it would be good not to break this. > > I am not familiar with CAPI. Does this affect CAPI? When a CAPI card experiences an EEH event, any cache lines it holds are filled with SUEs (Special UEs, interpreted by the kernel the same as regular UEs). When these are read, we get an MCE. Currently CAPI does not support virtualisation, but that is actively being worked on. There's a _very_ good chance it will then be backported to various distros, which could have old qemu. Therefore, I'd ideally like to make sure UEs in KVM guests work properly and continue to do so in all combinations of kernel and qemu. I'm following up the email from Mahesh that you linked: I'm not sure I quite follow his logic so I'll try to make sense of that and then get back to you. Regards, Daniel --=-=-= Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 Comment: GPGTools - https://gpgtools.org iQIcBAEBCgAGBQJWRQaFAAoJEPC3R3P2I92FBnoQALmoi05ujJ+iZXqG5UXD8JFb g+EB/pgDCARDhqUEal1jgy2bu1Je8c+YvcX7uhTFkDEMQ7khBFoklGue5kgCSJvh ijyMOxtA9ab+VD/p7SHMhVj5PLDVt8pGEBbd3MErxMooGuesCm9kH4tcQGrBh99l in7Hbx7Toko4oq7uc0U4Zo/pp2KwZSuKx99++VITZQfh3sC5dSItmhVEqH46saIp fFeXA9dnQ+HMyqCTmN9Zdef78r4uQ+S9rzVxBDnTennCwSPlO2O2uusV93xr3Lwb UzkVh4ocN2bSM6zhmBj5Sdt9CWcg3APGn71nuay1sXTdU1EqYft6xxS18w2q9tyW AkfBFtFxW8W3Akx+FSGV9JvRIu4ky1DE1WpaxUcqFpySmZT2Bc9UzF4Rcd4Jlvd9 R6g0qFkwvpzKi6lJEWwgbBW26PEDyiy4PjBIPba9i/6AHb9xL4PNnHR5FvVDLZ2z 8b7h2TNEiQzSNHxDK36Wwe8New2pqt03/imsTwQecrwDl7ip/FBKZ0CogeZModZq Y00uC9d5d/ewuQepVn/O+gBfjer5ecrPYC0FfjG0fCCfe+Azw8JaGCqiHlis0LZu D7GhokiDGJst4i6joN6S7ps8fT4TKfx3LU/yg+Fw6wV5KLHNFWo/EACWq7qv5yLd x8kZ8Ig1/IBrNcs6Ir7Y =H92U -----END PGP SIGNATURE----- --=-=-=-- From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Axtens Date: Thu, 12 Nov 2015 21:37:09 +0000 Subject: Re: [PATCH] KVM: PPC: Exit guest upon fatal machine check exception Message-Id: <87pozezypm.fsf@gamma.ozlabs.ibm.com> MIME-Version: 1 Content-Type: multipart/mixed; boundary="=-=-=" List-Id: References: <20151111165845.3721.98296.stgit@aravindap> <876118ymy4.fsf@gamma.ozlabs.ibm.com> <20151112033816.GJ5852@voom.redhat.com> <87bnazizkp.fsf@gamma.ozlabs.ibm.com> <5644CABE.3010905@linux.vnet.ibm.com> In-Reply-To: <5644CABE.3010905@linux.vnet.ibm.com> To: Aravinda Prasad Cc: David Gibson , kvm@vger.kernel.org, michaele@au1.ibm.com, mahesh@linux.vnet.ibm.com, agraf@suse.de, kvm-ppc@vger.kernel.org, linuxppc-dev@ozlabs.org --=-=-= Content-Type: text/plain Aravinda Prasad writes: >> Yeah, it would be good not to break this. > > I am not familiar with CAPI. Does this affect CAPI? When a CAPI card experiences an EEH event, any cache lines it holds are filled with SUEs (Special UEs, interpreted by the kernel the same as regular UEs). When these are read, we get an MCE. Currently CAPI does not support virtualisation, but that is actively being worked on. There's a _very_ good chance it will then be backported to various distros, which could have old qemu. Therefore, I'd ideally like to make sure UEs in KVM guests work properly and continue to do so in all combinations of kernel and qemu. I'm following up the email from Mahesh that you linked: I'm not sure I quite follow his logic so I'll try to make sense of that and then get back to you. Regards, Daniel --=-=-= Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 Comment: GPGTools - https://gpgtools.org iQIcBAEBCgAGBQJWRQaFAAoJEPC3R3P2I92FBnoQALmoi05ujJ+iZXqG5UXD8JFb g+EB/pgDCARDhqUEal1jgy2bu1Je8c+YvcX7uhTFkDEMQ7khBFoklGue5kgCSJvh ijyMOxtA9ab+VD/p7SHMhVj5PLDVt8pGEBbd3MErxMooGuesCm9kH4tcQGrBh99l in7Hbx7Toko4oq7uc0U4Zo/pp2KwZSuKx99++VITZQfh3sC5dSItmhVEqH46saIp fFeXA9dnQ+HMyqCTmN9Zdef78r4uQ+S9rzVxBDnTennCwSPlO2O2uusV93xr3Lwb UzkVh4ocN2bSM6zhmBj5Sdt9CWcg3APGn71nuay1sXTdU1EqYft6xxS18w2q9tyW AkfBFtFxW8W3Akx+FSGV9JvRIu4ky1DE1WpaxUcqFpySmZT2Bc9UzF4Rcd4Jlvd9 R6g0qFkwvpzKi6lJEWwgbBW26PEDyiy4PjBIPba9i/6AHb9xL4PNnHR5FvVDLZ2z 8b7h2TNEiQzSNHxDK36Wwe8New2pqt03/imsTwQecrwDl7ip/FBKZ0CogeZModZq Y00uC9d5d/ewuQepVn/O+gBfjer5ecrPYC0FfjG0fCCfe+Azw8JaGCqiHlis0LZu D7GhokiDGJst4i6joN6S7ps8fT4TKfx3LU/yg+Fw6wV5KLHNFWo/EACWq7qv5yLd x8kZ8Ig1/IBrNcs6Ir7Y =H92U -----END PGP SIGNATURE----- --=-=-=--