From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:45974) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a9s7H-0001Oq-C9 for qemu-devel@nongnu.org; Fri, 18 Dec 2015 05:18:28 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1a9s7E-0001ZP-Pt for qemu-devel@nongnu.org; Fri, 18 Dec 2015 05:18:27 -0500 Received: from e06smtp07.uk.ibm.com ([195.75.94.103]:53178) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a9s7E-0001ZB-HJ for qemu-devel@nongnu.org; Fri, 18 Dec 2015 05:18:24 -0500 Received: from localhost by e06smtp07.uk.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Fri, 18 Dec 2015 10:18:21 -0000 From: Greg Kurz Date: Fri, 18 Dec 2015 11:18:16 +0100 Message-ID: <20151218101815.26409.55642.stgit@bahia.huguette.org> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Subject: [Qemu-devel] [PATCH 0/7] target-ppc: endian fixes for KVM and gdbstub List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Alexander Graf Cc: Paolo Bonzini , qemu-ppc@nongnu.org, Anton Blanchard , qemu-devel@nongnu.org Hi, This series is a sequel to Anton's tentative at bringing VSX support in our gdbstub: http://patchwork.ozlabs.org/patch/453758/ Indeed, FP, SPE and Altivec registers need to be copied to memory with the appropriate ordering, like we already do for core registers. This series reuses the maybe_bswap_register() helper to do the job, since it already handles the user mode case where the target endianness is known at build time and don't need byteswap. This is covered by patches 2 to 6. I also found more serious issues that probably break more than gdbstub. First one is a bug in KVM, that completely breaks the KVM_GET_ONE_REG and KVM_SET_ONE_REG ioctls for Altivec registers. I've already sent a patch: http://patchwork.ozlabs.org/patch/557568/ Second one is a bug in QEMU that breaks synchronisation with KVM for FP, Altivec and VSX registers on little-endian hosts. I pushed the fix to patch 1 since it is needed for the gdbstub fixes to actually work, but it could even be handled separately. And finally, patch 7 is Anton's + the byteswapping for little-endian guests. Cheers. --- Anton Blanchard (1): target-ppc: gdbstub: Add VSX support Greg Kurz (6): target-ppc: kvm: fix floating point registers sync on little-endian hosts target-ppc: rename and export maybe_bswap_register() target-ppc: gdbstub: fix float registers for little-endian guests target-ppc: gdbstub: introduce avr_need_swap() target-ppc: gdbstub: fix altivec registers for little-endian guests target-ppc: gdbstub: fix spe registers for little-endian guests configure | 6 ++- gdb-xml/power-vsx.xml | 44 +++++++++++++++++++++++ target-ppc/cpu.h | 1 + target-ppc/gdbstub.c | 10 +++-- target-ppc/kvm.c | 12 ++++++ target-ppc/translate_init.c | 84 +++++++++++++++++++++++++++++++++++-------- 6 files changed, 134 insertions(+), 23 deletions(-) create mode 100644 gdb-xml/power-vsx.xml