From patchwork Mon Mar 11 09:36:59 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Peter Xu X-Patchwork-Id: 10847129 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 8AEA31823 for ; Mon, 11 Mar 2019 09:37:30 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 7869528FB5 for ; Mon, 11 Mar 2019 09:37:30 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 6BB2328FBD; Mon, 11 Mar 2019 09:37:30 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI autolearn=unavailable version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id DFF6928FB5 for ; Mon, 11 Mar 2019 09:37:29 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727225AbfCKJhX (ORCPT ); Mon, 11 Mar 2019 05:37:23 -0400 Received: from mx1.redhat.com ([209.132.183.28]:40352 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726969AbfCKJhX (ORCPT ); Mon, 11 Mar 2019 05:37:23 -0400 Received: from smtp.corp.redhat.com (int-mx05.intmail.prod.int.phx2.redhat.com [10.5.11.15]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 71CFD30821E2; Mon, 11 Mar 2019 09:37:22 +0000 (UTC) Received: from xz-x1.nay.redhat.com (dhcp-14-116.nay.redhat.com [10.66.14.116]) by smtp.corp.redhat.com (Postfix) with ESMTP id 595225D706; Mon, 11 Mar 2019 09:37:15 +0000 (UTC) From: Peter Xu To: linux-kernel@vger.kernel.org Cc: Paolo Bonzini , Hugh Dickins , Luis Chamberlain , Maxime Coquelin , kvm@vger.kernel.org, Jerome Glisse , Pavel Emelyanov , Johannes Weiner , peterx@redhat.com, Martin Cracauer , Denis Plotnikov , linux-mm@kvack.org, Marty McFadden , Maya Gokhale , Mike Kravetz , Andrea Arcangeli , Mike Rapoport , Kees Cook , Mel Gorman , "Kirill A . Shutemov" , linux-fsdevel@vger.kernel.org, "Dr . David Alan Gilbert" , Andrew Morton Subject: [PATCH 1/3] userfaultfd/sysctl: introduce unprivileged_userfaultfd Date: Mon, 11 Mar 2019 17:36:59 +0800 Message-Id: <20190311093701.15734-2-peterx@redhat.com> In-Reply-To: <20190311093701.15734-1-peterx@redhat.com> References: <20190311093701.15734-1-peterx@redhat.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.15 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.47]); Mon, 11 Mar 2019 09:37:22 +0000 (UTC) Sender: linux-fsdevel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-fsdevel@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Introduce a new sysctl called "vm.unprivileged_userfaultfd" that can be used to decide whether userfaultfd syscalls are allowed by unprivileged users. It'll allow three modes: - disabled: disallow unprivileged users to use uffd - enabled: allow unprivileged users to use uffd - kvm: allow unprivileged users to use uffd only if the user had enough permission to open /dev/kvm (this option only exists if the kernel turned on KVM). This patch only introduce the new interface but not yet applied it to the userfaultfd syscalls, which will be done in the follow up patch. Signed-off-by: Peter Xu --- fs/userfaultfd.c | 96 +++++++++++++++++++++++++++++++++++ include/linux/userfaultfd_k.h | 5 ++ init/Kconfig | 11 ++++ kernel/sysctl.c | 11 ++++ 4 files changed, 123 insertions(+) diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index 89800fc7dc9d..c2188464555a 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -29,6 +29,8 @@ #include #include #include +#include +#include static struct kmem_cache *userfaultfd_ctx_cachep __read_mostly; @@ -93,6 +95,95 @@ struct userfaultfd_wake_range { unsigned long len; }; +enum unprivileged_userfaultfd { + /* Disallow unprivileged users to use userfaultfd syscalls */ + UFFD_UNPRIV_DISABLED = 0, + /* Allow unprivileged users to use userfaultfd syscalls */ + UFFD_UNPRIV_ENABLED, +#if IS_ENABLED(CONFIG_KVM) + /* + * Allow unprivileged users to use userfaultfd syscalls only + * if the user had enough permission to open /dev/kvm + */ + UFFD_UNPRIV_KVM, +#endif + UFFD_UNPRIV_NUM, +}; + +static int unprivileged_userfaultfd __read_mostly; +static const char *unprivileged_userfaultfd_str[UFFD_UNPRIV_NUM] = { + "disabled", "enabled", +#if IS_ENABLED(CONFIG_KVM) + "kvm", +#endif +}; + +static int unprivileged_uffd_parse(char *buf, size_t size) +{ + int i; + + for (i = 0; i < UFFD_UNPRIV_NUM; i++) { + if (!strncmp(unprivileged_userfaultfd_str[i], buf, size)) { + unprivileged_userfaultfd = i; + return 0; + } + } + + return -EFAULT; +} + +static void unprivileged_uffd_dump(char *buf, size_t size) +{ + int i; + + *buf = 0x00; + for (i = 0; i < UFFD_UNPRIV_NUM; i++) { + if (i == unprivileged_userfaultfd) + strncat(buf, "[", size - strlen(buf)); + strncat(buf, unprivileged_userfaultfd_str[i], + size - strlen(buf)); + if (i == unprivileged_userfaultfd) + strncat(buf, "]", size - strlen(buf)); + strncat(buf, " ", size - strlen(buf)); + } + +} + +int proc_unprivileged_userfaultfd(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, + loff_t *ppos) +{ + struct ctl_table tmp_table = { .maxlen = 0 }; + int ret; + + if (write) { + tmp_table.maxlen = UFFD_UNPRIV_STRLEN; + tmp_table.data = kmalloc(UFFD_UNPRIV_STRLEN, GFP_KERNEL); + + ret = proc_dostring(&tmp_table, write, buffer, lenp, ppos); + if (ret) + goto out; + + ret = unprivileged_uffd_parse(tmp_table.data, + UFFD_UNPRIV_STRLEN); + } else { + /* Leave space for "[]" */ + int len = UFFD_UNPRIV_STRLEN * UFFD_UNPRIV_NUM + 2; + + tmp_table.maxlen = len; + tmp_table.data = kmalloc(len, GFP_KERNEL); + + unprivileged_uffd_dump(tmp_table.data, len); + + ret = proc_dostring(&tmp_table, write, buffer, lenp, ppos); + } + +out: + if (tmp_table.data) + kfree(tmp_table.data); + return ret; +} + static int userfaultfd_wake_function(wait_queue_entry_t *wq, unsigned mode, int wake_flags, void *key) { @@ -1955,6 +2046,11 @@ SYSCALL_DEFINE1(userfaultfd, int, flags) static int __init userfaultfd_init(void) { + char unpriv_uffd[UFFD_UNPRIV_STRLEN] = + CONFIG_USERFAULTFD_UNPRIVILEGED_DEFAULT; + + unprivileged_uffd_parse(unpriv_uffd, sizeof(unpriv_uffd)); + userfaultfd_ctx_cachep = kmem_cache_create("userfaultfd_ctx_cache", sizeof(struct userfaultfd_ctx), 0, diff --git a/include/linux/userfaultfd_k.h b/include/linux/userfaultfd_k.h index 37c9eba75c98..f53bc02ccffc 100644 --- a/include/linux/userfaultfd_k.h +++ b/include/linux/userfaultfd_k.h @@ -28,6 +28,11 @@ #define UFFD_SHARED_FCNTL_FLAGS (O_CLOEXEC | O_NONBLOCK) #define UFFD_FLAGS_SET (EFD_SHARED_FCNTL_FLAGS) +#define UFFD_UNPRIV_STRLEN 16 +int proc_unprivileged_userfaultfd(struct ctl_table *table, int write, + void __user *buffer, size_t *lenp, + loff_t *ppos); + extern vm_fault_t handle_userfault(struct vm_fault *vmf, unsigned long reason); extern ssize_t mcopy_atomic(struct mm_struct *dst_mm, unsigned long dst_start, diff --git a/init/Kconfig b/init/Kconfig index c9386a365eea..d90caa4fed17 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1512,6 +1512,17 @@ config USERFAULTFD Enable the userfaultfd() system call that allows to intercept and handle page faults in userland. +config USERFAULTFD_UNPRIVILEGED_DEFAULT + string "Default behavior for unprivileged userfault syscalls" + depends on USERFAULTFD + default "disabled" + help + Set this to "enabled" to allow userfaultfd syscalls from + unprivileged users. Set this to "disabled" to forbid + userfaultfd syscalls from unprivileged users. Set this to + "kvm" to forbid unpriviledged users but still allow users + who had enough permission to open /dev/kvm. + config ARCH_HAS_MEMBARRIER_CALLBACKS bool diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 7578e21a711b..5dc9f3d283dd 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -96,6 +96,9 @@ #ifdef CONFIG_LOCKUP_DETECTOR #include #endif +#ifdef CONFIG_USERFAULTFD +#include +#endif #if defined(CONFIG_SYSCTL) @@ -1704,6 +1707,14 @@ static struct ctl_table vm_table[] = { .extra1 = (void *)&mmap_rnd_compat_bits_min, .extra2 = (void *)&mmap_rnd_compat_bits_max, }, +#endif +#ifdef CONFIG_USERFAULTFD + { + .procname = "unprivileged_userfaultfd", + .maxlen = UFFD_UNPRIV_STRLEN, + .mode = 0644, + .proc_handler = proc_unprivileged_userfaultfd, + }, #endif { } }; From patchwork Mon Mar 11 09:37:00 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Peter Xu X-Patchwork-Id: 10847141 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 4E62D1823 for ; Mon, 11 Mar 2019 09:37:54 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 3C2BC28FD2 for ; Mon, 11 Mar 2019 09:37:54 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 2EEFE28FD4; Mon, 11 Mar 2019 09:37:54 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI autolearn=unavailable version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id D197528FD2 for ; Mon, 11 Mar 2019 09:37:53 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727309AbfCKJhf (ORCPT ); Mon, 11 Mar 2019 05:37:35 -0400 Received: from mx1.redhat.com ([209.132.183.28]:51420 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727263AbfCKJhe (ORCPT ); Mon, 11 Mar 2019 05:37:34 -0400 Received: from smtp.corp.redhat.com (int-mx05.intmail.prod.int.phx2.redhat.com [10.5.11.15]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 399093086265; Mon, 11 Mar 2019 09:37:34 +0000 (UTC) Received: from xz-x1.nay.redhat.com (dhcp-14-116.nay.redhat.com [10.66.14.116]) by smtp.corp.redhat.com (Postfix) with ESMTP id EAB725D705; Mon, 11 Mar 2019 09:37:22 +0000 (UTC) From: Peter Xu To: linux-kernel@vger.kernel.org Cc: Paolo Bonzini , Hugh Dickins , Luis Chamberlain , Maxime Coquelin , kvm@vger.kernel.org, Jerome Glisse , Pavel Emelyanov , Johannes Weiner , peterx@redhat.com, Martin Cracauer , Denis Plotnikov , linux-mm@kvack.org, Marty McFadden , Maya Gokhale , Mike Kravetz , Andrea Arcangeli , Mike Rapoport , Kees Cook , Mel Gorman , "Kirill A . Shutemov" , linux-fsdevel@vger.kernel.org, "Dr . David Alan Gilbert" , Andrew Morton Subject: [PATCH 2/3] kvm/mm: introduce MMF_USERFAULTFD_ALLOW flag Date: Mon, 11 Mar 2019 17:37:00 +0800 Message-Id: <20190311093701.15734-3-peterx@redhat.com> In-Reply-To: <20190311093701.15734-1-peterx@redhat.com> References: <20190311093701.15734-1-peterx@redhat.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.15 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.49]); Mon, 11 Mar 2019 09:37:34 +0000 (UTC) Sender: linux-fsdevel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-fsdevel@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Introduce a new MMF_USERFAULTFD_ALLOW flag and tag it upon the process memory address space as long as the process opened the /dev/kvm once. It'll be dropped automatically when fork() by MMF_INIT_TASK to reset the userfaultfd permission. Detecting the flag gives us a chance to open the green light for kvm upon using userfaultfd when we want to make sure all the existing kvm users will still be able to run their userspace programs without being affected by the new unprivileged userfaultfd switch. Suggested-by: Andrea Arcangeli Signed-off-by: Peter Xu --- include/linux/sched/coredump.h | 1 + virt/kvm/kvm_main.c | 7 +++++++ 2 files changed, 8 insertions(+) diff --git a/include/linux/sched/coredump.h b/include/linux/sched/coredump.h index ecdc6542070f..9f6e71182892 100644 --- a/include/linux/sched/coredump.h +++ b/include/linux/sched/coredump.h @@ -72,6 +72,7 @@ static inline int get_dumpable(struct mm_struct *mm) #define MMF_DISABLE_THP 24 /* disable THP for all VMAs */ #define MMF_OOM_VICTIM 25 /* mm is the oom victim */ #define MMF_OOM_REAP_QUEUED 26 /* mm was queued for oom_reaper */ +#define MMF_USERFAULTFD_ALLOW 27 /* allow userfaultfd syscall */ #define MMF_DISABLE_THP_MASK (1 << MMF_DISABLE_THP) #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK |\ diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index d237d3350a99..079f6ac00c36 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -3403,7 +3403,14 @@ static long kvm_dev_ioctl(struct file *filp, return r; } +static int kvm_dev_open(struct inode *inode, struct file *file) +{ + set_bit(MMF_USERFAULTFD_ALLOW, ¤t->mm->flags); + return 0; +} + static struct file_operations kvm_chardev_ops = { + .open = kvm_dev_open, .unlocked_ioctl = kvm_dev_ioctl, .llseek = noop_llseek, KVM_COMPAT(kvm_dev_ioctl), From patchwork Mon Mar 11 09:37:01 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Peter Xu X-Patchwork-Id: 10847137 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 8F6A714DE for ; Mon, 11 Mar 2019 09:37:49 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 7EFA528FBD for ; Mon, 11 Mar 2019 09:37:49 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 72C8628FD4; Mon, 11 Mar 2019 09:37:49 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI autolearn=unavailable version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 268C828FBD for ; Mon, 11 Mar 2019 09:37:49 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727355AbfCKJhm (ORCPT ); Mon, 11 Mar 2019 05:37:42 -0400 Received: from mx1.redhat.com ([209.132.183.28]:42232 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727323AbfCKJhl (ORCPT ); Mon, 11 Mar 2019 05:37:41 -0400 Received: from smtp.corp.redhat.com (int-mx05.intmail.prod.int.phx2.redhat.com [10.5.11.15]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id A7D2EC002965; Mon, 11 Mar 2019 09:37:41 +0000 (UTC) Received: from xz-x1.nay.redhat.com (dhcp-14-116.nay.redhat.com [10.66.14.116]) by smtp.corp.redhat.com (Postfix) with ESMTP id B3B195D705; Mon, 11 Mar 2019 09:37:34 +0000 (UTC) From: Peter Xu To: linux-kernel@vger.kernel.org Cc: Paolo Bonzini , Hugh Dickins , Luis Chamberlain , Maxime Coquelin , kvm@vger.kernel.org, Jerome Glisse , Pavel Emelyanov , Johannes Weiner , peterx@redhat.com, Martin Cracauer , Denis Plotnikov , linux-mm@kvack.org, Marty McFadden , Maya Gokhale , Mike Kravetz , Andrea Arcangeli , Mike Rapoport , Kees Cook , Mel Gorman , "Kirill A . Shutemov" , linux-fsdevel@vger.kernel.org, "Dr . David Alan Gilbert" , Andrew Morton Subject: [PATCH 3/3] userfaultfd: apply unprivileged_userfaultfd check Date: Mon, 11 Mar 2019 17:37:01 +0800 Message-Id: <20190311093701.15734-4-peterx@redhat.com> In-Reply-To: <20190311093701.15734-1-peterx@redhat.com> References: <20190311093701.15734-1-peterx@redhat.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.15 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.31]); Mon, 11 Mar 2019 09:37:41 +0000 (UTC) Sender: linux-fsdevel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-fsdevel@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Apply the unprivileged_userfaultfd check when doing userfaultfd syscall. We didn't check it in other paths of userfaultfd (e.g., the ioctl() path) because we don't want to drag down the fast path of userfaultfd, as suggested by Andrea. Suggested-by: Andrea Arcangeli Suggested-by: Mike Rapoport Signed-off-by: Peter Xu --- fs/userfaultfd.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index c2188464555a..effdcfc88629 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -951,6 +951,28 @@ void userfaultfd_unmap_complete(struct mm_struct *mm, struct list_head *uf) } } +/* Whether current process allows to use userfaultfd syscalls */ +static bool userfaultfd_allowed(void) +{ + bool allowed = false; + + switch (unprivileged_userfaultfd) { + case UFFD_UNPRIV_ENABLED: + allowed = true; + break; + case UFFD_UNPRIV_KVM: + allowed = !!test_bit(MMF_USERFAULTFD_ALLOW, + ¤t->mm->flags); + /* Fall through */ + case UFFD_UNPRIV_DISABLED: + allowed = allowed || ns_capable(current_user_ns(), + CAP_SYS_PTRACE); + break; + } + + return allowed; +} + static int userfaultfd_release(struct inode *inode, struct file *file) { struct userfaultfd_ctx *ctx = file->private_data; @@ -2018,6 +2040,9 @@ SYSCALL_DEFINE1(userfaultfd, int, flags) BUILD_BUG_ON(UFFD_CLOEXEC != O_CLOEXEC); BUILD_BUG_ON(UFFD_NONBLOCK != O_NONBLOCK); + if (!userfaultfd_allowed()) + return -EPERM; + if (flags & ~UFFD_SHARED_FCNTL_FLAGS) return -EINVAL;