From patchwork Fri Dec 4 09:31:00 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrey Gruzdev X-Patchwork-Id: 11951201 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=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,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 3A8D8C433FE for ; Fri, 4 Dec 2020 09:34:54 +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 AEC7B225AB for ; Fri, 4 Dec 2020 09:34:53 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org AEC7B225AB Authentication-Results: mail.kernel.org; dmarc=pass (p=none dis=none) header.from=nongnu.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Received: from localhost ([::1]:54116 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1kl7UK-0005Sl-AG for qemu-devel@archiver.kernel.org; Fri, 04 Dec 2020 04:34:52 -0500 Received: from eggs.gnu.org ([2001:470:142:3::10]:42088) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qp-0001gg-6d for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:17 -0500 Received: from relay.sw.ru ([185.231.240.75]:47040 helo=relay3.sw.ru) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qj-0000nI-Ns for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:14 -0500 Received: from [192.168.15.22] (helo=andrey-MS-7B54.sw.ru) by relay3.sw.ru with esmtp (Exim 4.94) (envelope-from ) id 1kl7Qc-00BfEl-87; Fri, 04 Dec 2020 12:31:02 +0300 To: qemu-devel@nongnu.org Cc: Den Lunev , Eric Blake , Paolo Bonzini , Juan Quintela , "Dr . David Alan Gilbert" , Markus Armbruster , Peter Xu , Andrey Gruzdev Subject: [PATCH v5 1/4] migration: introduce 'background-snapshot' migration capability Date: Fri, 4 Dec 2020 12:31:00 +0300 Message-Id: <20201204093103.9878-2-andrey.gruzdev@virtuozzo.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> References: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> MIME-Version: 1.0 Received-SPF: pass client-ip=185.231.240.75; envelope-from=andrey.gruzdev@virtuozzo.com; helo=relay3.sw.ru X-Spam_score_int: -18 X-Spam_score: -1.9 X-Spam_bar: - X-Spam_report: (-1.9 / 5.0 requ) BAYES_00=-1.9, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no 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: , Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: "Qemu-devel" Reply-to: Andrey Gruzdev X-Patchwork-Original-From: Andrey Gruzdev via From: Andrey Gruzdev Add new capability to 'qapi/migration.json' schema. Update migrate_caps_check() to validate enabled capability set against introduced one. Perform checks for required kernel features and compatibility with guest memory backends. Signed-off-by: Andrey Gruzdev Reviewed-by: Peter Xu --- migration/migration.c | 116 ++++++++++++++++++++++++++++++++++++++++++ migration/migration.h | 1 + migration/ram.c | 21 ++++++++ migration/ram.h | 4 ++ qapi/migration.json | 7 ++- 5 files changed, 148 insertions(+), 1 deletion(-) diff --git a/migration/migration.c b/migration/migration.c index 87a9b59f83..594a5b5d50 100644 --- a/migration/migration.c +++ b/migration/migration.c @@ -56,6 +56,7 @@ #include "net/announce.h" #include "qemu/queue.h" #include "multifd.h" +#include "sysemu/cpus.h" #ifdef CONFIG_VFIO #include "hw/vfio/vfio-common.h" @@ -134,6 +135,38 @@ enum mig_rp_message_type { MIG_RP_MSG_MAX }; +/* Migration capabilities set */ +struct MigrateCapsSet { + int size; /* Capability set size */ + MigrationCapability caps[]; /* Variadic array of capabilities */ +}; +typedef struct MigrateCapsSet MigrateCapsSet; + +/* Define and initialize MigrateCapsSet */ +#define INITIALIZE_MIGRATE_CAPS_SET(_name, ...) \ + MigrateCapsSet _name = { \ + .size = sizeof((int []) { __VA_ARGS__ }) / sizeof(int), \ + .caps = { __VA_ARGS__ } \ + } + +/* Background-snapshot compatibility check list */ +static const +INITIALIZE_MIGRATE_CAPS_SET(check_caps_background_snapshot, + MIGRATION_CAPABILITY_POSTCOPY_RAM, + MIGRATION_CAPABILITY_DIRTY_BITMAPS, + MIGRATION_CAPABILITY_POSTCOPY_BLOCKTIME, + MIGRATION_CAPABILITY_LATE_BLOCK_ACTIVATE, + MIGRATION_CAPABILITY_RETURN_PATH, + MIGRATION_CAPABILITY_MULTIFD, + MIGRATION_CAPABILITY_PAUSE_BEFORE_SWITCHOVER, + MIGRATION_CAPABILITY_AUTO_CONVERGE, + MIGRATION_CAPABILITY_RELEASE_RAM, + MIGRATION_CAPABILITY_RDMA_PIN_ALL, + MIGRATION_CAPABILITY_COMPRESS, + MIGRATION_CAPABILITY_XBZRLE, + MIGRATION_CAPABILITY_X_COLO, + MIGRATION_CAPABILITY_VALIDATE_UUID); + /* When we add fault tolerance, we could have several migrations at once. For now we don't need to add dynamic creation of migration */ @@ -1104,6 +1137,39 @@ static void fill_source_migration_info(MigrationInfo *info) info->status = s->state; } +#ifdef CONFIG_LINUX +typedef enum WriteTrackingSupport { + WT_SUPPORT_UNKNOWN = 0, + WT_SUPPORT_ABSENT, + WT_SUPPORT_AVAILABLE, + WT_SUPPORT_COMPATIBLE +} WriteTrackingSupport; + +static +WriteTrackingSupport migrate_query_write_tracking(void) +{ + static WriteTrackingSupport wt_support = WT_SUPPORT_UNKNOWN; + + /* Check UFFD-IO and memory backends for compatibility only once */ + if (wt_support == WT_SUPPORT_UNKNOWN) { + wt_support = WT_SUPPORT_ABSENT; + /* Check if kernel supports required UFFD features */ + if (ram_write_tracking_available()) { + wt_support = WT_SUPPORT_AVAILABLE; + /* + * Check if current memory configuration is + * compatible with required UFFD features. + */ + if (ram_write_tracking_compatible()) { + wt_support = WT_SUPPORT_COMPATIBLE; + } + } + } + + return wt_support; +} +#endif /* CONFIG_LINUX */ + /** * @migration_caps_check - check capability validity * @@ -1165,6 +1231,45 @@ static bool migrate_caps_check(bool *cap_list, } } + if (cap_list[MIGRATION_CAPABILITY_BACKGROUND_SNAPSHOT]) { +#ifdef CONFIG_LINUX + WriteTrackingSupport wt_support; + int idx; + /* + * Check if 'background-snapshot' capability is supported by + * host kernel and compatible with guest memory configuration. + */ + wt_support = migrate_query_write_tracking(); + if (wt_support < WT_SUPPORT_AVAILABLE) { + error_setg(errp, "Background-snapshot is not supported by host kernel"); + return false; + } + if (wt_support < WT_SUPPORT_COMPATIBLE) { + error_setg(errp, "Background-snapshot is not compatible " + "with guest memory configuration"); + return false; + } + + /* + * Check if there are any migration capabilities + * incompatible with 'background-snapshot'. + */ + for (idx = 0; idx < check_caps_background_snapshot.size; idx++) { + int incomp_cap = check_caps_background_snapshot.caps[idx]; + if (cap_list[incomp_cap]) { + error_setg(errp, + "Background-snapshot is not compatible with %s", + MigrationCapability_str(incomp_cap)); + return false; + } + } +#else + error_setg(errp, + "Background-snapshot is not supported on non-Linux hosts"); + return false; +#endif + } + return true; } @@ -2490,6 +2595,15 @@ bool migrate_use_block_incremental(void) return s->parameters.block_incremental; } +bool migrate_background_snapshot(void) +{ + MigrationState *s; + + s = migrate_get_current(); + + return s->enabled_capabilities[MIGRATION_CAPABILITY_BACKGROUND_SNAPSHOT]; +} + /* migration thread support */ /* * Something bad happened to the RP stream, mark an error @@ -3783,6 +3897,8 @@ static Property migration_properties[] = { DEFINE_PROP_MIG_CAP("x-block", MIGRATION_CAPABILITY_BLOCK), DEFINE_PROP_MIG_CAP("x-return-path", MIGRATION_CAPABILITY_RETURN_PATH), DEFINE_PROP_MIG_CAP("x-multifd", MIGRATION_CAPABILITY_MULTIFD), + DEFINE_PROP_MIG_CAP("x-background-snapshot", + MIGRATION_CAPABILITY_BACKGROUND_SNAPSHOT), DEFINE_PROP_END_OF_LIST(), }; diff --git a/migration/migration.h b/migration/migration.h index d096b77f74..f40338cfbf 100644 --- a/migration/migration.h +++ b/migration/migration.h @@ -341,6 +341,7 @@ int migrate_compress_wait_thread(void); int migrate_decompress_threads(void); bool migrate_use_events(void); bool migrate_postcopy_blocktime(void); +bool migrate_background_snapshot(void); /* Sending on the return path - generic and then for each message type */ void migrate_send_rp_shut(MigrationIncomingState *mis, diff --git a/migration/ram.c b/migration/ram.c index 7811cde643..ae8de17153 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -3788,6 +3788,27 @@ static int ram_resume_prepare(MigrationState *s, void *opaque) return 0; } +/* ram_write_tracking_available: check if kernel supports required UFFD features + * + * Returns true if supports, false otherwise + */ +bool ram_write_tracking_available(void) +{ + /* TODO: implement */ + return false; +} + +/* ram_write_tracking_compatible: check if guest configuration is + * compatible with 'write-tracking' + * + * Returns true if compatible, false otherwise + */ +bool ram_write_tracking_compatible(void) +{ + /* TODO: implement */ + return false; +} + static SaveVMHandlers savevm_ram_handlers = { .save_setup = ram_save_setup, .save_live_iterate = ram_save_iterate, diff --git a/migration/ram.h b/migration/ram.h index 011e85414e..1a9ff90304 100644 --- a/migration/ram.h +++ b/migration/ram.h @@ -79,4 +79,8 @@ void colo_flush_ram_cache(void); void colo_release_ram_cache(void); void colo_incoming_start_dirty_log(void); +/* Background snapshot */ +bool ram_write_tracking_available(void); +bool ram_write_tracking_compatible(void); + #endif diff --git a/qapi/migration.json b/qapi/migration.json index 3c75820527..6291143678 100644 --- a/qapi/migration.json +++ b/qapi/migration.json @@ -442,6 +442,11 @@ # @validate-uuid: Send the UUID of the source to allow the destination # to ensure it is the same. (since 4.2) # +# @background-snapshot: If enabled, the migration stream will be a snapshot +# of the VM exactly at the point when the migration +# procedure starts. The VM RAM is saved with running VM. +# (since 6.0) +# # Since: 1.2 ## { 'enum': 'MigrationCapability', @@ -449,7 +454,7 @@ 'compress', 'events', 'postcopy-ram', 'x-colo', 'release-ram', 'block', 'return-path', 'pause-before-switchover', 'multifd', 'dirty-bitmaps', 'postcopy-blocktime', 'late-block-activate', - 'x-ignore-shared', 'validate-uuid' ] } + 'x-ignore-shared', 'validate-uuid', 'background-snapshot'] } ## # @MigrationCapabilityStatus: From patchwork Fri Dec 4 09:31:01 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrey Gruzdev X-Patchwork-Id: 11951203 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=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,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 D001FC433FE for ; Fri, 4 Dec 2020 09:35:54 +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 3F85C227C3 for ; Fri, 4 Dec 2020 09:35:54 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 3F85C227C3 Authentication-Results: mail.kernel.org; dmarc=pass (p=none dis=none) header.from=nongnu.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Received: from localhost ([::1]:57366 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1kl7VI-0006tk-Tj for qemu-devel@archiver.kernel.org; Fri, 04 Dec 2020 04:35:52 -0500 Received: from eggs.gnu.org ([2001:470:142:3::10]:42090) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qp-0001gh-J5 for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:17 -0500 Received: from relay.sw.ru ([185.231.240.75]:47044 helo=relay3.sw.ru) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qk-0000nF-31 for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:15 -0500 Received: from [192.168.15.22] (helo=andrey-MS-7B54.sw.ru) by relay3.sw.ru with esmtp (Exim 4.94) (envelope-from ) id 1kl7Qc-00BfEl-BG; Fri, 04 Dec 2020 12:31:02 +0300 To: qemu-devel@nongnu.org Cc: Den Lunev , Eric Blake , Paolo Bonzini , Juan Quintela , "Dr . David Alan Gilbert" , Markus Armbruster , Peter Xu , Andrey Gruzdev Subject: [PATCH v5 2/4] migration: introduce UFFD-WP low-level interface helpers Date: Fri, 4 Dec 2020 12:31:01 +0300 Message-Id: <20201204093103.9878-3-andrey.gruzdev@virtuozzo.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> References: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> MIME-Version: 1.0 Received-SPF: pass client-ip=185.231.240.75; envelope-from=andrey.gruzdev@virtuozzo.com; helo=relay3.sw.ru X-Spam_score_int: -18 X-Spam_score: -1.9 X-Spam_bar: - X-Spam_report: (-1.9 / 5.0 requ) BAYES_00=-1.9, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no 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: , Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: "Qemu-devel" Reply-to: Andrey Gruzdev X-Patchwork-Original-From: Andrey Gruzdev via From: Andrey Gruzdev Glue code to the userfaultfd kernel implementation. Querying feature support, createing file descriptor, feature control, memory region registration, IOCTLs on registered registered regions. Signed-off-by: Andrey Gruzdev Reviewed-by: Peter Xu --- include/exec/memory.h | 1 + include/qemu/userfaultfd.h | 35 ++++ util/meson.build | 1 + util/trace-events | 9 + util/userfaultfd.c | 345 +++++++++++++++++++++++++++++++++++++ 5 files changed, 391 insertions(+) create mode 100644 include/qemu/userfaultfd.h create mode 100644 util/userfaultfd.c diff --git a/include/exec/memory.h b/include/exec/memory.h index 0f3e6bcd5e..4ea294500f 100644 --- a/include/exec/memory.h +++ b/include/exec/memory.h @@ -139,6 +139,7 @@ typedef struct IOMMUNotifier IOMMUNotifier; /* RAM is a persistent kind memory */ #define RAM_PMEM (1 << 5) + static inline void iommu_notifier_init(IOMMUNotifier *n, IOMMUNotify fn, IOMMUNotifierFlag flags, hwaddr start, hwaddr end, diff --git a/include/qemu/userfaultfd.h b/include/qemu/userfaultfd.h new file mode 100644 index 0000000000..6b74f92792 --- /dev/null +++ b/include/qemu/userfaultfd.h @@ -0,0 +1,35 @@ +/* + * Linux UFFD-WP support + * + * Copyright Virtuozzo GmbH, 2020 + * + * Authors: + * Andrey Gruzdev + * + * This work is licensed under the terms of the GNU GPL, version 2 or + * later. See the COPYING file in the top-level directory. + */ + +#ifndef USERFAULTFD_H +#define USERFAULTFD_H + +#include "qemu/osdep.h" +#include "exec/hwaddr.h" +#include + +int uffd_query_features(uint64_t *features); +int uffd_create_fd(uint64_t features, bool non_blocking); +void uffd_close_fd(int uffd_fd); +int uffd_register_memory(int uffd_fd, void *addr, uint64_t length, + uint64_t mode, uint64_t *ioctls); +int uffd_unregister_memory(int uffd_fd, void *addr, uint64_t length); +int uffd_change_protection(int uffd_fd, void *addr, uint64_t length, + bool wp, bool dont_wake); +int uffd_copy_page(int uffd_fd, void *dst_addr, void *src_addr, + uint64_t length, bool dont_wake); +int uffd_zero_page(int uffd_fd, void *addr, uint64_t length, bool dont_wake); +int uffd_wakeup(int uffd_fd, void *addr, uint64_t length); +int uffd_read_events(int uffd_fd, struct uffd_msg *msgs, int count); +bool uffd_poll_events(int uffd_fd, int tmo); + +#endif /* USERFAULTFD_H */ diff --git a/util/meson.build b/util/meson.build index f359af0d46..c64bfe94b3 100644 --- a/util/meson.build +++ b/util/meson.build @@ -50,6 +50,7 @@ endif if have_system util_ss.add(when: 'CONFIG_GIO', if_true: [files('dbus.c'), gio]) + util_ss.add(when: 'CONFIG_LINUX', if_true: files('userfaultfd.c')) endif if have_block diff --git a/util/trace-events b/util/trace-events index 61e0d4bcdf..bac0924899 100644 --- a/util/trace-events +++ b/util/trace-events @@ -91,3 +91,12 @@ qemu_vfio_pci_read_config(void *buf, int ofs, int size, uint64_t region_ofs, uin qemu_vfio_pci_write_config(void *buf, int ofs, int size, uint64_t region_ofs, uint64_t region_size) "write cfg ptr %p ofs 0x%x size 0x%x (region addr 0x%"PRIx64" size 0x%"PRIx64")" qemu_vfio_region_info(const char *desc, uint64_t region_ofs, uint64_t region_size, uint32_t cap_offset) "region '%s' addr 0x%"PRIx64" size 0x%"PRIx64" cap_ofs 0x%"PRIx32 qemu_vfio_pci_map_bar(int index, uint64_t region_ofs, uint64_t region_size, int ofs, void *host) "map region bar#%d addr 0x%"PRIx64" size 0x%"PRIx64" ofs 0x%x host %p" + +#userfaultfd.c +uffd_query_features_nosys(int err) "errno: %i" +uffd_query_features_api_failed(int err) "errno: %i" +uffd_create_fd_nosys(int err) "errno: %i" +uffd_create_fd_api_failed(int err) "errno: %i" +uffd_create_fd_api_noioctl(uint64_t ioctl_req, uint64_t ioctl_supp) "ioctl_req: 0x%" PRIx64 "ioctl_supp: 0x%" PRIx64 +uffd_register_memory_failed(void *addr, uint64_t length, uint64_t mode, int err) "addr: %p length: %" PRIu64 " mode: 0x%" PRIx64 " errno: %i" +uffd_unregister_memory_failed(void *addr, uint64_t length, int err) "addr: %p length: %" PRIu64 " errno: %i" diff --git a/util/userfaultfd.c b/util/userfaultfd.c new file mode 100644 index 0000000000..def50675b1 --- /dev/null +++ b/util/userfaultfd.c @@ -0,0 +1,345 @@ +/* + * Linux UFFD-WP support + * + * Copyright Virtuozzo GmbH, 2020 + * + * Authors: + * Andrey Gruzdev + * + * This work is licensed under the terms of the GNU GPL, version 2 or + * later. See the COPYING file in the top-level directory. + */ + +#include "qemu/osdep.h" +#include "qemu/bitops.h" +#include "qemu/error-report.h" +#include "qemu/userfaultfd.h" +#include "trace.h" +#include +#include +#include + +/** + * uffd_query_features: query UFFD features + * + * Returns: 0 on success, negative value in case of an error + * + * @features: parameter to receive 'uffdio_api.features' + */ +int uffd_query_features(uint64_t *features) +{ + int uffd_fd; + struct uffdio_api api_struct = { 0 }; + int ret = -1; + + uffd_fd = syscall(__NR_userfaultfd, O_CLOEXEC); + if (uffd_fd < 0) { + trace_uffd_query_features_nosys(errno); + return -1; + } + + api_struct.api = UFFD_API; + api_struct.features = 0; + + if (ioctl(uffd_fd, UFFDIO_API, &api_struct)) { + trace_uffd_query_features_api_failed(errno); + goto out; + } + *features = api_struct.features; + ret = 0; + +out: + close(uffd_fd); + return ret; +} + +/** + * uffd_create_fd: create UFFD file descriptor + * + * Returns non-negative file descriptor or negative value in case of an error + * + * @features: UFFD features to request + * @non_blocking: create UFFD file descriptor for non-blocking operation + */ +int uffd_create_fd(uint64_t features, bool non_blocking) +{ + int uffd_fd; + int flags; + struct uffdio_api api_struct = { 0 }; + uint64_t ioctl_mask = BIT(_UFFDIO_REGISTER) | BIT(_UFFDIO_UNREGISTER); + + flags = O_CLOEXEC | (non_blocking ? O_NONBLOCK : 0); + uffd_fd = syscall(__NR_userfaultfd, flags); + if (uffd_fd < 0) { + trace_uffd_create_fd_nosys(errno); + return -1; + } + + api_struct.api = UFFD_API; + api_struct.features = features; + if (ioctl(uffd_fd, UFFDIO_API, &api_struct)) { + trace_uffd_create_fd_api_failed(errno); + goto fail; + } + if ((api_struct.ioctls & ioctl_mask) != ioctl_mask) { + trace_uffd_create_fd_api_noioctl(ioctl_mask, api_struct.ioctls); + goto fail; + } + + return uffd_fd; + +fail: + close(uffd_fd); + return -1; +} + +/** + * uffd_close_fd: close UFFD file descriptor + * + * @uffd_fd: UFFD file descriptor + */ +void uffd_close_fd(int uffd_fd) +{ + assert(uffd_fd >= 0); + close(uffd_fd); +} + +/** + * uffd_register_memory: register memory range via UFFD-IO + * + * Returns 0 in case of success, negative value in case of an error + * + * @uffd_fd: UFFD file descriptor + * @addr: base address of memory range + * @length: length of memory range + * @mode: UFFD register mode (UFFDIO_REGISTER_MODE_MISSING, ...) + * @ioctls: optional pointer to receive supported IOCTL mask + */ +int uffd_register_memory(int uffd_fd, void *addr, uint64_t length, + uint64_t mode, uint64_t *ioctls) +{ + struct uffdio_register uffd_register; + + uffd_register.range.start = (uint64_t) addr; + uffd_register.range.len = length; + uffd_register.mode = mode; + + if (ioctl(uffd_fd, UFFDIO_REGISTER, &uffd_register)) { + trace_uffd_register_memory_failed(addr, length, mode, errno); + return -1; + } + if (ioctls) { + *ioctls = uffd_register.ioctls; + } + + return 0; +} + +/** + * uffd_unregister_memory: un-register memory range with UFFD-IO + * + * Returns 0 in case of success, negative value in case of an error + * + * @uffd_fd: UFFD file descriptor + * @addr: base address of memory range + * @length: length of memory range + */ +int uffd_unregister_memory(int uffd_fd, void *addr, uint64_t length) +{ + struct uffdio_range uffd_range; + + uffd_range.start = (uint64_t) addr; + uffd_range.len = length; + + if (ioctl(uffd_fd, UFFDIO_UNREGISTER, &uffd_range)) { + trace_uffd_unregister_memory_failed(addr, length, errno); + return -1; + } + + return 0; +} + +/** + * uffd_change_protection: protect/un-protect memory range for writes via UFFD-IO + * + * Returns 0 on success, negative value in case of error + * + * @uffd_fd: UFFD file descriptor + * @addr: base address of memory range + * @length: length of memory range + * @wp: write-protect/unprotect + * @dont_wake: do not wake threads waiting on wr-protected page + */ +int uffd_change_protection(int uffd_fd, void *addr, uint64_t length, + bool wp, bool dont_wake) +{ + struct uffdio_writeprotect uffd_writeprotect; + + uffd_writeprotect.range.start = (uint64_t) addr; + uffd_writeprotect.range.len = length; + if (!wp && dont_wake) { + /* DONTWAKE is meaningful only on protection release */ + uffd_writeprotect.mode = UFFDIO_WRITEPROTECT_MODE_DONTWAKE; + } else { + uffd_writeprotect.mode = (wp ? UFFDIO_WRITEPROTECT_MODE_WP : 0); + } + + if (ioctl(uffd_fd, UFFDIO_WRITEPROTECT, &uffd_writeprotect)) { + error_report("uffd_change_protection() failed: addr=%p len=%" PRIu64 + " mode=%" PRIx64 " errno=%i", addr, length, + (uint64_t) uffd_writeprotect.mode, errno); + return -1; + } + + return 0; +} + +/** + * uffd_copy_page: copy range of pages to destination via UFFD-IO + * + * Copy range of source pages to the destination to resolve + * missing page fault somewhere in the destination range. + * + * Returns 0 on success, negative value in case of an error + * + * @uffd_fd: UFFD file descriptor + * @dst_addr: destination base address + * @src_addr: source base address + * @length: length of the range to copy + * @dont_wake: do not wake threads waiting on missing page + */ +int uffd_copy_page(int uffd_fd, void *dst_addr, void *src_addr, + uint64_t length, bool dont_wake) +{ + struct uffdio_copy uffd_copy; + + uffd_copy.dst = (uint64_t) dst_addr; + uffd_copy.src = (uint64_t) src_addr; + uffd_copy.len = length; + uffd_copy.mode = dont_wake ? UFFDIO_COPY_MODE_DONTWAKE : 0; + + if (ioctl(uffd_fd, UFFDIO_COPY, &uffd_copy)) { + error_report("uffd_copy_page() failed: dst_addr=%p src_addr=%p length=%" PRIu64 + " mode=%" PRIx64 " errno=%i", dst_addr, src_addr, + length, (uint64_t) uffd_copy.mode, errno); + return -1; + } + + return 0; +} + +/** + * uffd_zero_page: fill range of pages with zeroes via UFFD-IO + * + * Fill range pages with zeroes to resolve missing page fault within the range. + * + * Returns 0 on success, negative value in case of an error + * + * @uffd_fd: UFFD file descriptor + * @addr: base address + * @length: length of the range to fill with zeroes + * @dont_wake: do not wake threads waiting on missing page + */ +int uffd_zero_page(int uffd_fd, void *addr, uint64_t length, bool dont_wake) +{ + struct uffdio_zeropage uffd_zeropage; + + uffd_zeropage.range.start = (uint64_t) addr; + uffd_zeropage.range.len = length; + uffd_zeropage.mode = dont_wake ? UFFDIO_ZEROPAGE_MODE_DONTWAKE : 0; + + if (ioctl(uffd_fd, UFFDIO_ZEROPAGE, &uffd_zeropage)) { + error_report("uffd_zero_page() failed: addr=%p length=%" PRIu64 + " mode=%" PRIx64 " errno=%i", addr, length, + (uint64_t) uffd_zeropage.mode, errno); + return -1; + } + + return 0; +} + +/** + * uffd_wakeup: wake up threads waiting on page UFFD-managed page fault resolution + * + * Wake up threads waiting on any page/pages from the designated range. + * The main use case is when during some period, page faults are resolved + * via UFFD-IO IOCTLs with MODE_DONTWAKE flag set, then after that all waits + * for the whole memory range are satisfied in a single call to uffd_wakeup(). + * + * Returns 0 on success, negative value in case of an error + * + * @uffd_fd: UFFD file descriptor + * @addr: base address + * @length: length of the range + */ +int uffd_wakeup(int uffd_fd, void *addr, uint64_t length) +{ + struct uffdio_range uffd_range; + + uffd_range.start = (uint64_t) addr; + uffd_range.len = length; + + if (ioctl(uffd_fd, UFFDIO_WAKE, &uffd_range)) { + error_report("uffd_wakeup() failed: addr=%p length=%" PRIu64 " errno=%i", + addr, length, errno); + return -1; + } + + return 0; +} + +/** + * uffd_read_events: read pending UFFD events + * + * Returns number of fetched messages, 0 if non is available or + * negative value in case of an error + * + * @uffd_fd: UFFD file descriptor + * @msgs: pointer to message buffer + * @count: number of messages that can fit in the buffer + */ +int uffd_read_events(int uffd_fd, struct uffd_msg *msgs, int count) +{ + ssize_t res; + do { + res = read(uffd_fd, msgs, count * sizeof(struct uffd_msg)); + } while (res < 0 && errno == EINTR); + + if ((res < 0 && errno == EAGAIN)) { + return 0; + } + if (res < 0) { + error_report("uffd_read_events() failed: errno=%i", errno); + return -1; + } + + return (int) (res / sizeof(struct uffd_msg)); +} + +/** + * uffd_poll_events: poll UFFD file descriptor for read + * + * Returns true if events are available for read, false otherwise + * + * @uffd_fd: UFFD file descriptor + * @tmo: timeout value + */ +bool uffd_poll_events(int uffd_fd, int tmo) +{ + int res; + struct pollfd poll_fd = { .fd = uffd_fd, .events = POLLIN, .revents = 0 }; + + do { + res = poll(&poll_fd, 1, tmo); + } while (res < 0 && errno == EINTR); + + if (res == 0) { + return false; + } + if (res < 0) { + error_report("uffd_poll_events() failed: errno=%i", errno); + return false; + } + + return (poll_fd.revents & POLLIN) != 0; +} From patchwork Fri Dec 4 09:31:02 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrey Gruzdev X-Patchwork-Id: 11951197 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=-16.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,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 1A6D6C4361A for ; Fri, 4 Dec 2020 09:32:58 +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 3465A227BF for ; Fri, 4 Dec 2020 09:32:57 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 3465A227BF Authentication-Results: mail.kernel.org; dmarc=pass (p=none dis=none) header.from=nongnu.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Received: from localhost ([::1]:48824 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1kl7SR-00037n-QZ for qemu-devel@archiver.kernel.org; Fri, 04 Dec 2020 04:32:55 -0500 Received: from eggs.gnu.org ([2001:470:142:3::10]:42086) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qo-0001gd-Vq for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:17 -0500 Received: from relay.sw.ru ([185.231.240.75]:47038 helo=relay3.sw.ru) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qk-0000nH-1Y for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:14 -0500 Received: from [192.168.15.22] (helo=andrey-MS-7B54.sw.ru) by relay3.sw.ru with esmtp (Exim 4.94) (envelope-from ) id 1kl7Qc-00BfEl-Ei; Fri, 04 Dec 2020 12:31:02 +0300 To: qemu-devel@nongnu.org Cc: Den Lunev , Eric Blake , Paolo Bonzini , Juan Quintela , "Dr . David Alan Gilbert" , Markus Armbruster , Peter Xu , Andrey Gruzdev Subject: [PATCH v5 3/4] migration: support UFFD write fault processing in ram_save_iterate() Date: Fri, 4 Dec 2020 12:31:02 +0300 Message-Id: <20201204093103.9878-4-andrey.gruzdev@virtuozzo.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> References: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> MIME-Version: 1.0 Received-SPF: pass client-ip=185.231.240.75; envelope-from=andrey.gruzdev@virtuozzo.com; helo=relay3.sw.ru X-Spam_score_int: -18 X-Spam_score: -1.9 X-Spam_bar: - X-Spam_report: (-1.9 / 5.0 requ) BAYES_00=-1.9, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no 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: , Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: "Qemu-devel" Reply-to: Andrey Gruzdev X-Patchwork-Original-From: Andrey Gruzdev via From: Andrey Gruzdev In this particular implementation the same single migration thread is responsible for both normal linear dirty page migration and procesing UFFD page fault events. Processing write faults includes reading UFFD file descriptor, finding respective RAM block and saving faulting page to the migration stream. After page has been saved, write protection can be removed. Since asynchronous version of qemu_put_buffer() is expected to be used to save pages, we also have to flush migraion stream prior to un-protecting saved memory range. Write protection is being removed for any previously protected memory chunk that has hit the migration stream. That's valid for pages from linear page scan along with write fault pages. Signed-off-by: Andrey Gruzdev Reviewed-by: Peter Xu --- include/exec/memory.h | 7 ++ migration/ram.c | 251 ++++++++++++++++++++++++++++++++++++++++- migration/ram.h | 2 + migration/trace-events | 2 + 4 files changed, 260 insertions(+), 2 deletions(-) diff --git a/include/exec/memory.h b/include/exec/memory.h index 4ea294500f..f201ede53e 100644 --- a/include/exec/memory.h +++ b/include/exec/memory.h @@ -140,6 +140,13 @@ typedef struct IOMMUNotifier IOMMUNotifier; #define RAM_PMEM (1 << 5) +/* + * UFFDIO_WRITEPROTECT is used on this RAMBlock to + * support 'write-tracking' migration type. + * Implies ram_state->ram_wt_enabled. + */ +#define RAM_UF_WRITEPROTECT (1 << 6) + static inline void iommu_notifier_init(IOMMUNotifier *n, IOMMUNotify fn, IOMMUNotifierFlag flags, hwaddr start, hwaddr end, diff --git a/migration/ram.c b/migration/ram.c index ae8de17153..72c932709a 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -56,6 +56,11 @@ #include "savevm.h" #include "qemu/iov.h" #include "multifd.h" +#include "sysemu/runstate.h" + +#ifdef CONFIG_LINUX +#include "qemu/userfaultfd.h" +#endif /***********************************************************/ /* ram save/restore */ @@ -298,6 +303,8 @@ struct RAMSrcPageRequest { struct RAMState { /* QEMUFile used for this migration */ QEMUFile *f; + /* UFFD file descriptor, used in 'write-tracking' migration */ + int uffdio_fd; /* Last block that we have visited searching for dirty pages */ RAMBlock *last_seen_block; /* Last block from where we have sent data */ @@ -1434,6 +1441,40 @@ static RAMBlock *unqueue_page(RAMState *rs, ram_addr_t *offset) return block; } +#ifdef CONFIG_LINUX +/** + * poll_fault_page: try to get next UFFD write fault page and, if pending fault + * is found, return RAM block pointer and page offset + * + * Returns pointer to the RAMBlock containing faulting page, + * NULL if no write faults are pending + * + * @rs: current RAM state + * @offset: page offset from the beginning of the block + */ +static RAMBlock *poll_fault_page(RAMState *rs, ram_addr_t *offset) +{ + struct uffd_msg uffd_msg; + void *page_address; + RAMBlock *bs; + int res; + + if (!migrate_background_snapshot()) { + return NULL; + } + + res = uffd_read_events(rs->uffdio_fd, &uffd_msg, 1); + if (res <= 0) { + return NULL; + } + + page_address = (void *) uffd_msg.arg.pagefault.address; + bs = qemu_ram_block_from_host(page_address, false, offset); + assert(bs && (bs->flags & RAM_UF_WRITEPROTECT) != 0); + return bs; +} +#endif /* CONFIG_LINUX */ + /** * get_queued_page: unqueue a page from the postcopy requests * @@ -1473,6 +1514,16 @@ static bool get_queued_page(RAMState *rs, PageSearchStatus *pss) } while (block && !dirty); +#ifdef CONFIG_LINUX + if (!block) { + /* + * Poll write faults too if background snapshot is enabled; that's + * when we have vcpus got blocked by the write protected pages. + */ + block = poll_fault_page(rs, &offset); + } +#endif /* CONFIG_LINUX */ + if (block) { /* * As soon as we start servicing pages out of order, then we have @@ -1746,6 +1797,53 @@ static int ram_save_host_page(RAMState *rs, PageSearchStatus *pss, return pages; } +/** + * ram_save_host_page_pre: ram_save_host_page() pre-notifier + * + * @rs: current RAM state + * @pss: page-search-status structure + * @opaque: pointer to receive opaque context value + */ +static inline +void ram_save_host_page_pre(RAMState *rs, PageSearchStatus *pss, void **opaque) +{ + *(unsigned long *) opaque = pss->page; +} + +/** + * ram_save_host_page_post: ram_save_host_page() post-notifier + * + * @rs: current RAM state + * @pss: page-search-status structure + * @opaque: opaque context value + * @res_override: pointer to the return value of ram_save_host_page(), + * overwritten in case of an error + */ +static void ram_save_host_page_post(RAMState *rs, PageSearchStatus *pss, + void *opaque, int *res_override) +{ +#ifdef CONFIG_LINUX + /* Check if page is from UFFD-managed region. */ + if (pss->block->flags & RAM_UF_WRITEPROTECT) { + unsigned long page_from = (unsigned long) opaque; + + void *page_address = pss->block->host + (page_from << TARGET_PAGE_BITS); + uint64_t run_length = (pss->page - page_from + 1) << TARGET_PAGE_BITS; + int res; + + /* Flush async buffers before un-protect. */ + qemu_fflush(rs->f); + /* Un-protect memory range. */ + res = uffd_change_protection(rs->uffdio_fd, page_address, run_length, + false, false); + /* We don't want to override existing error from ram_save_host_page(). */ + if (res < 0 && *res_override >= 0) { + *res_override = res; + } + } +#endif /* CONFIG_LINUX */ +} + /** * ram_find_and_save_block: finds a dirty page and sends it to f * @@ -1790,7 +1888,11 @@ static int ram_find_and_save_block(RAMState *rs, bool last_stage) } if (found) { + void *opaque; + + ram_save_host_page_pre(rs, &pss, &opaque); pages = ram_save_host_page(rs, &pss, last_stage); + ram_save_host_page_post(rs, &pss, opaque, &pages); } } while (!pages && again); @@ -3794,7 +3896,14 @@ static int ram_resume_prepare(MigrationState *s, void *opaque) */ bool ram_write_tracking_available(void) { - /* TODO: implement */ +#ifdef CONFIG_LINUX + uint64_t uffd_features; + int res; + + res = uffd_query_features(&uffd_features); + return (res == 0 && + (uffd_features & UFFD_FEATURE_PAGEFAULT_FLAG_WP) != 0); +#endif return false; } @@ -3805,10 +3914,148 @@ bool ram_write_tracking_available(void) */ bool ram_write_tracking_compatible(void) { - /* TODO: implement */ +#ifdef CONFIG_LINUX + const uint64_t uffd_ioctls_mask = BIT(_UFFDIO_WRITEPROTECT); + int uffd_fd; + RAMBlock *bs; + bool ret = false; + + /* Open UFFD file descriptor */ + uffd_fd = uffd_create_fd(UFFD_FEATURE_PAGEFAULT_FLAG_WP, false); + if (uffd_fd < 0) { + return false; + } + + RCU_READ_LOCK_GUARD(); + + RAMBLOCK_FOREACH_NOT_IGNORED(bs) { + uint64_t uffd_ioctls; + + /* Nothing to do with read-only and MMIO-writable regions */ + if (bs->mr->readonly || bs->mr->rom_device) { + continue; + } + /* Try to register block memory via UFFD-IO to track writes */ + if (uffd_register_memory(uffd_fd, bs->host, bs->max_length, + UFFDIO_REGISTER_MODE_WP, &uffd_ioctls)) { + goto out; + } + if ((uffd_ioctls & uffd_ioctls_mask) != uffd_ioctls_mask) { + goto out; + } + } + ret = true; + +out: + uffd_close_fd(uffd_fd); + return ret; +#endif return false; } +/* + * ram_write_tracking_start: start UFFD-WP memory tracking + * + * Returns 0 for success or negative value in case of error + */ +int ram_write_tracking_start(void) +{ +#ifdef CONFIG_LINUX + int uffd_fd; + RAMState *rs = ram_state; + RAMBlock *bs; + + /* Open UFFD file descriptor */ + uffd_fd = uffd_create_fd(UFFD_FEATURE_PAGEFAULT_FLAG_WP, true); + if (uffd_fd < 0) { + return uffd_fd; + } + rs->uffdio_fd = uffd_fd; + + RCU_READ_LOCK_GUARD(); + + RAMBLOCK_FOREACH_NOT_IGNORED(bs) { + /* Nothing to do with read-only and MMIO-writable regions */ + if (bs->mr->readonly || bs->mr->rom_device) { + continue; + } + + /* Register block memory with UFFD to track writes */ + if (uffd_register_memory(rs->uffdio_fd, bs->host, + bs->max_length, UFFDIO_REGISTER_MODE_WP, NULL)) { + goto fail; + } + /* Apply UFFD write protection to the block memory range */ + if (uffd_change_protection(rs->uffdio_fd, bs->host, + bs->max_length, true, false)) { + goto fail; + } + bs->flags |= RAM_UF_WRITEPROTECT; + memory_region_ref(bs->mr); + + trace_ram_write_tracking_ramblock_start(bs->idstr, bs->page_size, + bs->host, bs->max_length); + } + + return 0; + +fail: + error_report("ram_write_tracking_start() failed: restoring initial memory state"); + + RAMBLOCK_FOREACH_NOT_IGNORED(bs) { + if ((bs->flags & RAM_UF_WRITEPROTECT) == 0) { + continue; + } + /* + * In case some memory block failed to be write-protected + * remove protection and unregister all succeeded RAM blocks + */ + uffd_change_protection(rs->uffdio_fd, bs->host, bs->max_length, false, false); + uffd_unregister_memory(rs->uffdio_fd, bs->host, bs->max_length); + /* Cleanup flags and remove reference */ + bs->flags &= ~RAM_UF_WRITEPROTECT; + memory_region_unref(bs->mr); + } + + uffd_close_fd(uffd_fd); +#endif /* CONFIG_LINUX */ + rs->uffdio_fd = -1; + return -1; +} + +/** + * ram_write_tracking_stop: stop UFFD-WP memory tracking and remove protection + */ +void ram_write_tracking_stop(void) +{ +#ifdef CONFIG_LINUX + RAMState *rs = ram_state; + RAMBlock *bs; + + RCU_READ_LOCK_GUARD(); + + RAMBLOCK_FOREACH_NOT_IGNORED(bs) { + if ((bs->flags & RAM_UF_WRITEPROTECT) == 0) { + continue; + } + /* Remove protection and unregister all affected RAM blocks */ + uffd_change_protection(rs->uffdio_fd, bs->host, bs->max_length, false, false); + uffd_unregister_memory(rs->uffdio_fd, bs->host, bs->max_length); + + trace_ram_write_tracking_ramblock_stop(bs->idstr, bs->page_size, + bs->host, bs->max_length); + + /* Cleanup flags and remove reference */ + bs->flags &= ~RAM_UF_WRITEPROTECT; + memory_region_unref(bs->mr); + } + + /* Finally close UFFD file descriptor */ + uffd_close_fd(rs->uffdio_fd); + rs->uffdio_fd = -1; +#endif /* CONFIG_LINUX */ +} + static SaveVMHandlers savevm_ram_handlers = { .save_setup = ram_save_setup, .save_live_iterate = ram_save_iterate, diff --git a/migration/ram.h b/migration/ram.h index 1a9ff90304..c25540cb93 100644 --- a/migration/ram.h +++ b/migration/ram.h @@ -82,5 +82,7 @@ void colo_incoming_start_dirty_log(void); /* Background snapshot */ bool ram_write_tracking_available(void); bool ram_write_tracking_compatible(void); +int ram_write_tracking_start(void); +void ram_write_tracking_stop(void); #endif diff --git a/migration/trace-events b/migration/trace-events index 75de5004ac..668c562fed 100644 --- a/migration/trace-events +++ b/migration/trace-events @@ -111,6 +111,8 @@ save_xbzrle_page_skipping(void) "" save_xbzrle_page_overflow(void) "" ram_save_iterate_big_wait(uint64_t milliconds, int iterations) "big wait: %" PRIu64 " milliseconds, %d iterations" ram_load_complete(int ret, uint64_t seq_iter) "exit_code %d seq iteration %" PRIu64 +ram_write_tracking_ramblock_start(const char *block_id, size_t page_size, void *addr, size_t length) "%s: page_size: %zu addr: %p length: %zu" +ram_write_tracking_ramblock_stop(const char *block_id, size_t page_size, void *addr, size_t length) "%s: page_size: %zu addr: %p length: %zu" # multifd.c multifd_new_send_channel_async(uint8_t id) "channel %d" From patchwork Fri Dec 4 09:31:03 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrey Gruzdev X-Patchwork-Id: 11951195 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=-16.2 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URG_BIZ,URIBL_BLOCKED, 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 17550C433FE for ; Fri, 4 Dec 2020 09:32:53 +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 4F559225AB for ; Fri, 4 Dec 2020 09:32:52 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 4F559225AB Authentication-Results: mail.kernel.org; dmarc=pass (p=none dis=none) header.from=nongnu.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Received: from localhost ([::1]:48504 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1kl7SN-0002zo-0k for qemu-devel@archiver.kernel.org; Fri, 04 Dec 2020 04:32:51 -0500 Received: from eggs.gnu.org ([2001:470:142:3::10]:42084) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qo-0001gZ-BR for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:17 -0500 Received: from relay.sw.ru ([185.231.240.75]:47042 helo=relay3.sw.ru) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1kl7Qk-0000nG-Id for qemu-devel@nongnu.org; Fri, 04 Dec 2020 04:31:14 -0500 Received: from [192.168.15.22] (helo=andrey-MS-7B54.sw.ru) by relay3.sw.ru with esmtp (Exim 4.94) (envelope-from ) id 1kl7Qc-00BfEl-IG; Fri, 04 Dec 2020 12:31:02 +0300 To: qemu-devel@nongnu.org Cc: Den Lunev , Eric Blake , Paolo Bonzini , Juan Quintela , "Dr . David Alan Gilbert" , Markus Armbruster , Peter Xu , Andrey Gruzdev Subject: [PATCH v5 4/4] migration: implementation of background snapshot thread Date: Fri, 4 Dec 2020 12:31:03 +0300 Message-Id: <20201204093103.9878-5-andrey.gruzdev@virtuozzo.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> References: <20201204093103.9878-1-andrey.gruzdev@virtuozzo.com> MIME-Version: 1.0 Received-SPF: pass client-ip=185.231.240.75; envelope-from=andrey.gruzdev@virtuozzo.com; helo=relay3.sw.ru X-Spam_score_int: -12 X-Spam_score: -1.3 X-Spam_bar: - X-Spam_report: (-1.3 / 5.0 requ) BAYES_00=-1.9, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URG_BIZ=0.573 autolearn=no autolearn_force=no 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: , Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: "Qemu-devel" Reply-to: Andrey Gruzdev X-Patchwork-Original-From: Andrey Gruzdev via From: Andrey Gruzdev Introducing implementation of 'background' snapshot thread which in overall follows the logic of precopy migration while internally utilizes completely different mechanism to 'freeze' vmstate at the start of snapshot creation. This mechanism is based on userfault_fd with wr-protection support and is Linux-specific. Signed-off-by: Andrey Gruzdev Acked-by: Peter Xu --- migration/migration.c | 247 +++++++++++++++++++++++++++++++++++++++++- migration/migration.h | 3 + migration/ram.c | 2 + migration/savevm.c | 1 - migration/savevm.h | 2 + 5 files changed, 252 insertions(+), 3 deletions(-) diff --git a/migration/migration.c b/migration/migration.c index 594a5b5d50..b1d2a27d89 100644 --- a/migration/migration.c +++ b/migration/migration.c @@ -2032,6 +2032,7 @@ void migrate_init(MigrationState *s) * locks. */ s->cleanup_bh = 0; + s->vm_start_bh = 0; s->to_dst_file = NULL; s->state = MIGRATION_STATUS_NONE; s->rp_state.from_dst_file = NULL; @@ -3230,6 +3231,50 @@ fail: MIGRATION_STATUS_FAILED); } +/** + * bg_migration_completion: Used by bg_migration_thread when after all the + * RAM has been saved. The caller 'breaks' the loop when this returns. + * + * @s: Current migration state + */ +static void bg_migration_completion(MigrationState *s) +{ + int current_active_state = s->state; + + /* + * Stop tracking RAM writes - un-protect memory, un-register UFFD + * memory ranges, flush kernel wait queues and wake up threads + * waiting for write fault to be resolved. + */ + ram_write_tracking_stop(); + + if (s->state == MIGRATION_STATUS_ACTIVE) { + /* + * By this moment we have RAM content saved into the migration stream. + * The next step is to flush the non-RAM content (device state) + * right after the ram content. The device state has been stored into + * the temporary buffer before RAM saving started. + */ + qemu_put_buffer(s->to_dst_file, s->bioc->data, s->bioc->usage); + qemu_fflush(s->to_dst_file); + } else if (s->state == MIGRATION_STATUS_CANCELLING) { + goto fail; + } + + if (qemu_file_get_error(s->to_dst_file)) { + trace_migration_completion_file_err(); + goto fail; + } + + migrate_set_state(&s->state, current_active_state, + MIGRATION_STATUS_COMPLETED); + return; + +fail: + migrate_set_state(&s->state, current_active_state, + MIGRATION_STATUS_FAILED); +} + bool migrate_colo_enabled(void) { MigrationState *s = migrate_get_current(); @@ -3570,6 +3615,47 @@ static void migration_iteration_finish(MigrationState *s) qemu_mutex_unlock_iothread(); } +static void bg_migration_iteration_finish(MigrationState *s) +{ + qemu_mutex_lock_iothread(); + switch (s->state) { + case MIGRATION_STATUS_COMPLETED: + migration_calculate_complete(s); + break; + + case MIGRATION_STATUS_ACTIVE: + case MIGRATION_STATUS_FAILED: + case MIGRATION_STATUS_CANCELLED: + case MIGRATION_STATUS_CANCELLING: + break; + + default: + /* Should not reach here, but if so, forgive the VM. */ + error_report("%s: Unknown ending state %d", __func__, s->state); + break; + } + + migrate_fd_cleanup_schedule(s); + qemu_mutex_unlock_iothread(); +} + +/* + * Return true if continue to the next iteration directly, false + * otherwise. + */ +static MigIterateState bg_migration_iteration_run(MigrationState *s) +{ + int res; + + res = qemu_savevm_state_iterate(s->to_dst_file, false); + if (res > 0) { + bg_migration_completion(s); + return MIG_ITERATE_BREAK; + } + + return MIG_ITERATE_RESUME; +} + void migration_make_urgent_request(void) { qemu_sem_post(&migrate_get_current()->rate_limit_sem); @@ -3717,6 +3803,157 @@ static void *migration_thread(void *opaque) return NULL; } +static void bg_migration_vm_start_bh(void *opaque) +{ + MigrationState *s = opaque; + + qemu_bh_delete(s->vm_start_bh); + s->vm_start_bh = NULL; + + vm_start(); + s->downtime = qemu_clock_get_ms(QEMU_CLOCK_REALTIME) - s->downtime_start; +} + +/** + * Background snapshot thread, based on live migration code. + * This is an alternative implementation of live migration mechanism + * introduced specifically to support background snapshots. + * + * It takes advantage of userfault_fd write protection mechanism introduced + * in v5.7 kernel. Compared to existing dirty page logging migration much + * lesser stream traffic is produced resulting in smaller snapshot images, + * simply cause of no page duplicates can get into the stream. + * + * Another key point is that generated vmstate stream reflects machine state + * 'frozen' at the beginning of snapshot creation compared to dirty page logging + * mechanism, which effectively results in that saved snapshot is the state of VM + * at the end of the process. + */ +static void *bg_migration_thread(void *opaque) +{ + MigrationState *s = opaque; + int64_t setup_start; + MigThrError thr_error; + QEMUFile *fb; + bool early_fail = true; + + rcu_register_thread(); + object_ref(OBJECT(s)); + + qemu_file_set_rate_limit(s->to_dst_file, INT64_MAX); + + setup_start = qemu_clock_get_ms(QEMU_CLOCK_HOST); + /* + * We want to save vmstate for the moment when migration has been + * initiated but also we want to save RAM content while VM is running. + * The RAM content should appear first in the vmstate. So, we first + * stash the non-RAM part of the vmstate to the temporary buffer, + * then write RAM part of the vmstate to the migration stream + * with vCPUs running and, finally, write stashed non-RAM part of + * the vmstate from the buffer to the migration stream. + */ + s->bioc = qio_channel_buffer_new(128 * 1024); + qio_channel_set_name(QIO_CHANNEL(s->bioc), "vmstate-buffer"); + fb = qemu_fopen_channel_output(QIO_CHANNEL(s->bioc)); + object_unref(OBJECT(s->bioc)); + + update_iteration_initial_status(s); + + qemu_savevm_state_header(s->to_dst_file); + qemu_savevm_state_setup(s->to_dst_file); + + if (qemu_savevm_state_guest_unplug_pending()) { + migrate_set_state(&s->state, MIGRATION_STATUS_SETUP, + MIGRATION_STATUS_WAIT_UNPLUG); + + while (s->state == MIGRATION_STATUS_WAIT_UNPLUG && + qemu_savevm_state_guest_unplug_pending()) { + qemu_sem_timedwait(&s->wait_unplug_sem, 250); + } + + migrate_set_state(&s->state, MIGRATION_STATUS_WAIT_UNPLUG, + MIGRATION_STATUS_ACTIVE); + } + s->setup_time = qemu_clock_get_ms(QEMU_CLOCK_HOST) - setup_start; + + migrate_set_state(&s->state, MIGRATION_STATUS_SETUP, + MIGRATION_STATUS_ACTIVE); + trace_migration_thread_setup_complete(); + s->downtime_start = qemu_clock_get_ms(QEMU_CLOCK_REALTIME); + + qemu_mutex_lock_iothread(); + + if (global_state_store()) { + goto fail; + } + /* Forcibly stop VM before saving state of vCPUs and devices */ + if (vm_stop_force_state(RUN_STATE_PAUSED)) { + goto fail; + } + /* + * Put vCPUs in sync with shadow context structures, then + * save their state to channel-buffer along with devices. + */ + cpu_synchronize_all_states(); + if (qemu_savevm_state_complete_precopy_non_iterable(fb, false, false)) { + goto fail; + } + /* Now initialize UFFD context and start tracking RAM writes */ + if (ram_write_tracking_start()) { + goto fail; + } + early_fail = false; + + /* + * Start VM from BH handler to avoid write-fault lock here. + * UFFD-WP protection for the whole RAM is already enabled so + * calling VM state change notifiers from vm_start() would initiate + * writes to virtio VQs memory which is in write-protected region. + */ + s->vm_start_bh = qemu_bh_new(bg_migration_vm_start_bh, s); + qemu_bh_schedule(s->vm_start_bh); + + qemu_mutex_unlock_iothread(); + + while (migration_is_active(s)) { + MigIterateState iter_state = bg_migration_iteration_run(s); + if (iter_state == MIG_ITERATE_SKIP) { + continue; + } else if (iter_state == MIG_ITERATE_BREAK) { + break; + } + + /* + * Try to detect any kind of failures, and see whether we + * should stop the migration now. + */ + thr_error = migration_detect_error(s); + if (thr_error == MIG_THR_ERR_FATAL) { + /* Stop migration */ + break; + } + + migration_update_counters(s, qemu_clock_get_ms(QEMU_CLOCK_REALTIME)); + } + + trace_migration_thread_after_loop(); + +fail: + if (early_fail) { + migrate_set_state(&s->state, MIGRATION_STATUS_ACTIVE, + MIGRATION_STATUS_FAILED); + qemu_mutex_unlock_iothread(); + } + + bg_migration_iteration_finish(s); + + qemu_fclose(fb); + object_unref(OBJECT(s)); + rcu_unregister_thread(); + + return NULL; +} + void migrate_fd_connect(MigrationState *s, Error *error_in) { Error *local_err = NULL; @@ -3780,8 +4017,14 @@ void migrate_fd_connect(MigrationState *s, Error *error_in) migrate_fd_cleanup(s); return; } - qemu_thread_create(&s->thread, "live_migration", migration_thread, s, - QEMU_THREAD_JOINABLE); + + if (migrate_background_snapshot()) { + qemu_thread_create(&s->thread, "background_snapshot", + bg_migration_thread, s, QEMU_THREAD_JOINABLE); + } else { + qemu_thread_create(&s->thread, "live_migration", + migration_thread, s, QEMU_THREAD_JOINABLE); + } s->migration_thread_running = true; } diff --git a/migration/migration.h b/migration/migration.h index f40338cfbf..0723955cd7 100644 --- a/migration/migration.h +++ b/migration/migration.h @@ -20,6 +20,7 @@ #include "qemu/thread.h" #include "qemu/coroutine_int.h" #include "io/channel.h" +#include "io/channel-buffer.h" #include "net/announce.h" #include "qom/object.h" @@ -147,8 +148,10 @@ struct MigrationState { /*< public >*/ QemuThread thread; + QEMUBH *vm_start_bh; QEMUBH *cleanup_bh; QEMUFile *to_dst_file; + QIOChannelBuffer *bioc; /* * Protects to_dst_file pointer. We need to make sure we won't * yield or hang during the critical section, since this lock will diff --git a/migration/ram.c b/migration/ram.c index 72c932709a..559170df13 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -1471,6 +1471,7 @@ static RAMBlock *poll_fault_page(RAMState *rs, ram_addr_t *offset) page_address = (void *) uffd_msg.arg.pagefault.address; bs = qemu_ram_block_from_host(page_address, false, offset); assert(bs && (bs->flags & RAM_UF_WRITEPROTECT) != 0); + return bs; } #endif /* CONFIG_LINUX */ @@ -1836,6 +1837,7 @@ static void ram_save_host_page_post(RAMState *rs, PageSearchStatus *pss, /* Un-protect memory range. */ res = uffd_change_protection(rs->uffdio_fd, page_address, run_length, false, false); + /* We don't want to override existing error from ram_save_host_page(). */ if (res < 0 && *res_override >= 0) { *res_override = res; diff --git a/migration/savevm.c b/migration/savevm.c index 5f937a2762..62d5f8a869 100644 --- a/migration/savevm.c +++ b/migration/savevm.c @@ -1352,7 +1352,6 @@ int qemu_savevm_state_complete_precopy_iterable(QEMUFile *f, bool in_postcopy) return 0; } -static int qemu_savevm_state_complete_precopy_non_iterable(QEMUFile *f, bool in_postcopy, bool inactivate_disks) diff --git a/migration/savevm.h b/migration/savevm.h index ba64a7e271..aaee2528ed 100644 --- a/migration/savevm.h +++ b/migration/savevm.h @@ -64,5 +64,7 @@ int qemu_loadvm_state(QEMUFile *f); void qemu_loadvm_state_cleanup(void); int qemu_loadvm_state_main(QEMUFile *f, MigrationIncomingState *mis); int qemu_load_device_state(QEMUFile *f); +int qemu_savevm_state_complete_precopy_non_iterable(QEMUFile *f, + bool in_postcopy, bool inactivate_disks); #endif