From mboxrd@z Thu Jan 1 00:00:00 1970 Received: by 10.182.158.201 with SMTP id ww9csp3378023obb; Wed, 2 Dec 2015 14:55:45 -0800 (PST) X-Received: by 10.55.215.137 with SMTP id t9mr7141931qkt.23.1449096945281; Wed, 02 Dec 2015 14:55:45 -0800 (PST) Return-Path: Received: from lists.gnu.org (lists.gnu.org. [2001:4830:134:3::11]) by mx.google.com with ESMTPS id o190si5832045qhb.8.2015.12.02.14.55.45 for (version=TLS1 cipher=AES128-SHA bits=128/128); Wed, 02 Dec 2015 14:55:45 -0800 (PST) Received-SPF: pass (google.com: domain of qemu-arm-bounces+alex.bennee=linaro.org@nongnu.org designates 2001:4830:134:3::11 as permitted sender) client-ip=2001:4830:134:3::11; Authentication-Results: mx.google.com; spf=pass (google.com: domain of qemu-arm-bounces+alex.bennee=linaro.org@nongnu.org designates 2001:4830:134:3::11 as permitted sender) smtp.mailfrom=qemu-arm-bounces+alex.bennee=linaro.org@nongnu.org; dkim=fail header.i=@gmail.com; dmarc=fail (p=NONE dis=NONE) header.from=gmail.com Received: from localhost ([::1]:60679 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a4GJM-00066O-Us for alex.bennee@linaro.org; Wed, 02 Dec 2015 17:55:44 -0500 Received: from eggs.gnu.org ([2001:4830:134:3::10]:44140) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a4GJL-00066J-4b for qemu-arm@nongnu.org; Wed, 02 Dec 2015 17:55:44 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1a4GJF-0002BY-Aw for qemu-arm@nongnu.org; Wed, 02 Dec 2015 17:55:43 -0500 Received: from mail-qg0-x233.google.com ([2607:f8b0:400d:c04::233]:34747) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a4GJF-0002BT-6v; Wed, 02 Dec 2015 17:55:37 -0500 Received: by qgeb1 with SMTP id b1so47231449qge.1; Wed, 02 Dec 2015 14:55:36 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=message-id:date:from:user-agent:mime-version:to:cc:subject :references:in-reply-to:content-type:content-transfer-encoding; bh=tR5eBjEEz7Eji7087h0YtKsJk7Bx6W+1fjET3IQhOlo=; b=oWw7NT1u1zPgGO7nwpY/iLsrzZ+xDuIesrwcAYejgb0wZTSVFTdG+Kb04woYd6YMt8 4IEzM9G8/2x77hKsV1yT07ojvMqGA8Iv2JwPt3N5KFIYpf6qHAnw4kkJfUcZRwZ5c0hp dFfuF1qjqQoKA6PgQpKpzn3oWB6w5n1UZ2gu9h3tWm6F9/lpOvWfo9yBIzFoMfOuJWam 2t46VU11Pf1vy6F4tLr4nRD7a5fn7bWx6B6bms6oWAPiQnjvekk+j9BQittMqFLE97tP D3yUBqCcY2N+iQW8u2kJa2++LrJ7/z1mVq8wZXG5ZrKqV0OFaOkiWb5L7uhleur/ST49 LlhQ== X-Received: by 10.140.44.9 with SMTP id f9mr7277600qga.69.1449096936776; Wed, 02 Dec 2015 14:55:36 -0800 (PST) Received: from [10.142.1.2] (ool-182df582.dyn.optonline.net. [24.45.245.130]) by smtp.gmail.com with ESMTPSA id s131sm2132519qhs.11.2015.12.02.14.55.36 (version=TLSv1/SSLv3 cipher=OTHER); Wed, 02 Dec 2015 14:55:36 -0800 (PST) Message-ID: <565F76E7.4060305@gmail.com> Date: Wed, 02 Dec 2015 17:55:35 -0500 From: Michael Davidsaver User-Agent: Mozilla/5.0 (X11; Linux i686; rv:31.0) Gecko/20100101 Icedove/31.8.0 MIME-Version: 1.0 To: Peter Maydell References: <1447031505-12477-1-git-send-email-mdavidsaver@gmail.com> <1447031505-12477-5-git-send-email-mdavidsaver@gmail.com> In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-detected-operating-system: by eggs.gnu.org: GNU/Linux 2.2.x-3.x [generic] X-Received-From: 2607:f8b0:400d:c04::233 Cc: Peter Crosthwaite , qemu-arm@nongnu.org, QEMU Developers Subject: Re: [Qemu-arm] [PATCH 04/18] armv7m: Explicit error for bad vector table X-BeenThere: qemu-arm@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-arm-bounces+alex.bennee=linaro.org@nongnu.org Sender: qemu-arm-bounces+alex.bennee=linaro.org@nongnu.org X-TUID: xYGAOB+jisRT On 11/17/2015 12:33 PM, Peter Maydell wrote: > On 9 November 2015 at 01:11, Michael Davidsaver wrote: >> Give an explicit error and abort when a load >> from VECBASE fails. Otherwise would likely >> jump to 0, which for v7-m holds the reset stack >> pointer address. >> >> Signed-off-by: Michael Davidsaver >> --- >> target-arm/helper.c | 12 +++++++++++- >> 1 file changed, 11 insertions(+), 1 deletion(-) >> >> diff --git a/target-arm/helper.c b/target-arm/helper.c >> index 4178400..1d7ac43 100644 >> --- a/target-arm/helper.c >> +++ b/target-arm/helper.c >> @@ -5496,7 +5496,17 @@ void arm_v7m_cpu_do_interrupt(CPUState *cs) >> /* Clear IT bits */ >> env->condexec_bits = 0; >> env->regs[14] = lr; >> - addr = ldl_phys(cs->as, env->v7m.vecbase + env->v7m.exception * 4); >> + { >> + MemTxResult result; >> + addr = address_space_ldl(cs->as, >> + env->v7m.vecbase + env->v7m.exception * 4, >> + MEMTXATTRS_UNSPECIFIED, &result); >> + if (result != MEMTX_OK) { >> + cpu_abort(cs, "Failed to read from exception vector table " >> + "entry %08x\n", >> + env->v7m.vecbase + env->v7m.exception * 4); >> + } >> + } > > The behaviour on a failed vector table read is actually architecturally > specified: we should take a nested exception (escalated to HardFault). > If it happens while we're trying to take a HardFault in the first place > then we go into Lockup (where the CPU sits around repeatedly trying > to execute an instruction at 0xFFFFFFFE; it is technically possible > to get back out of Lockup by taking an NMI or a system reset). > > That said, trying to get nested exceptions and priority escalation > right is fairly involved, and implementing lockup is both involved > and an exercise in pointlessness. So I think this code is an > improvement overall. This is my thinking as well. One point against it is that abort() is inconvenient when using '-gdb'. I'm not sure if there is something else which could be done (cpu halt?). > I would suggest some small changes, though: > > (1) factor this out into its own function, something like: > static uint32_t v7m_read_vector(CPUARMState *env, int excnum) > so the calling code can just do > addr = v7m_read_vector(env, env->v7m.exception); > (2) use a local variable for "env->v7m.vecbase + excnum * 4" > rather than calculating it twice Done. From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:44155) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1a4GJN-00066k-DH for qemu-devel@nongnu.org; Wed, 02 Dec 2015 17:55:46 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1a4GJM-0002CO-Ab for qemu-devel@nongnu.org; Wed, 02 Dec 2015 17:55:45 -0500 Message-ID: <565F76E7.4060305@gmail.com> Date: Wed, 02 Dec 2015 17:55:35 -0500 From: Michael Davidsaver MIME-Version: 1.0 References: <1447031505-12477-1-git-send-email-mdavidsaver@gmail.com> <1447031505-12477-5-git-send-email-mdavidsaver@gmail.com> In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 04/18] armv7m: Explicit error for bad vector table List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell Cc: Peter Crosthwaite , qemu-arm@nongnu.org, QEMU Developers On 11/17/2015 12:33 PM, Peter Maydell wrote: > On 9 November 2015 at 01:11, Michael Davidsaver wrote: >> Give an explicit error and abort when a load >> from VECBASE fails. Otherwise would likely >> jump to 0, which for v7-m holds the reset stack >> pointer address. >> >> Signed-off-by: Michael Davidsaver >> --- >> target-arm/helper.c | 12 +++++++++++- >> 1 file changed, 11 insertions(+), 1 deletion(-) >> >> diff --git a/target-arm/helper.c b/target-arm/helper.c >> index 4178400..1d7ac43 100644 >> --- a/target-arm/helper.c >> +++ b/target-arm/helper.c >> @@ -5496,7 +5496,17 @@ void arm_v7m_cpu_do_interrupt(CPUState *cs) >> /* Clear IT bits */ >> env->condexec_bits = 0; >> env->regs[14] = lr; >> - addr = ldl_phys(cs->as, env->v7m.vecbase + env->v7m.exception * 4); >> + { >> + MemTxResult result; >> + addr = address_space_ldl(cs->as, >> + env->v7m.vecbase + env->v7m.exception * 4, >> + MEMTXATTRS_UNSPECIFIED, &result); >> + if (result != MEMTX_OK) { >> + cpu_abort(cs, "Failed to read from exception vector table " >> + "entry %08x\n", >> + env->v7m.vecbase + env->v7m.exception * 4); >> + } >> + } > > The behaviour on a failed vector table read is actually architecturally > specified: we should take a nested exception (escalated to HardFault). > If it happens while we're trying to take a HardFault in the first place > then we go into Lockup (where the CPU sits around repeatedly trying > to execute an instruction at 0xFFFFFFFE; it is technically possible > to get back out of Lockup by taking an NMI or a system reset). > > That said, trying to get nested exceptions and priority escalation > right is fairly involved, and implementing lockup is both involved > and an exercise in pointlessness. So I think this code is an > improvement overall. This is my thinking as well. One point against it is that abort() is inconvenient when using '-gdb'. I'm not sure if there is something else which could be done (cpu halt?). > I would suggest some small changes, though: > > (1) factor this out into its own function, something like: > static uint32_t v7m_read_vector(CPUARMState *env, int excnum) > so the calling code can just do > addr = v7m_read_vector(env, env->v7m.exception); > (2) use a local variable for "env->v7m.vecbase + excnum * 4" > rather than calculating it twice Done.