From patchwork Tue May 19 12:25:46 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Raphael Norwitz X-Patchwork-Id: 11557563 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id C1CB2913 for ; Tue, 19 May 2020 12:32:49 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 9761020825 for ; Tue, 19 May 2020 12:32:49 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (1024-bit key) header.d=sendgrid.net header.i=@sendgrid.net header.b="iPHqDY7z" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 9761020825 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=nutanix.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Received: from localhost ([::1]:39242 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jb1QO-0002zL-R1 for patchwork-qemu-devel@patchwork.kernel.org; Tue, 19 May 2020 08:32:48 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:51700) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jb1Jd-0000Bi-S1 for qemu-devel@nongnu.org; Tue, 19 May 2020 08:25:49 -0400 Received: from o1.dev.nutanix.com ([198.21.4.205]:42273) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jb1Jc-0003BW-Mt for qemu-devel@nongnu.org; Tue, 19 May 2020 08:25:49 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sendgrid.net; h=from:subject:in-reply-to:references:to:cc:content-type: content-transfer-encoding; s=smtpapi; bh=vwkmlnReXmrnXmx0K8MZishuO6jVYqTahHF5z/mitvs=; b=iPHqDY7zUXwChTM02sKFCf/pCh3heyXVnBRzK3m3cmnPSrlAiiUF42JpKCyRfW8JR3dq 2RQOMxquEXYOi5+ZnV3QdL+RTUiXAJrPbg1KuzV1dndb7446TD4lKshaPv1XIGuzO2ZOj8 KhjcTmirezjEWF5QN2FwqVOiGeGnIRLcw= Received: by filterdrecv-p3iad2-8ddf98858-f4h4l with SMTP id filterdrecv-p3iad2-8ddf98858-f4h4l-19-5EC3D04A-5E 2020-05-19 12:25:46.782585185 +0000 UTC m=+4706300.101817334 Received: from localhost.localdomain.com (unknown) by ismtpd0002p1lon1.sendgrid.net (SG) with ESMTP id mzgWj0hfT3qDXIYjNhZlFA Tue, 19 May 2020 12:25:46.505 +0000 (UTC) From: Raphael Norwitz Subject: [PATCH v3 08/10] Support adding individual regions in libvhost-user Date: Tue, 19 May 2020 12:25:46 +0000 (UTC) Message-Id: <1588473683-27067-9-git-send-email-raphael.norwitz@nutanix.com> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: <1588473683-27067-1-git-send-email-raphael.norwitz@nutanix.com> References: <1588473683-27067-1-git-send-email-raphael.norwitz@nutanix.com> X-SG-EID: YCLURHX+pjNDm1i7d69iKyMnQi/dvWah9veFa8nllaoUC0ScIWrCgiaWGu43VgxFdB4istXUBpN9H93OJgc8zWBwzjvdkQJxdrOj/1okDW0fm7tcyC8YuxkTPXtHIymnegKSUhglEFEOuzyfmVhWU9zjAa8bzu5KXQYXX7VgdJpG0TuxuVYzvXfiIHeQlKDyAEN2XhHg/D1TP8GFEGsKNiTwPtRldtCSz8djro8JANOB8juA/82vgBmex8B2Nw0NuhmAwU8HQvUm7wR5jzuZVw== To: qemu-devel@nongnu.org, mst@redhat.com, marcandre.lureau@redhat.com Received-SPF: pass client-ip=198.21.4.205; envelope-from=bounces+16159052-3d09-qemu-devel=nongnu.org@sendgrid.net; helo=o1.dev.nutanix.com X-detected-operating-system: by eggs.gnu.org: First seen = 2020/05/19 08:25:15 X-ACL-Warn: Detected OS = Linux 3.11 and newer [fuzzy] X-Spam_score_int: -17 X-Spam_score: -1.8 X-Spam_bar: - X-Spam_report: (-1.8 / 5.0 requ) BAYES_00=-1.9, DKIMWL_WL_MED=0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_EF=-0.1, HEADER_FROM_DIFFERENT_DOMAINS=0.249, RCVD_IN_DNSWL_NONE=-0.0001, RCVD_IN_MSPIKE_H2=-0.001, SPF_PASS=-0.001, UNPARSEABLE_RELAY=0.001, URIBL_BLOCKED=0.001 autolearn=_AUTOLEARN X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: raphael.s.norwitz@gmail.com, marcandre.lureau@gmail.com, Raphael Norwitz Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" When the VHOST_USER_PROTOCOL_F_CONFIGURE_MEM_SLOTS is enabled, qemu will transmit memory regions to a backend individually using the new message VHOST_USER_ADD_MEM_REG. With this change vhost-user backends built with libvhost-user can now map in new memory regions when VHOST_USER_ADD_MEM_REG messages are received. Qemu only sends VHOST_USER_ADD_MEM_REG messages when the VHOST_USER_PROTOCOL_F_CONFIGURE_MEM_SLOTS feature is negotiated, and since it is not yet supported in libvhost-user, this new functionality is not yet used. Signed-off-by: Raphael Norwitz --- contrib/libvhost-user/libvhost-user.c | 103 ++++++++++++++++++++++++++++++++++ contrib/libvhost-user/libvhost-user.h | 7 +++ 2 files changed, 110 insertions(+) diff --git a/contrib/libvhost-user/libvhost-user.c b/contrib/libvhost-user/libvhost-user.c index 9f039b7..2c2a8d9 100644 --- a/contrib/libvhost-user/libvhost-user.c +++ b/contrib/libvhost-user/libvhost-user.c @@ -138,6 +138,7 @@ vu_request_to_string(unsigned int req) REQ(VHOST_USER_GPU_SET_SOCKET), REQ(VHOST_USER_VRING_KICK), REQ(VHOST_USER_GET_MAX_MEM_SLOTS), + REQ(VHOST_USER_ADD_MEM_REG), REQ(VHOST_USER_MAX), }; #undef REQ @@ -663,6 +664,106 @@ generate_faults(VuDev *dev) { } static bool +vu_add_mem_reg(VuDev *dev, VhostUserMsg *vmsg) { + int i; + bool track_ramblocks = dev->postcopy_listening; + VhostUserMemoryRegion *msg_region = &vmsg->payload.memreg.region; + VuDevRegion *dev_region = &dev->regions[dev->nregions]; + void *mmap_addr; + + /* + * If we are in postcopy mode and we receive a u64 payload with a 0 value + * we know all the postcopy client bases have been recieved, and we + * should start generating faults. + */ + if (track_ramblocks && + vmsg->size == sizeof(vmsg->payload.u64) && + vmsg->payload.u64 == 0) { + (void)generate_faults(dev); + return false; + } + + DPRINT("Adding region: %d\n", dev->nregions); + DPRINT(" guest_phys_addr: 0x%016"PRIx64"\n", + msg_region->guest_phys_addr); + DPRINT(" memory_size: 0x%016"PRIx64"\n", + msg_region->memory_size); + DPRINT(" userspace_addr 0x%016"PRIx64"\n", + msg_region->userspace_addr); + DPRINT(" mmap_offset 0x%016"PRIx64"\n", + msg_region->mmap_offset); + + dev_region->gpa = msg_region->guest_phys_addr; + dev_region->size = msg_region->memory_size; + dev_region->qva = msg_region->userspace_addr; + dev_region->mmap_offset = msg_region->mmap_offset; + + /* + * We don't use offset argument of mmap() since the + * mapped address has to be page aligned, and we use huge + * pages. + */ + if (track_ramblocks) { + /* + * In postcopy we're using PROT_NONE here to catch anyone + * accessing it before we userfault. + */ + mmap_addr = mmap(0, dev_region->size + dev_region->mmap_offset, + PROT_NONE, MAP_SHARED, + vmsg->fds[0], 0); + } else { + mmap_addr = mmap(0, dev_region->size + dev_region->mmap_offset, + PROT_READ | PROT_WRITE, MAP_SHARED, vmsg->fds[0], + 0); + } + + if (mmap_addr == MAP_FAILED) { + vu_panic(dev, "region mmap error: %s", strerror(errno)); + } else { + dev_region->mmap_addr = (uint64_t)(uintptr_t)mmap_addr; + DPRINT(" mmap_addr: 0x%016"PRIx64"\n", + dev_region->mmap_addr); + } + + close(vmsg->fds[0]); + + if (track_ramblocks) { + /* + * Return the address to QEMU so that it can translate the ufd + * fault addresses back. + */ + msg_region->userspace_addr = (uintptr_t)(mmap_addr + + dev_region->mmap_offset); + + /* Send the message back to qemu with the addresses filled in. */ + vmsg->fd_num = 0; + if (!vu_send_reply(dev, dev->sock, vmsg)) { + vu_panic(dev, "failed to respond to add-mem-region for postcopy"); + return false; + } + + DPRINT("Successfully added new region in postcopy\n"); + dev->nregions++; + return false; + + } else { + for (i = 0; i < dev->max_queues; i++) { + if (dev->vq[i].vring.desc) { + if (map_ring(dev, &dev->vq[i])) { + vu_panic(dev, "remapping queue %d for new memory region", + i); + } + } + } + + DPRINT("Successfully added new region\n"); + dev->nregions++; + vmsg_set_reply_u64(vmsg, 0); + return true; + } +} + +static bool vu_set_mem_table_exec_postcopy(VuDev *dev, VhostUserMsg *vmsg) { int i; @@ -1668,6 +1769,8 @@ vu_process_message(VuDev *dev, VhostUserMsg *vmsg) return vu_handle_vring_kick(dev, vmsg); case VHOST_USER_GET_MAX_MEM_SLOTS: return vu_handle_get_max_memslots(dev, vmsg); + case VHOST_USER_ADD_MEM_REG: + return vu_add_mem_reg(dev, vmsg); default: vmsg_close_fds(vmsg); vu_panic(dev, "Unhandled request: %d", vmsg->request); diff --git a/contrib/libvhost-user/libvhost-user.h b/contrib/libvhost-user/libvhost-user.h index 88ef40d..60ef7fd 100644 --- a/contrib/libvhost-user/libvhost-user.h +++ b/contrib/libvhost-user/libvhost-user.h @@ -98,6 +98,7 @@ typedef enum VhostUserRequest { VHOST_USER_GPU_SET_SOCKET = 33, VHOST_USER_VRING_KICK = 35, VHOST_USER_GET_MAX_MEM_SLOTS = 36, + VHOST_USER_ADD_MEM_REG = 37, VHOST_USER_MAX } VhostUserRequest; @@ -124,6 +125,11 @@ typedef struct VhostUserMemory { VhostUserMemoryRegion regions[VHOST_MEMORY_MAX_NREGIONS]; } VhostUserMemory; +typedef struct VhostUserMemRegMsg { + uint32_t padding; + VhostUserMemoryRegion region; +} VhostUserMemRegMsg; + typedef struct VhostUserLog { uint64_t mmap_size; uint64_t mmap_offset; @@ -176,6 +182,7 @@ typedef struct VhostUserMsg { struct vhost_vring_state state; struct vhost_vring_addr addr; VhostUserMemory memory; + VhostUserMemRegMsg memreg; VhostUserLog log; VhostUserConfig config; VhostUserVringArea area;