From owner-soc-status@freebsd.org Mon Jul 13 20:34:35 2015 Return-Path: Delivered-To: soc-status@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id A73E299C9E0 for ; Mon, 13 Jul 2015 20:34:35 +0000 (UTC) (envelope-from mihai.carabas@gmail.com) Received: from mail-wi0-x22b.google.com (mail-wi0-x22b.google.com [IPv6:2a00:1450:400c:c05::22b]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority G2" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 3E97B10A8 for ; Mon, 13 Jul 2015 20:34:35 +0000 (UTC) (envelope-from mihai.carabas@gmail.com) Received: by wiga1 with SMTP id a1so80413722wig.0 for ; Mon, 13 Jul 2015 13:34:33 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; bh=96beqTTNRuc33Jw/zcCEnkusU3G2yFCzL4/JcEZJ6eE=; b=Vfy0tSOJReNhYRMA0bcnOR1zR3Ij4bQCk+gUGmsjnaA4CjOn8r4HQa9FgdBStxdBbr uFND7wHspEgpKc2PLzeXZHy9NWGxpHR7IcQbKa4uGIat5LIOeahETpZXemK4RugkChwJ SSLn37jSb4s/KKPz9jHLJwnvrMcDtMlfT7BOdSAldvUwFz139/OTSkMjmJj/SSki3rZl kxp+QKT8+ZBLUaPnFv5cNkmrvK40+v5GaF5qqMbx3k0P5wbnsoO1q19OVx6mbnHi8LH1 H41PlmK6UrUV9FyIdUSVwSpuwxsHCuK53ZydWiermRdvXO+zx5wspUVa7CAC3b08t0JU mjFg== MIME-Version: 1.0 X-Received: by 10.194.77.179 with SMTP id t19mr73408789wjw.30.1436819673425; Mon, 13 Jul 2015 13:34:33 -0700 (PDT) Received: by 10.28.21.134 with HTTP; Mon, 13 Jul 2015 13:34:33 -0700 (PDT) In-Reply-To: References: Date: Mon, 13 Jul 2015 23:34:33 +0300 Message-ID: Subject: Re: [GSOC] bhyve port on ARM - weekly status report From: Mihai Carabas To: soc-status@freebsd.org Content-Type: text/plain; charset=UTF-8 X-Content-Filtered-By: Mailman/MimeDel 2.1.20 X-BeenThere: soc-status@freebsd.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: Summer of Code Status Reports and Discussion List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 13 Jul 2015 20:34:35 -0000 Hi everyone, >> In the last week I've started writing low-level ASM code which is > saving/restoring the guest/host state. I've created a logic for a full > pass from the host to the guest and back [1]: > - when the host is calling the "HYP" instruction, causes a trap in > hyp-mode where it can run different functions specific to this mode. An > example is hyp_enter_guest function which saves the host state, loads the > guest state and jumps to the guest pc with an eret instruction (back to svc > mode basically) > - when the guest executes an instruction that causes a trap (e.g. wfi > which was configured by us) the code is saving the guest state, loading the > host state and than is returning to host svc-mode with the exception of the > guest to manage it > - I've also implemented a call to panic function in host if the abort > exceptions in hyp-mode are raised. > > There are still some configuration registers that need to be tackled (like > VCTBR - defines the guest configuration for the stage 2 address > translation). > > In the next week I have to write the C code which is calling the ASM > methods in arm_vmrun and the init code which allocates the context > structure (arm_vminit) and maps it in HYP-Mode. I will try to integrate > these operations with bhyve userspace tools to ease the development. > > In the last week I've written the glue C code which is calling the ASM methods to run a VM. The logic flow starts from userspace from bhyvearm program (it's a duplicate of the normal bhyve due to its high dependency of x86) using ioctl's to run the VM. I've also written a simplified bhyveloadarm which maps the desired memory from the GPA indicated as a parameter (here I had to write the entire flow from userspace to kernel because there were empty functions - the MMAP ioctl). Than it loads the raw kernel image in the memory at a given address. In the end I set up the PC using also an ioctl. I've created a new ramdisk containing the bhyvearm* programs and the dynamic libraries they depend on. Then I've managed to make bhyveloadarm to copy a file which contains the bytecode for the "wfi" instruction in the guest memory and now I'm playing with bhyvearm to make the "guest" run that instruction. I'm investigating a bug in restoring the host state after the VM has run (the VM throws a data abort exception because I haven't configured the VTCR yet and I'm trying to send this exception back in bhyvearm userspace). Thanks, Mihai