From patchwork Mon Aug 5 18:34:50 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Elliot Berman X-Patchwork-Id: 13753958 Received: from mx0b-0031df01.pphosted.com (mx0b-0031df01.pphosted.com [205.220.180.131]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 6B22A15AD96; Mon, 5 Aug 2024 18:35:27 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=205.220.180.131 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1722882929; cv=none; b=PXrOK2zNv54H7gm5QGOsEkwpWeJIzPiZ8mg3kdBG4Q17blj/TNusQaNx7XG598sMaSG8IUoOLgWS5B8NbkVWEiWuiwyAUOEaCVri6lk9a39PVj2FOFhF1VeZDTGKFI67kfVI9LQ+n4pFZ2Z/o8ekyXcBzwzqmlJz1pgUOuBbsxw= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1722882929; c=relaxed/simple; bh=Dc7xBPtMEVYkTHxbGqssJP61RmrrzdZj7DttRv9ZvEk=; h=From:Date:Subject:MIME-Version:Content-Type:Message-ID:References: In-Reply-To:To:CC; b=Tn+1OS+BCoPkizcTU1HIzC1NBz1CQ371EiHBLrIfvbNym6VVaFRmiHmzcMetr03v+RxmGqkOxMp9J1OfmgFPSP/5fk24sh5n+yKk7cEZOjYCNTDuTTCYz6ZwDau2zZP6ElvdasYXI4GKw/pVPoVwlQTJmcEdFX64NJ3qmRuBu+Q= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=quicinc.com; spf=pass smtp.mailfrom=quicinc.com; dkim=pass (2048-bit key) header.d=quicinc.com header.i=@quicinc.com header.b=TFJbHmIF; arc=none smtp.client-ip=205.220.180.131 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=quicinc.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=quicinc.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=quicinc.com header.i=@quicinc.com header.b="TFJbHmIF" Received: from pps.filterd (m0279873.ppops.net [127.0.0.1]) by mx0a-0031df01.pphosted.com (8.18.1.2/8.18.1.2) with ESMTP id 475IVP7L031919; Mon, 5 Aug 2024 18:35:16 GMT DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=quicinc.com; h= cc:content-transfer-encoding:content-type:date:from:in-reply-to :message-id:mime-version:references:subject:to; s=qcppdkim1; bh= VI8SceDSszYjbEPmaGh0xnvBavYMrOtuUVZ9U4i2+FQ=; b=TFJbHmIF42xpNejp uEiFEeP3sPD26D8spggjdCuJyfOu0jlxa8SYzLC2eQTPX7u9wJP4ia5QsXWmEWsS SmmKNHcrLc3pvEN18/ywtJj/pbGLEQCnQNIknjbgAONxWX8/vouT0oJpJsTqrjHI J9gmm9vFySGTdtaWdnKBAEB7N2/wzg46Yv+PGea5o6e+SE79cX8pw5ACket4tJVJ /46l1exQm31wG5TRfMLTxajnaKVpxbaSBGD4aUBEqV4hjewIyCoYwq4gqUdn0GO0 QWEzNjfgnFbOoQ2c53f7004mFKM1sJG/EDoM7V2ya+7A1keYJZCyH8BtLkBfygdu XmsY1w== Received: from nasanppmta04.qualcomm.com (i-global254.qualcomm.com [199.106.103.254]) by mx0a-0031df01.pphosted.com (PPS) with ESMTPS id 40tuhvsmeq-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 05 Aug 2024 18:35:16 +0000 (GMT) Received: from nasanex01b.na.qualcomm.com (nasanex01b.na.qualcomm.com [10.46.141.250]) by NASANPPMTA04.qualcomm.com (8.17.1.19/8.17.1.19) with ESMTPS id 475IZF4A029446 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 5 Aug 2024 18:35:15 GMT Received: from hu-eberman-lv.qualcomm.com (10.49.16.6) by nasanex01b.na.qualcomm.com (10.46.141.250) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.2.1544.9; Mon, 5 Aug 2024 11:35:14 -0700 From: Elliot Berman Date: Mon, 5 Aug 2024 11:34:50 -0700 Subject: [PATCH RFC 4/4] mm: guest_memfd: Add ability for mmap'ing pages Precedence: bulk X-Mailing-List: kvm@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Message-ID: <20240805-guest-memfd-lib-v1-4-e5a29a4ff5d7@quicinc.com> References: <20240805-guest-memfd-lib-v1-0-e5a29a4ff5d7@quicinc.com> In-Reply-To: <20240805-guest-memfd-lib-v1-0-e5a29a4ff5d7@quicinc.com> To: Andrew Morton , Paolo Bonzini , Sean Christopherson , Fuad Tabba , David Hildenbrand , Patrick Roy , , Ackerley Tng CC: , , , , , Elliot Berman X-Mailer: b4 0.13.0 X-ClientProxiedBy: nalasex01a.na.qualcomm.com (10.47.209.196) To nasanex01b.na.qualcomm.com (10.46.141.250) X-QCInternal: smtphost X-Proofpoint-Virus-Version: vendor=nai engine=6200 definitions=5800 signatures=585085 X-Proofpoint-ORIG-GUID: sfX1xQ5G9G6F0uigExaJMuPsByDlTZno X-Proofpoint-GUID: sfX1xQ5G9G6F0uigExaJMuPsByDlTZno X-Proofpoint-Virus-Version: vendor=baseguard engine=ICAP:2.0.293,Aquarius:18.0.1039,Hydra:6.0.680,FMLib:17.12.28.16 definitions=2024-08-05_07,2024-08-02_01,2024-05-17_01 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 lowpriorityscore=0 priorityscore=1501 spamscore=0 mlxlogscore=999 phishscore=0 clxscore=1015 suspectscore=0 adultscore=0 mlxscore=0 malwarescore=0 bulkscore=0 impostorscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.19.0-2407110000 definitions=main-2408050132 Confidential/protected guest virtual machines want to share some memory back with the host Linux. For example, virtqueues allow host and protected guest to exchange data. In MMU-only isolation of protected guest virtual machines, the transition between "shared" and "private" can be done in-place without a trusted hypervisor copying pages. Add support for this feature and allow Linux to mmap host-accessible pages. When the owner provides an ->accessible() callback in the struct guest_memfd_operations, guest_memfd allows folios to be mapped when the ->accessible() callback returns 0. To safely make inaccessible: ``` folio = guest_memfd_grab_folio(inode, index, flags); r = guest_memfd_make_inaccessible(inode, folio); if (r) goto err; hypervisor_does_guest_mapping(folio); folio_unlock(folio); ``` hypervisor_does_s2_mapping(folio) should make it so ops->accessible(...) on those folios fails. The folio lock ensures atomicity. Signed-off-by: Elliot Berman --- include/linux/guest_memfd.h | 7 ++++ mm/guest_memfd.c | 81 ++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 87 insertions(+), 1 deletion(-) diff --git a/include/linux/guest_memfd.h b/include/linux/guest_memfd.h index f9e4a27aed67..edcb4ba60cb0 100644 --- a/include/linux/guest_memfd.h +++ b/include/linux/guest_memfd.h @@ -16,12 +16,18 @@ * @invalidate_end: called after invalidate_begin returns success. Optional. * @prepare: called before a folio is mapped into the guest address space. * Optional. + * @accessible: called after prepare returns success and before it's mapped + * into the guest address space. Returns 0 if the folio can be + * accessed. + * Optional. If not present, assumes folios are never accessible. * @release: Called when releasing the guest_memfd file. Required. */ struct guest_memfd_operations { int (*invalidate_begin)(struct inode *inode, pgoff_t offset, unsigned long nr); void (*invalidate_end)(struct inode *inode, pgoff_t offset, unsigned long nr); int (*prepare)(struct inode *inode, pgoff_t offset, struct folio *folio); + int (*accessible)(struct inode *inode, struct folio *folio, + pgoff_t offset, unsigned long nr); int (*release)(struct inode *inode); }; @@ -48,5 +54,6 @@ struct file *guest_memfd_alloc(const char *name, const struct guest_memfd_operations *ops, loff_t size, unsigned long flags); bool is_guest_memfd(struct file *file, const struct guest_memfd_operations *ops); +int guest_memfd_make_inaccessible(struct file *file, struct folio *folio); #endif diff --git a/mm/guest_memfd.c b/mm/guest_memfd.c index e9d8cab72b28..6b5609932ca5 100644 --- a/mm/guest_memfd.c +++ b/mm/guest_memfd.c @@ -9,6 +9,8 @@ #include #include +#include "internal.h" + static inline int guest_memfd_folio_private(struct folio *folio) { unsigned long nr_pages = folio_nr_pages(folio); @@ -89,7 +91,7 @@ struct folio *guest_memfd_grab_folio(struct file *file, pgoff_t index, u32 flags goto out_err; } - if (gmem_flags & GUEST_MEMFD_FLAG_NO_DIRECT_MAP) { + if (!ops->accessible && (gmem_flags & GUEST_MEMFD_FLAG_NO_DIRECT_MAP)) { r = guest_memfd_folio_private(folio); if (r) goto out_err; @@ -107,6 +109,82 @@ struct folio *guest_memfd_grab_folio(struct file *file, pgoff_t index, u32 flags } EXPORT_SYMBOL_GPL(guest_memfd_grab_folio); +int guest_memfd_make_inaccessible(struct file *file, struct folio *folio) +{ + unsigned long gmem_flags = (unsigned long)file->private_data; + unsigned long i; + int r; + + unmap_mapping_folio(folio); + + /** + * We can't use the refcount. It might be elevated due to + * guest/vcpu trying to access same folio as another vcpu + * or because userspace is trying to access folio for same reason + * + * folio_lock serializes the transitions between (in)accessible + */ + if (folio_maybe_dma_pinned(folio)) + return -EBUSY; + + if (gmem_flags & GUEST_MEMFD_FLAG_NO_DIRECT_MAP) { + r = guest_memfd_folio_private(folio); + if (r) + return r; + } + + return 0; +} + +static vm_fault_t gmem_fault(struct vm_fault *vmf) +{ + struct file *file = vmf->vma->vm_file; + struct inode *inode = file_inode(file); + const struct guest_memfd_operations *ops = inode->i_private; + struct folio *folio; + pgoff_t off; + int r; + + folio = guest_memfd_grab_folio(file, vmf->pgoff, GUEST_MEMFD_GRAB_UPTODATE); + if (!folio) + return VM_FAULT_SIGBUS; + + off = vmf->pgoff & (folio_nr_pages(folio) - 1); + r = ops->accessible(inode, folio, off, 1); + if (r) { + folio_unlock(folio); + folio_put(folio); + return VM_FAULT_SIGBUS; + } + + guest_memfd_folio_clear_private(folio); + + vmf->page = folio_page(folio, off); + + return VM_FAULT_LOCKED; +} + +static const struct vm_operations_struct gmem_vm_ops = { + .fault = gmem_fault, +}; + +static int gmem_mmap(struct file *file, struct vm_area_struct *vma) +{ + const struct guest_memfd_operations *ops = file_inode(file)->i_private; + + if (!ops->accessible) + return -EPERM; + + /* No support for private mappings to avoid COW. */ + if ((vma->vm_flags & (VM_SHARED | VM_MAYSHARE)) != + (VM_SHARED | VM_MAYSHARE)) + return -EINVAL; + + file_accessed(file); + vma->vm_ops = &gmem_vm_ops; + return 0; +} + static long gmem_punch_hole(struct file *file, loff_t offset, loff_t len) { struct inode *inode = file_inode(file); @@ -220,6 +298,7 @@ static int gmem_release(struct inode *inode, struct file *file) static struct file_operations gmem_fops = { .open = generic_file_open, .llseek = generic_file_llseek, + .mmap = gmem_mmap, .release = gmem_release, .fallocate = gmem_fallocate, .owner = THIS_MODULE,