diff options
96 files changed, 1747 insertions, 634 deletions
diff --git a/.gitlab-ci.d/container-cross.yml b/.gitlab-ci.d/container-cross.yml index e0d75d5824..2848166ba3 100644 --- a/.gitlab-ci.d/container-cross.yml +++ b/.gitlab-ci.d/container-cross.yml @@ -95,6 +95,7 @@ riscv64-debian-cross-container: allow_failure: true variables: NAME: debian-riscv64-cross + QEMU_JOB_OPTIONAL: 1 # we can however build TCG tests using a non-sid base riscv64-debian-test-cross-container: diff --git a/MAINTAINERS b/MAINTAINERS index ceea4c2bf2..9b28cbb9e9 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -2891,6 +2891,7 @@ F: hw/mem/pc-dimm.c F: include/hw/mem/memory-device.h F: include/hw/mem/nvdimm.h F: include/hw/mem/pc-dimm.h +F: stubs/memory_device.c F: docs/nvdimm.txt SPICE @@ -3143,6 +3144,7 @@ M: Laurent Vivier <lvivier@redhat.com> R: Paolo Bonzini <pbonzini@redhat.com> S: Maintained F: system/qtest.c +F: include/sysemu/qtest.h F: accel/qtest/ F: tests/qtest/ F: docs/devel/qgraph.rst @@ -3864,7 +3866,7 @@ F: docs/block-replication.txt PVRDMA M: Yuval Shaia <yuval.shaia.ml@gmail.com> M: Marcel Apfelbaum <marcel.apfelbaum@gmail.com> -S: Maintained +S: Odd Fixes F: hw/rdma/* F: hw/rdma/vmw/* F: docs/pvrdma.txt diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c index 72e1d1141c..3f7eafe08c 100644 --- a/accel/kvm/kvm-all.c +++ b/accel/kvm/kvm-all.c @@ -174,13 +174,31 @@ void kvm_resample_fd_notify(int gsi) } } -int kvm_get_max_memslots(void) +unsigned int kvm_get_max_memslots(void) { KVMState *s = KVM_STATE(current_accel()); return s->nr_slots; } +unsigned int kvm_get_free_memslots(void) +{ + unsigned int used_slots = 0; + KVMState *s = kvm_state; + int i; + + kvm_slots_lock(); + for (i = 0; i < s->nr_as; i++) { + if (!s->as[i].ml) { + continue; + } + used_slots = MAX(used_slots, s->as[i].ml->nr_used_slots); + } + kvm_slots_unlock(); + + return s->nr_slots - used_slots; +} + /* Called with KVMMemoryListener.slots_lock held */ static KVMSlot *kvm_get_free_slot(KVMMemoryListener *kml) { @@ -196,19 +214,6 @@ static KVMSlot *kvm_get_free_slot(KVMMemoryListener *kml) return NULL; } -bool kvm_has_free_slot(MachineState *ms) -{ - KVMState *s = KVM_STATE(ms->accelerator); - bool result; - KVMMemoryListener *kml = &s->memory_listener; - - kvm_slots_lock(); - result = !!kvm_get_free_slot(kml); - kvm_slots_unlock(); - - return result; -} - /* Called with KVMMemoryListener.slots_lock held */ static KVMSlot *kvm_alloc_slot(KVMMemoryListener *kml) { @@ -1387,6 +1392,7 @@ static void kvm_set_phys_mem(KVMMemoryListener *kml, } start_addr += slot_size; size -= slot_size; + kml->nr_used_slots--; } while (size); return; } @@ -1412,6 +1418,7 @@ static void kvm_set_phys_mem(KVMMemoryListener *kml, ram_start_offset += slot_size; ram += slot_size; size -= slot_size; + kml->nr_used_slots++; } while (size); } diff --git a/accel/stubs/kvm-stub.c b/accel/stubs/kvm-stub.c index 235dc661bc..51f522e52e 100644 --- a/accel/stubs/kvm-stub.c +++ b/accel/stubs/kvm-stub.c @@ -109,9 +109,14 @@ int kvm_irqchip_remove_irqfd_notifier_gsi(KVMState *s, EventNotifier *n, return -ENOSYS; } -bool kvm_has_free_slot(MachineState *ms) +unsigned int kvm_get_max_memslots(void) { - return false; + return 0; +} + +unsigned int kvm_get_free_memslots(void) +{ + return 0; } void kvm_init_cpu_signals(CPUState *cpu) diff --git a/block.c b/block.c index af04c8ac6f..f9cf05ddcf 100644 --- a/block.c +++ b/block.c @@ -279,8 +279,9 @@ bool bdrv_is_read_only(BlockDriverState *bs) return !(bs->open_flags & BDRV_O_RDWR); } -static int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, - bool ignore_allow_rdw, Error **errp) +static int GRAPH_RDLOCK +bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, + bool ignore_allow_rdw, Error **errp) { IO_CODE(); @@ -371,8 +372,9 @@ char *bdrv_get_full_backing_filename_from_filename(const char *backed, * setting @errp. In all other cases, NULL will only be returned with * @errp set. */ -static char *bdrv_make_absolute_filename(BlockDriverState *relative_to, - const char *filename, Error **errp) +static char * GRAPH_RDLOCK +bdrv_make_absolute_filename(BlockDriverState *relative_to, + const char *filename, Error **errp) { char *dir, *full_name; @@ -1192,19 +1194,19 @@ static char *bdrv_child_get_parent_desc(BdrvChild *c) return g_strdup_printf("node '%s'", bdrv_get_node_name(parent)); } -static void bdrv_child_cb_drained_begin(BdrvChild *child) +static void GRAPH_RDLOCK bdrv_child_cb_drained_begin(BdrvChild *child) { BlockDriverState *bs = child->opaque; bdrv_do_drained_begin_quiesce(bs, NULL); } -static bool bdrv_child_cb_drained_poll(BdrvChild *child) +static bool GRAPH_RDLOCK bdrv_child_cb_drained_poll(BdrvChild *child) { BlockDriverState *bs = child->opaque; return bdrv_drain_poll(bs, NULL, false); } -static void bdrv_child_cb_drained_end(BdrvChild *child) +static void GRAPH_RDLOCK bdrv_child_cb_drained_end(BdrvChild *child) { BlockDriverState *bs = child->opaque; bdrv_drained_end(bs); @@ -1250,7 +1252,7 @@ static void bdrv_temp_snapshot_options(int *child_flags, QDict *child_options, *child_flags &= ~BDRV_O_NATIVE_AIO; } -static void bdrv_backing_attach(BdrvChild *c) +static void GRAPH_WRLOCK bdrv_backing_attach(BdrvChild *c) { BlockDriverState *parent = c->opaque; BlockDriverState *backing_hd = c->bs; @@ -1874,7 +1876,10 @@ static int bdrv_open_common(BlockDriverState *bs, BlockBackend *file, } if (file != NULL) { + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(blk_bs(file)); + bdrv_graph_rdunlock_main_loop(); + filename = blk_bs(file)->filename; } else { /* @@ -1901,7 +1906,9 @@ static int bdrv_open_common(BlockDriverState *bs, BlockBackend *file, if (use_bdrv_whitelist && !bdrv_is_whitelisted(drv, ro)) { if (!ro && bdrv_is_whitelisted(drv, true)) { + bdrv_graph_rdlock_main_loop(); ret = bdrv_apply_auto_read_only(bs, NULL, NULL); + bdrv_graph_rdunlock_main_loop(); } else { ret = -ENOTSUP; } @@ -2966,6 +2973,8 @@ static void bdrv_child_free(BdrvChild *child) { assert(!child->bs); GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + assert(!child->next.le_prev); /* not in children list */ g_free(child->name); @@ -3644,7 +3653,10 @@ int bdrv_open_backing_file(BlockDriverState *bs, QDict *parent_options, implicit_backing = !strcmp(bs->auto_backing_file, bs->backing_file); } + bdrv_graph_rdlock_main_loop(); backing_filename = bdrv_get_full_backing_filename(bs, &local_err); + bdrv_graph_rdunlock_main_loop(); + if (local_err) { ret = -EINVAL; error_propagate(errp, local_err); @@ -3675,7 +3687,9 @@ int bdrv_open_backing_file(BlockDriverState *bs, QDict *parent_options, } if (implicit_backing) { + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(backing_hd); + bdrv_graph_rdunlock_main_loop(); pstrcpy(bs->auto_backing_file, sizeof(bs->auto_backing_file), backing_hd->filename); } @@ -4314,8 +4328,8 @@ static int bdrv_reset_options_allowed(BlockDriverState *bs, /* * Returns true if @child can be reached recursively from @bs */ -static bool bdrv_recurse_has_child(BlockDriverState *bs, - BlockDriverState *child) +static bool GRAPH_RDLOCK +bdrv_recurse_has_child(BlockDriverState *bs, BlockDriverState *child) { BdrvChild *c; @@ -4356,15 +4370,12 @@ static bool bdrv_recurse_has_child(BlockDriverState *bs, * * To be called with bs->aio_context locked. */ -static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, - BlockDriverState *bs, - QDict *options, - const BdrvChildClass *klass, - BdrvChildRole role, - bool parent_is_format, - QDict *parent_options, - int parent_flags, - bool keep_old_opts) +static BlockReopenQueue * GRAPH_RDLOCK +bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, BlockDriverState *bs, + QDict *options, const BdrvChildClass *klass, + BdrvChildRole role, bool parent_is_format, + QDict *parent_options, int parent_flags, + bool keep_old_opts) { assert(bs != NULL); @@ -4376,6 +4387,11 @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, GLOBAL_STATE_CODE(); + /* + * Strictly speaking, draining is illegal under GRAPH_RDLOCK. We know that + * we've been called with bdrv_graph_rdlock_main_loop(), though, so it's ok + * in practice. + */ bdrv_drained_begin(bs); if (bs_queue == NULL) { @@ -4517,6 +4533,7 @@ BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue, QDict *options, bool keep_old_opts) { GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); return bdrv_reopen_queue_child(bs_queue, bs, options, NULL, 0, false, NULL, 0, keep_old_opts); @@ -4736,9 +4753,10 @@ int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only, * Callers must make sure that their AioContext locking is still correct after * this. */ -static int bdrv_reopen_parse_file_or_backing(BDRVReopenState *reopen_state, - bool is_backing, Transaction *tran, - Error **errp) +static int GRAPH_UNLOCKED +bdrv_reopen_parse_file_or_backing(BDRVReopenState *reopen_state, + bool is_backing, Transaction *tran, + Error **errp) { BlockDriverState *bs = reopen_state->bs; BlockDriverState *new_child_bs; @@ -4748,6 +4766,7 @@ static int bdrv_reopen_parse_file_or_backing(BDRVReopenState *reopen_state, QObject *value; const char *str; AioContext *ctx, *old_ctx; + bool has_child; int ret; GLOBAL_STATE_CODE(); @@ -4767,7 +4786,13 @@ static int bdrv_reopen_parse_file_or_backing(BDRVReopenState *reopen_state, new_child_bs = bdrv_lookup_bs(NULL, str, errp); if (new_child_bs == NULL) { return -EINVAL; - } else if (bdrv_recurse_has_child(new_child_bs, bs)) { + } + + bdrv_graph_rdlock_main_loop(); + has_child = bdrv_recurse_has_child(new_child_bs, bs); + bdrv_graph_rdunlock_main_loop(); + + if (has_child) { error_setg(errp, "Making '%s' a %s child of '%s' would create a " "cycle", str, child_name, bs->node_name); return -EINVAL; @@ -4866,9 +4891,9 @@ static int bdrv_reopen_parse_file_or_backing(BDRVReopenState *reopen_state, * After calling this function, the transaction @change_child_tran may only be * completed while holding a writer lock for the graph. */ -static int bdrv_reopen_prepare(BDRVReopenState *reopen_state, - BlockReopenQueue *queue, - Transaction *change_child_tran, Error **errp) +static int GRAPH_UNLOCKED +bdrv_reopen_prepare(BDRVReopenState *reopen_state, BlockReopenQueue *queue, + Transaction *change_child_tran, Error **errp) { int ret = -1; int old_flags; @@ -4930,7 +4955,10 @@ static int bdrv_reopen_prepare(BDRVReopenState *reopen_state, * to r/w. Attempting to set to r/w may fail if either BDRV_O_ALLOW_RDWR is * not set, or if the BDS still has copy_on_read enabled */ read_only = !(reopen_state->flags & BDRV_O_RDWR); + + bdrv_graph_rdlock_main_loop(); ret = bdrv_can_set_read_only(reopen_state->bs, read_only, true, &local_err); + bdrv_graph_rdunlock_main_loop(); if (local_err) { error_propagate(errp, local_err); goto error; @@ -4953,7 +4981,9 @@ static int bdrv_reopen_prepare(BDRVReopenState *reopen_state, if (local_err != NULL) { error_propagate(errp, local_err); } else { + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(reopen_state->bs); + bdrv_graph_rdunlock_main_loop(); error_setg(errp, "failed while preparing to reopen image '%s'", reopen_state->bs->filename); } @@ -4962,9 +4992,11 @@ static int bdrv_reopen_prepare(BDRVReopenState *reopen_state, } else { /* It is currently mandatory to have a bdrv_reopen_prepare() * handler for each supported drv. */ + bdrv_graph_rdlock_main_loop(); error_setg(errp, "Block format '%s' used by node '%s' " "does not support reopening files", drv->format_name, bdrv_get_device_or_node_name(reopen_state->bs)); + bdrv_graph_rdunlock_main_loop(); ret = -1; goto error; } @@ -5010,6 +5042,8 @@ static int bdrv_reopen_prepare(BDRVReopenState *reopen_state, if (qdict_size(reopen_state->options)) { const QDictEntry *entry = qdict_first(reopen_state->options); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + do { QObject *new = entry->value; QObject *old = qdict_get(reopen_state->bs->options, entry->key); @@ -5083,7 +5117,7 @@ error: * makes them final by swapping the staging BlockDriverState contents into * the active BlockDriverState contents. */ -static void bdrv_reopen_commit(BDRVReopenState *reopen_state) +static void GRAPH_UNLOCKED bdrv_reopen_commit(BDRVReopenState *reopen_state) { BlockDriver *drv; BlockDriverState *bs; @@ -5100,6 +5134,8 @@ static void bdrv_reopen_commit(BDRVReopenState *reopen_state) drv->bdrv_reopen_commit(reopen_state); } + GRAPH_RDLOCK_GUARD_MAINLOOP(); + /* set BDS specific flags now */ qobject_unref(bs->explicit_options); qobject_unref(bs->options); @@ -5121,9 +5157,7 @@ static void bdrv_reopen_commit(BDRVReopenState *reopen_state) qdict_del(bs->explicit_options, "backing"); qdict_del(bs->options, "backing"); - bdrv_graph_rdlock_main_loop(); bdrv_refresh_limits(bs, NULL, NULL); - bdrv_graph_rdunlock_main_loop(); bdrv_refresh_total_sectors(bs, bs->total_sectors); } @@ -5131,7 +5165,7 @@ static void bdrv_reopen_commit(BDRVReopenState *reopen_state) * Abort the reopen, and delete and free the staged changes in * reopen_state */ -static void bdrv_reopen_abort(BDRVReopenState *reopen_state) +static void GRAPH_UNLOCKED bdrv_reopen_abort(BDRVReopenState *reopen_state) { BlockDriver *drv; @@ -5918,6 +5952,7 @@ int bdrv_drop_intermediate(BlockDriverState *top, BlockDriverState *base, bdrv_ref(top); bdrv_drained_begin(base); + bdrv_graph_rdlock_main_loop(); if (!top->drv || !base->drv) { goto exit; @@ -5942,11 +5977,9 @@ int bdrv_drop_intermediate(BlockDriverState *top, BlockDriverState *base, backing_file_str = base->filename; } - bdrv_graph_rdlock_main_loop(); QLIST_FOREACH(c, &top->parents, next_parent) { updated_children = g_slist_prepend(updated_children, c); } - bdrv_graph_rdunlock_main_loop(); /* * It seems correct to pass detach_subchain=true here, but it triggers @@ -5992,6 +6025,7 @@ int bdrv_drop_intermediate(BlockDriverState *top, BlockDriverState *base, ret = 0; exit: + bdrv_graph_rdunlock_main_loop(); bdrv_drained_end(base); bdrv_unref(top); return ret; @@ -6282,6 +6316,7 @@ BlockDeviceInfoList *bdrv_named_nodes_list(bool flat, BlockDriverState *bs; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); list = NULL; QTAILQ_FOREACH(bs, &graph_bdrv_states, node_list) { @@ -6667,7 +6702,8 @@ void coroutine_fn bdrv_co_debug_event(BlockDriverState *bs, BlkdebugEvent event) bs->drv->bdrv_co_debug_event(bs, event); } -static BlockDriverState *bdrv_find_debug_node(BlockDriverState *bs) +static BlockDriverState * GRAPH_RDLOCK +bdrv_find_debug_node(BlockDriverState *bs) { GLOBAL_STATE_CODE(); while (bs && bs->drv && !bs->drv->bdrv_debug_breakpoint) { @@ -6686,6 +6722,8 @@ int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event, const char *tag) { GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = bdrv_find_debug_node(bs); if (bs) { return bs->drv->bdrv_debug_breakpoint(bs, event, tag); @@ -6697,6 +6735,8 @@ int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event, int bdrv_debug_remove_breakpoint(BlockDriverState *bs, const char *tag) { GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = bdrv_find_debug_node(bs); if (bs) { return bs->drv->bdrv_debug_remove_breakpoint(bs, tag); @@ -6708,6 +6748,8 @@ int bdrv_debug_remove_breakpoint(BlockDriverState *bs, const char *tag) int bdrv_debug_resume(BlockDriverState *bs, const char *tag) { GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + while (bs && (!bs->drv || !bs->drv->bdrv_debug_resume)) { bs = bdrv_primary_bs(bs); } @@ -6722,6 +6764,8 @@ int bdrv_debug_resume(BlockDriverState *bs, const char *tag) bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag) { GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + while (bs && bs->drv && !bs->drv->bdrv_debug_is_suspended) { bs = bdrv_primary_bs(bs); } @@ -6750,6 +6794,7 @@ BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs, BlockDriverState *bs_below; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (!bs || !bs->drv || !backing_file) { return NULL; @@ -6961,6 +7006,7 @@ void bdrv_activate_all(Error **errp) BdrvNextIterator it; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { AioContext *aio_context = bdrv_get_aio_context(bs); @@ -6976,7 +7022,8 @@ void bdrv_activate_all(Error **errp) } } -static bool bdrv_has_bds_parent(BlockDriverState *bs, bool only_active) +static bool GRAPH_RDLOCK +bdrv_has_bds_parent(BlockDriverState *bs, bool only_active) { BdrvChild *parent; GLOBAL_STATE_CODE(); @@ -6993,14 +7040,13 @@ static bool bdrv_has_bds_parent(BlockDriverState *bs, bool only_active) return false; } -static int bdrv_inactivate_recurse(BlockDriverState *bs) +static int GRAPH_RDLOCK bdrv_inactivate_recurse(BlockDriverState *bs) { BdrvChild *child, *parent; int ret; uint64_t cumulative_perms, cumulative_shared_perms; GLOBAL_STATE_CODE(); - GRAPH_RDLOCK_GUARD_MAINLOOP(); if (!bs->drv) { return -ENOMEDIUM; @@ -7066,6 +7112,7 @@ int bdrv_inactivate_all(void) GSList *aio_ctxs = NULL, *ctx; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { AioContext *aio_context = bdrv_get_aio_context(bs); @@ -7205,6 +7252,7 @@ bool bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp) { BdrvOpBlocker *blocker; GLOBAL_STATE_CODE(); + assert((int) op >= 0 && op < BLOCK_OP_TYPE_MAX); if (!QLIST_EMPTY(&bs->op_blockers[op])) { blocker = QLIST_FIRST(&bs->op_blockers[op]); diff --git a/block/backup.c b/block/backup.c index db3791f4d1..9a3c4bdc82 100644 --- a/block/backup.c +++ b/block/backup.c @@ -374,6 +374,7 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs, assert(bs); assert(target); GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); /* QMP interface protects us from these cases */ assert(sync_mode != MIRROR_SYNC_MODE_INCREMENTAL); diff --git a/block/block-backend.c b/block/block-backend.c index efe2e7cbf8..39aac1bbce 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -780,11 +780,12 @@ BlockDriverState *blk_bs(BlockBackend *blk) return blk->root ? blk->root->bs : NULL; } -static BlockBackend *bdrv_first_blk(BlockDriverState *bs) +static BlockBackend * GRAPH_RDLOCK bdrv_first_blk(BlockDriverState *bs) { BdrvChild *child; GLOBAL_STATE_CODE(); + assert_bdrv_graph_readable(); QLIST_FOREACH(child, &bs->parents, next_parent) { if (child->klass == &child_root) { @@ -812,6 +813,8 @@ bool bdrv_is_root_node(BlockDriverState *bs) BdrvChild *c; GLOBAL_STATE_CODE(); + assert_bdrv_graph_readable(); + QLIST_FOREACH(c, &bs->parents, next_parent) { if (c->klass != &child_root) { return false; @@ -2259,6 +2262,7 @@ void blk_activate(BlockBackend *blk, Error **errp) if (qemu_in_coroutine()) { bdrv_co_activate(bs, errp); } else { + GRAPH_RDLOCK_GUARD_MAINLOOP(); bdrv_activate(bs, errp); } } @@ -2384,6 +2388,7 @@ bool blk_op_is_blocked(BlockBackend *blk, BlockOpType op, Error **errp) { BlockDriverState *bs = blk_bs(blk); GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (!bs) { return false; @@ -2901,6 +2906,8 @@ const BdrvChild *blk_root(BlockBackend *blk) int blk_make_empty(BlockBackend *blk, Error **errp) { GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (!blk_is_available(blk)) { error_setg(errp, "No medium inserted"); return -ENOMEDIUM; diff --git a/block/bochs.c b/block/bochs.c index 66e7a58e5e..8c659fa9b9 100644 --- a/block/bochs.c +++ b/block/bochs.c @@ -106,7 +106,9 @@ static int bochs_open(BlockDriverState *bs, QDict *options, int flags, int ret; /* No write support yet */ + bdrv_graph_rdlock_main_loop(); ret = bdrv_apply_auto_read_only(bs, NULL, errp); + bdrv_graph_rdunlock_main_loop(); if (ret < 0) { return ret; } diff --git a/block/cloop.c b/block/cloop.c index 835a0fe3da..773d7918be 100644 --- a/block/cloop.c +++ b/block/cloop.c @@ -67,7 +67,9 @@ static int cloop_open(BlockDriverState *bs, QDict *options, int flags, uint32_t offsets_size, max_compressed_block_size = 1, i; int ret; + bdrv_graph_rdlock_main_loop(); ret = bdrv_apply_auto_read_only(bs, NULL, errp); + bdrv_graph_rdunlock_main_loop(); if (ret < 0) { return ret; } diff --git a/block/commit.c b/block/commit.c index aa45beb0f0..43d1de7577 100644 --- a/block/commit.c +++ b/block/commit.c @@ -434,6 +434,7 @@ int bdrv_commit(BlockDriverState *bs) Error *local_err = NULL; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (!drv) return -ENOMEDIUM; diff --git a/block/copy-before-write.c b/block/copy-before-write.c index aeaff3bb82..4ffabc5ca2 100644 --- a/block/copy-before-write.c +++ b/block/copy-before-write.c @@ -305,7 +305,7 @@ cbw_co_snapshot_block_status(BlockDriverState *bs, return -EACCES; } - ret = bdrv_block_status(child->bs, offset, cur_bytes, pnum, map, file); + ret = bdrv_co_block_status(child->bs, offset, cur_bytes, pnum, map, file); if (child == s->target) { /* * We refer to s->target only for areas that we've written to it. diff --git a/block/copy-on-read.c b/block/copy-on-read.c index b4d6b7efc3..5149fcf63a 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -146,11 +146,11 @@ cor_co_preadv_part(BlockDriverState *bs, int64_t offset, int64_t bytes, local_flags = flags; /* In case of failure, try to copy-on-read anyway */ - ret = bdrv_is_allocated(bs->file->bs, offset, bytes, &n); + ret = bdrv_co_is_allocated(bs->file->bs, offset, bytes, &n); if (ret <= 0) { - ret = bdrv_is_allocated_above(bdrv_backing_chain_next(bs->file->bs), - state->bottom_bs, true, offset, - n, &n); + ret = bdrv_co_is_allocated_above(bdrv_backing_chain_next(bs->file->bs), + state->bottom_bs, true, offset, + n, &n); if (ret > 0 || ret < 0) { local_flags |= BDRV_REQ_COPY_ON_READ; } diff --git a/block/crypto.c b/block/crypto.c index c9c9a39fa3..b3f0233d53 100644 --- a/block/crypto.c +++ b/block/crypto.c @@ -828,7 +828,7 @@ block_crypto_amend_options_generic_luks(BlockDriverState *bs, errp); } -static int +static int GRAPH_RDLOCK block_crypto_amend_options_luks(BlockDriverState *bs, QemuOpts *opts, BlockDriverAmendStatusCB *status_cb, @@ -841,8 +841,6 @@ block_crypto_amend_options_luks(BlockDriverState *bs, QCryptoBlockAmendOptions *amend_options = NULL; int ret = -EINVAL; - assume_graph_lock(); /* FIXME */ - assert(crypto); assert(crypto->block); diff --git a/block/curl.c b/block/curl.c index 0fc42d03d7..419f7c89ef 100644 --- a/block/curl.c +++ b/block/curl.c @@ -696,8 +696,10 @@ static int curl_open(BlockDriverState *bs, QDict *options, int flags, const char *protocol_delimiter; int ret; + bdrv_graph_rdlock_main_loop(); ret = bdrv_apply_auto_read_only(bs, "curl driver does not support writes", errp); + bdrv_graph_rdunlock_main_loop(); if (ret < 0) { return ret; } diff --git a/block/dmg.c b/block/dmg.c index 06a0244a9c..38ee72bbe5 100644 --- a/block/dmg.c +++ b/block/dmg.c @@ -452,7 +452,9 @@ static int dmg_open(BlockDriverState *bs, QDict *options, int flags, int64_t offset; int ret; + bdrv_graph_rdlock_main_loop(); ret = bdrv_apply_auto_read_only(bs, NULL, errp); + bdrv_graph_rdunlock_main_loop(); if (ret < 0) { return ret; } diff --git a/block/export/export.c b/block/export/export.c index 10316b43c5..a8f274e526 100644 --- a/block/export/export.c +++ b/block/export/export.c @@ -83,6 +83,8 @@ BlockExport *blk_exp_add(BlockExportOptions *export, Error **errp) uint64_t perm; int ret; + GLOBAL_STATE_CODE(); + if (!id_wellformed(export->id)) { error_setg(errp, "Invalid block export id"); return NULL; @@ -145,7 +147,9 @@ BlockExport *blk_exp_add(BlockExportOptions *export, Error **errp) * access since the export could be available before migration handover. * ctx was acquired in the caller. */ + bdrv_graph_rdlock_main_loop(); bdrv_activate(bs, NULL); + bdrv_graph_rdunlock_main_loop(); perm = BLK_PERM_CONSISTENT_READ; if (export->writable) { diff --git a/block/gluster.c b/block/gluster.c index ad5fadbe79..cc74af06dc 100644 --- a/block/gluster.c +++ b/block/gluster.c @@ -863,11 +863,13 @@ static int qemu_gluster_open(BlockDriverState *bs, QDict *options, if (ret == -EACCES || ret == -EROFS) { /* Try to degrade to read-only, but if it doesn't work, still use the * normal error message. */ + bdrv_graph_rdlock_main_loop(); if (bdrv_apply_auto_read_only(bs, NULL, NULL) == 0) { open_flags = (open_flags & ~O_RDWR) | O_RDONLY; s->fd = glfs_open(s->glfs, gconf->path, open_flags); ret = s->fd ? 0 : -errno; } + bdrv_graph_rdunlock_main_loop(); } s->supports_seek_data = qemu_gluster_test_seek(s->fd); diff --git a/block/graph-lock.c b/block/graph-lock.c index 58a799065f..e5525ee2db 100644 --- a/block/graph-lock.c +++ b/block/graph-lock.c @@ -106,12 +106,13 @@ static uint32_t reader_count(void) return rd; } -void bdrv_graph_wrlock(BlockDriverState *bs) +void no_coroutine_fn bdrv_graph_wrlock(BlockDriverState *bs) { AioContext *ctx = NULL; GLOBAL_STATE_CODE(); assert(!qatomic_read(&has_writer)); + assert(!qemu_in_coroutine()); /* * Release only non-mainloop AioContext. The mainloop often relies on the diff --git a/block/io.c b/block/io.c index e7f9448d5a..527a1de04e 100644 --- a/block/io.c +++ b/block/io.c @@ -42,13 +42,18 @@ /* Maximum bounce buffer for copy-on-read and write zeroes, in bytes */ #define MAX_BOUNCE_BUFFER (32768 << BDRV_SECTOR_BITS) -static void bdrv_parent_cb_resize(BlockDriverState *bs); +static void coroutine_fn GRAPH_RDLOCK +bdrv_parent_cb_resize(BlockDriverState *bs); + static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, BdrvRequestFlags flags); -static void bdrv_parent_drained_begin(BlockDriverState *bs, BdrvChild *ignore) +static void GRAPH_RDLOCK +bdrv_parent_drained_begin(BlockDriverState *bs, BdrvChild *ignore) { BdrvChild *c, *next; + IO_OR_GS_CODE(); + assert_bdrv_graph_readable(); QLIST_FOREACH_SAFE(c, &bs->parents, next_parent, next) { if (c == ignore) { @@ -70,9 +75,12 @@ void bdrv_parent_drained_end_single(BdrvChild *c) } } -static void bdrv_parent_drained_end(BlockDriverState *bs, BdrvChild *ignore) +static void GRAPH_RDLOCK +bdrv_parent_drained_end(BlockDriverState *bs, BdrvChild *ignore) { BdrvChild *c; + IO_OR_GS_CODE(); + assert_bdrv_graph_readable(); QLIST_FOREACH(c, &bs->parents, next_parent) { if (c == ignore) { @@ -84,17 +92,22 @@ static void bdrv_parent_drained_end(BlockDriverState *bs, BdrvChild *ignore) bool bdrv_parent_drained_poll_single(BdrvChild *c) { + IO_OR_GS_CODE(); + if (c->klass->drained_poll) { return c->klass->drained_poll(c); } return false; } -static bool bdrv_parent_drained_poll(BlockDriverState *bs, BdrvChild *ignore, - bool ignore_bds_parents) +static bool GRAPH_RDLOCK +bdrv_parent_drained_poll(BlockDriverState *bs, BdrvChild *ignore, + bool ignore_bds_parents) { BdrvChild *c, *next; bool busy = false; + IO_OR_GS_CODE(); + assert_bdrv_graph_readable(); QLIST_FOREACH_SAFE(c, &bs->parents, next_parent, next) { if (c == ignore || (ignore_bds_parents && c->klass->parent_is_bds)) { @@ -114,6 +127,7 @@ void bdrv_parent_drained_begin_single(BdrvChild *c) c->quiesced_parent = true; if (c->klass->drained_begin) { + /* called with rdlock taken, but it doesn't really need it. */ c->klass->drained_begin(c); } } @@ -263,6 +277,9 @@ bool bdrv_drain_poll(BlockDriverState *bs, BdrvChild *ignore_parent, static bool bdrv_drain_poll_top_level(BlockDriverState *bs, BdrvChild *ignore_parent) { + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + return bdrv_drain_poll(bs, ignore_parent, false); } @@ -362,6 +379,7 @@ static void bdrv_do_drained_begin(BlockDriverState *bs, BdrvChild *parent, /* Stop things in parent-to-child order */ if (qatomic_fetch_inc(&bs->quiesce_counter) == 0) { + GRAPH_RDLOCK_GUARD_MAINLOOP(); bdrv_parent_drained_begin(bs, parent); if (bs->drv && bs->drv->bdrv_drain_begin) { bs->drv->bdrv_drain_begin(bs); @@ -408,12 +426,16 @@ static void bdrv_do_drained_end(BlockDriverState *bs, BdrvChild *parent) bdrv_co_yield_to_drain(bs, false, parent, false); return; } + + /* At this point, we should be always running in the main loop. */ + GLOBAL_STATE_CODE(); assert(bs->quiesce_counter > 0); GLOBAL_STATE_CODE(); /* Re-enable things in child-to-parent order */ old_quiesce_counter = qatomic_fetch_dec(&bs->quiesce_counter); if (old_quiesce_counter == 1) { + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (bs->drv && bs->drv->bdrv_drain_end) { bs->drv->bdrv_drain_end(bs); } @@ -437,6 +459,8 @@ void bdrv_drain(BlockDriverState *bs) static void bdrv_drain_assert_idle(BlockDriverState *bs) { BdrvChild *child, *next; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); assert(qatomic_read(&bs->in_flight) == 0); QLIST_FOREACH_SAFE(child, &bs->children, next, next) { @@ -450,7 +474,9 @@ static bool bdrv_drain_all_poll(void) { BlockDriverState *bs = NULL; bool result = false; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); /* bdrv_drain_poll() can't make changes to the graph and we are holding the * main AioContext lock, so iterating bdrv_next_all_states() is safe. */ @@ -1223,8 +1249,8 @@ bdrv_co_do_copy_on_readv(BdrvChild *child, int64_t offset, int64_t bytes, ret = 1; /* "already allocated", so nothing will be copied */ pnum = MIN(align_bytes, max_transfer); } else { - ret = bdrv_is_allocated(bs, align_offset, - MIN(align_bytes, max_transfer), &pnum); + ret = bdrv_co_is_allocated(bs, align_offset, + MIN(align_bytes, max_transfer), &pnum); if (ret < 0) { /* * Safe to treat errors in querying allocation as if @@ -1371,7 +1397,7 @@ bdrv_aligned_preadv(BdrvChild *child, BdrvTrackedRequest *req, /* The flag BDRV_REQ_COPY_ON_READ has reached its addressee */ flags &= ~BDRV_REQ_COPY_ON_READ; - ret = bdrv_is_allocated(bs, offset, bytes, &pnum); + ret = bdrv_co_is_allocated(bs, offset, bytes, &pnum); if (ret < 0) { goto out; } @@ -2003,7 +2029,7 @@ bdrv_co_write_req_prepare(BdrvChild *child, int64_t offset, int64_t bytes, } } -static inline void coroutine_fn +static inline void coroutine_fn GRAPH_RDLOCK bdrv_co_write_req_finish(BdrvChild *child, int64_t offset, int64_t bytes, BdrvTrackedRequest *req, int ret) { @@ -2330,6 +2356,7 @@ int bdrv_flush_all(void) int result = 0; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); /* * bdrv queue is managed by record/replay, @@ -2383,9 +2410,9 @@ int bdrv_flush_all(void) * set to the host mapping and BDS corresponding to the guest offset. */ static int coroutine_fn GRAPH_RDLOCK -bdrv_co_block_status(BlockDriverState *bs, bool want_zero, - int64_t offset, int64_t bytes, - int64_t *pnum, int64_t *map, BlockDriverState **file) +bdrv_co_do_block_status(BlockDriverState *bs, bool want_zero, + int64_t offset, int64_t bytes, + int64_t *pnum, int64_t *map, BlockDriverState **file) { int64_t total_size; int64_t n; /* bytes */ @@ -2544,8 +2571,8 @@ bdrv_co_block_status(BlockDriverState *bs, bool want_zero, if (ret & BDRV_BLOCK_RAW) { assert(ret & BDRV_BLOCK_OFFSET_VALID && local_file); - ret = bdrv_co_block_status(local_file, want_zero, local_map, - *pnum, pnum, &local_map, &local_file); + ret = bdrv_co_do_block_status(local_file, want_zero, local_map, + *pnum, pnum, &local_map, &local_file); goto out; } @@ -2572,8 +2599,8 @@ bdrv_co_block_status(BlockDriverState *bs, bool want_zero, int64_t file_pnum; int ret2; - ret2 = bdrv_co_block_status(local_file, want_zero, local_map, - *pnum, &file_pnum, NULL, NULL); + ret2 = bdrv_co_do_block_status(local_file, want_zero, local_map, + *pnum, &file_pnum, NULL, NULL); if (ret2 >= 0) { /* Ignore errors. This is just providing extra information, it * is useful but not necessary. @@ -2640,7 +2667,8 @@ bdrv_co_common_block_status_above(BlockDriverState *bs, return 0; } - ret = bdrv_co_block_status(bs, want_zero, offset, bytes, pnum, map, file); + ret = bdrv_co_do_block_status(bs, want_zero, offset, bytes, pnum, + map, file); ++*depth; if (ret < 0 || *pnum == 0 || ret & BDRV_BLOCK_ALLOCATED || bs == base) { return ret; @@ -2656,8 +2684,8 @@ bdrv_co_common_block_status_above(BlockDriverState *bs, for (p = bdrv_filter_or_cow_bs(bs); include_base || p != base; p = bdrv_filter_or_cow_bs(p)) { - ret = bdrv_co_block_status(p, want_zero, offset, bytes, pnum, map, - file); + ret = bdrv_co_do_block_status(p, want_zero, offset, bytes, pnum, + map, file); ++*depth; if (ret < 0) { return ret; @@ -2723,21 +2751,13 @@ int coroutine_fn bdrv_co_block_status_above(BlockDriverState *bs, bytes, pnum, map, file, NULL); } -int bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, - int64_t offset, int64_t bytes, int64_t *pnum, - int64_t *map, BlockDriverState **file) -{ - IO_CODE(); - return bdrv_common_block_status_above(bs, base, false, true, offset, bytes, - pnum, map, file, NULL); -} - -int bdrv_block_status(BlockDriverState *bs, int64_t offset, int64_t bytes, - int64_t *pnum, int64_t *map, BlockDriverState **file) +int coroutine_fn bdrv_co_block_status(BlockDriverState *bs, int64_t offset, + int64_t bytes, int64_t *pnum, + int64_t *map, BlockDriverState **file) { IO_CODE(); - return bdrv_block_status_above(bs, bdrv_filter_or_cow_bs(bs), - offset, bytes, pnum, map, file); + return bdrv_co_block_status_above(bs, bdrv_filter_or_cow_bs(bs), + offset, bytes, pnum, map, file); } /* @@ -2784,45 +2804,6 @@ int coroutine_fn bdrv_co_is_allocated(BlockDriverState *bs, int64_t offset, return !!(ret & BDRV_BLOCK_ALLOCATED); } -int bdrv_is_allocated(BlockDriverState *bs, int64_t offset, int64_t bytes, - int64_t *pnum) -{ - int ret; - int64_t dummy; - IO_CODE(); - - ret = bdrv_common_block_status_above(bs, bs, true, false, offset, - bytes, pnum ? pnum : &dummy, NULL, - NULL, NULL); - if (ret < 0) { - return ret; - } - return !!(ret & BDRV_BLOCK_ALLOCATED); -} - -/* See bdrv_is_allocated_above for documentation */ -int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *top, - BlockDriverState *base, - bool include_base, int64_t offset, - int64_t bytes, int64_t *pnum) -{ - int depth; - int ret; - IO_CODE(); - - ret = bdrv_co_common_block_status_above(top, base, include_base, false, - offset, bytes, pnum, NULL, NULL, - &depth); - if (ret < 0) { - return ret; - } - - if (ret & BDRV_BLOCK_ALLOCATED) { - return depth; - } - return 0; -} - /* * Given an image chain: ... -> [BASE] -> [INTER1] -> [INTER2] -> [TOP] * @@ -2840,18 +2821,18 @@ int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *top, * words, the result is not necessarily the maximum possible range); * but 'pnum' will only be 0 when end of file is reached. */ -int bdrv_is_allocated_above(BlockDriverState *top, - BlockDriverState *base, - bool include_base, int64_t offset, - int64_t bytes, int64_t *pnum) +int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *bs, + BlockDriverState *base, + bool include_base, int64_t offset, + int64_t bytes, int64_t *pnum) { int depth; int ret; IO_CODE(); - ret = bdrv_common_block_status_above(top, base, include_base, false, - offset, bytes, pnum, NULL, NULL, - &depth); + ret = bdrv_co_common_block_status_above(bs, base, include_base, false, + offset, bytes, pnum, NULL, NULL, + &depth); if (ret < 0) { return ret; } @@ -3551,9 +3532,13 @@ int coroutine_fn bdrv_co_copy_range(BdrvChild *src, int64_t src_offset, bytes, read_flags, write_flags); } -static void bdrv_parent_cb_resize(BlockDriverState *bs) +static void coroutine_fn GRAPH_RDLOCK +bdrv_parent_cb_resize(BlockDriverState *bs) { BdrvChild *c; + + assert_bdrv_graph_readable(); + QLIST_FOREACH(c, &bs->parents, next_parent) { if (c->klass->resize) { c->klass->resize(c); diff --git a/block/iscsi.c b/block/iscsi.c index 5640c8b565..2ff14b7472 100644 --- a/block/iscsi.c +++ b/block/iscsi.c @@ -1925,7 +1925,9 @@ static int iscsi_open(BlockDriverState *bs, QDict *options, int flags, /* Check the write protect flag of the LUN if we want to write */ if (iscsilun->type == TYPE_DISK && (flags & BDRV_O_RDWR) && iscsilun->write_protected) { + bdrv_graph_rdlock_main_loop(); ret = bdrv_apply_auto_read_only(bs, "LUN is write protected", errp); + bdrv_graph_rdunlock_main_loop(); if (ret < 0) { goto out; } diff --git a/block/mirror.c b/block/mirror.c index 3cc0757a03..dcd88de2e3 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -559,9 +559,9 @@ static void coroutine_fn mirror_iteration(MirrorBlockJob *s) assert(!(offset % s->granularity)); WITH_GRAPH_RDLOCK_GUARD() { - ret = bdrv_block_status_above(source, NULL, offset, - nb_chunks * s->granularity, - &io_bytes, NULL, NULL); + ret = bdrv_co_block_status_above(source, NULL, offset, + nb_chunks * s->granularity, + &io_bytes, NULL, NULL); } if (ret < 0) { io_bytes = MIN(nb_chunks * s->granularity, max_io_bytes); @@ -879,8 +879,8 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s) } WITH_GRAPH_RDLOCK_GUARD() { - ret = bdrv_is_allocated_above(bs, s->base_overlay, true, offset, - bytes, &count); + ret = bdrv_co_is_allocated_above(bs, s->base_overlay, true, offset, + bytes, &count); } if (ret < 0) { return ret; diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c index ca2599de44..7645c7e5fb 100644 --- a/block/monitor/block-hmp-cmds.c +++ b/block/monitor/block-hmp-cmds.c @@ -144,6 +144,9 @@ void hmp_drive_del(Monitor *mon, const QDict *qdict) AioContext *aio_context; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = bdrv_find_node(id); if (bs) { qmp_blockdev_del(id, &local_err); @@ -896,6 +899,8 @@ void hmp_info_snapshots(Monitor *mon, const QDict *qdict) SnapshotEntry *snapshot_entry; Error *err = NULL; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = bdrv_all_find_vmstate_bs(NULL, false, NULL, &err); if (!bs) { error_report_err(err); diff --git a/block/nbd.c b/block/nbd.c index 52ebc8b2f5..b9d4f935e0 100644 --- a/block/nbd.c +++ b/block/nbd.c @@ -275,7 +275,8 @@ static bool nbd_client_will_reconnect(BDRVNBDState *s) * Return failure if the server's advertised options are incompatible with the * client's needs. */ -static int nbd_handle_updated_info(BlockDriverState *bs, Error **errp) +static int coroutine_fn GRAPH_RDLOCK +nbd_handle_updated_info(BlockDriverState *bs, Error **errp) { BDRVNBDState *s = (BDRVNBDState *)bs->opaque; int ret; diff --git a/block/nfs.c b/block/nfs.c index c24df49747..f737e19cd3 100644 --- a/block/nfs.c +++ b/block/nfs.c @@ -843,7 +843,7 @@ static void nfs_refresh_filename(BlockDriverState *bs) } } -static char *nfs_dirname(BlockDriverState *bs, Error **errp) +static char * GRAPH_RDLOCK nfs_dirname(BlockDriverState *bs, Error **errp) { NFSClient *client = bs->opaque; diff --git a/block/parallels.c b/block/parallels.c index d026ce9e2f..6b46623241 100644 --- a/block/parallels.c +++ b/block/parallels.c @@ -1363,9 +1363,12 @@ static int parallels_open(BlockDriverState *bs, QDict *options, int flags, bitmap_new(DIV_ROUND_UP(s->header_size, s->bat_dirty_block)); /* Disable migration until bdrv_activate method is added */ + bdrv_graph_rdlock_main_loop(); error_setg(&s->migration_blocker, "The Parallels format used by node '%s' " "does not support live migration", bdrv_get_device_or_node_name(bs)); + bdrv_graph_rdunlock_main_loop(); + ret = migrate_add_blocker(s->migration_blocker, errp); if (ret < 0) { error_setg(errp, "Migration blocker error"); diff --git a/block/qapi-sysemu.c b/block/qapi-sysemu.c index ef07151892..3f614cbc04 100644 --- a/block/qapi-sysemu.c +++ b/block/qapi-sysemu.c @@ -169,14 +169,16 @@ void qmp_blockdev_close_tray(const char *device, } } -static void blockdev_remove_medium(const char *device, const char *id, - Error **errp) +static void GRAPH_UNLOCKED +blockdev_remove_medium(const char *device, const char *id, Error **errp) { BlockBackend *blk; BlockDriverState *bs; AioContext *aio_context; bool has_attached_device; + GLOBAL_STATE_CODE(); + blk = qmp_get_blk(device, id, errp); if (!blk) { return; @@ -205,9 +207,12 @@ static void blockdev_remove_medium(const char *device, const char *id, aio_context = bdrv_get_aio_context(bs); aio_context_acquire(aio_context); + bdrv_graph_rdlock_main_loop(); if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_EJECT, errp)) { + bdrv_graph_rdunlock_main_loop(); goto out; } + bdrv_graph_rdunlock_main_loop(); blk_remove_bs(blk); @@ -279,6 +284,8 @@ static void blockdev_insert_medium(const char *device, const char *id, BlockBackend *blk; BlockDriverState *bs; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + blk = qmp_get_blk(device, id, errp); if (!blk) { return; diff --git a/block/qapi.c b/block/qapi.c index 1cbb0935ff..82a30b38fe 100644 --- a/block/qapi.c +++ b/block/qapi.c @@ -225,9 +225,8 @@ int bdrv_query_snapshot_info_list(BlockDriverState *bs, * Helper function for other query info functions. Store information about @bs * in @info, setting @errp on error. */ -static void bdrv_do_query_node_info(BlockDriverState *bs, - BlockNodeInfo *info, - Error **errp) +static void GRAPH_RDLOCK +bdrv_do_query_node_info(BlockDriverState *bs, BlockNodeInfo *info, Error **errp) { int64_t size; const char *backing_filename; @@ -423,8 +422,8 @@ fail: } /* @p_info will be set only on success. */ -static void bdrv_query_info(BlockBackend *blk, BlockInfo **p_info, - Error **errp) +static void GRAPH_RDLOCK +bdrv_query_info(BlockBackend *blk, BlockInfo **p_info, Error **errp) { BlockInfo *info = g_malloc0(sizeof(*info)); BlockDriverState *bs = blk_bs(blk); @@ -672,6 +671,8 @@ BlockInfoList *qmp_query_block(Error **errp) BlockBackend *blk; Error *local_err = NULL; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + for (blk = blk_all_next(NULL); blk; blk = blk_all_next(blk)) { BlockInfoList *info; diff --git a/block/qcow.c b/block/qcow.c index d56d24ab6d..38a16253b8 100644 --- a/block/qcow.c +++ b/block/qcow.c @@ -301,9 +301,12 @@ static int qcow_open(BlockDriverState *bs, QDict *options, int flags, } /* Disable migration when qcow images are used */ + bdrv_graph_rdlock_main_loop(); error_setg(&s->migration_blocker, "The qcow format used by node '%s' " "does not support live migration", bdrv_get_device_or_node_name(bs)); + bdrv_graph_rdunlock_main_loop(); + ret = migrate_add_blocker(s->migration_blocker, errp); if (ret < 0) { error_free(s->migration_blocker); diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c index ffd5cd3b23..3058309c47 100644 --- a/block/qcow2-bitmap.c +++ b/block/qcow2-bitmap.c @@ -156,10 +156,9 @@ static int64_t get_bitmap_bytes_needed(int64_t len, uint32_t granularity) return DIV_ROUND_UP(num_bits, 8); } -static int check_constraints_on_bitmap(BlockDriverState *bs, - const char *name, - uint32_t granularity, - Error **errp) +static int GRAPH_RDLOCK +check_constraints_on_bitmap(BlockDriverState *bs, const char *name, + uint32_t granularity, Error **errp) { BDRVQcow2State *s = bs->opaque; int granularity_bits = ctz32(granularity); @@ -204,8 +203,9 @@ static int check_constraints_on_bitmap(BlockDriverState *bs, return 0; } -static void clear_bitmap_table(BlockDriverState *bs, uint64_t *bitmap_table, - uint32_t bitmap_table_size) +static void GRAPH_RDLOCK +clear_bitmap_table(BlockDriverState *bs, uint64_t *bitmap_table, + uint32_t bitmap_table_size) { BDRVQcow2State *s = bs->opaque; int i; @@ -259,7 +259,8 @@ fail: return ret; } -static int free_bitmap_clusters(BlockDriverState *bs, Qcow2BitmapTable *tb) +static int GRAPH_RDLOCK +free_bitmap_clusters(BlockDriverState *bs, Qcow2BitmapTable *tb) { int ret; uint64_t *bitmap_table; @@ -730,8 +731,9 @@ out: * Store bitmap list to qcow2 image as a bitmap directory. * Everything is checked. */ -static int bitmap_list_store(BlockDriverState *bs, Qcow2BitmapList *bm_list, - uint64_t *offset, uint64_t *size, bool in_place) +static int GRAPH_RDLOCK +bitmap_list_store(BlockDriverState *bs, Qcow2BitmapList *bm_list, + uint64_t *offset, uint64_t *size, bool in_place) { int ret; uint8_t *dir; @@ -829,8 +831,9 @@ fail: * Bitmap List end */ -static int update_ext_header_and_dir_in_place(BlockDriverState *bs, - Qcow2BitmapList *bm_list) +static int GRAPH_RDLOCK +update_ext_header_and_dir_in_place(BlockDriverState *bs, + Qcow2BitmapList *bm_list) { BDRVQcow2State *s = bs->opaque; int ret; @@ -877,8 +880,8 @@ static int update_ext_header_and_dir_in_place(BlockDriverState *bs, */ } -static int update_ext_header_and_dir(BlockDriverState *bs, - Qcow2BitmapList *bm_list) +static int GRAPH_RDLOCK +update_ext_header_and_dir(BlockDriverState *bs, Qcow2BitmapList *bm_list) { BDRVQcow2State *s = bs->opaque; int ret; @@ -1271,9 +1274,9 @@ out: /* store_bitmap_data() * Store bitmap to image, filling bitmap table accordingly. */ -static uint64_t *store_bitmap_data(BlockDriverState *bs, - BdrvDirtyBitmap *bitmap, - uint32_t *bitmap_table_size, Error **errp) +static uint64_t * GRAPH_RDLOCK +store_bitmap_data(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, + uint32_t *bitmap_table_size, Error **errp) { int ret; BDRVQcow2State *s = bs->opaque; @@ -1370,7 +1373,8 @@ fail: * Store bm->dirty_bitmap to qcow2. * Set bm->table_offset and bm->table_size accordingly. */ -static int store_bitmap(BlockDriverState *bs, Qcow2Bitmap *bm, Error **errp) +static int GRAPH_RDLOCK +store_bitmap(BlockDriverState *bs, Qcow2Bitmap *bm, Error **errp) { int ret; uint64_t *tb; diff --git a/block/qcow2-cache.c b/block/qcow2-cache.c index 01c67bdddc..23d9588b08 100644 --- a/block/qcow2-cache.c +++ b/block/qcow2-cache.c @@ -163,7 +163,8 @@ int qcow2_cache_destroy(Qcow2Cache *c) return 0; } -static int qcow2_cache_flush_dependency(BlockDriverState *bs, Qcow2Cache *c) +static int GRAPH_RDLOCK +qcow2_cache_flush_dependency(BlockDriverState *bs, Qcow2Cache *c) { int ret; @@ -178,7 +179,8 @@ static int qcow2_cache_flush_dependency(BlockDriverState *bs, Qcow2Cache *c) return 0; } -static int qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i) +static int GRAPH_RDLOCK +qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i) { BDRVQcow2State *s = bs->opaque; int ret = 0; @@ -318,8 +320,9 @@ int qcow2_cache_empty(BlockDriverState *bs, Qcow2Cache *c) return 0; } -static int qcow2_cache_do_get(BlockDriverState *bs, Qcow2Cache *c, - uint64_t offset, void **table, bool read_from_disk) +static int GRAPH_RDLOCK +qcow2_cache_do_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset, + void **table, bool read_from_disk) { BDRVQcow2State *s = bs->opaque; int i; diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c index f4f6cd6ad0..904f00d1b3 100644 --- a/block/qcow2-cluster.c +++ b/block/qcow2-cluster.c @@ -207,8 +207,9 @@ int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size, * the cache is used; otherwise the L2 slice is loaded from the image * file. */ -static int l2_load(BlockDriverState *bs, uint64_t offset, - uint64_t l2_offset, uint64_t **l2_slice) +static int GRAPH_RDLOCK +l2_load(BlockDriverState *bs, uint64_t offset, + uint64_t l2_offset, uint64_t **l2_slice) { BDRVQcow2State *s = bs->opaque; int start_of_slice = l2_entry_size(s) * @@ -269,7 +270,7 @@ int qcow2_write_l1_entry(BlockDriverState *bs, int l1_index) * */ -static int l2_allocate(BlockDriverState *bs, int l1_index) +static int GRAPH_RDLOCK l2_allocate(BlockDriverState *bs, int l1_index) { BDRVQcow2State *s = bs->opaque; uint64_t old_l2_offset; @@ -751,9 +752,9 @@ fail: * * Returns 0 on success, -errno in failure case */ -static int get_cluster_table(BlockDriverState *bs, uint64_t offset, - uint64_t **new_l2_slice, - int *new_l2_index) +static int GRAPH_RDLOCK +get_cluster_table(BlockDriverState *bs, uint64_t offset, + uint64_t **new_l2_slice, int *new_l2_index) { BDRVQcow2State *s = bs->opaque; unsigned int l2_index; @@ -1155,11 +1156,10 @@ void coroutine_fn qcow2_alloc_cluster_abort(BlockDriverState *bs, QCowL2Meta *m) * * Returns 0 on success, -errno on failure. */ -static int coroutine_fn calculate_l2_meta(BlockDriverState *bs, - uint64_t host_cluster_offset, - uint64_t guest_offset, unsigned bytes, - uint64_t *l2_slice, QCowL2Meta **m, - bool keep_old) +static int coroutine_fn GRAPH_RDLOCK +calculate_l2_meta(BlockDriverState *bs, uint64_t host_cluster_offset, + uint64_t guest_offset, unsigned bytes, uint64_t *l2_slice, + QCowL2Meta **m, bool keep_old) { BDRVQcow2State *s = bs->opaque; int sc_index, l2_index = offset_to_l2_slice_index(s, guest_offset); @@ -1490,9 +1490,9 @@ static int coroutine_fn handle_dependencies(BlockDriverState *bs, * * -errno: in error cases */ -static int coroutine_fn handle_copied(BlockDriverState *bs, - uint64_t guest_offset, uint64_t *host_offset, uint64_t *bytes, - QCowL2Meta **m) +static int coroutine_fn GRAPH_RDLOCK +handle_copied(BlockDriverState *bs, uint64_t guest_offset, + uint64_t *host_offset, uint64_t *bytes, QCowL2Meta **m) { BDRVQcow2State *s = bs->opaque; int l2_index; @@ -1600,10 +1600,9 @@ out: * function has been waiting for another request and the allocation must be * restarted, but the whole request should not be failed. */ -static int coroutine_fn do_alloc_cluster_offset(BlockDriverState *bs, - uint64_t guest_offset, - uint64_t *host_offset, - uint64_t *nb_clusters) +static int coroutine_fn GRAPH_RDLOCK +do_alloc_cluster_offset(BlockDriverState *bs, uint64_t guest_offset, + uint64_t *host_offset, uint64_t *nb_clusters) { BDRVQcow2State *s = bs->opaque; @@ -1658,9 +1657,9 @@ static int coroutine_fn do_alloc_cluster_offset(BlockDriverState *bs, * * -errno: in error cases */ -static int coroutine_fn handle_alloc(BlockDriverState *bs, - uint64_t guest_offset, uint64_t *host_offset, uint64_t *bytes, - QCowL2Meta **m) +static int coroutine_fn GRAPH_RDLOCK +handle_alloc(BlockDriverState *bs, uint64_t guest_offset, + uint64_t *host_offset, uint64_t *bytes, QCowL2Meta **m) { BDRVQcow2State *s = bs->opaque; int l2_index; @@ -1898,9 +1897,9 @@ again: * all clusters in the same L2 slice) and returns the number of discarded * clusters. */ -static int discard_in_l2_slice(BlockDriverState *bs, uint64_t offset, - uint64_t nb_clusters, - enum qcow2_discard_type type, bool full_discard) +static int GRAPH_RDLOCK +discard_in_l2_slice(BlockDriverState *bs, uint64_t offset, uint64_t nb_clusters, + enum qcow2_discard_type type, bool full_discard) { BDRVQcow2State *s = bs->opaque; uint64_t *l2_slice; @@ -2037,7 +2036,7 @@ fail: * all clusters in the same L2 slice) and returns the number of zeroed * clusters. */ -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK zero_in_l2_slice(BlockDriverState *bs, uint64_t offset, uint64_t nb_clusters, int flags) { @@ -2093,7 +2092,7 @@ zero_in_l2_slice(BlockDriverState *bs, uint64_t offset, return nb_clusters; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK zero_l2_subclusters(BlockDriverState *bs, uint64_t offset, unsigned nb_subclusters) { @@ -2231,11 +2230,12 @@ fail: * status_cb(). l1_entries contains the total number of L1 entries and * *visited_l1_entries counts all visited L1 entries. */ -static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table, - int l1_size, int64_t *visited_l1_entries, - int64_t l1_entries, - BlockDriverAmendStatusCB *status_cb, - void *cb_opaque) +static int GRAPH_RDLOCK +expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table, + int l1_size, int64_t *visited_l1_entries, + int64_t l1_entries, + BlockDriverAmendStatusCB *status_cb, + void *cb_opaque) { BDRVQcow2State *s = bs->opaque; bool is_active_l1 = (l1_table == s->l1_table); diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c index 996d1217d0..0266542cee 100644 --- a/block/qcow2-refcount.c +++ b/block/qcow2-refcount.c @@ -229,9 +229,9 @@ static void set_refcount_ro6(void *refcount_array, uint64_t index, } -static int load_refcount_block(BlockDriverState *bs, - int64_t refcount_block_offset, - void **refcount_block) +static int GRAPH_RDLOCK +load_refcount_block(BlockDriverState *bs, int64_t refcount_block_offset, + void **refcount_block) { BDRVQcow2State *s = bs->opaque; @@ -302,8 +302,9 @@ static int in_same_refcount_block(BDRVQcow2State *s, uint64_t offset_a, * * Returns 0 on success or -errno in error case */ -static int alloc_refcount_block(BlockDriverState *bs, - int64_t cluster_index, void **refcount_block) +static int GRAPH_RDLOCK +alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index, + void **refcount_block) { BDRVQcow2State *s = bs->opaque; unsigned int refcount_table_index; @@ -806,12 +807,9 @@ found: /* XXX: cache several refcount block clusters ? */ /* @addend is the absolute value of the addend; if @decrease is set, @addend * will be subtracted from the current refcount, otherwise it will be added */ -static int update_refcount(BlockDriverState *bs, - int64_t offset, - int64_t length, - uint64_t addend, - bool decrease, - enum qcow2_discard_type type) +static int GRAPH_RDLOCK +update_refcount(BlockDriverState *bs, int64_t offset, int64_t length, + uint64_t addend, bool decrease, enum qcow2_discard_type type) { BDRVQcow2State *s = bs->opaque; int64_t start, last, cluster_offset; @@ -967,8 +965,8 @@ int qcow2_update_cluster_refcount(BlockDriverState *bs, /* return < 0 if error */ -static int64_t alloc_clusters_noref(BlockDriverState *bs, uint64_t size, - uint64_t max) +static int64_t GRAPH_RDLOCK +alloc_clusters_noref(BlockDriverState *bs, uint64_t size, uint64_t max) { BDRVQcow2State *s = bs->opaque; uint64_t i, nb_clusters, refcount; @@ -2302,7 +2300,7 @@ calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res, * Compares the actual reference count for each cluster in the image against the * refcount as reported by the refcount structures on-disk. */ -static void coroutine_fn +static void coroutine_fn GRAPH_RDLOCK compare_refcounts(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix, bool *rebuild, int64_t *highest_cluster, @@ -3103,20 +3101,22 @@ int qcow2_pre_write_overlap_check(BlockDriverState *bs, int ign, int64_t offset, * * @allocated should be set to true if a new cluster has been allocated. */ -typedef int (RefblockFinishOp)(BlockDriverState *bs, uint64_t **reftable, - uint64_t reftable_index, uint64_t *reftable_size, - void *refblock, bool refblock_empty, - bool *allocated, Error **errp); +typedef int /* GRAPH_RDLOCK_PTR */ + (RefblockFinishOp)(BlockDriverState *bs, uint64_t **reftable, + uint64_t reftable_index, uint64_t *reftable_size, + void *refblock, bool refblock_empty, + bool *allocated, Error **errp); /** * This "operation" for walk_over_reftable() allocates the refblock on disk (if * it is not empty) and inserts its offset into the new reftable. The size of * this new reftable is increased as required. */ -static int alloc_refblock(BlockDriverState *bs, uint64_t **reftable, - uint64_t reftable_index, uint64_t *reftable_size, - void *refblock, bool refblock_empty, bool *allocated, - Error **errp) +static int GRAPH_RDLOCK +alloc_refblock(BlockDriverState *bs, uint64_t **reftable, + uint64_t reftable_index, uint64_t *reftable_size, + void *refblock, bool refblock_empty, bool *allocated, + Error **errp) { BDRVQcow2State *s = bs->opaque; int64_t offset; @@ -3166,10 +3166,11 @@ static int alloc_refblock(BlockDriverState *bs, uint64_t **reftable, * offset specified by the new reftable's entry. It does not modify the new * reftable or change any refcounts. */ -static int flush_refblock(BlockDriverState *bs, uint64_t **reftable, - uint64_t reftable_index, uint64_t *reftable_size, - void *refblock, bool refblock_empty, bool *allocated, - Error **errp) +static int GRAPH_RDLOCK +flush_refblock(BlockDriverState *bs, uint64_t **reftable, + uint64_t reftable_index, uint64_t *reftable_size, + void *refblock, bool refblock_empty, bool *allocated, + Error **errp) { BDRVQcow2State *s = bs->opaque; int64_t offset; @@ -3210,16 +3211,17 @@ static int flush_refblock(BlockDriverState *bs, uint64_t **reftable, * * @allocated is set to true if a new cluster has been allocated. */ -static int walk_over_reftable(BlockDriverState *bs, uint64_t **new_reftable, - uint64_t *new_reftable_index, - uint64_t *new_reftable_size, - void *new_refblock, int new_refblock_size, - int new_refcount_bits, - RefblockFinishOp *operation, bool *allocated, - Qcow2SetRefcountFunc *new_set_refcount, - BlockDriverAmendStatusCB *status_cb, - void *cb_opaque, int index, int total, - Error **errp) +static int GRAPH_RDLOCK +walk_over_reftable(BlockDriverState *bs, uint64_t **new_reftable, + uint64_t *new_reftable_index, + uint64_t *new_reftable_size, + void *new_refblock, int new_refblock_size, + int new_refcount_bits, + RefblockFinishOp *operation, bool *allocated, + Qcow2SetRefcountFunc *new_set_refcount, + BlockDriverAmendStatusCB *status_cb, + void *cb_opaque, int index, int total, + Error **errp) { BDRVQcow2State *s = bs->opaque; uint64_t reftable_index; @@ -3545,8 +3547,8 @@ done: return ret; } -static int64_t coroutine_fn get_refblock_offset(BlockDriverState *bs, - uint64_t offset) +static int64_t coroutine_fn GRAPH_RDLOCK +get_refblock_offset(BlockDriverState *bs, uint64_t offset) { BDRVQcow2State *s = bs->opaque; uint32_t index = offset_to_reftable_index(s, offset); @@ -3565,7 +3567,7 @@ static int64_t coroutine_fn get_refblock_offset(BlockDriverState *bs, return covering_refblock_offset; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK qcow2_discard_refcount_block(BlockDriverState *bs, uint64_t discard_block_offs) { BDRVQcow2State *s = bs->opaque; diff --git a/block/qcow2.c b/block/qcow2.c index 5a3c537f14..aa01d9e7b5 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -536,7 +536,7 @@ int qcow2_mark_dirty(BlockDriverState *bs) * function when there are no pending requests, it does not guard against * concurrent requests dirtying the image. */ -static int qcow2_mark_clean(BlockDriverState *bs) +static int GRAPH_RDLOCK qcow2_mark_clean(BlockDriverState *bs) { BDRVQcow2State *s = bs->opaque; @@ -570,7 +570,8 @@ int qcow2_mark_corrupt(BlockDriverState *bs) * Marks the image as consistent, i.e., unsets the corrupt bit, and flushes * before if necessary. */ -static int coroutine_fn qcow2_mark_consistent(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK +qcow2_mark_consistent(BlockDriverState *bs) { BDRVQcow2State *s = bs->opaque; @@ -980,10 +981,9 @@ typedef struct Qcow2ReopenState { QCryptoBlockOpenOptions *crypto_opts; /* Disk encryption runtime options */ } Qcow2ReopenState; -static int qcow2_update_options_prepare(BlockDriverState *bs, - Qcow2ReopenState *r, - QDict *options, int flags, - Error **errp) +static int GRAPH_RDLOCK +qcow2_update_options_prepare(BlockDriverState *bs, Qcow2ReopenState *r, + QDict *options, int flags, Error **errp) { BDRVQcow2State *s = bs->opaque; QemuOpts *opts = NULL; @@ -1260,7 +1260,7 @@ static void qcow2_update_options_abort(BlockDriverState *bs, qapi_free_QCryptoBlockOpenOptions(r->crypto_opts); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK qcow2_update_options(BlockDriverState *bs, QDict *options, int flags, Error **errp) { @@ -1969,13 +1969,17 @@ static void qcow2_refresh_limits(BlockDriverState *bs, Error **errp) bs->bl.pdiscard_alignment = s->cluster_size; } -static int qcow2_reopen_prepare(BDRVReopenState *state, - BlockReopenQueue *queue, Error **errp) +static int GRAPH_UNLOCKED +qcow2_reopen_prepare(BDRVReopenState *state,BlockReopenQueue *queue, + Error **errp) { BDRVQcow2State *s = state->bs->opaque; Qcow2ReopenState *r; int ret; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + r = g_new0(Qcow2ReopenState, 1); state->opaque = r; @@ -2038,6 +2042,8 @@ static void qcow2_reopen_commit(BDRVReopenState *state) static void qcow2_reopen_commit_post(BDRVReopenState *state) { + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (state->flags & BDRV_O_RDWR) { Error *local_err = NULL; @@ -2731,7 +2737,7 @@ fail_nometa: return ret; } -static int qcow2_inactivate(BlockDriverState *bs) +static int GRAPH_RDLOCK qcow2_inactivate(BlockDriverState *bs) { BDRVQcow2State *s = bs->opaque; int ret, result = 0; @@ -2766,7 +2772,8 @@ static int qcow2_inactivate(BlockDriverState *bs) return result; } -static void qcow2_do_close(BlockDriverState *bs, bool close_data_file) +static void coroutine_mixed_fn GRAPH_RDLOCK +qcow2_do_close(BlockDriverState *bs, bool close_data_file) { BDRVQcow2State *s = bs->opaque; qemu_vfree(s->l1_table); @@ -2793,18 +2800,24 @@ static void qcow2_do_close(BlockDriverState *bs, bool close_data_file) g_free(s->image_backing_format); if (close_data_file && has_data_file(bs)) { + GLOBAL_STATE_CODE(); + bdrv_graph_rdunlock_main_loop(); bdrv_graph_wrlock(NULL); bdrv_unref_child(bs, s->data_file); bdrv_graph_wrunlock(); s->data_file = NULL; + bdrv_graph_rdlock_main_loop(); } qcow2_refcount_close(bs); qcow2_free_snapshots(bs); } -static void qcow2_close(BlockDriverState *bs) +static void GRAPH_UNLOCKED qcow2_close(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + qcow2_do_close(bs, true); } @@ -3991,7 +4004,8 @@ finish: } -static bool is_zero(BlockDriverState *bs, int64_t offset, int64_t bytes) +static bool coroutine_fn GRAPH_RDLOCK +is_zero(BlockDriverState *bs, int64_t offset, int64_t bytes) { int64_t nr; int res; @@ -4012,7 +4026,7 @@ static bool is_zero(BlockDriverState *bs, int64_t offset, int64_t bytes) * backing file. So, we need a loop. */ do { - res = bdrv_block_status_above(bs, NULL, offset, bytes, &nr, NULL, NULL); + res = bdrv_co_block_status_above(bs, NULL, offset, bytes, &nr, NULL, NULL); offset += nr; bytes -= nr; } while (res >= 0 && (res & BDRV_BLOCK_ZERO) && nr && bytes); @@ -4076,8 +4090,8 @@ qcow2_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, return ret; } -static coroutine_fn int qcow2_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { int ret; BDRVQcow2State *s = bs->opaque; @@ -4822,7 +4836,7 @@ fail: return ret; } -static int make_completely_empty(BlockDriverState *bs) +static int GRAPH_RDLOCK make_completely_empty(BlockDriverState *bs) { BDRVQcow2State *s = bs->opaque; Error *local_err = NULL; @@ -4973,7 +4987,7 @@ fail: return ret; } -static int qcow2_make_empty(BlockDriverState *bs) +static int GRAPH_RDLOCK qcow2_make_empty(BlockDriverState *bs) { BDRVQcow2State *s = bs->opaque; uint64_t offset, end_offset; @@ -5017,7 +5031,7 @@ static int qcow2_make_empty(BlockDriverState *bs) return ret; } -static coroutine_fn int qcow2_co_flush_to_os(BlockDriverState *bs) +static coroutine_fn GRAPH_RDLOCK int qcow2_co_flush_to_os(BlockDriverState *bs) { BDRVQcow2State *s = bs->opaque; int ret; @@ -5366,7 +5380,7 @@ qcow2_co_load_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) return bs->drv->bdrv_co_preadv_part(bs, offset, qiov->size, qiov, 0, 0); } -static int qcow2_has_compressed_clusters(BlockDriverState *bs) +static int GRAPH_RDLOCK qcow2_has_compressed_clusters(BlockDriverState *bs) { int64_t offset = 0; int64_t bytes = bdrv_getlength(bs); @@ -5402,9 +5416,10 @@ static int qcow2_has_compressed_clusters(BlockDriverState *bs) * Downgrades an image's version. To achieve this, any incompatible features * have to be removed. */ -static int qcow2_downgrade(BlockDriverState *bs, int target_version, - BlockDriverAmendStatusCB *status_cb, void *cb_opaque, - Error **errp) +static int GRAPH_RDLOCK +qcow2_downgrade(BlockDriverState *bs, int target_version, + BlockDriverAmendStatusCB *status_cb, void *cb_opaque, + Error **errp) { BDRVQcow2State *s = bs->opaque; int current_version = s->qcow_version; @@ -5512,9 +5527,10 @@ static int qcow2_downgrade(BlockDriverState *bs, int target_version, * features of older versions, some things may have to be presented * differently. */ -static int qcow2_upgrade(BlockDriverState *bs, int target_version, - BlockDriverAmendStatusCB *status_cb, void *cb_opaque, - Error **errp) +static int GRAPH_RDLOCK +qcow2_upgrade(BlockDriverState *bs, int target_version, + BlockDriverAmendStatusCB *status_cb, void *cb_opaque, + Error **errp) { BDRVQcow2State *s = bs->opaque; bool need_snapshot_update; @@ -5640,11 +5656,10 @@ static void qcow2_amend_helper_cb(BlockDriverState *bs, info->original_cb_opaque); } -static int qcow2_amend_options(BlockDriverState *bs, QemuOpts *opts, - BlockDriverAmendStatusCB *status_cb, - void *cb_opaque, - bool force, - Error **errp) +static int GRAPH_RDLOCK +qcow2_amend_options(BlockDriverState *bs, QemuOpts *opts, + BlockDriverAmendStatusCB *status_cb, void *cb_opaque, + bool force, Error **errp) { BDRVQcow2State *s = bs->opaque; int old_version = s->qcow_version, new_version = old_version; diff --git a/block/qcow2.h b/block/qcow2.h index f789ce3ae0..29958c512b 100644 --- a/block/qcow2.h +++ b/block/qcow2.h @@ -838,9 +838,10 @@ int qcow2_mark_dirty(BlockDriverState *bs); int qcow2_mark_corrupt(BlockDriverState *bs); int qcow2_update_header(BlockDriverState *bs); -void qcow2_signal_corruption(BlockDriverState *bs, bool fatal, int64_t offset, - int64_t size, const char *message_format, ...) - G_GNUC_PRINTF(5, 6); +void GRAPH_RDLOCK +qcow2_signal_corruption(BlockDriverState *bs, bool fatal, int64_t offset, + int64_t size, const char *message_format, ...) + G_GNUC_PRINTF(5, 6); int qcow2_validate_table(BlockDriverState *bs, uint64_t offset, uint64_t entries, size_t entry_len, @@ -851,33 +852,41 @@ int qcow2_validate_table(BlockDriverState *bs, uint64_t offset, int coroutine_fn GRAPH_RDLOCK qcow2_refcount_init(BlockDriverState *bs); void qcow2_refcount_close(BlockDriverState *bs); -int qcow2_get_refcount(BlockDriverState *bs, int64_t cluster_index, - uint64_t *refcount); +int GRAPH_RDLOCK qcow2_get_refcount(BlockDriverState *bs, int64_t cluster_index, + uint64_t *refcount); -int qcow2_update_cluster_refcount(BlockDriverState *bs, int64_t cluster_index, - uint64_t addend, bool decrease, - enum qcow2_discard_type type); +int GRAPH_RDLOCK +qcow2_update_cluster_refcount(BlockDriverState *bs, int64_t cluster_index, + uint64_t addend, bool decrease, + enum qcow2_discard_type type); -int64_t qcow2_refcount_area(BlockDriverState *bs, uint64_t offset, - uint64_t additional_clusters, bool exact_size, - int new_refblock_index, - uint64_t new_refblock_offset); +int64_t GRAPH_RDLOCK +qcow2_refcount_area(BlockDriverState *bs, uint64_t offset, + uint64_t additional_clusters, bool exact_size, + int new_refblock_index, + uint64_t new_refblock_offset); -int64_t qcow2_alloc_clusters(BlockDriverState *bs, uint64_t size); -int64_t coroutine_fn qcow2_alloc_clusters_at(BlockDriverState *bs, uint64_t offset, - int64_t nb_clusters); -int64_t coroutine_fn GRAPH_RDLOCK qcow2_alloc_bytes(BlockDriverState *bs, int size); -void qcow2_free_clusters(BlockDriverState *bs, - int64_t offset, int64_t size, - enum qcow2_discard_type type); -void qcow2_free_any_cluster(BlockDriverState *bs, uint64_t l2_entry, - enum qcow2_discard_type type); +int64_t GRAPH_RDLOCK +qcow2_alloc_clusters(BlockDriverState *bs, uint64_t size); -int qcow2_update_snapshot_refcount(BlockDriverState *bs, - int64_t l1_table_offset, int l1_size, int addend); +int64_t GRAPH_RDLOCK coroutine_fn +qcow2_alloc_clusters_at(BlockDriverState *bs, uint64_t offset, + int64_t nb_clusters); -int qcow2_flush_caches(BlockDriverState *bs); -int qcow2_write_caches(BlockDriverState *bs); +int64_t coroutine_fn GRAPH_RDLOCK qcow2_alloc_bytes(BlockDriverState *bs, int size); +void GRAPH_RDLOCK qcow2_free_clusters(BlockDriverState *bs, + int64_t offset, int64_t size, + enum qcow2_discard_type type); +void GRAPH_RDLOCK +qcow2_free_any_cluster(BlockDriverState *bs, uint64_t l2_entry, + enum qcow2_discard_type type); + +int GRAPH_RDLOCK +qcow2_update_snapshot_refcount(BlockDriverState *bs, int64_t l1_table_offset, + int l1_size, int addend); + +int GRAPH_RDLOCK qcow2_flush_caches(BlockDriverState *bs); +int GRAPH_RDLOCK qcow2_write_caches(BlockDriverState *bs); int coroutine_fn qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix); @@ -885,39 +894,48 @@ void qcow2_process_discards(BlockDriverState *bs, int ret); int qcow2_check_metadata_overlap(BlockDriverState *bs, int ign, int64_t offset, int64_t size); -int qcow2_pre_write_overlap_check(BlockDriverState *bs, int ign, int64_t offset, - int64_t size, bool data_file); +int GRAPH_RDLOCK +qcow2_pre_write_overlap_check(BlockDriverState *bs, int ign, int64_t offset, + int64_t size, bool data_file); + int coroutine_fn qcow2_inc_refcounts_imrt(BlockDriverState *bs, BdrvCheckResult *res, void **refcount_table, int64_t *refcount_table_size, int64_t offset, int64_t size); -int qcow2_change_refcount_order(BlockDriverState *bs, int refcount_order, - BlockDriverAmendStatusCB *status_cb, - void *cb_opaque, Error **errp); +int GRAPH_RDLOCK +qcow2_change_refcount_order(BlockDriverState *bs, int refcount_order, + BlockDriverAmendStatusCB *status_cb, + void *cb_opaque, Error **errp); int coroutine_fn GRAPH_RDLOCK qcow2_shrink_reftable(BlockDriverState *bs); -int64_t coroutine_fn qcow2_get_last_cluster(BlockDriverState *bs, int64_t size); + +int64_t coroutine_fn GRAPH_RDLOCK +qcow2_get_last_cluster(BlockDriverState *bs, int64_t size); int coroutine_fn GRAPH_RDLOCK qcow2_detect_metadata_preallocation(BlockDriverState *bs); /* qcow2-cluster.c functions */ -int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size, - bool exact_size); +int GRAPH_RDLOCK +qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size, bool exact_size); int coroutine_fn GRAPH_RDLOCK qcow2_shrink_l1_table(BlockDriverState *bs, uint64_t max_size); -int qcow2_write_l1_entry(BlockDriverState *bs, int l1_index); +int GRAPH_RDLOCK qcow2_write_l1_entry(BlockDriverState *bs, int l1_index); int qcow2_encrypt_sectors(BDRVQcow2State *s, int64_t sector_num, uint8_t *buf, int nb_sectors, bool enc, Error **errp); -int qcow2_get_host_offset(BlockDriverState *bs, uint64_t offset, - unsigned int *bytes, uint64_t *host_offset, - QCow2SubclusterType *subcluster_type); -int coroutine_fn qcow2_alloc_host_offset(BlockDriverState *bs, uint64_t offset, - unsigned int *bytes, - uint64_t *host_offset, QCowL2Meta **m); +int GRAPH_RDLOCK +qcow2_get_host_offset(BlockDriverState *bs, uint64_t offset, + unsigned int *bytes, uint64_t *host_offset, + QCow2SubclusterType *subcluster_type); + +int coroutine_fn GRAPH_RDLOCK +qcow2_alloc_host_offset(BlockDriverState *bs, uint64_t offset, + unsigned int *bytes, uint64_t *host_offset, + QCowL2Meta **m); + int coroutine_fn GRAPH_RDLOCK qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs, uint64_t offset, int compressed_size, uint64_t *host_offset); @@ -927,26 +945,33 @@ void qcow2_parse_compressed_l2_entry(BlockDriverState *bs, uint64_t l2_entry, int coroutine_fn GRAPH_RDLOCK qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m); -void coroutine_fn qcow2_alloc_cluster_abort(BlockDriverState *bs, QCowL2Meta *m); -int qcow2_cluster_discard(BlockDriverState *bs, uint64_t offset, - uint64_t bytes, enum qcow2_discard_type type, - bool full_discard); +void coroutine_fn GRAPH_RDLOCK +qcow2_alloc_cluster_abort(BlockDriverState *bs, QCowL2Meta *m); + +int GRAPH_RDLOCK +qcow2_cluster_discard(BlockDriverState *bs, uint64_t offset, uint64_t bytes, + enum qcow2_discard_type type, bool full_discard); int coroutine_fn GRAPH_RDLOCK qcow2_subcluster_zeroize(BlockDriverState *bs, uint64_t offset, uint64_t bytes, int flags); -int qcow2_expand_zero_clusters(BlockDriverState *bs, - BlockDriverAmendStatusCB *status_cb, - void *cb_opaque); +int GRAPH_RDLOCK +qcow2_expand_zero_clusters(BlockDriverState *bs, + BlockDriverAmendStatusCB *status_cb, + void *cb_opaque); /* qcow2-snapshot.c functions */ -int qcow2_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info); -int qcow2_snapshot_goto(BlockDriverState *bs, const char *snapshot_id); -int qcow2_snapshot_delete(BlockDriverState *bs, - const char *snapshot_id, - const char *name, - Error **errp); +int GRAPH_RDLOCK +qcow2_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info); + +int GRAPH_RDLOCK +qcow2_snapshot_goto(BlockDriverState *bs, const char *snapshot_id); + +int GRAPH_RDLOCK +qcow2_snapshot_delete(BlockDriverState *bs, const char *snapshot_id, + const char *name, Error **errp); + int qcow2_snapshot_list(BlockDriverState *bs, QEMUSnapshotInfo **psn_tab); int qcow2_snapshot_load_tmp(BlockDriverState *bs, const char *snapshot_id, @@ -956,15 +981,15 @@ int qcow2_snapshot_load_tmp(BlockDriverState *bs, void qcow2_free_snapshots(BlockDriverState *bs); int coroutine_fn GRAPH_RDLOCK qcow2_read_snapshots(BlockDriverState *bs, Error **errp); -int qcow2_write_snapshots(BlockDriverState *bs); +int GRAPH_RDLOCK qcow2_write_snapshots(BlockDriverState *bs); int coroutine_fn GRAPH_RDLOCK qcow2_check_read_snapshot_table(BlockDriverState *bs, BdrvCheckResult *result, BdrvCheckMode fix); -int coroutine_fn qcow2_check_fix_snapshot_table(BlockDriverState *bs, - BdrvCheckResult *result, - BdrvCheckMode fix); +int coroutine_fn GRAPH_RDLOCK +qcow2_check_fix_snapshot_table(BlockDriverState *bs, BdrvCheckResult *result, + BdrvCheckMode fix); /* qcow2-cache.c functions */ Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables, @@ -972,19 +997,23 @@ Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables, int qcow2_cache_destroy(Qcow2Cache *c); void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table); -int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c); -int qcow2_cache_write(BlockDriverState *bs, Qcow2Cache *c); -int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c, - Qcow2Cache *dependency); +int GRAPH_RDLOCK qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c); +int GRAPH_RDLOCK qcow2_cache_write(BlockDriverState *bs, Qcow2Cache *c); +int GRAPH_RDLOCK qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c, + Qcow2Cache *dependency); void qcow2_cache_depends_on_flush(Qcow2Cache *c); void qcow2_cache_clean_unused(Qcow2Cache *c); -int qcow2_cache_empty(BlockDriverState *bs, Qcow2Cache *c); +int GRAPH_RDLOCK qcow2_cache_empty(BlockDriverState *bs, Qcow2Cache *c); + +int GRAPH_RDLOCK +qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset, + void **table); + +int GRAPH_RDLOCK +qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset, + void **table); -int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset, - void **table); -int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset, - void **table); void qcow2_cache_put(Qcow2Cache *c, void **table); void *qcow2_cache_is_table_offset(Qcow2Cache *c, uint64_t offset); void qcow2_cache_discard(Qcow2Cache *c, void *table); @@ -998,18 +1027,22 @@ bool coroutine_fn GRAPH_RDLOCK qcow2_load_dirty_bitmaps(BlockDriverState *bs, bool *header_updated, Error **errp); bool qcow2_get_bitmap_info_list(BlockDriverState *bs, Qcow2BitmapInfoList **info_list, Error **errp); -int qcow2_reopen_bitmaps_rw(BlockDriverState *bs, Error **errp); +int GRAPH_RDLOCK qcow2_reopen_bitmaps_rw(BlockDriverState *bs, Error **errp); +int GRAPH_RDLOCK qcow2_reopen_bitmaps_ro(BlockDriverState *bs, Error **errp); int coroutine_fn qcow2_truncate_bitmaps_check(BlockDriverState *bs, Error **errp); -bool qcow2_store_persistent_dirty_bitmaps(BlockDriverState *bs, - bool release_stored, Error **errp); -int qcow2_reopen_bitmaps_ro(BlockDriverState *bs, Error **errp); -bool coroutine_fn qcow2_co_can_store_new_dirty_bitmap(BlockDriverState *bs, - const char *name, - uint32_t granularity, - Error **errp); -int coroutine_fn qcow2_co_remove_persistent_dirty_bitmap(BlockDriverState *bs, - const char *name, - Error **errp); + +bool GRAPH_RDLOCK +qcow2_store_persistent_dirty_bitmaps(BlockDriverState *bs, bool release_stored, + Error **errp); + +bool coroutine_fn GRAPH_RDLOCK +qcow2_co_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, + uint32_t granularity, Error **errp); + +int coroutine_fn GRAPH_RDLOCK +qcow2_co_remove_persistent_dirty_bitmap(BlockDriverState *bs, const char *name, + Error **errp); + bool qcow2_supports_persistent_dirty_bitmap(BlockDriverState *bs); uint64_t qcow2_get_persistent_dirty_bitmap_size(BlockDriverState *bs, uint32_t cluster_size); diff --git a/block/quorum.c b/block/quorum.c index 05220cab7f..d3ffc2ee33 100644 --- a/block/quorum.c +++ b/block/quorum.c @@ -206,7 +206,7 @@ static void quorum_report_bad(QuorumOpType type, uint64_t offset, end_sector - start_sector); } -static void quorum_report_failure(QuorumAIOCB *acb) +static void GRAPH_RDLOCK quorum_report_failure(QuorumAIOCB *acb) { const char *reference = bdrv_get_device_or_node_name(acb->bs); int64_t start_sector = acb->offset / BDRV_SECTOR_SIZE; @@ -219,7 +219,7 @@ static void quorum_report_failure(QuorumAIOCB *acb) static int quorum_vote_error(QuorumAIOCB *acb); -static bool quorum_has_too_much_io_failed(QuorumAIOCB *acb) +static bool GRAPH_RDLOCK quorum_has_too_much_io_failed(QuorumAIOCB *acb) { BDRVQuorumState *s = acb->bs->opaque; diff --git a/block/raw-format.c b/block/raw-format.c index a8bdee5279..8ff03adfa4 100644 --- a/block/raw-format.c +++ b/block/raw-format.c @@ -505,7 +505,9 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags, BDRV_REQ_ZERO_WRITE; if (bs->probed && !bdrv_is_read_only(bs)) { + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(bs->file->bs); + bdrv_graph_rdunlock_main_loop(); fprintf(stderr, "WARNING: Image format was not specified for '%s' and probing " "guessed raw.\n" diff --git a/block/rbd.c b/block/rbd.c index 472ca05cba..84bb2fa5d7 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -1168,7 +1168,9 @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, /* If we are using an rbd snapshot, we must be r/o, otherwise * leave as-is */ if (s->snap != NULL) { + bdrv_graph_rdlock_main_loop(); r = bdrv_apply_auto_read_only(bs, "rbd snapshots are read-only", errp); + bdrv_graph_rdunlock_main_loop(); if (r < 0) { goto failed_post_open; } @@ -1208,6 +1210,8 @@ static int qemu_rbd_reopen_prepare(BDRVReopenState *state, BDRVRBDState *s = state->bs->opaque; int ret = 0; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (s->snap && state->flags & BDRV_O_RDWR) { error_setg(errp, "Cannot change node '%s' to r/w when using RBD snapshot", diff --git a/block/replication.c b/block/replication.c index dd166d2d82..d522c7396f 100644 --- a/block/replication.c +++ b/block/replication.c @@ -276,10 +276,10 @@ replication_co_writev(BlockDriverState *bs, int64_t sector_num, while (remaining_sectors > 0) { int64_t count; - ret = bdrv_is_allocated_above(top->bs, base->bs, false, - sector_num * BDRV_SECTOR_SIZE, - remaining_sectors * BDRV_SECTOR_SIZE, - &count); + ret = bdrv_co_is_allocated_above(top->bs, base->bs, false, + sector_num * BDRV_SECTOR_SIZE, + remaining_sectors * BDRV_SECTOR_SIZE, + &count); if (ret < 0) { goto out1; } @@ -307,13 +307,16 @@ out: return ret; } -static void secondary_do_checkpoint(BlockDriverState *bs, Error **errp) +static void GRAPH_UNLOCKED +secondary_do_checkpoint(BlockDriverState *bs, Error **errp) { BDRVReplicationState *s = bs->opaque; BdrvChild *active_disk = bs->file; Error *local_err = NULL; int ret; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (!s->backup_job) { error_setg(errp, "Backup job was cancelled unexpectedly"); return; @@ -427,7 +430,8 @@ static void backup_job_completed(void *opaque, int ret) backup_job_cleanup(bs); } -static bool check_top_bs(BlockDriverState *top_bs, BlockDriverState *bs) +static bool GRAPH_RDLOCK +check_top_bs(BlockDriverState *top_bs, BlockDriverState *bs) { BdrvChild *child; @@ -458,6 +462,8 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode, Error *local_err = NULL; BackupPerf perf = { .use_copy_range = true, .max_workers = 1 }; + GLOBAL_STATE_CODE(); + aio_context = bdrv_get_aio_context(bs); aio_context_acquire(aio_context); s = bs->opaque; @@ -504,12 +510,15 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode, return; } + bdrv_graph_rdlock_main_loop(); secondary_disk = hidden_disk->bs->backing; if (!secondary_disk->bs || !bdrv_has_blk(secondary_disk->bs)) { error_setg(errp, "The secondary disk doesn't have block backend"); + bdrv_graph_rdunlock_main_loop(); aio_context_release(aio_context); return; } + bdrv_graph_rdunlock_main_loop(); /* verify the length */ active_length = bdrv_getlength(active_disk->bs); @@ -526,13 +535,16 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode, /* Must be true, or the bdrv_getlength() calls would have failed */ assert(active_disk->bs->drv && hidden_disk->bs->drv); + bdrv_graph_rdlock_main_loop(); if (!active_disk->bs->drv->bdrv_make_empty || !hidden_disk->bs->drv->bdrv_make_empty) { error_setg(errp, "Active disk or hidden disk doesn't support make_empty"); aio_context_release(aio_context); + bdrv_graph_rdunlock_main_loop(); return; } + bdrv_graph_rdunlock_main_loop(); /* reopen the backing file in r/w mode */ reopen_backing_file(bs, true, &local_err); @@ -566,8 +578,6 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode, return; } - bdrv_graph_wrunlock(); - /* start backup job now */ error_setg(&s->blocker, "Block device is in use by internal backup job"); @@ -576,6 +586,7 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode, if (!top_bs || !bdrv_is_root_node(top_bs) || !check_top_bs(top_bs, bs)) { error_setg(errp, "No top_bs or it is invalid"); + bdrv_graph_wrunlock(); reopen_backing_file(bs, false, NULL); aio_context_release(aio_context); return; @@ -583,6 +594,8 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode, bdrv_op_block_all(top_bs, s->blocker); bdrv_op_unblock(top_bs, BLOCK_OP_TYPE_DATAPLANE, s->blocker); + bdrv_graph_wrunlock(); + s->backup_job = backup_job_create( NULL, s->secondary_disk->bs, s->hidden_disk->bs, 0, MIRROR_SYNC_MODE_NONE, NULL, 0, false, NULL, diff --git a/block/snapshot.c b/block/snapshot.c index b86b5b24ad..6e16eb803a 100644 --- a/block/snapshot.c +++ b/block/snapshot.c @@ -155,11 +155,15 @@ bool bdrv_snapshot_find_by_id_and_name(BlockDriverState *bs, * back if the given BDS does not support snapshots. * Return NULL if there is no BDS to (safely) fall back to. */ -static BdrvChild *bdrv_snapshot_fallback_child(BlockDriverState *bs) +static BdrvChild * GRAPH_RDLOCK +bdrv_snapshot_fallback_child(BlockDriverState *bs) { BdrvChild *fallback = bdrv_primary_child(bs); BdrvChild *child; + GLOBAL_STATE_CODE(); + assert_bdrv_graph_readable(); + /* We allow fallback only to primary child */ if (!fallback) { return NULL; @@ -182,8 +186,10 @@ static BdrvChild *bdrv_snapshot_fallback_child(BlockDriverState *bs) return fallback; } -static BlockDriverState *bdrv_snapshot_fallback(BlockDriverState *bs) +static BlockDriverState * GRAPH_RDLOCK +bdrv_snapshot_fallback(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); return child_bs(bdrv_snapshot_fallback_child(bs)); } @@ -254,7 +260,10 @@ int bdrv_snapshot_goto(BlockDriverState *bs, return ret; } + bdrv_graph_rdlock_main_loop(); fallback = bdrv_snapshot_fallback_child(bs); + bdrv_graph_rdunlock_main_loop(); + if (fallback) { QDict *options; QDict *file_options; @@ -302,7 +311,10 @@ int bdrv_snapshot_goto(BlockDriverState *bs, * respective option (with the qdict_put_str() call above). * Assert that .bdrv_open() has attached the right BDS as primary child. */ + bdrv_graph_rdlock_main_loop(); assert(bdrv_primary_bs(bs) == fallback_bs); + bdrv_graph_rdunlock_main_loop(); + bdrv_unref(fallback_bs); return ret; } @@ -374,10 +386,12 @@ int bdrv_snapshot_delete(BlockDriverState *bs, int bdrv_snapshot_list(BlockDriverState *bs, QEMUSnapshotInfo **psn_info) { + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + BlockDriver *drv = bs->drv; BlockDriverState *fallback_bs = bdrv_snapshot_fallback(bs); - GLOBAL_STATE_CODE(); if (!drv) { return -ENOMEDIUM; } @@ -418,6 +432,7 @@ int bdrv_snapshot_load_tmp(BlockDriverState *bs, BlockDriver *drv = bs->drv; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (!drv) { error_setg(errp, QERR_DEVICE_HAS_NO_MEDIUM, bdrv_get_device_name(bs)); @@ -462,9 +477,9 @@ int bdrv_snapshot_load_tmp_by_id_or_name(BlockDriverState *bs, } -static int bdrv_all_get_snapshot_devices(bool has_devices, strList *devices, - GList **all_bdrvs, - Error **errp) +static int GRAPH_RDLOCK +bdrv_all_get_snapshot_devices(bool has_devices, strList *devices, + GList **all_bdrvs, Error **errp) { g_autoptr(GList) bdrvs = NULL; @@ -496,8 +511,11 @@ static int bdrv_all_get_snapshot_devices(bool has_devices, strList *devices, } -static bool bdrv_all_snapshots_includes_bs(BlockDriverState *bs) +static bool GRAPH_RDLOCK bdrv_all_snapshots_includes_bs(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); + assert_bdrv_graph_readable(); + if (!bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) { return false; } @@ -518,6 +536,7 @@ bool bdrv_all_can_snapshot(bool has_devices, strList *devices, GList *iterbdrvs; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return false; @@ -554,6 +573,7 @@ int bdrv_all_delete_snapshot(const char *name, GList *iterbdrvs; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return -1; @@ -593,10 +613,15 @@ int bdrv_all_goto_snapshot(const char *name, { g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + int ret; GLOBAL_STATE_CODE(); - if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { + bdrv_graph_rdlock_main_loop(); + ret = bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp); + bdrv_graph_rdunlock_main_loop(); + + if (ret < 0) { return -1; } @@ -605,15 +630,22 @@ int bdrv_all_goto_snapshot(const char *name, BlockDriverState *bs = iterbdrvs->data; AioContext *ctx = bdrv_get_aio_context(bs); int ret = 0; + bool all_snapshots_includes_bs; aio_context_acquire(ctx); - if (devices || bdrv_all_snapshots_includes_bs(bs)) { + bdrv_graph_rdlock_main_loop(); + all_snapshots_includes_bs = bdrv_all_snapshots_includes_bs(bs); + bdrv_graph_rdunlock_main_loop(); + + if (devices || all_snapshots_includes_bs) { ret = bdrv_snapshot_goto(bs, name, errp); } aio_context_release(ctx); if (ret < 0) { + bdrv_graph_rdlock_main_loop(); error_prepend(errp, "Could not load snapshot '%s' on '%s': ", name, bdrv_get_device_or_node_name(bs)); + bdrv_graph_rdunlock_main_loop(); return -1; } @@ -631,6 +663,7 @@ int bdrv_all_has_snapshot(const char *name, GList *iterbdrvs; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return -1; @@ -673,7 +706,9 @@ int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn, { g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return -1; @@ -715,6 +750,7 @@ BlockDriverState *bdrv_all_find_vmstate_bs(const char *vmstate_bs, GList *iterbdrvs; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return NULL; diff --git a/block/stream.c b/block/stream.c index 133cb72fb4..ddaab7dbbd 100644 --- a/block/stream.c +++ b/block/stream.c @@ -172,7 +172,7 @@ static int coroutine_fn stream_run(Job *job, Error **errp) copy = false; WITH_GRAPH_RDLOCK_GUARD() { - ret = bdrv_is_allocated(unfiltered_bs, offset, STREAM_CHUNK, &n); + ret = bdrv_co_is_allocated(unfiltered_bs, offset, STREAM_CHUNK, &n); if (ret == 1) { /* Allocated in the top, no need to copy. */ } else if (ret >= 0) { @@ -180,9 +180,9 @@ static int coroutine_fn stream_run(Job *job, Error **errp) * Copy if allocated in the intermediate images. Limit to the * known-unallocated area [offset, offset+n*BDRV_SECTOR_SIZE). */ - ret = bdrv_is_allocated_above(bdrv_cow_bs(unfiltered_bs), - s->base_overlay, true, - offset, n, &n); + ret = bdrv_co_is_allocated_above(bdrv_cow_bs(unfiltered_bs), + s->base_overlay, true, + offset, n, &n); /* Finish early if end of backing file has been reached */ if (ret == 0 && n == 0) { n = len - offset; diff --git a/block/vdi.c b/block/vdi.c index 934e1b849b..3ed43b6f35 100644 --- a/block/vdi.c +++ b/block/vdi.c @@ -492,9 +492,12 @@ static int vdi_open(BlockDriverState *bs, QDict *options, int flags, } /* Disable migration when vdi images are used */ + bdrv_graph_rdlock_main_loop(); error_setg(&s->migration_blocker, "The vdi format used by node '%s' " "does not support live migration", bdrv_get_device_or_node_name(bs)); + bdrv_graph_rdunlock_main_loop(); + ret = migrate_add_blocker(s->migration_blocker, errp); if (ret < 0) { error_free(s->migration_blocker); diff --git a/block/vhdx.c b/block/vhdx.c index a67edcc03e..73cb214fb4 100644 --- a/block/vhdx.c +++ b/block/vhdx.c @@ -1001,11 +1001,15 @@ static int vhdx_open(BlockDriverState *bs, QDict *options, int flags, uint64_t signature; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + ret = bdrv_open_file_child(NULL, options, "file", bs, errp); if (ret < 0) { return ret; } + GRAPH_RDLOCK_GUARD_MAINLOOP(); + s->bat = NULL; s->first_visible_write = true; diff --git a/block/vhdx.h b/block/vhdx.h index 455a627a46..85594a5380 100644 --- a/block/vhdx.h +++ b/block/vhdx.h @@ -410,8 +410,9 @@ uint32_t vhdx_checksum_calc(uint32_t crc, uint8_t *buf, size_t size, bool vhdx_checksum_is_valid(uint8_t *buf, size_t size, int crc_offset); -int vhdx_parse_log(BlockDriverState *bs, BDRVVHDXState *s, bool *flushed, - Error **errp); +int GRAPH_RDLOCK +vhdx_parse_log(BlockDriverState *bs, BDRVVHDXState *s, bool *flushed, + Error **errp); int coroutine_fn GRAPH_RDLOCK vhdx_log_write_and_flush(BlockDriverState *bs, BDRVVHDXState *s, diff --git a/block/vmdk.c b/block/vmdk.c index e90649c8bf..8a3b152798 100644 --- a/block/vmdk.c +++ b/block/vmdk.c @@ -578,8 +578,8 @@ static int vmdk_add_extent(BlockDriverState *bs, return 0; } -static int vmdk_init_tables(BlockDriverState *bs, VmdkExtent *extent, - Error **errp) +static int GRAPH_RDLOCK +vmdk_init_tables(BlockDriverState *bs, VmdkExtent *extent, Error **errp) { int ret; size_t l1_size; @@ -641,9 +641,9 @@ static int vmdk_init_tables(BlockDriverState *bs, VmdkExtent *extent, return ret; } -static int vmdk_open_vmfs_sparse(BlockDriverState *bs, - BdrvChild *file, - int flags, Error **errp) +static int GRAPH_RDLOCK +vmdk_open_vmfs_sparse(BlockDriverState *bs, BdrvChild *file, int flags, + Error **errp) { int ret; uint32_t magic; @@ -797,9 +797,9 @@ static int check_se_sparse_volatile_header(VMDKSESparseVolatileHeader *header, return 0; } -static int vmdk_open_se_sparse(BlockDriverState *bs, - BdrvChild *file, - int flags, Error **errp) +static int GRAPH_RDLOCK +vmdk_open_se_sparse(BlockDriverState *bs, BdrvChild *file, int flags, + Error **errp) { int ret; VMDKSESparseConstHeader const_header; @@ -913,9 +913,9 @@ static char *vmdk_read_desc(BdrvChild *file, uint64_t desc_offset, Error **errp) return buf; } -static int vmdk_open_vmdk4(BlockDriverState *bs, - BdrvChild *file, - int flags, QDict *options, Error **errp) +static int GRAPH_RDLOCK +vmdk_open_vmdk4(BlockDriverState *bs, BdrvChild *file, int flags, + QDict *options, Error **errp) { int ret; uint32_t magic; @@ -1095,8 +1095,9 @@ static int vmdk_parse_description(const char *desc, const char *opt_name, } /* Open an extent file and append to bs array */ -static int vmdk_open_sparse(BlockDriverState *bs, BdrvChild *file, int flags, - char *buf, QDict *options, Error **errp) +static int GRAPH_RDLOCK +vmdk_open_sparse(BlockDriverState *bs, BdrvChild *file, int flags, + char *buf, QDict *options, Error **errp) { uint32_t magic; @@ -1123,8 +1124,9 @@ static const char *next_line(const char *s) return s; } -static int vmdk_parse_extents(const char *desc, BlockDriverState *bs, - QDict *options, Error **errp) +static int GRAPH_RDLOCK +vmdk_parse_extents(const char *desc, BlockDriverState *bs, QDict *options, + Error **errp) { int ret; int matches; @@ -1143,6 +1145,8 @@ static int vmdk_parse_extents(const char *desc, BlockDriverState *bs, char extent_opt_prefix[32]; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + for (p = desc; *p; p = next_line(p)) { /* parse extent line in one of below formats: * @@ -1223,9 +1227,11 @@ static int vmdk_parse_extents(const char *desc, BlockDriverState *bs, ret = vmdk_add_extent(bs, extent_file, true, sectors, 0, 0, 0, 0, 0, &extent, errp); if (ret < 0) { + bdrv_graph_rdunlock_main_loop(); bdrv_graph_wrlock(NULL); bdrv_unref_child(bs, extent_file); bdrv_graph_wrunlock(); + bdrv_graph_rdlock_main_loop(); goto out; } extent->flat_start_offset = flat_offset << 9; @@ -1240,26 +1246,32 @@ static int vmdk_parse_extents(const char *desc, BlockDriverState *bs, } g_free(buf); if (ret) { + bdrv_graph_rdunlock_main_loop(); bdrv_graph_wrlock(NULL); bdrv_unref_child(bs, extent_file); bdrv_graph_wrunlock(); + bdrv_graph_rdlock_main_loop(); goto out; } extent = &s->extents[s->num_extents - 1]; } else if (!strcmp(type, "SESPARSE")) { ret = vmdk_open_se_sparse(bs, extent_file, bs->open_flags, errp); if (ret) { + bdrv_graph_rdunlock_main_loop(); bdrv_graph_wrlock(NULL); bdrv_unref_child(bs, extent_file); bdrv_graph_wrunlock(); + bdrv_graph_rdlock_main_loop(); goto out; } extent = &s->extents[s->num_extents - 1]; } else { error_setg(errp, "Unsupported extent type '%s'", type); + bdrv_graph_rdunlock_main_loop(); bdrv_graph_wrlock(NULL); bdrv_unref_child(bs, extent_file); bdrv_graph_wrunlock(); + bdrv_graph_rdlock_main_loop(); ret = -ENOTSUP; goto out; } @@ -1283,8 +1295,9 @@ out: return ret; } -static int vmdk_open_desc_file(BlockDriverState *bs, int flags, char *buf, - QDict *options, Error **errp) +static int GRAPH_RDLOCK +vmdk_open_desc_file(BlockDriverState *bs, int flags, char *buf, QDict *options, + Error **errp) { int ret; char ct[128]; @@ -2900,7 +2913,7 @@ static int vmdk_has_zero_init(BlockDriverState *bs) return 1; } -static VmdkExtentInfo *vmdk_get_extent_info(VmdkExtent *extent) +static VmdkExtentInfo * GRAPH_RDLOCK vmdk_get_extent_info(VmdkExtent *extent) { VmdkExtentInfo *info = g_new0(VmdkExtentInfo, 1); @@ -2977,8 +2990,8 @@ vmdk_co_check(BlockDriverState *bs, BdrvCheckResult *result, BdrvCheckMode fix) return ret; } -static ImageInfoSpecific *vmdk_get_specific_info(BlockDriverState *bs, - Error **errp) +static ImageInfoSpecific * GRAPH_RDLOCK +vmdk_get_specific_info(BlockDriverState *bs, Error **errp) { int i; BDRVVmdkState *s = bs->opaque; diff --git a/block/vpc.c b/block/vpc.c index ceb87dd3d8..945847fe4a 100644 --- a/block/vpc.c +++ b/block/vpc.c @@ -446,9 +446,12 @@ static int vpc_open(BlockDriverState *bs, QDict *options, int flags, } /* Disable migration when VHD images are used */ + bdrv_graph_rdlock_main_loop(); error_setg(&s->migration_blocker, "The vpc format used by node '%s' " "does not support live migration", bdrv_get_device_or_node_name(bs)); + bdrv_graph_rdunlock_main_loop(); + ret = migrate_add_blocker(s->migration_blocker, errp); if (ret < 0) { error_free(s->migration_blocker); diff --git a/block/vvfat.c b/block/vvfat.c index 856b479c91..b0415798c0 100644 --- a/block/vvfat.c +++ b/block/vvfat.c @@ -1144,6 +1144,8 @@ static int vvfat_open(BlockDriverState *bs, QDict *options, int flags, QemuOpts *opts; int ret; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + #ifdef DEBUG vvv = s; #endif @@ -1480,8 +1482,8 @@ vvfat_read(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, int nb_sector if (s->qcow) { int64_t n; int ret; - ret = bdrv_is_allocated(s->qcow->bs, sector_num * BDRV_SECTOR_SIZE, - (nb_sectors - i) * BDRV_SECTOR_SIZE, &n); + ret = bdrv_co_is_allocated(s->qcow->bs, sector_num * BDRV_SECTOR_SIZE, + (nb_sectors - i) * BDRV_SECTOR_SIZE, &n); if (ret < 0) { return ret; } @@ -1806,10 +1808,10 @@ cluster_was_modified(BDRVVVFATState *s, uint32_t cluster_num) } for (i = 0; !was_modified && i < s->sectors_per_cluster; i++) { - was_modified = bdrv_is_allocated(s->qcow->bs, - (cluster2sector(s, cluster_num) + - i) * BDRV_SECTOR_SIZE, - BDRV_SECTOR_SIZE, NULL); + was_modified = bdrv_co_is_allocated(s->qcow->bs, + (cluster2sector(s, cluster_num) + + i) * BDRV_SECTOR_SIZE, + BDRV_SECTOR_SIZE, NULL); } /* @@ -1967,9 +1969,9 @@ get_cluster_count_for_direntry(BDRVVVFATState* s, direntry_t* direntry, const ch for (i = 0; i < s->sectors_per_cluster; i++) { int res; - res = bdrv_is_allocated(s->qcow->bs, - (offs + i) * BDRV_SECTOR_SIZE, - BDRV_SECTOR_SIZE, NULL); + res = bdrv_co_is_allocated(s->qcow->bs, + (offs + i) * BDRV_SECTOR_SIZE, + BDRV_SECTOR_SIZE, NULL); if (res < 0) { return -1; } diff --git a/blockdev.c b/blockdev.c index 325b7a3bef..a01c62596b 100644 --- a/blockdev.c +++ b/blockdev.c @@ -1041,6 +1041,8 @@ static BlockDriverState *qmp_get_root_bs(const char *name, Error **errp) BlockDriverState *bs; AioContext *aio_context; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = bdrv_lookup_bs(name, name, errp); if (bs == NULL) { return NULL; @@ -1136,6 +1138,9 @@ SnapshotInfo *qmp_blockdev_snapshot_delete_internal_sync(const char *device, SnapshotInfo *info = NULL; int ret; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = qmp_get_root_bs(device, errp); if (!bs) { return NULL; @@ -1221,6 +1226,9 @@ static void internal_snapshot_action(BlockdevSnapshotInternal *internal, AioContext *aio_context; int ret1; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + tran_add(tran, &internal_snapshot_drv, state); device = internal->device; @@ -1309,6 +1317,9 @@ static void internal_snapshot_abort(void *opaque) AioContext *aio_context; Error *local_error = NULL; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (!state->created) { return; } @@ -1654,6 +1665,8 @@ static void drive_backup_action(DriveBackup *backup, bool set_backing_hd = false; int ret; + GLOBAL_STATE_CODE(); + tran_add(tran, &drive_backup_drv, state); if (!backup->has_mode) { @@ -1683,9 +1696,12 @@ static void drive_backup_action(DriveBackup *backup, } /* Early check to avoid creating target */ + bdrv_graph_rdlock_main_loop(); if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp)) { + bdrv_graph_rdunlock_main_loop(); goto out; } + bdrv_graph_rdunlock_main_loop(); flags = bs->open_flags | BDRV_O_RDWR; @@ -1724,7 +1740,10 @@ static void drive_backup_action(DriveBackup *backup, BlockDriverState *explicit_backing = bdrv_skip_implicit_filters(source); + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(explicit_backing); + bdrv_graph_rdunlock_main_loop(); + bdrv_img_create(backup->target, format, explicit_backing->filename, explicit_backing->drv->format_name, NULL, @@ -2344,10 +2363,13 @@ void coroutine_fn qmp_block_resize(const char *device, const char *node_name, return; } + bdrv_graph_co_rdlock(); if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_RESIZE, NULL)) { error_setg(errp, QERR_DEVICE_IN_USE, device); + bdrv_graph_co_rdunlock(); return; } + bdrv_graph_co_rdunlock(); blk = blk_co_new_with_bs(bs, BLK_PERM_RESIZE, BLK_PERM_ALL, errp); if (!blk) { @@ -2387,6 +2409,8 @@ void qmp_block_stream(const char *job_id, const char *device, Error *local_err = NULL; int job_flags = JOB_DEFAULT; + GLOBAL_STATE_CODE(); + if (base && base_node) { error_setg(errp, "'base' and 'base-node' cannot be specified " "at the same time"); @@ -2437,7 +2461,10 @@ void qmp_block_stream(const char *job_id, const char *device, goto out; } assert(bdrv_get_aio_context(base_bs) == aio_context); + + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(base_bs); + bdrv_graph_rdunlock_main_loop(); } if (bottom) { @@ -2466,13 +2493,16 @@ void qmp_block_stream(const char *job_id, const char *device, * Check for op blockers in the whole chain between bs and base (or bottom) */ iter_end = bottom ? bdrv_filter_or_cow_bs(bottom_bs) : base_bs; + bdrv_graph_rdlock_main_loop(); for (iter = bs; iter && iter != iter_end; iter = bdrv_filter_or_cow_bs(iter)) { if (bdrv_op_is_blocked(iter, BLOCK_OP_TYPE_STREAM, errp)) { + bdrv_graph_rdunlock_main_loop(); goto out; } } + bdrv_graph_rdunlock_main_loop(); /* if we are streaming the entire chain, the result will have no backing * file, and specifying one is therefore an error */ @@ -2835,6 +2865,8 @@ BlockDeviceInfoList *qmp_query_named_block_nodes(bool has_flat, XDbgBlockGraph *qmp_x_debug_query_block_graph(Error **errp) { + GRAPH_RDLOCK_GUARD_MAINLOOP(); + return bdrv_get_xdbg_block_graph(errp); } @@ -2998,9 +3030,12 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp) } /* Early check to avoid creating target */ + bdrv_graph_rdlock_main_loop(); if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_MIRROR_SOURCE, errp)) { + bdrv_graph_rdunlock_main_loop(); return; } + bdrv_graph_rdunlock_main_loop(); aio_context = bdrv_get_aio_context(bs); aio_context_acquire(aio_context); @@ -3063,7 +3098,10 @@ void qmp_drive_mirror(DriveMirror *arg, Error **errp) break; case NEW_IMAGE_MODE_ABSOLUTE_PATHS: /* create new image with backing file */ + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(explicit_backing); + bdrv_graph_rdunlock_main_loop(); + bdrv_img_create(arg->target, format, explicit_backing->filename, explicit_backing->drv->format_name, @@ -3383,9 +3421,12 @@ void qmp_change_backing_file(const char *device, /* even though we are not necessarily operating on bs, we need it to * determine if block ops are currently prohibited on the chain */ + bdrv_graph_rdlock_main_loop(); if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_CHANGE, errp)) { + bdrv_graph_rdunlock_main_loop(); goto out; } + bdrv_graph_rdunlock_main_loop(); /* final sanity check */ if (!bdrv_chain_contains(bs, image_bs)) { @@ -3509,6 +3550,7 @@ void qmp_blockdev_del(const char *node_name, Error **errp) BlockDriverState *bs; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); bs = bdrv_find_node(node_name); if (!bs) { @@ -3636,6 +3678,8 @@ void qmp_x_blockdev_set_iothread(const char *node_name, StrOrNull *iothread, AioContext *new_context; BlockDriverState *bs; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = bdrv_find_node(node_name); if (!bs) { error_setg(errp, "Failed to find node with node-name='%s'", node_name); diff --git a/blockjob.c b/blockjob.c index 58c5d64539..807f992b59 100644 --- a/blockjob.c +++ b/blockjob.c @@ -485,6 +485,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver, BlockJob *job; int ret; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); if (job_id == NULL && !(flags & JOB_INTERNAL)) { job_id = bdrv_get_device_name(bs); diff --git a/contrib/vhost-user-gpu/vhost-user-gpu.c b/contrib/vhost-user-gpu/vhost-user-gpu.c index aa304475a0..bb41758e34 100644 --- a/contrib/vhost-user-gpu/vhost-user-gpu.c +++ b/contrib/vhost-user-gpu/vhost-user-gpu.c @@ -834,7 +834,7 @@ vg_resource_flush(VuGpu *g, .width = width, .height = height, }; - pixman_image_t *i = + pixman_image_t *img = pixman_image_create_bits(pixman_image_get_format(res->image), msg->payload.update.width, msg->payload.update.height, @@ -842,11 +842,11 @@ vg_resource_flush(VuGpu *g, payload.update.data), width * bpp); pixman_image_composite(PIXMAN_OP_SRC, - res->image, NULL, i, + res->image, NULL, img, extents->x1, extents->y1, 0, 0, 0, 0, width, height); - pixman_image_unref(i); + pixman_image_unref(img); vg_send_msg(g, msg, -1); g_free(msg); } diff --git a/contrib/vhost-user-gpu/vugpu.h b/contrib/vhost-user-gpu/vugpu.h index 509b679f03..654c392fbb 100644 --- a/contrib/vhost-user-gpu/vugpu.h +++ b/contrib/vhost-user-gpu/vugpu.h @@ -164,12 +164,12 @@ struct virtio_gpu_ctrl_command { }; #define VUGPU_FILL_CMD(out) do { \ - size_t s; \ - s = iov_to_buf(cmd->elem.out_sg, cmd->elem.out_num, 0, \ + size_t vugpufillcmd_s_ = \ + iov_to_buf(cmd->elem.out_sg, cmd->elem.out_num, 0, \ &out, sizeof(out)); \ - if (s != sizeof(out)) { \ + if (vugpufillcmd_s_ != sizeof(out)) { \ g_critical("%s: command size incorrect %zu vs %zu", \ - __func__, s, sizeof(out)); \ + __func__, vugpufillcmd_s_, sizeof(out)); \ return; \ } \ } while (0) diff --git a/docs/about/deprecated.rst b/docs/about/deprecated.rst index 1c4d7f36f0..2febd2d12f 100644 --- a/docs/about/deprecated.rst +++ b/docs/about/deprecated.rst @@ -363,6 +363,15 @@ recommending to switch to their stable counterparts: - "Zve64f" should be replaced with "zve64f" - "Zve64d" should be replaced with "zve64d" +``-device pvrdma`` and the rdma subsystem (since 8.2) +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +The pvrdma device and the whole rdma subsystem are in a bad shape and +without active maintenance. The QEMU project intends to remove this +device and subsystem from the code base in a future release without +replacement unless somebody steps up and improves the situation. + + Block device options '''''''''''''''''''' diff --git a/hw/mem/memory-device.c b/hw/mem/memory-device.c index 667d56bd29..ae38f48f16 100644 --- a/hw/mem/memory-device.c +++ b/hw/mem/memory-device.c @@ -52,19 +52,135 @@ static int memory_device_build_list(Object *obj, void *opaque) return 0; } -static void memory_device_check_addable(MachineState *ms, MemoryRegion *mr, - Error **errp) +static unsigned int memory_device_get_memslots(MemoryDeviceState *md) { + const MemoryDeviceClass *mdc = MEMORY_DEVICE_GET_CLASS(md); + + if (mdc->get_memslots) { + return mdc->get_memslots(md); + } + return 1; +} + +/* + * Memslots that are reserved by memory devices (required but still reported + * as free from KVM / vhost). + */ +static unsigned int get_reserved_memslots(MachineState *ms) +{ + if (ms->device_memory->used_memslots > + ms->device_memory->required_memslots) { + /* This is unexpected, and we warned already in the memory notifier. */ + return 0; + } + return ms->device_memory->required_memslots - + ms->device_memory->used_memslots; +} + +unsigned int memory_devices_get_reserved_memslots(void) +{ + if (!current_machine->device_memory) { + return 0; + } + return get_reserved_memslots(current_machine); +} + +bool memory_devices_memslot_auto_decision_active(void) +{ + if (!current_machine->device_memory) { + return false; + } + + return current_machine->device_memory->memslot_auto_decision_active; +} + +static unsigned int memory_device_memslot_decision_limit(MachineState *ms, + MemoryRegion *mr) +{ + const unsigned int reserved = get_reserved_memslots(ms); + const uint64_t size = memory_region_size(mr); + unsigned int max = vhost_get_max_memslots(); + unsigned int free = vhost_get_free_memslots(); + uint64_t available_space; + unsigned int memslots; + + if (kvm_enabled()) { + max = MIN(max, kvm_get_max_memslots()); + free = MIN(free, kvm_get_free_memslots()); + } + + /* + * If we only have less overall memslots than what we consider reasonable, + * just keep it to a minimum. + */ + if (max < MEMORY_DEVICES_SAFE_MAX_MEMSLOTS) { + return 1; + } + + /* + * Consider our soft-limit across all memory devices. We don't really + * expect to exceed this limit in reasonable configurations. + */ + if (MEMORY_DEVICES_SOFT_MEMSLOT_LIMIT <= + ms->device_memory->required_memslots) { + return 1; + } + memslots = MEMORY_DEVICES_SOFT_MEMSLOT_LIMIT - + ms->device_memory->required_memslots; + + /* + * Consider the actually still free memslots. This is only relevant if + * other memslot consumers would consume *significantly* more memslots than + * what we prepared for (> 253). Unlikely, but let's just handle it + * cleanly. + */ + memslots = MIN(memslots, free - reserved); + if (memslots < 1 || unlikely(free < reserved)) { + return 1; + } + + /* We cannot have any other memory devices? So give all to this device. */ + if (size == ms->maxram_size - ms->ram_size) { + return memslots; + } + + /* + * Simple heuristic: equally distribute the memslots over the space + * still available for memory devices. + */ + available_space = ms->maxram_size - ms->ram_size - + ms->device_memory->used_region_size; + memslots = (double)memslots * size / available_space; + return memslots < 1 ? 1 : memslots; +} + +static void memory_device_check_addable(MachineState *ms, MemoryDeviceState *md, + MemoryRegion *mr, Error **errp) +{ + const MemoryDeviceClass *mdc = MEMORY_DEVICE_GET_CLASS(md); const uint64_t used_region_size = ms->device_memory->used_region_size; const uint64_t size = memory_region_size(mr); + const unsigned int reserved_memslots = get_reserved_memslots(ms); + unsigned int required_memslots, memslot_limit; + + /* + * Instruct the device to decide how many memslots to use, if applicable, + * before we query the number of required memslots the first time. + */ + if (mdc->decide_memslots) { + memslot_limit = memory_device_memslot_decision_limit(ms, mr); + mdc->decide_memslots(md, memslot_limit); + } + required_memslots = memory_device_get_memslots(md); - /* we will need a new memory slot for kvm and vhost */ - if (kvm_enabled() && !kvm_has_free_slot(ms)) { - error_setg(errp, "hypervisor has no free memory slots left"); + /* we will need memory slots for kvm and vhost */ + if (kvm_enabled() && + kvm_get_free_memslots() < required_memslots + reserved_memslots) { + error_setg(errp, "hypervisor has not enough free memory slots left"); return; } - if (!vhost_has_free_slot()) { - error_setg(errp, "a used vhost backend has no free memory slots left"); + if (vhost_get_free_memslots() < required_memslots + reserved_memslots) { + error_setg(errp, "a used vhost backend has not enough free memory slots left"); return; } @@ -233,7 +349,7 @@ void memory_device_pre_plug(MemoryDeviceState *md, MachineState *ms, goto out; } - memory_device_check_addable(ms, mr, &local_err); + memory_device_check_addable(ms, md, mr, &local_err); if (local_err) { goto out; } @@ -264,6 +380,7 @@ out: void memory_device_plug(MemoryDeviceState *md, MachineState *ms) { const MemoryDeviceClass *mdc = MEMORY_DEVICE_GET_CLASS(md); + const unsigned int memslots = memory_device_get_memslots(md); const uint64_t addr = mdc->get_addr(md); MemoryRegion *mr; @@ -275,6 +392,11 @@ void memory_device_plug(MemoryDeviceState *md, MachineState *ms) g_assert(ms->device_memory); ms->device_memory->used_region_size += memory_region_size(mr); + ms->device_memory->required_memslots += memslots; + if (mdc->decide_memslots && memslots > 1) { + ms->device_memory->memslot_auto_decision_active++; + } + memory_region_add_subregion(&ms->device_memory->mr, addr - ms->device_memory->base, mr); trace_memory_device_plug(DEVICE(md)->id ? DEVICE(md)->id : "", addr); @@ -283,6 +405,7 @@ void memory_device_plug(MemoryDeviceState *md, MachineState *ms) void memory_device_unplug(MemoryDeviceState *md, MachineState *ms) { const MemoryDeviceClass *mdc = MEMORY_DEVICE_GET_CLASS(md); + const unsigned int memslots = memory_device_get_memslots(md); MemoryRegion *mr; /* @@ -293,7 +416,12 @@ void memory_device_unplug(MemoryDeviceState *md, MachineState *ms) g_assert(ms->device_memory); memory_region_del_subregion(&ms->device_memory->mr, mr); + + if (mdc->decide_memslots && memslots > 1) { + ms->device_memory->memslot_auto_decision_active--; + } ms->device_memory->used_region_size -= memory_region_size(mr); + ms->device_memory->required_memslots -= memslots; trace_memory_device_unplug(DEVICE(md)->id ? DEVICE(md)->id : "", mdc->get_addr(md)); } @@ -313,6 +441,50 @@ uint64_t memory_device_get_region_size(const MemoryDeviceState *md, return memory_region_size(mr); } +static void memory_devices_region_mod(MemoryListener *listener, + MemoryRegionSection *mrs, bool add) +{ + DeviceMemoryState *dms = container_of(listener, DeviceMemoryState, + listener); + + if (!memory_region_is_ram(mrs->mr)) { + warn_report("Unexpected memory region mapped into device memory region."); + return; + } + + /* + * The expectation is that each distinct RAM memory region section in + * our region for memory devices consumes exactly one memslot in KVM + * and in vhost. For vhost, this is true, except: + * * ROM memory regions don't consume a memslot. These get used very + * rarely for memory devices (R/O NVDIMMs). + * * Memslots without a fd (memory-backend-ram) don't necessarily + * consume a memslot. Such setups are quite rare and possibly bogus: + * the memory would be inaccessible by such vhost devices. + * + * So for vhost, in corner cases we might over-estimate the number of + * memslots that are currently used or that might still be reserved + * (required - used). + */ + dms->used_memslots += add ? 1 : -1; + + if (dms->used_memslots > dms->required_memslots) { + warn_report("Memory devices use more memory slots than indicated as required."); + } +} + +static void memory_devices_region_add(MemoryListener *listener, + MemoryRegionSection *mrs) +{ + return memory_devices_region_mod(listener, mrs, true); +} + +static void memory_devices_region_del(MemoryListener *listener, + MemoryRegionSection *mrs) +{ + return memory_devices_region_mod(listener, mrs, false); +} + void machine_memory_devices_init(MachineState *ms, hwaddr base, uint64_t size) { g_assert(size); @@ -322,8 +494,16 @@ void machine_memory_devices_init(MachineState *ms, hwaddr base, uint64_t size) memory_region_init(&ms->device_memory->mr, OBJECT(ms), "device-memory", size); + address_space_init(&ms->device_memory->as, &ms->device_memory->mr, + "device-memory"); memory_region_add_subregion(get_system_memory(), ms->device_memory->base, &ms->device_memory->mr); + + /* Track the number of memslots used by memory devices. */ + ms->device_memory->listener.region_add = memory_devices_region_add; + ms->device_memory->listener.region_del = memory_devices_region_del; + memory_listener_register(&ms->device_memory->listener, + &ms->device_memory->as); } static const TypeInfo memory_device_info = { diff --git a/hw/misc/Kconfig b/hw/misc/Kconfig index 858277bb60..dba41afe67 100644 --- a/hw/misc/Kconfig +++ b/hw/misc/Kconfig @@ -34,6 +34,11 @@ config PCA9552 bool depends on I2C +config I2C_ECHO + bool + default y if TEST_DEVICES + depends on I2C + config PL310 bool diff --git a/hw/misc/i2c-echo.c b/hw/misc/i2c-echo.c index 5705ab5d73..5ae3d0817e 100644 --- a/hw/misc/i2c-echo.c +++ b/hw/misc/i2c-echo.c @@ -1,3 +1,13 @@ +/* + * Example I2C device using asynchronous I2C send. + * + * Copyright (C) 2023 Samsung Electronics Co., Ltd. All Rights Reserved. + * + * This work is licensed under the terms of the GNU GPL, version 2. See + * the COPYING file in the top-level directory. + * + */ + #include "qemu/osdep.h" #include "qemu/timer.h" #include "qemu/main-loop.h" diff --git a/hw/misc/meson.build b/hw/misc/meson.build index 33659313b4..f60de33f9a 100644 --- a/hw/misc/meson.build +++ b/hw/misc/meson.build @@ -138,7 +138,7 @@ system_ss.add(when: 'CONFIG_NRF51_SOC', if_true: files('nrf51_rng.c')) system_ss.add(when: 'CONFIG_GRLIB', if_true: files('grlib_ahb_apb_pnp.c')) -system_ss.add(when: 'CONFIG_I2C', if_true: files('i2c-echo.c')) +system_ss.add(when: 'CONFIG_I2C_ECHO', if_true: files('i2c-echo.c')) specific_ss.add(when: 'CONFIG_AVR_POWER', if_true: files('avr_power.c')) diff --git a/hw/rdma/vmw/pvrdma_main.c b/hw/rdma/vmw/pvrdma_main.c index 4fc6712025..e735ff97eb 100644 --- a/hw/rdma/vmw/pvrdma_main.c +++ b/hw/rdma/vmw/pvrdma_main.c @@ -91,19 +91,33 @@ static int init_dev_ring(PvrdmaRing *ring, PvrdmaRingState **ring_state, dma_addr_t dir_addr, uint32_t num_pages) { uint64_t *dir, *tbl; - int rc = 0; + int max_pages, rc = 0; if (!num_pages) { rdma_error_report("Ring pages count must be strictly positive"); return -EINVAL; } + /* + * Make sure we can satisfy the requested number of pages in a single + * TARGET_PAGE_SIZE sized page table (taking into account that first entry + * is reserved for ring-state) + */ + max_pages = TARGET_PAGE_SIZE / sizeof(dma_addr_t) - 1; + if (num_pages > max_pages) { + rdma_error_report("Maximum pages on a single directory must not exceed %d\n", + max_pages); + return -EINVAL; + } + dir = rdma_pci_dma_map(pci_dev, dir_addr, TARGET_PAGE_SIZE); if (!dir) { rdma_error_report("Failed to map to page directory (ring %s)", name); rc = -ENOMEM; goto out; } + + /* We support only one page table for a ring */ tbl = rdma_pci_dma_map(pci_dev, dir[0], TARGET_PAGE_SIZE); if (!tbl) { rdma_error_report("Failed to map to page table (ring %s)", name); @@ -601,6 +615,8 @@ static void pvrdma_realize(PCIDevice *pdev, Error **errp) bool ram_shared = false; PCIDevice *func0; + warn_report_once("pvrdma is deprecated and will be removed in a future release"); + rdma_info_report("Initializing device %s %x.%x", pdev->name, PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn)); diff --git a/hw/virtio/vhost-stub.c b/hw/virtio/vhost-stub.c index aa858ef3fb..52d42adab2 100644 --- a/hw/virtio/vhost-stub.c +++ b/hw/virtio/vhost-stub.c @@ -2,9 +2,14 @@ #include "hw/virtio/vhost.h" #include "hw/virtio/vhost-user.h" -bool vhost_has_free_slot(void) +unsigned int vhost_get_max_memslots(void) { - return true; + return UINT_MAX; +} + +unsigned int vhost_get_free_memslots(void) +{ + return UINT_MAX; } bool vhost_user_init(VhostUserState *user, CharBackend *chr, Error **errp) diff --git a/hw/virtio/vhost-user.c b/hw/virtio/vhost-user.c index 3766b415f8..68eb1f0c99 100644 --- a/hw/virtio/vhost-user.c +++ b/hw/virtio/vhost-user.c @@ -2327,19 +2327,6 @@ static int vhost_user_migration_done(struct vhost_dev *dev, char* mac_addr) return -ENOTSUP; } -static bool vhost_user_can_merge(struct vhost_dev *dev, - uint64_t start1, uint64_t size1, - uint64_t start2, uint64_t size2) -{ - ram_addr_t offset; - int mfd, rfd; - - (void)vhost_user_get_mr_data(start1, &offset, &mfd); - (void)vhost_user_get_mr_data(start2, &offset, &rfd); - - return mfd == rfd; -} - static int vhost_user_net_set_mtu(struct vhost_dev *dev, uint16_t mtu) { VhostUserMsg msg; @@ -2622,10 +2609,9 @@ vhost_user_crypto_close_session(struct vhost_dev *dev, uint64_t session_id) return 0; } -static bool vhost_user_mem_section_filter(struct vhost_dev *dev, - MemoryRegionSection *section) +static bool vhost_user_no_private_memslots(struct vhost_dev *dev) { - return memory_region_get_fd(section->mr) >= 0; + return true; } static int vhost_user_get_inflight_fd(struct vhost_dev *dev, @@ -2868,6 +2854,7 @@ const VhostOps user_ops = { .vhost_backend_init = vhost_user_backend_init, .vhost_backend_cleanup = vhost_user_backend_cleanup, .vhost_backend_memslots_limit = vhost_user_memslots_limit, + .vhost_backend_no_private_memslots = vhost_user_no_private_memslots, .vhost_set_log_base = vhost_user_set_log_base, .vhost_set_mem_table = vhost_user_set_mem_table, .vhost_set_vring_addr = vhost_user_set_vring_addr, @@ -2886,7 +2873,6 @@ const VhostOps user_ops = { .vhost_set_vring_enable = vhost_user_set_vring_enable, .vhost_requires_shm_log = vhost_user_requires_shm_log, .vhost_migration_done = vhost_user_migration_done, - .vhost_backend_can_merge = vhost_user_can_merge, .vhost_net_set_mtu = vhost_user_net_set_mtu, .vhost_set_iotlb_callback = vhost_user_set_iotlb_callback, .vhost_send_device_iotlb_msg = vhost_user_send_device_iotlb_msg, @@ -2894,7 +2880,6 @@ const VhostOps user_ops = { .vhost_set_config = vhost_user_set_config, .vhost_crypto_create_session = vhost_user_crypto_create_session, .vhost_crypto_close_session = vhost_user_crypto_close_session, - .vhost_backend_mem_section_filter = vhost_user_mem_section_filter, .vhost_get_inflight_fd = vhost_user_get_inflight_fd, .vhost_set_inflight_fd = vhost_user_set_inflight_fd, .vhost_dev_start = vhost_user_dev_start, diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c index 89ff02a999..819b2d811a 100644 --- a/hw/virtio/vhost-vdpa.c +++ b/hw/virtio/vhost-vdpa.c @@ -1512,7 +1512,6 @@ const VhostOps vdpa_ops = { .vhost_set_config = vhost_vdpa_set_config, .vhost_requires_shm_log = NULL, .vhost_migration_done = NULL, - .vhost_backend_can_merge = NULL, .vhost_net_set_mtu = NULL, .vhost_set_iotlb_callback = NULL, .vhost_send_device_iotlb_msg = NULL, diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c index 9cfac40fde..9f37206ba0 100644 --- a/hw/virtio/vhost.c +++ b/hw/virtio/vhost.c @@ -23,6 +23,7 @@ #include "qemu/log.h" #include "standard-headers/linux/vhost_types.h" #include "hw/virtio/virtio-bus.h" +#include "hw/mem/memory-device.h" #include "migration/blocker.h" #include "migration/qemu-file-types.h" #include "sysemu/dma.h" @@ -45,20 +46,44 @@ static struct vhost_log *vhost_log; static struct vhost_log *vhost_log_shm; +/* Memslots used by backends that support private memslots (without an fd). */ static unsigned int used_memslots; + +/* Memslots used by backends that only support shared memslots (with an fd). */ +static unsigned int used_shared_memslots; + static QLIST_HEAD(, vhost_dev) vhost_devices = QLIST_HEAD_INITIALIZER(vhost_devices); -bool vhost_has_free_slot(void) +unsigned int vhost_get_max_memslots(void) +{ + unsigned int max = UINT_MAX; + struct vhost_dev *hdev; + + QLIST_FOREACH(hdev, &vhost_devices, entry) { + max = MIN(max, hdev->vhost_ops->vhost_backend_memslots_limit(hdev)); + } + return max; +} + +unsigned int vhost_get_free_memslots(void) { - unsigned int slots_limit = ~0U; + unsigned int free = UINT_MAX; struct vhost_dev *hdev; QLIST_FOREACH(hdev, &vhost_devices, entry) { unsigned int r = hdev->vhost_ops->vhost_backend_memslots_limit(hdev); - slots_limit = MIN(slots_limit, r); + unsigned int cur_free; + + if (hdev->vhost_ops->vhost_backend_no_private_memslots && + hdev->vhost_ops->vhost_backend_no_private_memslots(hdev)) { + cur_free = r - used_shared_memslots; + } else { + cur_free = r - used_memslots; + } + free = MIN(free, cur_free); } - return slots_limit > used_memslots; + return free; } static void vhost_dev_sync_region(struct vhost_dev *dev, @@ -474,8 +499,7 @@ static int vhost_verify_ring_mappings(struct vhost_dev *dev, * vhost_section: identify sections needed for vhost access * * We only care about RAM sections here (where virtqueue and guest - * internals accessed by virtio might live). If we find one we still - * allow the backend to potentially filter it out of our list. + * internals accessed by virtio might live). */ static bool vhost_section(struct vhost_dev *dev, MemoryRegionSection *section) { @@ -502,8 +526,16 @@ static bool vhost_section(struct vhost_dev *dev, MemoryRegionSection *section) return false; } - if (dev->vhost_ops->vhost_backend_mem_section_filter && - !dev->vhost_ops->vhost_backend_mem_section_filter(dev, section)) { + /* + * Some backends (like vhost-user) can only handle memory regions + * that have an fd (can be mapped into a different process). Filter + * the ones without an fd out, if requested. + * + * TODO: we might have to limit to MAP_SHARED as well. + */ + if (memory_region_get_fd(section->mr) < 0 && + dev->vhost_ops->vhost_backend_no_private_memslots && + dev->vhost_ops->vhost_backend_no_private_memslots(dev)) { trace_vhost_reject_section(mr->name, 2); return false; } @@ -568,7 +600,14 @@ static void vhost_commit(MemoryListener *listener) dev->n_mem_sections * sizeof dev->mem->regions[0]; dev->mem = g_realloc(dev->mem, regions_size); dev->mem->nregions = dev->n_mem_sections; - used_memslots = dev->mem->nregions; + + if (dev->vhost_ops->vhost_backend_no_private_memslots && + dev->vhost_ops->vhost_backend_no_private_memslots(dev)) { + used_shared_memslots = dev->mem->nregions; + } else { + used_memslots = dev->mem->nregions; + } + for (i = 0; i < dev->n_mem_sections; i++) { struct vhost_memory_region *cur_vmr = dev->mem->regions + i; struct MemoryRegionSection *mrs = dev->mem_sections + i; @@ -668,7 +707,7 @@ static void vhost_region_add_section(struct vhost_dev *dev, mrs_size, mrs_host); } - if (dev->n_tmp_sections) { + if (dev->n_tmp_sections && !section->unmergeable) { /* Since we already have at least one section, lets see if * this extends it; since we're scanning in order, we only * have to look at the last one, and the FlatView that calls @@ -701,11 +740,7 @@ static void vhost_region_add_section(struct vhost_dev *dev, size_t offset = mrs_gpa - prev_gpa_start; if (prev_host_start + offset == mrs_host && - section->mr == prev_sec->mr && - (!dev->vhost_ops->vhost_backend_can_merge || - dev->vhost_ops->vhost_backend_can_merge(dev, - mrs_host, mrs_size, - prev_host_start, prev_size))) { + section->mr == prev_sec->mr && !prev_sec->unmergeable) { uint64_t max_end = MAX(prev_host_end, mrs_host + mrs_size); need_add = false; prev_sec->offset_within_address_space = @@ -1400,6 +1435,7 @@ int vhost_dev_init(struct vhost_dev *hdev, void *opaque, VhostBackendType backend_type, uint32_t busyloop_timeout, Error **errp) { + unsigned int used, reserved, limit; uint64_t features; int i, r, n_initialized_vqs = 0; @@ -1426,6 +1462,19 @@ int vhost_dev_init(struct vhost_dev *hdev, void *opaque, goto fail; } + limit = hdev->vhost_ops->vhost_backend_memslots_limit(hdev); + if (limit < MEMORY_DEVICES_SAFE_MAX_MEMSLOTS && + memory_devices_memslot_auto_decision_active()) { + error_setg(errp, "some memory device (like virtio-mem)" + " decided how many memory slots to use based on the overall" + " number of memory slots; this vhost backend would further" + " restricts the overall number of memory slots"); + error_append_hint(errp, "Try plugging this vhost backend before" + " plugging such memory devices.\n"); + r = -EINVAL; + goto fail; + } + for (i = 0; i < hdev->nvqs; ++i, ++n_initialized_vqs) { r = vhost_virtqueue_init(hdev, hdev->vqs + i, hdev->vq_index + i); if (r < 0) { @@ -1495,9 +1544,27 @@ int vhost_dev_init(struct vhost_dev *hdev, void *opaque, memory_listener_register(&hdev->memory_listener, &address_space_memory); QLIST_INSERT_HEAD(&vhost_devices, hdev, entry); - if (used_memslots > hdev->vhost_ops->vhost_backend_memslots_limit(hdev)) { - error_setg(errp, "vhost backend memory slots limit is less" - " than current number of present memory slots"); + /* + * The listener we registered properly updated the corresponding counter. + * So we can trust that these values are accurate. + */ + if (hdev->vhost_ops->vhost_backend_no_private_memslots && + hdev->vhost_ops->vhost_backend_no_private_memslots(hdev)) { + used = used_shared_memslots; + } else { + used = used_memslots; + } + /* + * We assume that all reserved memslots actually require a real memslot + * in our vhost backend. This might not be true, for example, if the + * memslot would be ROM. If ever relevant, we can optimize for that -- + * but we'll need additional information about the reservations. + */ + reserved = memory_devices_get_reserved_memslots(); + if (used + reserved > limit) { + error_setg(errp, "vhost backend memory slots limit (%d) is less" + " than current number of used (%d) and reserved (%d)" + " memory slots for memory devices.", limit, used, reserved); r = -EINVAL; goto fail_busyloop; } diff --git a/hw/virtio/virtio-mem-pci.c b/hw/virtio/virtio-mem-pci.c index c4597e029e..1b4e9a3284 100644 --- a/hw/virtio/virtio-mem-pci.c +++ b/hw/virtio/virtio-mem-pci.c @@ -48,6 +48,25 @@ static MemoryRegion *virtio_mem_pci_get_memory_region(MemoryDeviceState *md, return vmc->get_memory_region(vmem, errp); } +static void virtio_mem_pci_decide_memslots(MemoryDeviceState *md, + unsigned int limit) +{ + VirtIOMEMPCI *pci_mem = VIRTIO_MEM_PCI(md); + VirtIOMEM *vmem = VIRTIO_MEM(&pci_mem->vdev); + VirtIOMEMClass *vmc = VIRTIO_MEM_GET_CLASS(vmem); + + vmc->decide_memslots(vmem, limit); +} + +static unsigned int virtio_mem_pci_get_memslots(MemoryDeviceState *md) +{ + VirtIOMEMPCI *pci_mem = VIRTIO_MEM_PCI(md); + VirtIOMEM *vmem = VIRTIO_MEM(&pci_mem->vdev); + VirtIOMEMClass *vmc = VIRTIO_MEM_GET_CLASS(vmem); + + return vmc->get_memslots(vmem); +} + static uint64_t virtio_mem_pci_get_plugged_size(const MemoryDeviceState *md, Error **errp) { @@ -150,6 +169,8 @@ static void virtio_mem_pci_class_init(ObjectClass *klass, void *data) mdc->set_addr = virtio_mem_pci_set_addr; mdc->get_plugged_size = virtio_mem_pci_get_plugged_size; mdc->get_memory_region = virtio_mem_pci_get_memory_region; + mdc->decide_memslots = virtio_mem_pci_decide_memslots; + mdc->get_memslots = virtio_mem_pci_get_memslots; mdc->fill_device_info = virtio_mem_pci_fill_device_info; mdc->get_min_alignment = virtio_mem_pci_get_min_alignment; diff --git a/hw/virtio/virtio-mem.c b/hw/virtio/virtio-mem.c index da5b09cefc..9dc3c61b5a 100644 --- a/hw/virtio/virtio-mem.c +++ b/hw/virtio/virtio-mem.c @@ -67,6 +67,13 @@ static uint32_t virtio_mem_default_thp_size(void) } /* + * The minimum memslot size depends on this setting ("sane default"), the + * device block size, and the memory backend page size. The last (or single) + * memslot might be smaller than this constant. + */ +#define VIRTIO_MEM_MIN_MEMSLOT_SIZE (1 * GiB) + +/* * We want to have a reasonable default block size such that * 1. We avoid splitting THPs when unplugging memory, which degrades * performance. @@ -177,10 +184,10 @@ static bool virtio_mem_is_busy(void) return migration_in_incoming_postcopy() || !migration_is_idle(); } -typedef int (*virtio_mem_range_cb)(const VirtIOMEM *vmem, void *arg, +typedef int (*virtio_mem_range_cb)(VirtIOMEM *vmem, void *arg, uint64_t offset, uint64_t size); -static int virtio_mem_for_each_unplugged_range(const VirtIOMEM *vmem, void *arg, +static int virtio_mem_for_each_unplugged_range(VirtIOMEM *vmem, void *arg, virtio_mem_range_cb cb) { unsigned long first_zero_bit, last_zero_bit; @@ -204,7 +211,7 @@ static int virtio_mem_for_each_unplugged_range(const VirtIOMEM *vmem, void *arg, return ret; } -static int virtio_mem_for_each_plugged_range(const VirtIOMEM *vmem, void *arg, +static int virtio_mem_for_each_plugged_range(VirtIOMEM *vmem, void *arg, virtio_mem_range_cb cb) { unsigned long first_bit, last_bit; @@ -483,6 +490,96 @@ static bool virtio_mem_valid_range(const VirtIOMEM *vmem, uint64_t gpa, return true; } +static void virtio_mem_activate_memslot(VirtIOMEM *vmem, unsigned int idx) +{ + const uint64_t memslot_offset = idx * vmem->memslot_size; + + assert(vmem->memslots); + + /* + * Instead of enabling/disabling memslots, we add/remove them. This should + * make address space updates faster, because we don't have to loop over + * many disabled subregions. + */ + if (memory_region_is_mapped(&vmem->memslots[idx])) { + return; + } + memory_region_add_subregion(vmem->mr, memslot_offset, &vmem->memslots[idx]); +} + +static void virtio_mem_deactivate_memslot(VirtIOMEM *vmem, unsigned int idx) +{ + assert(vmem->memslots); + + if (!memory_region_is_mapped(&vmem->memslots[idx])) { + return; + } + memory_region_del_subregion(vmem->mr, &vmem->memslots[idx]); +} + +static void virtio_mem_activate_memslots_to_plug(VirtIOMEM *vmem, + uint64_t offset, uint64_t size) +{ + const unsigned int start_idx = offset / vmem->memslot_size; + const unsigned int end_idx = (offset + size + vmem->memslot_size - 1) / + vmem->memslot_size; + unsigned int idx; + + if (!vmem->dynamic_memslots) { + return; + } + + /* Activate all involved memslots in a single transaction. */ + memory_region_transaction_begin(); + for (idx = start_idx; idx < end_idx; idx++) { + virtio_mem_activate_memslot(vmem, idx); + } + memory_region_transaction_commit(); +} + +static void virtio_mem_deactivate_unplugged_memslots(VirtIOMEM *vmem, + uint64_t offset, + uint64_t size) +{ + const uint64_t region_size = memory_region_size(&vmem->memdev->mr); + const unsigned int start_idx = offset / vmem->memslot_size; + const unsigned int end_idx = (offset + size + vmem->memslot_size - 1) / + vmem->memslot_size; + unsigned int idx; + + if (!vmem->dynamic_memslots) { + return; + } + + /* Deactivate all memslots with unplugged blocks in a single transaction. */ + memory_region_transaction_begin(); + for (idx = start_idx; idx < end_idx; idx++) { + const uint64_t memslot_offset = idx * vmem->memslot_size; + uint64_t memslot_size = vmem->memslot_size; + + /* The size of the last memslot might be smaller. */ + if (idx == vmem->nb_memslots - 1) { + memslot_size = region_size - memslot_offset; + } + + /* + * Partially covered memslots might still have some blocks plugged and + * have to remain active if that's the case. + */ + if (offset > memslot_offset || + offset + size < memslot_offset + memslot_size) { + const uint64_t gpa = vmem->addr + memslot_offset; + + if (!virtio_mem_is_range_unplugged(vmem, gpa, memslot_size)) { + continue; + } + } + + virtio_mem_deactivate_memslot(vmem, idx); + } + memory_region_transaction_commit(); +} + static int virtio_mem_set_block_state(VirtIOMEM *vmem, uint64_t start_gpa, uint64_t size, bool plug) { @@ -500,6 +597,8 @@ static int virtio_mem_set_block_state(VirtIOMEM *vmem, uint64_t start_gpa, } virtio_mem_notify_unplug(vmem, offset, size); virtio_mem_set_range_unplugged(vmem, start_gpa, size); + /* Deactivate completely unplugged memslots after updating the state. */ + virtio_mem_deactivate_unplugged_memslots(vmem, offset, size); return 0; } @@ -527,7 +626,20 @@ static int virtio_mem_set_block_state(VirtIOMEM *vmem, uint64_t start_gpa, } if (!ret) { + /* + * Activate before notifying and rollback in case of any errors. + * + * When activating a yet inactive memslot, memory notifiers will get + * notified about the added memory region and can register with the + * RamDiscardManager; this will traverse all plugged blocks and skip the + * blocks we are plugging here. The following notification will inform + * registered listeners about the blocks we're plugging. + */ + virtio_mem_activate_memslots_to_plug(vmem, offset, size); ret = virtio_mem_notify_plug(vmem, offset, size); + if (ret) { + virtio_mem_deactivate_unplugged_memslots(vmem, offset, size); + } } if (ret) { /* Could be preallocation or a notifier populated memory. */ @@ -620,6 +732,7 @@ static void virtio_mem_resize_usable_region(VirtIOMEM *vmem, static int virtio_mem_unplug_all(VirtIOMEM *vmem) { + const uint64_t region_size = memory_region_size(&vmem->memdev->mr); RAMBlock *rb = vmem->memdev->mr.ram_block; if (vmem->size) { @@ -634,6 +747,9 @@ static int virtio_mem_unplug_all(VirtIOMEM *vmem) bitmap_clear(vmem->bitmap, 0, vmem->bitmap_size); vmem->size = 0; notifier_list_notify(&vmem->size_change_notifiers, &vmem->size); + + /* Deactivate all memslots after updating the state. */ + virtio_mem_deactivate_unplugged_memslots(vmem, 0, region_size); } trace_virtio_mem_unplugged_all(); @@ -790,6 +906,49 @@ static void virtio_mem_system_reset(void *opaque) virtio_mem_unplug_all(vmem); } +static void virtio_mem_prepare_mr(VirtIOMEM *vmem) +{ + const uint64_t region_size = memory_region_size(&vmem->memdev->mr); + + assert(!vmem->mr && vmem->dynamic_memslots); + vmem->mr = g_new0(MemoryRegion, 1); + memory_region_init(vmem->mr, OBJECT(vmem), "virtio-mem", + region_size); + vmem->mr->align = memory_region_get_alignment(&vmem->memdev->mr); +} + +static void virtio_mem_prepare_memslots(VirtIOMEM *vmem) +{ + const uint64_t region_size = memory_region_size(&vmem->memdev->mr); + unsigned int idx; + + g_assert(!vmem->memslots && vmem->nb_memslots && vmem->dynamic_memslots); + vmem->memslots = g_new0(MemoryRegion, vmem->nb_memslots); + + /* Initialize our memslots, but don't map them yet. */ + for (idx = 0; idx < vmem->nb_memslots; idx++) { + const uint64_t memslot_offset = idx * vmem->memslot_size; + uint64_t memslot_size = vmem->memslot_size; + char name[20]; + + /* The size of the last memslot might be smaller. */ + if (idx == vmem->nb_memslots - 1) { + memslot_size = region_size - memslot_offset; + } + + snprintf(name, sizeof(name), "memslot-%u", idx); + memory_region_init_alias(&vmem->memslots[idx], OBJECT(vmem), name, + &vmem->memdev->mr, memslot_offset, + memslot_size); + /* + * We want to be able to atomically and efficiently activate/deactivate + * individual memslots without affecting adjacent memslots in memory + * notifiers. + */ + memory_region_set_unmergeable(&vmem->memslots[idx], true); + } +} + static void virtio_mem_device_realize(DeviceState *dev, Error **errp) { MachineState *ms = MACHINE(qdev_get_machine()); @@ -861,6 +1020,14 @@ static void virtio_mem_device_realize(DeviceState *dev, Error **errp) vmem->unplugged_inaccessible = ON_OFF_AUTO_ON; #endif /* VIRTIO_MEM_HAS_LEGACY_GUESTS */ + if (vmem->dynamic_memslots && + vmem->unplugged_inaccessible != ON_OFF_AUTO_ON) { + error_setg(errp, "'%s' property set to 'on' requires '%s' to be 'on'", + VIRTIO_MEM_DYNAMIC_MEMSLOTS_PROP, + VIRTIO_MEM_UNPLUGGED_INACCESSIBLE_PROP); + return; + } + /* * If the block size wasn't configured by the user, use a sane default. This * allows using hugetlbfs backends of any page size without manual @@ -930,6 +1097,25 @@ static void virtio_mem_device_realize(DeviceState *dev, Error **errp) virtio_init(vdev, VIRTIO_ID_MEM, sizeof(struct virtio_mem_config)); vmem->vq = virtio_add_queue(vdev, 128, virtio_mem_handle_request); + /* + * With "dynamic-memslots=off" (old behavior) we always map the whole + * RAM memory region directly. + */ + if (vmem->dynamic_memslots) { + if (!vmem->mr) { + virtio_mem_prepare_mr(vmem); + } + if (vmem->nb_memslots <= 1) { + vmem->nb_memslots = 1; + vmem->memslot_size = memory_region_size(&vmem->memdev->mr); + } + if (!vmem->memslots) { + virtio_mem_prepare_memslots(vmem); + } + } else { + assert(!vmem->mr && !vmem->nb_memslots && !vmem->memslots); + } + host_memory_backend_set_mapped(vmem->memdev, true); vmstate_register_ram(&vmem->memdev->mr, DEVICE(vmem)); if (vmem->early_migration) { @@ -969,7 +1155,7 @@ static void virtio_mem_device_unrealize(DeviceState *dev) ram_block_coordinated_discard_require(false); } -static int virtio_mem_discard_range_cb(const VirtIOMEM *vmem, void *arg, +static int virtio_mem_discard_range_cb(VirtIOMEM *vmem, void *arg, uint64_t offset, uint64_t size) { RAMBlock *rb = vmem->memdev->mr.ram_block; @@ -984,13 +1170,32 @@ static int virtio_mem_restore_unplugged(VirtIOMEM *vmem) virtio_mem_discard_range_cb); } -static int virtio_mem_post_load(void *opaque, int version_id) +static int virtio_mem_activate_memslot_range_cb(VirtIOMEM *vmem, void *arg, + uint64_t offset, uint64_t size) +{ + virtio_mem_activate_memslots_to_plug(vmem, offset, size); + return 0; +} + +static int virtio_mem_post_load_bitmap(VirtIOMEM *vmem) { - VirtIOMEM *vmem = VIRTIO_MEM(opaque); RamDiscardListener *rdl; int ret; /* + * We restored the bitmap and updated the requested size; activate all + * memslots (so listeners register) before notifying about plugged blocks. + */ + if (vmem->dynamic_memslots) { + /* + * We don't expect any active memslots at this point to deactivate: no + * memory was plugged on the migration destination. + */ + virtio_mem_for_each_plugged_range(vmem, NULL, + virtio_mem_activate_memslot_range_cb); + } + + /* * We started out with all memory discarded and our memory region is mapped * into an address space. Replay, now that we updated the bitmap. */ @@ -1001,6 +1206,20 @@ static int virtio_mem_post_load(void *opaque, int version_id) return ret; } } + return 0; +} + +static int virtio_mem_post_load(void *opaque, int version_id) +{ + VirtIOMEM *vmem = VIRTIO_MEM(opaque); + int ret; + + if (!vmem->early_migration) { + ret = virtio_mem_post_load_bitmap(vmem); + if (ret) { + return ret; + } + } /* * If shared RAM is migrated using the file content and not using QEMU, @@ -1021,7 +1240,7 @@ static int virtio_mem_post_load(void *opaque, int version_id) return virtio_mem_restore_unplugged(vmem); } -static int virtio_mem_prealloc_range_cb(const VirtIOMEM *vmem, void *arg, +static int virtio_mem_prealloc_range_cb(VirtIOMEM *vmem, void *arg, uint64_t offset, uint64_t size) { void *area = memory_region_get_ram_ptr(&vmem->memdev->mr) + offset; @@ -1043,7 +1262,7 @@ static int virtio_mem_post_load_early(void *opaque, int version_id) int ret; if (!vmem->prealloc) { - return 0; + goto post_load_bitmap; } /* @@ -1051,7 +1270,7 @@ static int virtio_mem_post_load_early(void *opaque, int version_id) * don't mess with preallocation and postcopy. */ if (migrate_ram_is_ignored(rb)) { - return 0; + goto post_load_bitmap; } /* @@ -1084,7 +1303,10 @@ static int virtio_mem_post_load_early(void *opaque, int version_id) return -EBUSY; } } - return 0; + +post_load_bitmap: + /* Finally, update any other state to be consistent with the new bitmap. */ + return virtio_mem_post_load_bitmap(vmem); } typedef struct VirtIOMEMMigSanityChecks { @@ -1235,11 +1457,79 @@ static MemoryRegion *virtio_mem_get_memory_region(VirtIOMEM *vmem, Error **errp) if (!vmem->memdev) { error_setg(errp, "'%s' property must be set", VIRTIO_MEM_MEMDEV_PROP); return NULL; + } else if (vmem->dynamic_memslots) { + if (!vmem->mr) { + virtio_mem_prepare_mr(vmem); + } + return vmem->mr; } return &vmem->memdev->mr; } +static void virtio_mem_decide_memslots(VirtIOMEM *vmem, unsigned int limit) +{ + uint64_t region_size, memslot_size, min_memslot_size; + unsigned int memslots; + RAMBlock *rb; + + if (!vmem->dynamic_memslots) { + return; + } + + /* We're called exactly once, before realizing the device. */ + assert(!vmem->nb_memslots); + + /* If realizing the device will fail, just assume a single memslot. */ + if (limit <= 1 || !vmem->memdev || !vmem->memdev->mr.ram_block) { + vmem->nb_memslots = 1; + return; + } + + rb = vmem->memdev->mr.ram_block; + region_size = memory_region_size(&vmem->memdev->mr); + + /* + * Determine the default block size now, to determine the minimum memslot + * size. We want the minimum slot size to be at least the device block size. + */ + if (!vmem->block_size) { + vmem->block_size = virtio_mem_default_block_size(rb); + } + /* If realizing the device will fail, just assume a single memslot. */ + if (vmem->block_size < qemu_ram_pagesize(rb) || + !QEMU_IS_ALIGNED(region_size, vmem->block_size)) { + vmem->nb_memslots = 1; + return; + } + + /* + * All memslots except the last one have a reasonable minimum size, and + * and all memslot sizes are aligned to the device block size. + */ + memslot_size = QEMU_ALIGN_UP(region_size / limit, vmem->block_size); + min_memslot_size = MAX(vmem->block_size, VIRTIO_MEM_MIN_MEMSLOT_SIZE); + memslot_size = MAX(memslot_size, min_memslot_size); + + memslots = QEMU_ALIGN_UP(region_size, memslot_size) / memslot_size; + if (memslots != 1) { + vmem->memslot_size = memslot_size; + } + vmem->nb_memslots = memslots; +} + +static unsigned int virtio_mem_get_memslots(VirtIOMEM *vmem) +{ + if (!vmem->dynamic_memslots) { + /* Exactly one static RAM memory region. */ + return 1; + } + + /* We're called after instructed to make a decision. */ + g_assert(vmem->nb_memslots); + return vmem->nb_memslots; +} + static void virtio_mem_add_size_change_notifier(VirtIOMEM *vmem, Notifier *notifier) { @@ -1377,6 +1667,21 @@ static void virtio_mem_instance_init(Object *obj) NULL, NULL); } +static void virtio_mem_instance_finalize(Object *obj) +{ + VirtIOMEM *vmem = VIRTIO_MEM(obj); + + /* + * Note: the core already dropped the references on all memory regions + * (it's passed as the owner to memory_region_init_*()) and finalized + * these objects. We can simply free the memory. + */ + g_free(vmem->memslots); + vmem->memslots = NULL; + g_free(vmem->mr); + vmem->mr = NULL; +} + static Property virtio_mem_properties[] = { DEFINE_PROP_UINT64(VIRTIO_MEM_ADDR_PROP, VirtIOMEM, addr, 0), DEFINE_PROP_UINT32(VIRTIO_MEM_NODE_PROP, VirtIOMEM, node, 0), @@ -1389,6 +1694,8 @@ static Property virtio_mem_properties[] = { #endif DEFINE_PROP_BOOL(VIRTIO_MEM_EARLY_MIGRATION_PROP, VirtIOMEM, early_migration, true), + DEFINE_PROP_BOOL(VIRTIO_MEM_DYNAMIC_MEMSLOTS_PROP, VirtIOMEM, + dynamic_memslots, false), DEFINE_PROP_END_OF_LIST(), }; @@ -1556,6 +1863,8 @@ static void virtio_mem_class_init(ObjectClass *klass, void *data) vmc->fill_device_info = virtio_mem_fill_device_info; vmc->get_memory_region = virtio_mem_get_memory_region; + vmc->decide_memslots = virtio_mem_decide_memslots; + vmc->get_memslots = virtio_mem_get_memslots; vmc->add_size_change_notifier = virtio_mem_add_size_change_notifier; vmc->remove_size_change_notifier = virtio_mem_remove_size_change_notifier; vmc->unplug_request_check = virtio_mem_unplug_request_check; @@ -1573,6 +1882,7 @@ static const TypeInfo virtio_mem_info = { .parent = TYPE_VIRTIO_DEVICE, .instance_size = sizeof(VirtIOMEM), .instance_init = virtio_mem_instance_init, + .instance_finalize = virtio_mem_instance_finalize, .class_init = virtio_mem_class_init, .class_size = sizeof(VirtIOMEMClass), .interfaces = (InterfaceInfo[]) { diff --git a/include/block/block-common.h b/include/block/block-common.h index 2d2af7230d..d7599564db 100644 --- a/include/block/block-common.h +++ b/include/block/block-common.h @@ -66,13 +66,16 @@ * function. The coroutine yields after scheduling the BH and is reentered when * the wrapped function returns. * - * A no_co_wrapper_bdrv_wrlock function is a no_co_wrapper function that - * automatically takes the graph wrlock when calling the wrapped function. + * A no_co_wrapper_bdrv_rdlock function is a no_co_wrapper function that + * automatically takes the graph rdlock when calling the wrapped function. In + * the same way, no_co_wrapper_bdrv_wrlock functions automatically take the + * graph wrlock. * * If the first parameter of the function is a BlockDriverState, BdrvChild or * BlockBackend pointer, the AioContext lock for it is taken in the wrapper. */ #define no_co_wrapper +#define no_co_wrapper_bdrv_rdlock #define no_co_wrapper_bdrv_wrlock #include "block/blockjob.h" diff --git a/include/block/block-global-state.h b/include/block/block-global-state.h index 6061220a6c..6bfafe781d 100644 --- a/include/block/block-global-state.h +++ b/include/block/block-global-state.h @@ -132,13 +132,13 @@ int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only, Error **errp); BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs, const char *backing_file); -void bdrv_refresh_filename(BlockDriverState *bs); +void GRAPH_RDLOCK bdrv_refresh_filename(BlockDriverState *bs); void GRAPH_RDLOCK bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp); int bdrv_commit(BlockDriverState *bs); -int bdrv_make_empty(BdrvChild *c, Error **errp); +int GRAPH_RDLOCK bdrv_make_empty(BdrvChild *c, Error **errp); int bdrv_change_backing_file(BlockDriverState *bs, const char *backing_file, const char *backing_fmt, bool warn); void bdrv_register(BlockDriver *bdrv); @@ -160,19 +160,20 @@ void bdrv_unfreeze_backing_chain(BlockDriverState *bs, BlockDriverState *base); */ typedef void BlockDriverAmendStatusCB(BlockDriverState *bs, int64_t offset, int64_t total_work_size, void *opaque); -int bdrv_amend_options(BlockDriverState *bs_new, QemuOpts *opts, - BlockDriverAmendStatusCB *status_cb, void *cb_opaque, - bool force, - Error **errp); +int GRAPH_RDLOCK +bdrv_amend_options(BlockDriverState *bs_new, QemuOpts *opts, + BlockDriverAmendStatusCB *status_cb, void *cb_opaque, + bool force, Error **errp); /* check if a named node can be replaced when doing drive-mirror */ BlockDriverState * GRAPH_RDLOCK check_to_replace_node(BlockDriverState *parent_bs, const char *node_name, Error **errp); -int no_coroutine_fn bdrv_activate(BlockDriverState *bs, Error **errp); +int no_coroutine_fn GRAPH_RDLOCK +bdrv_activate(BlockDriverState *bs, Error **errp); -int coroutine_fn no_co_wrapper +int coroutine_fn no_co_wrapper_bdrv_rdlock bdrv_co_activate(BlockDriverState *bs, Error **errp); void bdrv_activate_all(Error **errp); @@ -191,7 +192,7 @@ int bdrv_has_zero_init_1(BlockDriverState *bs); int bdrv_has_zero_init(BlockDriverState *bs); BlockDriverState *bdrv_find_node(const char *node_name); BlockDeviceInfoList *bdrv_named_nodes_list(bool flat, Error **errp); -XDbgBlockGraph *bdrv_get_xdbg_block_graph(Error **errp); +XDbgBlockGraph * GRAPH_RDLOCK bdrv_get_xdbg_block_graph(Error **errp); BlockDriverState *bdrv_lookup_bs(const char *device, const char *node_name, Error **errp); @@ -208,15 +209,18 @@ typedef struct BdrvNextIterator { BlockDriverState *bs; } BdrvNextIterator; -BlockDriverState *bdrv_first(BdrvNextIterator *it); -BlockDriverState *bdrv_next(BdrvNextIterator *it); +BlockDriverState * GRAPH_RDLOCK bdrv_first(BdrvNextIterator *it); +BlockDriverState * GRAPH_RDLOCK bdrv_next(BdrvNextIterator *it); void bdrv_next_cleanup(BdrvNextIterator *it); BlockDriverState *bdrv_next_monitor_owned(BlockDriverState *bs); void bdrv_iterate_format(void (*it)(void *opaque, const char *name), void *opaque, bool read_only); -char *bdrv_get_full_backing_filename(BlockDriverState *bs, Error **errp); -char *bdrv_dirname(BlockDriverState *bs, Error **errp); + +char * GRAPH_RDLOCK +bdrv_get_full_backing_filename(BlockDriverState *bs, Error **errp); + +char * GRAPH_RDLOCK bdrv_dirname(BlockDriverState *bs, Error **errp); void bdrv_img_create(const char *filename, const char *fmt, const char *base_filename, const char *base_fmt, @@ -242,7 +246,9 @@ bdrv_attach_child(BlockDriverState *parent_bs, BdrvChildRole child_role, Error **errp); -bool bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp); +bool GRAPH_RDLOCK +bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp); + void bdrv_op_block(BlockDriverState *bs, BlockOpType op, Error *reason); void bdrv_op_unblock(BlockDriverState *bs, BlockOpType op, Error *reason); void bdrv_op_block_all(BlockDriverState *bs, Error *reason); diff --git a/include/block/block-io.h b/include/block/block-io.h index f1c796a1ce..ad270b6ad2 100644 --- a/include/block/block-io.h +++ b/include/block/block-io.h @@ -127,37 +127,46 @@ int coroutine_fn GRAPH_RDLOCK bdrv_co_zone_append(BlockDriverState *bs, BdrvRequestFlags flags); bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs); -int bdrv_block_status(BlockDriverState *bs, int64_t offset, - int64_t bytes, int64_t *pnum, int64_t *map, - BlockDriverState **file); + +int coroutine_fn GRAPH_RDLOCK +bdrv_co_block_status(BlockDriverState *bs, int64_t offset, int64_t bytes, + int64_t *pnum, int64_t *map, BlockDriverState **file); +int co_wrapper_mixed_bdrv_rdlock +bdrv_block_status(BlockDriverState *bs, int64_t offset, int64_t bytes, + int64_t *pnum, int64_t *map, BlockDriverState **file); int coroutine_fn GRAPH_RDLOCK bdrv_co_block_status_above(BlockDriverState *bs, BlockDriverState *base, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file); -int bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, - int64_t offset, int64_t bytes, int64_t *pnum, - int64_t *map, BlockDriverState **file); +int co_wrapper_mixed_bdrv_rdlock +bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, + int64_t offset, int64_t bytes, int64_t *pnum, + int64_t *map, BlockDriverState **file); int coroutine_fn GRAPH_RDLOCK bdrv_co_is_allocated(BlockDriverState *bs, int64_t offset, int64_t bytes, int64_t *pnum); -int bdrv_is_allocated(BlockDriverState *bs, int64_t offset, int64_t bytes, - int64_t *pnum); +int co_wrapper_mixed_bdrv_rdlock +bdrv_is_allocated(BlockDriverState *bs, int64_t offset, + int64_t bytes, int64_t *pnum); int coroutine_fn GRAPH_RDLOCK bdrv_co_is_allocated_above(BlockDriverState *top, BlockDriverState *base, bool include_base, int64_t offset, int64_t bytes, int64_t *pnum); -int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base, - bool include_base, int64_t offset, int64_t bytes, - int64_t *pnum); +int co_wrapper_mixed_bdrv_rdlock +bdrv_is_allocated_above(BlockDriverState *bs, BlockDriverState *base, + bool include_base, int64_t offset, + int64_t bytes, int64_t *pnum); int coroutine_fn GRAPH_RDLOCK bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, int64_t bytes); -int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg, - Error **errp); +int GRAPH_RDLOCK +bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg, + Error **errp); + bool bdrv_is_read_only(BlockDriverState *bs); bool bdrv_is_writable(BlockDriverState *bs); bool bdrv_is_sg(BlockDriverState *bs); @@ -176,8 +185,12 @@ const char *bdrv_get_format_name(BlockDriverState *bs); bool bdrv_supports_compressed_writes(BlockDriverState *bs); const char *bdrv_get_node_name(const BlockDriverState *bs); -const char *bdrv_get_device_name(const BlockDriverState *bs); -const char *bdrv_get_device_or_node_name(const BlockDriverState *bs); + +const char * GRAPH_RDLOCK +bdrv_get_device_name(const BlockDriverState *bs); + +const char * GRAPH_RDLOCK +bdrv_get_device_or_node_name(const BlockDriverState *bs); int coroutine_fn GRAPH_RDLOCK bdrv_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); @@ -185,8 +198,9 @@ bdrv_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); int co_wrapper_mixed_bdrv_rdlock bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); -ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs, - Error **errp); +ImageInfoSpecific * GRAPH_RDLOCK +bdrv_get_specific_info(BlockDriverState *bs, Error **errp); + BlockStatsSpecific *bdrv_get_specific_stats(BlockDriverState *bs); void bdrv_round_to_subclusters(BlockDriverState *bs, int64_t offset, int64_t bytes, @@ -363,7 +377,7 @@ bdrv_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); * * Begin a quiesced section for the parent of @c. */ -void bdrv_parent_drained_begin_single(BdrvChild *c); +void GRAPH_RDLOCK bdrv_parent_drained_begin_single(BdrvChild *c); /** * bdrv_parent_drained_poll_single: @@ -371,14 +385,14 @@ void bdrv_parent_drained_begin_single(BdrvChild *c); * Returns true if there is any pending activity to cease before @c can be * called quiesced, false otherwise. */ -bool bdrv_parent_drained_poll_single(BdrvChild *c); +bool GRAPH_RDLOCK bdrv_parent_drained_poll_single(BdrvChild *c); /** * bdrv_parent_drained_end_single: * * End a quiesced section for the parent of @c. */ -void bdrv_parent_drained_end_single(BdrvChild *c); +void GRAPH_RDLOCK bdrv_parent_drained_end_single(BdrvChild *c); /** * bdrv_drain_poll: @@ -391,8 +405,9 @@ void bdrv_parent_drained_end_single(BdrvChild *c); * * This is part of bdrv_drained_begin. */ -bool bdrv_drain_poll(BlockDriverState *bs, BdrvChild *ignore_parent, - bool ignore_bds_parents); +bool GRAPH_RDLOCK +bdrv_drain_poll(BlockDriverState *bs, BdrvChild *ignore_parent, + bool ignore_bds_parents); /** * bdrv_drained_begin: @@ -400,6 +415,12 @@ bool bdrv_drain_poll(BlockDriverState *bs, BdrvChild *ignore_parent, * Begin a quiesced section for exclusive access to the BDS, by disabling * external request sources including NBD server, block jobs, and device model. * + * This function can only be invoked by the main loop or a coroutine + * (regardless of the AioContext where it is running). + * If the coroutine is running in an Iothread AioContext, this function will + * just schedule a BH to run in the main loop. + * However, it cannot be directly called by an Iothread. + * * This function can be recursive. */ void bdrv_drained_begin(BlockDriverState *bs); @@ -416,6 +437,12 @@ void bdrv_do_drained_begin_quiesce(BlockDriverState *bs, BdrvChild *parent); * bdrv_drained_end: * * End a quiescent section started by bdrv_drained_begin(). + * + * This function can only be invoked by the main loop or a coroutine + * (regardless of the AioContext where it is running). + * If the coroutine is running in an Iothread AioContext, this function will + * just schedule a BH to run in the main loop. + * However, it cannot be directly called by an Iothread. */ void bdrv_drained_end(BlockDriverState *bs); diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index 2ca3758cb8..b8d9d24f39 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -235,11 +235,14 @@ struct BlockDriver { Error **errp); /* For handling image reopen for split or non-split files. */ - int (*bdrv_reopen_prepare)(BDRVReopenState *reopen_state, - BlockReopenQueue *queue, Error **errp); - void (*bdrv_reopen_commit)(BDRVReopenState *reopen_state); - void (*bdrv_reopen_commit_post)(BDRVReopenState *reopen_state); - void (*bdrv_reopen_abort)(BDRVReopenState *reopen_state); + int GRAPH_UNLOCKED_PTR (*bdrv_reopen_prepare)( + BDRVReopenState *reopen_state, BlockReopenQueue *queue, Error **errp); + void GRAPH_UNLOCKED_PTR (*bdrv_reopen_commit)( + BDRVReopenState *reopen_state); + void GRAPH_UNLOCKED_PTR (*bdrv_reopen_commit_post)( + BDRVReopenState *reopen_state); + void GRAPH_UNLOCKED_PTR (*bdrv_reopen_abort)( + BDRVReopenState *reopen_state); void (*bdrv_join_options)(QDict *options, QDict *old_options); int GRAPH_UNLOCKED_PTR (*bdrv_open)( @@ -256,20 +259,18 @@ struct BlockDriver { int coroutine_fn GRAPH_UNLOCKED_PTR (*bdrv_co_create_opts)( BlockDriver *drv, const char *filename, QemuOpts *opts, Error **errp); - int (*bdrv_amend_options)(BlockDriverState *bs, - QemuOpts *opts, - BlockDriverAmendStatusCB *status_cb, - void *cb_opaque, - bool force, - Error **errp); + int GRAPH_RDLOCK_PTR (*bdrv_amend_options)( + BlockDriverState *bs, QemuOpts *opts, + BlockDriverAmendStatusCB *status_cb, void *cb_opaque, + bool force, Error **errp); - int (*bdrv_make_empty)(BlockDriverState *bs); + int GRAPH_RDLOCK_PTR (*bdrv_make_empty)(BlockDriverState *bs); /* * Refreshes the bs->exact_filename field. If that is impossible, * bs->exact_filename has to be left empty. */ - void (*bdrv_refresh_filename)(BlockDriverState *bs); + void GRAPH_RDLOCK_PTR (*bdrv_refresh_filename)(BlockDriverState *bs); /* * Gathers the open options for all children into @target. @@ -292,15 +293,15 @@ struct BlockDriver { * block driver which implements it is probably doing something * shady regarding its runtime option structure. */ - void (*bdrv_gather_child_options)(BlockDriverState *bs, QDict *target, - bool backing_overridden); + void GRAPH_RDLOCK_PTR (*bdrv_gather_child_options)( + BlockDriverState *bs, QDict *target, bool backing_overridden); /* * Returns an allocated string which is the directory name of this BDS: It * will be used to make relative filenames absolute by prepending this * function's return value to them. */ - char *(*bdrv_dirname)(BlockDriverState *bs, Error **errp); + char * GRAPH_RDLOCK_PTR (*bdrv_dirname)(BlockDriverState *bs, Error **errp); /* * This informs the driver that we are no longer interested in the result @@ -313,14 +314,16 @@ struct BlockDriver { int GRAPH_RDLOCK_PTR (*bdrv_inactivate)(BlockDriverState *bs); - int (*bdrv_snapshot_create)(BlockDriverState *bs, - QEMUSnapshotInfo *sn_info); - int (*bdrv_snapshot_goto)(BlockDriverState *bs, - const char *snapshot_id); - int (*bdrv_snapshot_delete)(BlockDriverState *bs, - const char *snapshot_id, - const char *name, - Error **errp); + int GRAPH_RDLOCK_PTR (*bdrv_snapshot_create)( + BlockDriverState *bs, QEMUSnapshotInfo *sn_info); + + int GRAPH_UNLOCKED_PTR (*bdrv_snapshot_goto)( + BlockDriverState *bs, const char *snapshot_id); + + int GRAPH_RDLOCK_PTR (*bdrv_snapshot_delete)( + BlockDriverState *bs, const char *snapshot_id, const char *name, + Error **errp); + int (*bdrv_snapshot_list)(BlockDriverState *bs, QEMUSnapshotInfo **psn_info); int (*bdrv_snapshot_load_tmp)(BlockDriverState *bs, @@ -725,8 +728,8 @@ struct BlockDriver { int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_get_info)( BlockDriverState *bs, BlockDriverInfo *bdi); - ImageInfoSpecific *(*bdrv_get_specific_info)(BlockDriverState *bs, - Error **errp); + ImageInfoSpecific * GRAPH_RDLOCK_PTR (*bdrv_get_specific_info)( + BlockDriverState *bs, Error **errp); BlockStatsSpecific *(*bdrv_get_specific_stats)(BlockDriverState *bs); int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_save_vmstate)( @@ -963,15 +966,15 @@ struct BdrvChildClass { * Note that this can be nested. If drained_begin() was called twice, new * I/O is allowed only after drained_end() was called twice, too. */ - void (*drained_begin)(BdrvChild *child); - void (*drained_end)(BdrvChild *child); + void GRAPH_RDLOCK_PTR (*drained_begin)(BdrvChild *child); + void GRAPH_RDLOCK_PTR (*drained_end)(BdrvChild *child); /* * Returns whether the parent has pending requests for the child. This * callback is polled after .drained_begin() has been called until all * activity on the child has stopped. */ - bool (*drained_poll)(BdrvChild *child); + bool GRAPH_RDLOCK_PTR (*drained_poll)(BdrvChild *child); /* * Notifies the parent that the filename of its child has changed (e.g. @@ -1039,8 +1042,8 @@ struct BdrvChild { */ bool quiesced_parent; - QLIST_ENTRY(BdrvChild) next; - QLIST_ENTRY(BdrvChild) next_parent; + QLIST_ENTRY(BdrvChild GRAPH_RDLOCK_PTR) next; + QLIST_ENTRY(BdrvChild GRAPH_RDLOCK_PTR) next_parent; }; /* @@ -1173,11 +1176,11 @@ struct BlockDriverState { * See also comment in include/block/block.h, to learn how backing and file * are connected with BdrvChildRole. */ - QLIST_HEAD(, BdrvChild) children; + QLIST_HEAD(, BdrvChild GRAPH_RDLOCK_PTR) children; BdrvChild *backing; BdrvChild *file; - QLIST_HEAD(, BdrvChild) parents; + QLIST_HEAD(, BdrvChild GRAPH_RDLOCK_PTR) parents; QDict *options; QDict *explicit_options; diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index eb0da7232e..34eac72d7a 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -99,7 +99,7 @@ BlockDriver *bdrv_probe_all(const uint8_t *buf, int buf_size, */ void bdrv_wakeup(BlockDriverState *bs); -const char *bdrv_get_parent_name(const BlockDriverState *bs); +const char * GRAPH_RDLOCK bdrv_get_parent_name(const BlockDriverState *bs); bool blk_dev_has_tray(BlockBackend *blk); bool blk_dev_is_tray_open(BlockBackend *blk); @@ -133,7 +133,7 @@ bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint); BdrvChild *bdrv_cow_child(BlockDriverState *bs); BdrvChild *bdrv_filter_child(BlockDriverState *bs); BdrvChild *bdrv_filter_or_cow_child(BlockDriverState *bs); -BdrvChild *bdrv_primary_child(BlockDriverState *bs); +BdrvChild * GRAPH_RDLOCK bdrv_primary_child(BlockDriverState *bs); BlockDriverState *bdrv_skip_filters(BlockDriverState *bs); BlockDriverState *bdrv_backing_chain_next(BlockDriverState *bs); @@ -155,7 +155,8 @@ static inline BlockDriverState *bdrv_filter_or_cow_bs(BlockDriverState *bs) return child_bs(bdrv_filter_or_cow_child(bs)); } -static inline BlockDriverState *bdrv_primary_bs(BlockDriverState *bs) +static inline BlockDriverState * GRAPH_RDLOCK +bdrv_primary_bs(BlockDriverState *bs) { IO_CODE(); return child_bs(bdrv_primary_child(bs)); diff --git a/include/block/graph-lock.h b/include/block/graph-lock.h index 7e04f98ff0..6f1cd12745 100644 --- a/include/block/graph-lock.h +++ b/include/block/graph-lock.h @@ -116,7 +116,8 @@ void unregister_aiocontext(AioContext *ctx); * This function polls. Callers must not hold the lock of any AioContext other * than the current one and the one of @bs. */ -void bdrv_graph_wrlock(BlockDriverState *bs) TSA_ACQUIRE(graph_lock) TSA_NO_TSA; +void no_coroutine_fn TSA_ACQUIRE(graph_lock) TSA_NO_TSA +bdrv_graph_wrlock(BlockDriverState *bs); /* * bdrv_graph_wrunlock: diff --git a/include/block/qapi.h b/include/block/qapi.h index 8663971c58..54c48de26a 100644 --- a/include/block/qapi.h +++ b/include/block/qapi.h @@ -29,18 +29,17 @@ #include "block/snapshot.h" #include "qapi/qapi-types-block-core.h" -BlockDeviceInfo *bdrv_block_device_info(BlockBackend *blk, - BlockDriverState *bs, - bool flat, - Error **errp); -int bdrv_query_snapshot_info_list(BlockDriverState *bs, - SnapshotInfoList **p_list, - Error **errp); -void bdrv_query_image_info(BlockDriverState *bs, - ImageInfo **p_info, - bool flat, - bool skip_implicit_filters, - Error **errp); +BlockDeviceInfo * GRAPH_RDLOCK +bdrv_block_device_info(BlockBackend *blk, BlockDriverState *bs, + bool flat, Error **errp); + +int GRAPH_RDLOCK +bdrv_query_snapshot_info_list(BlockDriverState *bs, + SnapshotInfoList **p_list, + Error **errp); +void GRAPH_RDLOCK +bdrv_query_image_info(BlockDriverState *bs, ImageInfo **p_info, bool flat, + bool skip_implicit_filters, Error **errp); void GRAPH_RDLOCK bdrv_query_block_graph_info(BlockDriverState *bs, BlockGraphInfo **p_info, Error **errp); diff --git a/include/block/snapshot.h b/include/block/snapshot.h index 50ff924710..d49c5599d9 100644 --- a/include/block/snapshot.h +++ b/include/block/snapshot.h @@ -25,6 +25,7 @@ #ifndef SNAPSHOT_H #define SNAPSHOT_H +#include "block/graph-lock.h" #include "qapi/qapi-builtin-types.h" #define SNAPSHOT_OPT_BASE "snapshot." @@ -59,16 +60,19 @@ bool bdrv_snapshot_find_by_id_and_name(BlockDriverState *bs, const char *name, QEMUSnapshotInfo *sn_info, Error **errp); -int bdrv_can_snapshot(BlockDriverState *bs); -int bdrv_snapshot_create(BlockDriverState *bs, - QEMUSnapshotInfo *sn_info); -int bdrv_snapshot_goto(BlockDriverState *bs, - const char *snapshot_id, - Error **errp); -int bdrv_snapshot_delete(BlockDriverState *bs, - const char *snapshot_id, - const char *name, - Error **errp); + +int GRAPH_RDLOCK bdrv_can_snapshot(BlockDriverState *bs); + +int GRAPH_RDLOCK +bdrv_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info); + +int GRAPH_UNLOCKED +bdrv_snapshot_goto(BlockDriverState *bs, const char *snapshot_id, Error **errp); + +int GRAPH_RDLOCK +bdrv_snapshot_delete(BlockDriverState *bs, const char *snapshot_id, + const char *name, Error **errp); + int bdrv_snapshot_list(BlockDriverState *bs, QEMUSnapshotInfo **psn_info); int bdrv_snapshot_load_tmp(BlockDriverState *bs, diff --git a/include/exec/cpu-common.h b/include/exec/cpu-common.h index 605b160a7e..30c376a4de 100644 --- a/include/exec/cpu-common.h +++ b/include/exec/cpu-common.h @@ -83,6 +83,21 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length); ram_addr_t qemu_ram_addr_from_host(void *ptr); ram_addr_t qemu_ram_addr_from_host_nofail(void *ptr); RAMBlock *qemu_ram_block_by_name(const char *name); + +/* + * Translates a host ptr back to a RAMBlock and an offset in that RAMBlock. + * + * @ptr: The host pointer to translate. + * @round_offset: Whether to round the result offset down to a target page + * @offset: Will be set to the offset within the returned RAMBlock. + * + * Returns: RAMBlock (or NULL if not found) + * + * By the time this function returns, the returned pointer is not protected + * by RCU anymore. If the caller is not within an RCU critical section and + * does not hold the iothread lock, it must have other means of protecting the + * pointer, such as a reference to the memory region that owns the RAMBlock. + */ RAMBlock *qemu_ram_block_from_host(void *ptr, bool round_offset, ram_addr_t *offset); ram_addr_t qemu_ram_block_host_offset(RAMBlock *rb, void *host); diff --git a/include/exec/memory.h b/include/exec/memory.h index c99842d2fc..653a32ea10 100644 --- a/include/exec/memory.h +++ b/include/exec/memory.h @@ -95,6 +95,7 @@ struct ReservedRegion { * relative to the region's address space * @readonly: writes to this section are ignored * @nonvolatile: this section is non-volatile + * @unmergeable: this section should not get merged with adjacent sections */ struct MemoryRegionSection { Int128 size; @@ -104,6 +105,7 @@ struct MemoryRegionSection { hwaddr offset_within_address_space; bool readonly; bool nonvolatile; + bool unmergeable; }; typedef struct IOMMUTLBEntry IOMMUTLBEntry; @@ -599,8 +601,9 @@ typedef void (*ReplayRamDiscard)(MemoryRegionSection *section, void *opaque); * populated (consuming memory), to be used/accessed by the VM. * * A #RamDiscardManager can only be set for a RAM #MemoryRegion while the - * #MemoryRegion isn't mapped yet; it cannot change while the #MemoryRegion is - * mapped. + * #MemoryRegion isn't mapped into an address space yet (either directly + * or via an alias); it cannot change while the #MemoryRegion is + * mapped into an address space. * * The #RamDiscardManager is intended to be used by technologies that are * incompatible with discarding of RAM (e.g., VFIO, which may pin all @@ -772,6 +775,7 @@ struct MemoryRegion { bool nonvolatile; bool rom_device; bool flush_coalesced_mmio; + bool unmergeable; uint8_t dirty_log_mask; bool is_iommu; RAMBlock *ram_block; @@ -2350,6 +2354,25 @@ void memory_region_set_size(MemoryRegion *mr, uint64_t size); void memory_region_set_alias_offset(MemoryRegion *mr, hwaddr offset); +/* + * memory_region_set_unmergeable: Set a memory region unmergeable + * + * Mark a memory region unmergeable, resulting in the memory region (or + * everything contained in a memory region container) not getting merged when + * simplifying the address space and notifying memory listeners. Consequently, + * memory listeners will never get notified about ranges that are larger than + * the original memory regions. + * + * This is primarily useful when multiple aliases to a RAM memory region are + * mapped into a memory region container, and updates (e.g., enable/disable or + * map/unmap) of individual memory region aliases are not supposed to affect + * other memory regions in the same container. + * + * @mr: the #MemoryRegion to be updated + * @unmergeable: whether to mark the #MemoryRegion unmergeable + */ +void memory_region_set_unmergeable(MemoryRegion *mr, bool unmergeable); + /** * memory_region_present: checks if an address relative to a @container * translates into #MemoryRegion within @container diff --git a/include/hw/boards.h b/include/hw/boards.h index 55a64a13fd..43a56dc51e 100644 --- a/include/hw/boards.h +++ b/include/hw/boards.h @@ -297,15 +297,27 @@ struct MachineClass { * DeviceMemoryState: * @base: address in guest physical address space where the memory * address space for memory devices starts - * @mr: address space container for memory devices + * @mr: memory region container for memory devices + * @as: address space for memory devices + * @listener: memory listener used to track used memslots in the address space * @dimm_size: the sum of plugged DIMMs' sizes * @used_region_size: the part of @mr already used by memory devices + * @required_memslots: the number of memslots required by memory devices + * @used_memslots: the number of memslots currently used by memory devices + * @memslot_auto_decision_active: whether any plugged memory device + * automatically decided to use more than + * one memslot */ typedef struct DeviceMemoryState { hwaddr base; MemoryRegion mr; + AddressSpace as; + MemoryListener listener; uint64_t dimm_size; uint64_t used_region_size; + unsigned int required_memslots; + unsigned int used_memslots; + unsigned int memslot_auto_decision_active; } DeviceMemoryState; /** diff --git a/include/hw/mem/memory-device.h b/include/hw/mem/memory-device.h index 48d2611fc5..3354d6c166 100644 --- a/include/hw/mem/memory-device.h +++ b/include/hw/mem/memory-device.h @@ -14,6 +14,7 @@ #define MEMORY_DEVICE_H #include "hw/qdev-core.h" +#include "qemu/typedefs.h" #include "qapi/qapi-types-machine.h" #include "qom/object.h" @@ -41,6 +42,17 @@ typedef struct MemoryDeviceState MemoryDeviceState; * successive memory regions are used, a covering memory region has to * be provided. Scattered memory regions are not supported for single * devices. + * + * The device memory region returned via @get_memory_region may either be a + * single RAM memory region or a memory region container with subregions + * that are RAM memory regions or aliases to RAM memory regions. Other + * memory regions or subregions are not supported. + * + * If the device memory region returned via @get_memory_region is a + * memory region container, it's supported to dynamically (un)map subregions + * as long as the number of memslots returned by @get_memslots() won't + * be exceeded and as long as all memory regions are of the same kind (e.g., + * all RAM or all ROM). */ struct MemoryDeviceClass { /* private */ @@ -89,6 +101,28 @@ struct MemoryDeviceClass { MemoryRegion *(*get_memory_region)(MemoryDeviceState *md, Error **errp); /* + * Optional: Instruct the memory device to decide how many memory slots + * it requires, not exceeding the given limit. + * + * Called exactly once when pre-plugging the memory device, before + * querying the number of memslots using @get_memslots the first time. + */ + void (*decide_memslots)(MemoryDeviceState *md, unsigned int limit); + + /* + * Optional for memory devices that require only a single memslot, + * required for all other memory devices: Return the number of memslots + * (distinct RAM memory regions in the device memory region) that are + * required by the device. + * + * If this function is not implemented, the assumption is "1". + * + * Called when (un)plugging the memory device, to check if the requirements + * can be satisfied, and to do proper accounting. + */ + unsigned int (*get_memslots)(MemoryDeviceState *md); + + /* * Optional: Return the desired minimum alignment of the device in guest * physical address space. The final alignment is computed based on this * alignment and the alignment requirements of the memory region. @@ -105,8 +139,31 @@ struct MemoryDeviceClass { MemoryDeviceInfo *info); }; +/* + * Traditionally, KVM/vhost in many setups supported 509 memslots, whereby + * 253 memslots were "reserved" for boot memory and other devices (such + * as PCI BARs, which can get mapped dynamically) and 256 memslots were + * dedicated for DIMMs. These magic numbers worked reliably in the past. + * + * Further, using many memslots can negatively affect performance, so setting + * the soft-limit of memslots used by memory devices to the traditional + * DIMM limit of 256 sounds reasonable. + * + * If we have less than 509 memslots, we will instruct memory devices that + * support automatically deciding how many memslots to use to only use a single + * one. + * + * Hotplugging vhost devices with at least 509 memslots is not expected to + * cause problems, not even when devices automatically decided how many memslots + * to use. + */ +#define MEMORY_DEVICES_SOFT_MEMSLOT_LIMIT 256 +#define MEMORY_DEVICES_SAFE_MAX_MEMSLOTS 509 + MemoryDeviceInfoList *qmp_memory_device_list(void); uint64_t get_plugged_memory_size(void); +unsigned int memory_devices_get_reserved_memslots(void); +bool memory_devices_memslot_auto_decision_active(void); void memory_device_pre_plug(MemoryDeviceState *md, MachineState *ms, const uint64_t *legacy_align, Error **errp); void memory_device_plug(MemoryDeviceState *md, MachineState *ms); diff --git a/include/hw/virtio/vhost-backend.h b/include/hw/virtio/vhost-backend.h index 1860b541d8..96ccc18cd3 100644 --- a/include/hw/virtio/vhost-backend.h +++ b/include/hw/virtio/vhost-backend.h @@ -86,9 +86,6 @@ typedef int (*vhost_set_vring_enable_op)(struct vhost_dev *dev, typedef bool (*vhost_requires_shm_log_op)(struct vhost_dev *dev); typedef int (*vhost_migration_done_op)(struct vhost_dev *dev, char *mac_addr); -typedef bool (*vhost_backend_can_merge_op)(struct vhost_dev *dev, - uint64_t start1, uint64_t size1, - uint64_t start2, uint64_t size2); typedef int (*vhost_vsock_set_guest_cid_op)(struct vhost_dev *dev, uint64_t guest_cid); typedef int (*vhost_vsock_set_running_op)(struct vhost_dev *dev, int start); @@ -108,8 +105,7 @@ typedef int (*vhost_crypto_create_session_op)(struct vhost_dev *dev, typedef int (*vhost_crypto_close_session_op)(struct vhost_dev *dev, uint64_t session_id); -typedef bool (*vhost_backend_mem_section_filter_op)(struct vhost_dev *dev, - MemoryRegionSection *section); +typedef bool (*vhost_backend_no_private_memslots_op)(struct vhost_dev *dev); typedef int (*vhost_get_inflight_fd_op)(struct vhost_dev *dev, uint16_t queue_size, @@ -138,6 +134,7 @@ typedef struct VhostOps { vhost_backend_init vhost_backend_init; vhost_backend_cleanup vhost_backend_cleanup; vhost_backend_memslots_limit vhost_backend_memslots_limit; + vhost_backend_no_private_memslots_op vhost_backend_no_private_memslots; vhost_net_set_backend_op vhost_net_set_backend; vhost_net_set_mtu_op vhost_net_set_mtu; vhost_scsi_set_endpoint_op vhost_scsi_set_endpoint; @@ -163,7 +160,6 @@ typedef struct VhostOps { vhost_set_vring_enable_op vhost_set_vring_enable; vhost_requires_shm_log_op vhost_requires_shm_log; vhost_migration_done_op vhost_migration_done; - vhost_backend_can_merge_op vhost_backend_can_merge; vhost_vsock_set_guest_cid_op vhost_vsock_set_guest_cid; vhost_vsock_set_running_op vhost_vsock_set_running; vhost_set_iotlb_callback_op vhost_set_iotlb_callback; @@ -172,7 +168,6 @@ typedef struct VhostOps { vhost_set_config_op vhost_set_config; vhost_crypto_create_session_op vhost_crypto_create_session; vhost_crypto_close_session_op vhost_crypto_close_session; - vhost_backend_mem_section_filter_op vhost_backend_mem_section_filter; vhost_get_inflight_fd_op vhost_get_inflight_fd; vhost_set_inflight_fd_op vhost_set_inflight_fd; vhost_dev_start_op vhost_dev_start; diff --git a/include/hw/virtio/vhost.h b/include/hw/virtio/vhost.h index 6a173cb9fa..c7e5467693 100644 --- a/include/hw/virtio/vhost.h +++ b/include/hw/virtio/vhost.h @@ -315,7 +315,8 @@ uint64_t vhost_get_features(struct vhost_dev *hdev, const int *feature_bits, */ void vhost_ack_features(struct vhost_dev *hdev, const int *feature_bits, uint64_t features); -bool vhost_has_free_slot(void); +unsigned int vhost_get_max_memslots(void); +unsigned int vhost_get_free_memslots(void); int vhost_net_set_backend(struct vhost_dev *hdev, struct vhost_vring_file *file); diff --git a/include/hw/virtio/virtio-gpu.h b/include/hw/virtio/virtio-gpu.h index 390c4642b8..4739fa4689 100644 --- a/include/hw/virtio/virtio-gpu.h +++ b/include/hw/virtio/virtio-gpu.h @@ -225,13 +225,13 @@ struct VhostUserGPU { }; #define VIRTIO_GPU_FILL_CMD(out) do { \ - size_t s; \ - s = iov_to_buf(cmd->elem.out_sg, cmd->elem.out_num, 0, \ + size_t virtiogpufillcmd_s_ = \ + iov_to_buf(cmd->elem.out_sg, cmd->elem.out_num, 0, \ &out, sizeof(out)); \ - if (s != sizeof(out)) { \ + if (virtiogpufillcmd_s_ != sizeof(out)) { \ qemu_log_mask(LOG_GUEST_ERROR, \ "%s: command size incorrect %zu vs %zu\n", \ - __func__, s, sizeof(out)); \ + __func__, virtiogpufillcmd_s_, sizeof(out)); \ return; \ } \ } while (0) diff --git a/include/hw/virtio/virtio-mem.h b/include/hw/virtio/virtio-mem.h index ab0fe2b4f2..5f5b02b8f9 100644 --- a/include/hw/virtio/virtio-mem.h +++ b/include/hw/virtio/virtio-mem.h @@ -33,6 +33,7 @@ OBJECT_DECLARE_TYPE(VirtIOMEM, VirtIOMEMClass, #define VIRTIO_MEM_UNPLUGGED_INACCESSIBLE_PROP "unplugged-inaccessible" #define VIRTIO_MEM_EARLY_MIGRATION_PROP "x-early-migration" #define VIRTIO_MEM_PREALLOC_PROP "prealloc" +#define VIRTIO_MEM_DYNAMIC_MEMSLOTS_PROP "dynamic-memslots" struct VirtIOMEM { VirtIODevice parent_obj; @@ -44,7 +45,28 @@ struct VirtIOMEM { int32_t bitmap_size; unsigned long *bitmap; - /* assigned memory backend and memory region */ + /* + * With "dynamic-memslots=on": Device memory region in which we dynamically + * map the memslots. + */ + MemoryRegion *mr; + + /* + * With "dynamic-memslots=on": The individual memslots (aliases into the + * memory backend). + */ + MemoryRegion *memslots; + + /* With "dynamic-memslots=on": The total number of memslots. */ + uint16_t nb_memslots; + + /* + * With "dynamic-memslots=on": Size of one memslot (the size of the + * last one can differ). + */ + uint64_t memslot_size; + + /* Assigned memory backend with the RAM memory region. */ HostMemoryBackend *memdev; /* NUMA node */ @@ -82,6 +104,12 @@ struct VirtIOMEM { */ bool early_migration; + /* + * Whether we dynamically map (multiple, if possible) memslots instead of + * statically mapping the whole RAM memory region. + */ + bool dynamic_memslots; + /* notifiers to notify when "size" changes */ NotifierList size_change_notifiers; @@ -96,6 +124,8 @@ struct VirtIOMEMClass { /* public */ void (*fill_device_info)(const VirtIOMEM *vmen, VirtioMEMDeviceInfo *vi); MemoryRegion *(*get_memory_region)(VirtIOMEM *vmem, Error **errp); + void (*decide_memslots)(VirtIOMEM *vmem, unsigned int limit); + unsigned int (*get_memslots)(VirtIOMEM *vmem); void (*add_size_change_notifier)(VirtIOMEM *vmem, Notifier *notifier); void (*remove_size_change_notifier)(VirtIOMEM *vmem, Notifier *notifier); void (*unplug_request_check)(VirtIOMEM *vmem, Error **errp); diff --git a/include/sysemu/block-backend-global-state.h b/include/sysemu/block-backend-global-state.h index d5f675493a..49c12b0fa9 100644 --- a/include/sysemu/block-backend-global-state.h +++ b/include/sysemu/block-backend-global-state.h @@ -59,8 +59,8 @@ BlockBackend *blk_by_public(BlockBackendPublic *public); void blk_remove_bs(BlockBackend *blk); int blk_insert_bs(BlockBackend *blk, BlockDriverState *bs, Error **errp); int blk_replace_bs(BlockBackend *blk, BlockDriverState *new_bs, Error **errp); -bool bdrv_has_blk(BlockDriverState *bs); -bool bdrv_is_root_node(BlockDriverState *bs); +bool GRAPH_RDLOCK bdrv_has_blk(BlockDriverState *bs); +bool GRAPH_RDLOCK bdrv_is_root_node(BlockDriverState *bs); int GRAPH_UNLOCKED blk_set_perm(BlockBackend *blk, uint64_t perm, uint64_t shared_perm, Error **errp); void blk_get_perm(BlockBackend *blk, uint64_t *perm, uint64_t *shared_perm); diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h index ee9025f8e9..97a8a4f201 100644 --- a/include/sysemu/kvm.h +++ b/include/sysemu/kvm.h @@ -215,7 +215,8 @@ typedef struct KVMRouteChange { /* external API */ -bool kvm_has_free_slot(MachineState *ms); +unsigned int kvm_get_max_memslots(void); +unsigned int kvm_get_free_memslots(void); bool kvm_has_sync_mmu(void); int kvm_has_vcpu_events(void); int kvm_has_robust_singlestep(void); @@ -552,7 +553,6 @@ int kvm_set_one_reg(CPUState *cs, uint64_t id, void *source); */ int kvm_get_one_reg(CPUState *cs, uint64_t id, void *target); struct ppc_radix_page_info *kvm_get_radix_page_info(void); -int kvm_get_max_memslots(void); /* Notify resamplefd for EOI of specific interrupts. */ void kvm_resample_fd_notify(int gsi); diff --git a/include/sysemu/kvm_int.h b/include/sysemu/kvm_int.h index a5b9122cb8..075939a3c4 100644 --- a/include/sysemu/kvm_int.h +++ b/include/sysemu/kvm_int.h @@ -40,6 +40,7 @@ typedef struct KVMMemoryUpdate { typedef struct KVMMemoryListener { MemoryListener listener; KVMSlot *slots; + unsigned int nr_used_slots; int as_id; QSIMPLEQ_HEAD(, KVMMemoryUpdate) transaction_add; QSIMPLEQ_HEAD(, KVMMemoryUpdate) transaction_del; diff --git a/migration/block.c b/migration/block.c index 5f930870a5..d115e1cfa5 100644 --- a/migration/block.c +++ b/migration/block.c @@ -388,6 +388,8 @@ static int init_blk_migration(QEMUFile *f) Error *local_err = NULL; int ret; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + block_mig_state.submitted = 0; block_mig_state.read_done = 0; block_mig_state.transferred = 0; diff --git a/migration/migration-hmp-cmds.c b/migration/migration-hmp-cmds.c index c115ef2d23..5b25ba24f7 100644 --- a/migration/migration-hmp-cmds.c +++ b/migration/migration-hmp-cmds.c @@ -794,6 +794,8 @@ static void vm_completion(ReadLineState *rs, const char *str) BlockDriverState *bs; BdrvNextIterator it; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + len = strlen(str); readline_set_completion_index(rs, len); diff --git a/qemu-img.c b/qemu-img.c index 6068ab0d27..585b65640f 100644 --- a/qemu-img.c +++ b/qemu-img.c @@ -3165,7 +3165,9 @@ static int get_block_status(BlockDriverState *bs, int64_t offset, has_offset = !!(ret & BDRV_BLOCK_OFFSET_VALID); if (file && has_offset) { + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(file); + bdrv_graph_rdunlock_main_loop(); filename = file->filename; } @@ -3470,7 +3472,10 @@ static int img_snapshot(int argc, char **argv) sn.date_sec = rt / G_USEC_PER_SEC; sn.date_nsec = (rt % G_USEC_PER_SEC) * 1000; + bdrv_graph_rdlock_main_loop(); ret = bdrv_snapshot_create(bs, &sn); + bdrv_graph_rdunlock_main_loop(); + if (ret) { error_report("Could not create snapshot '%s': %s", snapshot_name, strerror(-ret)); @@ -3486,6 +3491,7 @@ static int img_snapshot(int argc, char **argv) break; case SNAPSHOT_DELETE: + bdrv_graph_rdlock_main_loop(); ret = bdrv_snapshot_find(bs, &sn, snapshot_name); if (ret < 0) { error_report("Could not delete snapshot '%s': snapshot not " @@ -3499,6 +3505,7 @@ static int img_snapshot(int argc, char **argv) ret = 1; } } + bdrv_graph_rdunlock_main_loop(); break; } @@ -3683,7 +3690,9 @@ static int img_rebase(int argc, char **argv) qdict_put_bool(options, BDRV_OPT_FORCE_SHARE, true); } + bdrv_graph_rdlock_main_loop(); bdrv_refresh_filename(bs); + bdrv_graph_rdunlock_main_loop(); overlay_filename = bs->exact_filename[0] ? bs->exact_filename : bs->filename; out_real_path = @@ -4120,6 +4129,8 @@ static int print_amend_option_help(const char *format) { BlockDriver *drv; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + /* Find driver and parse its options */ drv = bdrv_find_format(format); if (!drv) { @@ -4258,9 +4269,11 @@ static int img_amend(int argc, char **argv) goto out; } + bdrv_graph_rdlock_main_loop(); if (!bs->drv->bdrv_amend_options) { error_report("Format driver '%s' does not support option amendment", fmt); + bdrv_graph_rdunlock_main_loop(); ret = -1; goto out; } @@ -4280,6 +4293,7 @@ static int img_amend(int argc, char **argv) "This option is only supported for image creation\n"); } + bdrv_graph_rdunlock_main_loop(); error_report_err(err); ret = -1; goto out; @@ -4289,6 +4303,8 @@ static int img_amend(int argc, char **argv) qemu_progress_print(0.f, 0); ret = bdrv_amend_options(bs, opts, &amend_status_cb, NULL, force, &err); qemu_progress_print(100.f, 0); + bdrv_graph_rdunlock_main_loop(); + if (ret < 0) { error_report_err(err); goto out; diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index 3f75d2f5a6..f5d7202a13 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -2037,6 +2037,9 @@ static int info_f(BlockBackend *blk, int argc, char **argv) char s1[64], s2[64]; int ret; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (bs->drv && bs->drv->format_name) { printf("format name: %s\n", bs->drv->format_name); } diff --git a/scripts/block-coroutine-wrapper.py b/scripts/block-coroutine-wrapper.py index 685d0b4ed4..a601c3c672 100644 --- a/scripts/block-coroutine-wrapper.py +++ b/scripts/block-coroutine-wrapper.py @@ -87,8 +87,9 @@ class FuncDecl: raise ValueError(f"Invalid no_co function name: {self.name}") if not self.create_only_co: raise ValueError(f"no_co function can't be mixed: {self.name}") - if self.graph_rdlock: - raise ValueError(f"no_co function can't be rdlock: {self.name}") + if self.graph_rdlock and self.graph_wrlock: + raise ValueError("function can't be both rdlock and wrlock: " + f"{self.name}") self.target_name = f'{subsystem}_{subname}' self.ctx = self.gen_ctx() @@ -256,7 +257,10 @@ def gen_no_co_wrapper(func: FuncDecl) -> str: graph_lock='' graph_unlock='' - if func.graph_wrlock: + if func.graph_rdlock: + graph_lock=' bdrv_graph_rdlock_main_loop();' + graph_unlock=' bdrv_graph_rdunlock_main_loop();' + elif func.graph_wrlock: graph_lock=' bdrv_graph_wrlock(NULL);' graph_unlock=' bdrv_graph_wrunlock();' diff --git a/stubs/qmp_memory_device.c b/stubs/memory_device.c index e75cac62dc..15fd93ff67 100644 --- a/stubs/qmp_memory_device.c +++ b/stubs/memory_device.c @@ -10,3 +10,13 @@ uint64_t get_plugged_memory_size(void) { return (uint64_t)-1; } + +unsigned int memory_devices_get_reserved_memslots(void) +{ + return 0; +} + +bool memory_devices_memslot_auto_decision_active(void) +{ + return false; +} diff --git a/stubs/meson.build b/stubs/meson.build index ef6e39a64d..cde44972bf 100644 --- a/stubs/meson.build +++ b/stubs/meson.build @@ -32,7 +32,7 @@ stub_ss.add(files('monitor.c')) stub_ss.add(files('monitor-core.c')) stub_ss.add(files('physmem.c')) stub_ss.add(files('qemu-timer-notify-cb.c')) -stub_ss.add(files('qmp_memory_device.c')) +stub_ss.add(files('memory_device.c')) stub_ss.add(files('qmp-command-available.c')) stub_ss.add(files('qmp-quit.c')) stub_ss.add(files('qtest.c')) diff --git a/subprojects/libvduse/libvduse.c b/subprojects/libvduse/libvduse.c index 377959a0b4..21ffbb5b8d 100644 --- a/subprojects/libvduse/libvduse.c +++ b/subprojects/libvduse/libvduse.c @@ -1031,7 +1031,7 @@ int vduse_dev_handler(VduseDev *dev) /* The iova will be updated by iova_to_va() later, so just remove it */ vduse_iova_remove_region(dev, req.iova.start, req.iova.last); for (i = 0; i < dev->num_queues; i++) { - VduseVirtq *vq = &dev->vqs[i]; + vq = &dev->vqs[i]; if (vq->ready) { if (vduse_queue_update_vring(vq, vq->vring.desc_addr, vq->vring.avail_addr, diff --git a/subprojects/libvhost-user/libvhost-user.c b/subprojects/libvhost-user/libvhost-user.c index ac6d5d01d3..6684057370 100644 --- a/subprojects/libvhost-user/libvhost-user.c +++ b/subprojects/libvhost-user/libvhost-user.c @@ -870,10 +870,10 @@ vu_rem_mem_reg(VuDev *dev, VhostUserMsg *vmsg) { for (i = 0; i < dev->nregions; i++) { if (reg_equal(&dev->regions[i], msg_region)) { VuDevRegion *r = &dev->regions[i]; - void *m = (void *) (uintptr_t) r->mmap_addr; + void *ma = (void *) (uintptr_t) r->mmap_addr; - if (m) { - munmap(m, r->size + r->mmap_offset); + if (ma) { + munmap(ma, r->size + r->mmap_offset); } /* @@ -1005,10 +1005,10 @@ vu_set_mem_table_exec(VuDev *dev, VhostUserMsg *vmsg) for (i = 0; i < dev->nregions; i++) { VuDevRegion *r = &dev->regions[i]; - void *m = (void *) (uintptr_t) r->mmap_addr; + void *ma = (void *) (uintptr_t) r->mmap_addr; - if (m) { - munmap(m, r->size + r->mmap_offset); + if (ma) { + munmap(ma, r->size + r->mmap_offset); } } dev->nregions = memory->nregions; diff --git a/system/memory.c b/system/memory.c index fa1c99f9ba..a800fbc9e5 100644 --- a/system/memory.c +++ b/system/memory.c @@ -224,6 +224,7 @@ struct FlatRange { bool romd_mode; bool readonly; bool nonvolatile; + bool unmergeable; }; #define FOR_EACH_FLAT_RANGE(var, view) \ @@ -240,6 +241,7 @@ section_from_flat_range(FlatRange *fr, FlatView *fv) .offset_within_address_space = int128_get64(fr->addr.start), .readonly = fr->readonly, .nonvolatile = fr->nonvolatile, + .unmergeable = fr->unmergeable, }; } @@ -250,7 +252,8 @@ static bool flatrange_equal(FlatRange *a, FlatRange *b) && a->offset_in_region == b->offset_in_region && a->romd_mode == b->romd_mode && a->readonly == b->readonly - && a->nonvolatile == b->nonvolatile; + && a->nonvolatile == b->nonvolatile + && a->unmergeable == b->unmergeable; } static FlatView *flatview_new(MemoryRegion *mr_root) @@ -323,7 +326,8 @@ static bool can_merge(FlatRange *r1, FlatRange *r2) && r1->dirty_log_mask == r2->dirty_log_mask && r1->romd_mode == r2->romd_mode && r1->readonly == r2->readonly - && r1->nonvolatile == r2->nonvolatile; + && r1->nonvolatile == r2->nonvolatile + && !r1->unmergeable && !r2->unmergeable; } /* Attempt to simplify a view by merging adjacent ranges */ @@ -599,7 +603,8 @@ static void render_memory_region(FlatView *view, Int128 base, AddrRange clip, bool readonly, - bool nonvolatile) + bool nonvolatile, + bool unmergeable) { MemoryRegion *subregion; unsigned i; @@ -616,6 +621,7 @@ static void render_memory_region(FlatView *view, int128_addto(&base, int128_make64(mr->addr)); readonly |= mr->readonly; nonvolatile |= mr->nonvolatile; + unmergeable |= mr->unmergeable; tmp = addrrange_make(base, mr->size); @@ -629,14 +635,14 @@ static void render_memory_region(FlatView *view, int128_subfrom(&base, int128_make64(mr->alias->addr)); int128_subfrom(&base, int128_make64(mr->alias_offset)); render_memory_region(view, mr->alias, base, clip, - readonly, nonvolatile); + readonly, nonvolatile, unmergeable); return; } /* Render subregions in priority order. */ QTAILQ_FOREACH(subregion, &mr->subregions, subregions_link) { render_memory_region(view, subregion, base, clip, - readonly, nonvolatile); + readonly, nonvolatile, unmergeable); } if (!mr->terminates) { @@ -652,6 +658,7 @@ static void render_memory_region(FlatView *view, fr.romd_mode = mr->romd_mode; fr.readonly = readonly; fr.nonvolatile = nonvolatile; + fr.unmergeable = unmergeable; /* Render the region itself into any gaps left by the current view. */ for (i = 0; i < view->nr && int128_nz(remain); ++i) { @@ -753,7 +760,7 @@ static FlatView *generate_memory_topology(MemoryRegion *mr) if (mr) { render_memory_region(view, mr, int128_zero(), addrrange_make(int128_zero(), int128_2_64()), - false, false); + false, false, false); } flatview_simplify(view); @@ -2085,7 +2092,7 @@ int memory_region_iommu_num_indexes(IOMMUMemoryRegion *iommu_mr) RamDiscardManager *memory_region_get_ram_discard_manager(MemoryRegion *mr) { - if (!memory_region_is_mapped(mr) || !memory_region_is_ram(mr)) { + if (!memory_region_is_ram(mr)) { return NULL; } return mr->rdm; @@ -2094,7 +2101,7 @@ RamDiscardManager *memory_region_get_ram_discard_manager(MemoryRegion *mr) void memory_region_set_ram_discard_manager(MemoryRegion *mr, RamDiscardManager *rdm) { - g_assert(memory_region_is_ram(mr) && !memory_region_is_mapped(mr)); + g_assert(memory_region_is_ram(mr)); g_assert(!rdm || !mr->rdm); mr->rdm = rdm; } @@ -2755,6 +2762,18 @@ void memory_region_set_alias_offset(MemoryRegion *mr, hwaddr offset) memory_region_transaction_commit(); } +void memory_region_set_unmergeable(MemoryRegion *mr, bool unmergeable) +{ + if (unmergeable == mr->unmergeable) { + return; + } + + memory_region_transaction_begin(); + mr->unmergeable = unmergeable; + memory_region_update_pending |= mr->enabled; + memory_region_transaction_commit(); +} + uint64_t memory_region_get_alignment(const MemoryRegion *mr) { return mr->align; diff --git a/system/physmem.c b/system/physmem.c index edc3ed8ab9..fc2b0fee01 100644 --- a/system/physmem.c +++ b/system/physmem.c @@ -2221,23 +2221,6 @@ ram_addr_t qemu_ram_block_host_offset(RAMBlock *rb, void *host) return res; } -/* - * Translates a host ptr back to a RAMBlock, a ram_addr and an offset - * in that RAMBlock. - * - * ptr: Host pointer to look up - * round_offset: If true round the result offset down to a page boundary - * *ram_addr: set to result ram_addr - * *offset: set to result offset within the RAMBlock - * - * Returns: RAMBlock (or NULL if not found) - * - * By the time this function returns, the returned pointer is not protected - * by RCU anymore. If the caller is not within an RCU critical section and - * does not hold the iothread lock, it must have other means of protecting the - * pointer, such as a reference to the region that includes the incoming - * ram_addr_t. - */ RAMBlock *qemu_ram_block_from_host(void *ptr, bool round_offset, ram_addr_t *offset) { diff --git a/target/i386/tcg/seg_helper.c b/target/i386/tcg/seg_helper.c index 2b92aee207..eb29a1fd4e 100644 --- a/target/i386/tcg/seg_helper.c +++ b/target/i386/tcg/seg_helper.c @@ -229,7 +229,7 @@ static void tss_load_seg(CPUX86State *env, X86Seg seg_reg, int selector, static void tss_set_busy(CPUX86State *env, int tss_selector, bool value, uintptr_t retaddr) { - target_ulong ptr = env->gdt.base + (env->tr.selector & ~7); + target_ulong ptr = env->gdt.base + (tss_selector & ~7); uint32_t e2 = cpu_ldl_kernel_ra(env, ptr + 4, retaddr); if (value) { diff --git a/tests/qtest/npcm7xx_timer-test.c b/tests/qtest/npcm7xx_timer-test.c index 43711049ca..58f58c2f71 100644 --- a/tests/qtest/npcm7xx_timer-test.c +++ b/tests/qtest/npcm7xx_timer-test.c @@ -465,6 +465,7 @@ static void test_periodic_interrupt(gconstpointer test_data) int i; tim_reset(td); + clock_step_next(); tim_write_ticr(td, count); tim_write_tcsr(td, CEN | IE | MODE_PERIODIC | PRESCALE(ps)); diff --git a/tests/unit/test-bdrv-drain.c b/tests/unit/test-bdrv-drain.c index 0b603e7c57..f67e9df01c 100644 --- a/tests/unit/test-bdrv-drain.c +++ b/tests/unit/test-bdrv-drain.c @@ -1034,9 +1034,13 @@ static void coroutine_fn test_co_delete_by_drain(void *opaque) blk_co_unref(blk); } else { BdrvChild *c, *next_c; + bdrv_graph_co_rdlock(); QLIST_FOREACH_SAFE(c, &bs->children, next, next_c) { + bdrv_graph_co_rdunlock(); bdrv_co_unref_child(bs, c); + bdrv_graph_co_rdlock(); } + bdrv_graph_co_rdunlock(); } dbdd->done = true; @@ -1168,7 +1172,7 @@ struct detach_by_parent_data { }; static struct detach_by_parent_data detach_by_parent_data; -static void detach_indirect_bh(void *opaque) +static void no_coroutine_fn detach_indirect_bh(void *opaque) { struct detach_by_parent_data *data = opaque; @@ -1184,18 +1188,19 @@ static void detach_indirect_bh(void *opaque) bdrv_graph_wrunlock(); } -static void detach_by_parent_aio_cb(void *opaque, int ret) +static void coroutine_mixed_fn detach_by_parent_aio_cb(void *opaque, int ret) { struct detach_by_parent_data *data = &detach_by_parent_data; g_assert_cmpint(ret, ==, 0); if (data->by_parent_cb) { bdrv_inc_in_flight(data->child_b->bs); - detach_indirect_bh(data); + aio_bh_schedule_oneshot(qemu_get_current_aio_context(), + detach_indirect_bh, &detach_by_parent_data); } } -static void detach_by_driver_cb_drained_begin(BdrvChild *child) +static void GRAPH_RDLOCK detach_by_driver_cb_drained_begin(BdrvChild *child) { struct detach_by_parent_data *data = &detach_by_parent_data; @@ -1232,7 +1237,7 @@ static BdrvChildClass detach_by_driver_cb_class; * state is messed up, but if it is only polled in the single * BDRV_POLL_WHILE() at the end of the drain, this should work fine. */ -static void test_detach_indirect(bool by_parent_cb) +static void TSA_NO_TSA test_detach_indirect(bool by_parent_cb) { BlockBackend *blk; BlockDriverState *parent_a, *parent_b, *a, *b, *c; diff --git a/tests/unit/test-block-iothread.c b/tests/unit/test-block-iothread.c index 9155547313..9b15d2768c 100644 --- a/tests/unit/test-block-iothread.c +++ b/tests/unit/test-block-iothread.c @@ -383,6 +383,9 @@ static void test_sync_op_check(BdrvChild *c) static void test_sync_op_activate(BdrvChild *c) { + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + /* Early success: Image is not inactive */ bdrv_activate(c->bs, NULL); } @@ -468,11 +471,16 @@ static void test_sync_op(const void *opaque) BlockDriverState *bs; BdrvChild *c; + GLOBAL_STATE_CODE(); + blk = blk_new(qemu_get_aio_context(), BLK_PERM_ALL, BLK_PERM_ALL); bs = bdrv_new_open_driver(&bdrv_test, "base", BDRV_O_RDWR, &error_abort); bs->total_sectors = 65536 / BDRV_SECTOR_SIZE; blk_insert_bs(blk, bs, &error_abort); + + bdrv_graph_rdlock_main_loop(); c = QLIST_FIRST(&bs->parents); + bdrv_graph_rdunlock_main_loop(); blk_set_aio_context(blk, ctx, &error_abort); aio_context_acquire(ctx); |