From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-9.0 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 27756C10F14 for ; Fri, 12 Apr 2019 12:06:48 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 002822082E for ; Fri, 12 Apr 2019 12:06:47 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727403AbfDLMGh (ORCPT ); Fri, 12 Apr 2019 08:06:37 -0400 Received: from szxga04-in.huawei.com ([45.249.212.190]:6175 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1727104AbfDLMGf (ORCPT ); Fri, 12 Apr 2019 08:06:35 -0400 Received: from DGGEMS405-HUB.china.huawei.com (unknown [172.30.72.60]) by Forcepoint Email with ESMTP id 283C5269CD2F03AD3FD8; Fri, 12 Apr 2019 20:06:30 +0800 (CST) Received: from linux-ibm.site (10.175.102.37) by DGGEMS405-HUB.china.huawei.com (10.3.19.205) with Microsoft SMTP Server id 14.3.408.0; Fri, 12 Apr 2019 20:06:20 +0800 From: Xiongfeng Wang To: , , , , CC: , Subject: [RFC PATCH 2/3] sdei: enable dbg in '_sdei_handler' Date: Fri, 12 Apr 2019 20:04:58 +0800 Message-ID: <1555070699-3685-3-git-send-email-wangxiongfeng2@huawei.com> X-Mailer: git-send-email 1.7.12.4 In-Reply-To: <1555070699-3685-1-git-send-email-wangxiongfeng2@huawei.com> References: <1555070699-3685-1-git-send-email-wangxiongfeng2@huawei.com> MIME-Version: 1.0 Content-Type: text/plain X-Originating-IP: [10.175.102.37] X-CFilter-Loop: Reflected Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org When we monitor a sdei_event callback using 'kprobe', the singlestep handler can not be called because dbg is masked in sdei_handler. This patch enable dbg in '_sdei_handler'. When SDEI events interrupt the userspace, 'vbar_el1' contains 'tramp_vectors' if CONFIG_UNMAP_KERNEL_AT_EL0 is enabled. So we need to restore 'vbar_el1' with kernel vectors, otherwise we will go to the wrong place when brk exception or dbg exception happens. SDEI events may interrupt 'kernel_entry' before we save 'spsr_el1' and 'elr_el1', and dbg exception will corrupts these two registers. So we also need to save and restore these two registers. If SDEI events interrupt an instruction being singlestepped, the instruction in '_sdei_handler' will begin to be singlestepped after we enable dbg. So we need to disable singlestep in the beginning of _sdei_handler if we find out we interrupt a singlestep process. Signed-off-by: Xiongfeng Wang --- arch/arm64/include/asm/debug-monitors.h | 1 + arch/arm64/kernel/debug-monitors.c | 8 ++++++ arch/arm64/kernel/sdei.c | 43 ++++++++++++++++++++++++++------- 3 files changed, 43 insertions(+), 9 deletions(-) diff --git a/arch/arm64/include/asm/debug-monitors.h b/arch/arm64/include/asm/debug-monitors.h index a44cf52..a730266 100644 --- a/arch/arm64/include/asm/debug-monitors.h +++ b/arch/arm64/include/asm/debug-monitors.h @@ -121,6 +121,7 @@ enum dbg_active_el { void user_fastforward_single_step(struct task_struct *task); void kernel_enable_single_step(struct pt_regs *regs); +void kernel_enable_single_step_noregs(void); void kernel_disable_single_step(void); int kernel_active_single_step(void); diff --git a/arch/arm64/kernel/debug-monitors.c b/arch/arm64/kernel/debug-monitors.c index d7bb6ae..d6074f4 100644 --- a/arch/arm64/kernel/debug-monitors.c +++ b/arch/arm64/kernel/debug-monitors.c @@ -404,6 +404,14 @@ void kernel_enable_single_step(struct pt_regs *regs) } NOKPROBE_SYMBOL(kernel_enable_single_step); +void kernel_enable_single_step_noregs(void) +{ + WARN_ON(!irqs_disabled()); + mdscr_write(mdscr_read() | DBG_MDSCR_SS); + enable_debug_monitors(DBG_ACTIVE_EL1); +} +NOKPROBE_SYMBOL(kernel_enable_single_step_noregs); + void kernel_disable_single_step(void) { WARN_ON(!irqs_disabled()); diff --git a/arch/arm64/kernel/sdei.c b/arch/arm64/kernel/sdei.c index ea94cf8..9dd9cf6 100644 --- a/arch/arm64/kernel/sdei.c +++ b/arch/arm64/kernel/sdei.c @@ -9,6 +9,7 @@ #include #include +#include #include #include #include @@ -176,6 +177,8 @@ unsigned long sdei_arch_get_entry_point(int conduit) } +extern char vectors[]; /* kernel exception vectors */ + /* * __sdei_handler() returns one of: * SDEI_EV_HANDLED - success, return to the interrupted context. @@ -189,8 +192,10 @@ static __kprobes unsigned long _sdei_handler(struct pt_regs *regs, int i, err = 0; int clobbered_registers = 4; u64 elr = read_sysreg(elr_el1); + u64 spsr = read_sysreg(spsr_el1); u32 kernel_mode = read_sysreg(CurrentEL) | 1; /* +SPSel */ unsigned long vbar = read_sysreg(vbar_el1); + int ss_active = 0; if (arm64_kernel_unmapped_at_el0()) clobbered_registers++; @@ -207,19 +212,39 @@ static __kprobes unsigned long _sdei_handler(struct pt_regs *regs, */ __uaccess_enable_hw_pan(); + /* + * Enable dbg here so that we can kprobe a sdei event handler. Before we + * enable dbg, we need to restore vbar_el1 with kernel vectors + */ +#ifdef CONFIG_UNMAP_KERNEL_AT_EL0 + write_sysreg(vectors, vbar_el1); + isb(); +#endif + ss_active = kernel_active_single_step(); + if (ss_active) + kernel_disable_single_step(); + local_dbg_enable(); + err = sdei_event_handler(regs, arg); + + local_dbg_disable(); + if (ss_active) + kernel_enable_single_step_noregs(); + + /* + * brk exception will corrupt elr_el1 and spsr_el1, and trust firmware + * doesn't save it for us. So we need to restore these two registers + * after 'sdei_event_handler'. + */ + write_sysreg(elr, elr_el1); + write_sysreg(spsr, spsr_el1); +#ifdef CONFIG_UNMAP_KERNEL_AT_EL0 + write_sysreg(vbar, vbar_el1); +#endif + if (err) return SDEI_EV_FAILED; - if (elr != read_sysreg(elr_el1)) { - /* - * We took a synchronous exception from the SDEI handler. - * This could deadlock, and if you interrupt KVM it will - * hyp-panic instead. - */ - pr_warn("unsafe: exception during handler\n"); - } - mode = regs->pstate & (PSR_MODE32_BIT | PSR_MODE_MASK); /* -- 1.7.12.4