diff options
Diffstat (limited to 'drivers/firewire')
| -rw-r--r-- | drivers/firewire/Kconfig | 2 | ||||
| -rw-r--r-- | drivers/firewire/core-card.c | 510 | ||||
| -rw-r--r-- | drivers/firewire/core-cdev.c | 172 | ||||
| -rw-r--r-- | drivers/firewire/core-device.c | 244 | ||||
| -rw-r--r-- | drivers/firewire/core-iso.c | 103 | ||||
| -rw-r--r-- | drivers/firewire/core-topology.c | 94 | ||||
| -rw-r--r-- | drivers/firewire/core-transaction.c | 321 | ||||
| -rw-r--r-- | drivers/firewire/core.h | 41 | ||||
| -rw-r--r-- | drivers/firewire/device-attribute-test.c | 2 | ||||
| -rw-r--r-- | drivers/firewire/init_ohci1394_dma.c | 10 | ||||
| -rw-r--r-- | drivers/firewire/net.c | 15 | ||||
| -rw-r--r-- | drivers/firewire/nosy.c | 14 | ||||
| -rw-r--r-- | drivers/firewire/ohci.c | 834 | ||||
| -rw-r--r-- | drivers/firewire/sbp2.c | 25 |
14 files changed, 1275 insertions, 1112 deletions
diff --git a/drivers/firewire/Kconfig b/drivers/firewire/Kconfig index 905c82e26ce7..a5f5e250223a 100644 --- a/drivers/firewire/Kconfig +++ b/drivers/firewire/Kconfig @@ -83,7 +83,7 @@ config FIREWIRE_KUNIT_SELF_ID_SEQUENCE_HELPER_TEST config FIREWIRE_OHCI tristate "OHCI-1394 controllers" - depends on PCI && FIREWIRE && MMU + depends on PCI && FIREWIRE help Enable this driver if you have a FireWire controller based on the OHCI specification. For all practical purposes, this diff --git a/drivers/firewire/core-card.c b/drivers/firewire/core-card.c index 01354b9de8b2..a754c6366b97 100644 --- a/drivers/firewire/core-card.c +++ b/drivers/firewire/core-card.c @@ -86,8 +86,6 @@ static size_t config_rom_length = 1 + 4 + 1 + 1; */ #define DEFAULT_SPLIT_TIMEOUT (2 * 8000) -#define CANON_OUI 0x000085 - static void generate_config_rom(struct fw_card *card, __be32 *config_rom) { struct fw_descriptor *desc; @@ -229,22 +227,21 @@ void fw_schedule_bus_reset(struct fw_card *card, bool delayed, bool short_reset) /* Use an arbitrary short delay to combine multiple reset requests. */ fw_card_get(card); - if (!queue_delayed_work(fw_workqueue, &card->br_work, - delayed ? DIV_ROUND_UP(HZ, 100) : 0)) + if (!queue_delayed_work(fw_workqueue, &card->br_work, delayed ? msecs_to_jiffies(10) : 0)) fw_card_put(card); } EXPORT_SYMBOL(fw_schedule_bus_reset); static void br_work(struct work_struct *work) { - struct fw_card *card = container_of(work, struct fw_card, br_work.work); + struct fw_card *card = from_work(card, work, br_work.work); /* Delay for 2s after last reset per IEEE 1394 clause 8.2.1. */ if (card->reset_jiffies != 0 && - time_before64(get_jiffies_64(), card->reset_jiffies + 2 * HZ)) { + time_is_after_jiffies64(card->reset_jiffies + secs_to_jiffies(2))) { trace_bus_reset_postpone(card->index, card->generation, card->br_short); - if (!queue_delayed_work(fw_workqueue, &card->br_work, 2 * HZ)) + if (!queue_delayed_work(fw_workqueue, &card->br_work, secs_to_jiffies(2))) fw_card_put(card); return; } @@ -273,10 +270,6 @@ static void allocate_broadcast_channel(struct fw_card *card, int generation) fw_device_set_broadcast_channel); } -static const char gap_count_table[] = { - 63, 5, 7, 8, 10, 13, 16, 18, 21, 24, 26, 29, 32, 35, 37, 40 -}; - void fw_schedule_bm_work(struct fw_card *card, unsigned long delay) { fw_card_get(card); @@ -284,222 +277,249 @@ void fw_schedule_bm_work(struct fw_card *card, unsigned long delay) fw_card_put(card); } -static void bm_work(struct work_struct *work) -{ - struct fw_card *card = container_of(work, struct fw_card, bm_work.work); - struct fw_device *root_device, *irm_device; - struct fw_node *root_node; - int root_id, new_root_id, irm_id, bm_id, local_id; - int gap_count, generation, grace, rcode; - bool do_reset = false; - bool root_device_is_running; - bool root_device_is_cmc; - bool irm_is_1394_1995_only; - bool keep_this_irm; - __be32 transaction_data[2]; - - spin_lock_irq(&card->lock); +enum bm_contention_outcome { + // The bus management contention window is not expired. + BM_CONTENTION_OUTCOME_WITHIN_WINDOW = 0, + // The IRM node has link off. + BM_CONTENTION_OUTCOME_IRM_HAS_LINK_OFF, + // The IRM node complies IEEE 1394:1994 only. + BM_CONTENTION_OUTCOME_IRM_COMPLIES_1394_1995_ONLY, + // Another bus reset, BM work has been rescheduled. + BM_CONTENTION_OUTCOME_AT_NEW_GENERATION, + // We have been unable to send the lock request to IRM node due to some local problem. + BM_CONTENTION_OUTCOME_LOCAL_PROBLEM_AT_TRANSACTION, + // The lock request failed, maybe the IRM isn't really IRM capable after all. + BM_CONTENTION_OUTCOME_IRM_IS_NOT_CAPABLE_FOR_IRM, + // Somebody else is BM. + BM_CONTENTION_OUTCOME_IRM_HOLDS_ANOTHER_NODE_AS_BM, + // The local node succeeds after contending for bus manager. + BM_CONTENTION_OUTCOME_IRM_HOLDS_LOCAL_NODE_AS_BM, +}; - if (card->local_node == NULL) { - spin_unlock_irq(&card->lock); - goto out_put_card; +static enum bm_contention_outcome contend_for_bm(struct fw_card *card) +__must_hold(&card->lock) +{ + int generation = card->generation; + int local_id = card->local_node->node_id; + __be32 data[2] = { + cpu_to_be32(BUS_MANAGER_ID_NOT_REGISTERED), + cpu_to_be32(local_id), + }; + bool grace = time_is_before_jiffies64(card->reset_jiffies + msecs_to_jiffies(125)); + struct fw_node *irm_node; + struct fw_device *irm_device; + int irm_node_id, irm_device_quirks = 0; + int rcode; + + lockdep_assert_held(&card->lock); + + if (!grace) { + if (!is_next_generation(generation, card->bm_generation) || card->bm_abdicate) + return BM_CONTENTION_OUTCOME_WITHIN_WINDOW; } - generation = card->generation; - - root_node = card->root_node; - fw_node_get(root_node); - root_device = root_node->data; - root_device_is_running = root_device && - atomic_read(&root_device->state) == FW_DEVICE_RUNNING; - root_device_is_cmc = root_device && root_device->cmc; + irm_node = card->irm_node; + if (!irm_node->link_on) { + fw_notice(card, "IRM has link off, making local node (%02x) root\n", local_id); + return BM_CONTENTION_OUTCOME_IRM_HAS_LINK_OFF; + } - irm_device = card->irm_node->data; - irm_is_1394_1995_only = irm_device && irm_device->config_rom && - (irm_device->config_rom[2] & 0x000000f0) == 0; + // NOTE: It is likely that the quirk detection for IRM device has not done yet. + irm_device = fw_node_get_device(irm_node); + if (irm_device) + irm_device_quirks = READ_ONCE(irm_device->quirks); + if ((irm_device_quirks & FW_DEVICE_QUIRK_IRM_IS_1394_1995_ONLY) && + !(irm_device_quirks & FW_DEVICE_QUIRK_IRM_IGNORES_BUS_MANAGER)) { + fw_notice(card, "IRM is not 1394a compliant, making local node (%02x) root\n", + local_id); + return BM_CONTENTION_OUTCOME_IRM_COMPLIES_1394_1995_ONLY; + } - /* Canon MV5i works unreliably if it is not root node. */ - keep_this_irm = irm_device && irm_device->config_rom && - irm_device->config_rom[3] >> 8 == CANON_OUI; + irm_node_id = irm_node->node_id; - root_id = root_node->node_id; - irm_id = card->irm_node->node_id; - local_id = card->local_node->node_id; + spin_unlock_irq(&card->lock); - grace = time_after64(get_jiffies_64(), - card->reset_jiffies + DIV_ROUND_UP(HZ, 8)); + rcode = fw_run_transaction(card, TCODE_LOCK_COMPARE_SWAP, irm_node_id, generation, + SCODE_100, CSR_REGISTER_BASE + CSR_BUS_MANAGER_ID, data, + sizeof(data)); - if ((is_next_generation(generation, card->bm_generation) && - !card->bm_abdicate) || - (card->bm_generation != generation && grace)) { - /* - * This first step is to figure out who is IRM and - * then try to become bus manager. If the IRM is not - * well defined (e.g. does not have an active link - * layer or does not responds to our lock request, we - * will have to do a little vigilante bus management. - * In that case, we do a goto into the gap count logic - * so that when we do the reset, we still optimize the - * gap count. That could well save a reset in the - * next generation. - */ + spin_lock_irq(&card->lock); - if (!card->irm_node->link_on) { - new_root_id = local_id; - fw_notice(card, "%s, making local node (%02x) root\n", - "IRM has link off", new_root_id); - goto pick_me; + switch (rcode) { + case RCODE_GENERATION: + return BM_CONTENTION_OUTCOME_AT_NEW_GENERATION; + case RCODE_SEND_ERROR: + return BM_CONTENTION_OUTCOME_LOCAL_PROBLEM_AT_TRANSACTION; + case RCODE_COMPLETE: + { + int bm_id = be32_to_cpu(data[0]); + + // Used by cdev layer for "struct fw_cdev_event_bus_reset". + if (bm_id != BUS_MANAGER_ID_NOT_REGISTERED) + card->bm_node_id = 0xffc0 & bm_id; + else + card->bm_node_id = local_id; + + if (bm_id != BUS_MANAGER_ID_NOT_REGISTERED) + return BM_CONTENTION_OUTCOME_IRM_HOLDS_ANOTHER_NODE_AS_BM; + else + return BM_CONTENTION_OUTCOME_IRM_HOLDS_LOCAL_NODE_AS_BM; + } + default: + if (!(irm_device_quirks & FW_DEVICE_QUIRK_IRM_IGNORES_BUS_MANAGER)) { + fw_notice(card, "BM lock failed (%s), making local node (%02x) root\n", + fw_rcode_string(rcode), local_id); + return BM_CONTENTION_OUTCOME_IRM_COMPLIES_1394_1995_ONLY; + } else { + return BM_CONTENTION_OUTCOME_IRM_IS_NOT_CAPABLE_FOR_IRM; } + } +} - if (irm_is_1394_1995_only && !keep_this_irm) { - new_root_id = local_id; - fw_notice(card, "%s, making local node (%02x) root\n", - "IRM is not 1394a compliant", new_root_id); - goto pick_me; - } +DEFINE_FREE(node_unref, struct fw_node *, if (_T) fw_node_put(_T)) +DEFINE_FREE(card_unref, struct fw_card *, if (_T) fw_card_put(_T)) - transaction_data[0] = cpu_to_be32(0x3f); - transaction_data[1] = cpu_to_be32(local_id); +static void bm_work(struct work_struct *work) +{ + static const char gap_count_table[] = { + 63, 5, 7, 8, 10, 13, 16, 18, 21, 24, 26, 29, 32, 35, 37, 40 + }; + struct fw_card *card __free(card_unref) = from_work(card, work, bm_work.work); + struct fw_node *root_node __free(node_unref) = NULL; + int root_id, new_root_id, irm_id, local_id; + int expected_gap_count, generation; + bool stand_for_root = false; + spin_lock_irq(&card->lock); + + if (card->local_node == NULL) { spin_unlock_irq(&card->lock); + return; + } - rcode = fw_run_transaction(card, TCODE_LOCK_COMPARE_SWAP, - irm_id, generation, SCODE_100, - CSR_REGISTER_BASE + CSR_BUS_MANAGER_ID, - transaction_data, 8); + generation = card->generation; - if (rcode == RCODE_GENERATION) - /* Another bus reset, BM work has been rescheduled. */ - goto out; + root_node = fw_node_get(card->root_node); - bm_id = be32_to_cpu(transaction_data[0]); + root_id = root_node->node_id; + irm_id = card->irm_node->node_id; + local_id = card->local_node->node_id; - scoped_guard(spinlock_irq, &card->lock) { - if (rcode == RCODE_COMPLETE && generation == card->generation) - card->bm_node_id = - bm_id == 0x3f ? local_id : 0xffc0 | bm_id; - } + if (card->bm_generation != generation) { + enum bm_contention_outcome result = contend_for_bm(card); - if (rcode == RCODE_COMPLETE && bm_id != 0x3f) { - /* Somebody else is BM. Only act as IRM. */ - if (local_id == irm_id) + switch (result) { + case BM_CONTENTION_OUTCOME_WITHIN_WINDOW: + spin_unlock_irq(&card->lock); + fw_schedule_bm_work(card, msecs_to_jiffies(125)); + return; + case BM_CONTENTION_OUTCOME_IRM_HAS_LINK_OFF: + stand_for_root = true; + break; + case BM_CONTENTION_OUTCOME_IRM_COMPLIES_1394_1995_ONLY: + stand_for_root = true; + break; + case BM_CONTENTION_OUTCOME_AT_NEW_GENERATION: + // BM work has been rescheduled. + spin_unlock_irq(&card->lock); + return; + case BM_CONTENTION_OUTCOME_LOCAL_PROBLEM_AT_TRANSACTION: + // Let's try again later and hope that the local problem has gone away by + // then. + spin_unlock_irq(&card->lock); + fw_schedule_bm_work(card, msecs_to_jiffies(125)); + return; + case BM_CONTENTION_OUTCOME_IRM_IS_NOT_CAPABLE_FOR_IRM: + // Let's do a bus reset and pick the local node as root, and thus, IRM. + stand_for_root = true; + break; + case BM_CONTENTION_OUTCOME_IRM_HOLDS_ANOTHER_NODE_AS_BM: + if (local_id == irm_id) { + // Only acts as IRM. + spin_unlock_irq(&card->lock); allocate_broadcast_channel(card, generation); - - goto out; + spin_lock_irq(&card->lock); + } + fallthrough; + case BM_CONTENTION_OUTCOME_IRM_HOLDS_LOCAL_NODE_AS_BM: + default: + card->bm_generation = generation; + break; } + } - if (rcode == RCODE_SEND_ERROR) { - /* - * We have been unable to send the lock request due to - * some local problem. Let's try again later and hope - * that the problem has gone away by then. - */ - fw_schedule_bm_work(card, DIV_ROUND_UP(HZ, 8)); - goto out; - } - - spin_lock_irq(&card->lock); - - if (rcode != RCODE_COMPLETE && !keep_this_irm) { - /* - * The lock request failed, maybe the IRM - * isn't really IRM capable after all. Let's - * do a bus reset and pick the local node as - * root, and thus, IRM. - */ - new_root_id = local_id; - fw_notice(card, "BM lock failed (%s), making local node (%02x) root\n", - fw_rcode_string(rcode), new_root_id); - goto pick_me; + // We're bus manager for this generation, so next step is to make sure we have an active + // cycle master and do gap count optimization. + if (!stand_for_root) { + if (card->gap_count == GAP_COUNT_MISMATCHED) { + // If self IDs have inconsistent gap counts, do a + // bus reset ASAP. The config rom read might never + // complete, so don't wait for it. However, still + // send a PHY configuration packet prior to the + // bus reset. The PHY configuration packet might + // fail, but 1394-2008 8.4.5.2 explicitly permits + // it in this case, so it should be safe to try. + stand_for_root = true; + + // We must always send a bus reset if the gap count + // is inconsistent, so bypass the 5-reset limit. + card->bm_retries = 0; + } else { + // Now investigate root node. + struct fw_device *root_device = fw_node_get_device(root_node); + + if (root_device == NULL) { + // Either link_on is false, or we failed to read the + // config rom. In either case, pick another root. + stand_for_root = true; + } else { + bool root_device_is_running = + atomic_read(&root_device->state) == FW_DEVICE_RUNNING; + + if (!root_device_is_running) { + // If we haven't probed this device yet, bail out now + // and let's try again once that's done. + spin_unlock_irq(&card->lock); + return; + } else if (!root_device->cmc) { + // Current root has an active link layer and we + // successfully read the config rom, but it's not + // cycle master capable. + stand_for_root = true; + } + } } - } else if (card->bm_generation != generation) { - /* - * We weren't BM in the last generation, and the last - * bus reset is less than 125ms ago. Reschedule this job. - */ - spin_unlock_irq(&card->lock); - fw_schedule_bm_work(card, DIV_ROUND_UP(HZ, 8)); - goto out; } - /* - * We're bus manager for this generation, so next step is to - * make sure we have an active cycle master and do gap count - * optimization. - */ - card->bm_generation = generation; - - if (card->gap_count == 0) { - /* - * If self IDs have inconsistent gap counts, do a - * bus reset ASAP. The config rom read might never - * complete, so don't wait for it. However, still - * send a PHY configuration packet prior to the - * bus reset. The PHY configuration packet might - * fail, but 1394-2008 8.4.5.2 explicitly permits - * it in this case, so it should be safe to try. - */ + if (stand_for_root) { new_root_id = local_id; - /* - * We must always send a bus reset if the gap count - * is inconsistent, so bypass the 5-reset limit. - */ - card->bm_retries = 0; - } else if (root_device == NULL) { - /* - * Either link_on is false, or we failed to read the - * config rom. In either case, pick another root. - */ - new_root_id = local_id; - } else if (!root_device_is_running) { - /* - * If we haven't probed this device yet, bail out now - * and let's try again once that's done. - */ - spin_unlock_irq(&card->lock); - goto out; - } else if (root_device_is_cmc) { - /* - * We will send out a force root packet for this - * node as part of the gap count optimization. - */ - new_root_id = root_id; } else { - /* - * Current root has an active link layer and we - * successfully read the config rom, but it's not - * cycle master capable. - */ - new_root_id = local_id; + // We will send out a force root packet for this node as part of the gap count + // optimization on behalf of the node. + new_root_id = root_id; } - pick_me: /* * Pick a gap count from 1394a table E-1. The table doesn't cover * the typically much larger 1394b beta repeater delays though. */ if (!card->beta_repeaters_present && root_node->max_hops < ARRAY_SIZE(gap_count_table)) - gap_count = gap_count_table[root_node->max_hops]; + expected_gap_count = gap_count_table[root_node->max_hops]; else - gap_count = 63; + expected_gap_count = 63; - /* - * Finally, figure out if we should do a reset or not. If we have - * done less than 5 resets with the same physical topology and we - * have either a new root or a new gap count setting, let's do it. - */ - - if (card->bm_retries++ < 5 && - (card->gap_count != gap_count || new_root_id != root_id)) - do_reset = true; + // Finally, figure out if we should do a reset or not. If we have done less than 5 resets + // with the same physical topology and we have either a new root or a new gap count + // setting, let's do it. + if (card->bm_retries++ < 5 && (card->gap_count != expected_gap_count || new_root_id != root_id)) { + int card_gap_count = card->gap_count; - spin_unlock_irq(&card->lock); + spin_unlock_irq(&card->lock); - if (do_reset) { fw_notice(card, "phy config: new root=%x, gap_count=%d\n", - new_root_id, gap_count); - fw_send_phy_config(card, new_root_id, generation, gap_count); + new_root_id, expected_gap_count); + fw_send_phy_config(card, new_root_id, generation, expected_gap_count); /* * Where possible, use a short bus reset to minimize * disruption to isochronous transfers. But in the event @@ -512,31 +532,27 @@ static void bm_work(struct work_struct *work) * may treat it as two, causing a gap count inconsistency * again. Using a long bus reset prevents this. */ - reset_bus(card, card->gap_count != 0); + reset_bus(card, card_gap_count != 0); /* Will allocate broadcast channel after the reset. */ - goto out; - } + } else { + struct fw_device *root_device = fw_node_get_device(root_node); - if (root_device_is_cmc) { - /* - * Make sure that the cycle master sends cycle start packets. - */ - transaction_data[0] = cpu_to_be32(CSR_STATE_BIT_CMSTR); - rcode = fw_run_transaction(card, TCODE_WRITE_QUADLET_REQUEST, - root_id, generation, SCODE_100, - CSR_REGISTER_BASE + CSR_STATE_SET, - transaction_data, 4); - if (rcode == RCODE_GENERATION) - goto out; - } + spin_unlock_irq(&card->lock); - if (local_id == irm_id) - allocate_broadcast_channel(card, generation); + if (root_device && root_device->cmc) { + // Make sure that the cycle master sends cycle start packets. + __be32 data = cpu_to_be32(CSR_STATE_BIT_CMSTR); + int rcode = fw_run_transaction(card, TCODE_WRITE_QUADLET_REQUEST, + root_id, generation, SCODE_100, + CSR_REGISTER_BASE + CSR_STATE_SET, + &data, sizeof(data)); + if (rcode == RCODE_GENERATION) + return; + } - out: - fw_node_put(root_node); - out_put_card: - fw_card_put(card); + if (local_id == irm_id) + allocate_broadcast_channel(card, generation); + } } void fw_card_initialize(struct fw_card *card, @@ -548,20 +564,26 @@ void fw_card_initialize(struct fw_card *card, card->index = atomic_inc_return(&index); card->driver = driver; card->device = device; - card->current_tlabel = 0; - card->tlabel_mask = 0; - card->split_timeout_hi = DEFAULT_SPLIT_TIMEOUT / 8000; - card->split_timeout_lo = (DEFAULT_SPLIT_TIMEOUT % 8000) << 19; - card->split_timeout_cycles = DEFAULT_SPLIT_TIMEOUT; - card->split_timeout_jiffies = - DIV_ROUND_UP(DEFAULT_SPLIT_TIMEOUT * HZ, 8000); + + card->transactions.current_tlabel = 0; + card->transactions.tlabel_mask = 0; + INIT_LIST_HEAD(&card->transactions.list); + spin_lock_init(&card->transactions.lock); + + spin_lock_init(&card->topology_map.lock); + + card->split_timeout.hi = DEFAULT_SPLIT_TIMEOUT / 8000; + card->split_timeout.lo = (DEFAULT_SPLIT_TIMEOUT % 8000) << 19; + card->split_timeout.cycles = DEFAULT_SPLIT_TIMEOUT; + card->split_timeout.jiffies = isoc_cycles_to_jiffies(DEFAULT_SPLIT_TIMEOUT); + spin_lock_init(&card->split_timeout.lock); + card->color = 0; card->broadcast_channel = BROADCAST_CHANNEL_INITIAL; kref_init(&card->kref); init_completion(&card->done); - INIT_LIST_HEAD(&card->transaction_list); - INIT_LIST_HEAD(&card->phy_receiver_list); + spin_lock_init(&card->lock); card->local_node = NULL; @@ -571,10 +593,13 @@ void fw_card_initialize(struct fw_card *card, } EXPORT_SYMBOL(fw_card_initialize); +DEFINE_FREE(workqueue_destroy, struct workqueue_struct *, if (_T) destroy_workqueue(_T)) + int fw_card_add(struct fw_card *card, u32 max_receive, u32 link_speed, u64 guid, unsigned int supported_isoc_contexts) { - struct workqueue_struct *isoc_wq; + struct workqueue_struct *isoc_wq __free(workqueue_destroy) = NULL; + struct workqueue_struct *async_wq __free(workqueue_destroy) = NULL; int ret; // This workqueue should be: @@ -595,22 +620,41 @@ int fw_card_add(struct fw_card *card, u32 max_receive, u32 link_speed, u64 guid, if (!isoc_wq) return -ENOMEM; + // This workqueue should be: + // * != WQ_BH Sleepable. + // * == WQ_UNBOUND Any core can process data for asynchronous context. + // * == WQ_MEM_RECLAIM Used for any backend of block device. + // * == WQ_FREEZABLE The target device would not be available when being freezed. + // * == WQ_HIGHPRI High priority to process semi-realtime timestamped data. + // * == WQ_SYSFS Parameters are available via sysfs. + // * max_active == 4 A hardIRQ could notify events for a pair of requests and + // response AR/AT contexts. + async_wq = alloc_workqueue("firewire-async-card%u", + WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_FREEZABLE | WQ_HIGHPRI | WQ_SYSFS, + 4, card->index); + if (!async_wq) + return -ENOMEM; + + card->isoc_wq = isoc_wq; + card->async_wq = async_wq; card->max_receive = max_receive; card->link_speed = link_speed; card->guid = guid; - guard(mutex)(&card_mutex); + scoped_guard(mutex, &card_mutex) { + generate_config_rom(card, tmp_config_rom); + ret = card->driver->enable(card, tmp_config_rom, config_rom_length); + if (ret < 0) { + card->isoc_wq = NULL; + card->async_wq = NULL; + return ret; + } + retain_and_null_ptr(isoc_wq); + retain_and_null_ptr(async_wq); - generate_config_rom(card, tmp_config_rom); - ret = card->driver->enable(card, tmp_config_rom, config_rom_length); - if (ret < 0) { - destroy_workqueue(isoc_wq); - return ret; + list_add_tail(&card->link, &card_list); } - card->isoc_wq = isoc_wq; - list_add_tail(&card->link, &card_list); - return 0; } EXPORT_SYMBOL(fw_card_add); @@ -660,8 +704,8 @@ static int dummy_enable_phys_dma(struct fw_card *card, return -ENODEV; } -static struct fw_iso_context *dummy_allocate_iso_context(struct fw_card *card, - int type, int channel, size_t header_size) +static struct fw_iso_context *dummy_allocate_iso_context(struct fw_card *card, int type, + int channel, size_t header_size, size_t header_storage_size) { return ERR_PTR(-ENODEV); } @@ -742,8 +786,13 @@ void fw_core_remove_card(struct fw_card *card) /* Switch off most of the card driver interface. */ dummy_driver.free_iso_context = card->driver->free_iso_context; dummy_driver.stop_iso = card->driver->stop_iso; + dummy_driver.disable = card->driver->disable; card->driver = &dummy_driver; + drain_workqueue(card->isoc_wq); + drain_workqueue(card->async_wq); + card->driver->disable(card); + fw_cancel_pending_transactions(card); scoped_guard(spinlock_irqsave, &card->lock) fw_destroy_nodes(card); @@ -753,8 +802,9 @@ void fw_core_remove_card(struct fw_card *card) wait_for_completion(&card->done); destroy_workqueue(card->isoc_wq); + destroy_workqueue(card->async_wq); - WARN_ON(!list_empty(&card->transaction_list)); + WARN_ON(!list_empty(&card->transactions.list)); } EXPORT_SYMBOL(fw_core_remove_card); diff --git a/drivers/firewire/core-cdev.c b/drivers/firewire/core-cdev.c index b360dca2c69e..f791db4c8dff 100644 --- a/drivers/firewire/core-cdev.c +++ b/drivers/firewire/core-cdev.c @@ -41,12 +41,15 @@ /* * ABI version history is documented in linux/firewire-cdev.h. */ -#define FW_CDEV_KERNEL_VERSION 5 +#define FW_CDEV_KERNEL_VERSION 6 #define FW_CDEV_VERSION_EVENT_REQUEST2 4 #define FW_CDEV_VERSION_ALLOCATE_REGION_END 4 #define FW_CDEV_VERSION_AUTO_FLUSH_ISO_OVERFLOW 5 #define FW_CDEV_VERSION_EVENT_ASYNC_TSTAMP 6 +static DEFINE_SPINLOCK(phy_receiver_list_lock); +static LIST_HEAD(phy_receiver_list); + struct client { u32 version; struct fw_device *device; @@ -60,10 +63,10 @@ struct client { u64 bus_reset_closure; struct fw_iso_context *iso_context; + struct mutex iso_context_mutex; u64 iso_closure; struct fw_iso_buffer buffer; unsigned long vm_start; - bool buffer_is_mapped; struct list_head phy_receiver_link; u64 phy_receiver_closure; @@ -288,7 +291,7 @@ static int fw_device_op_open(struct inode *inode, struct file *file) return -ENODEV; } - client = kzalloc(sizeof(*client), GFP_KERNEL); + client = kzalloc_obj(*client); if (client == NULL) { fw_device_put(device); return -ENOMEM; @@ -303,6 +306,7 @@ static int fw_device_op_open(struct inode *inode, struct file *file) INIT_LIST_HEAD(&client->phy_receiver_link); INIT_LIST_HEAD(&client->link); kref_init(&client->kref); + mutex_init(&client->iso_context_mutex); file->private_data = client; @@ -408,7 +412,7 @@ static void queue_bus_reset_event(struct client *client) struct client_resource *resource; unsigned long index; - e = kzalloc(sizeof(*e), GFP_KERNEL); + e = kzalloc_obj(*e); if (e == NULL) return; @@ -741,8 +745,8 @@ static void handle_request(struct fw_card *card, struct fw_request *request, if (is_fcp) fw_request_get(request); - r = kmalloc(sizeof(*r), GFP_ATOMIC); - e = kmalloc(sizeof(*e), GFP_ATOMIC); + r = kmalloc_obj(*r, GFP_ATOMIC); + e = kmalloc_obj(*e, GFP_ATOMIC); if (r == NULL || e == NULL) goto failed; @@ -833,7 +837,7 @@ static int ioctl_allocate(struct client *client, union ioctl_arg *arg) struct fw_address_region region; int ret; - r = kmalloc(sizeof(*r), GFP_KERNEL); + r = kmalloc_obj(*r); if (r == NULL) return -ENOMEM; @@ -937,11 +941,12 @@ static int ioctl_add_descriptor(struct client *client, union ioctl_arg *arg) if (a->length > 256) return -EINVAL; - r = kmalloc(sizeof(*r) + a->length * 4, GFP_KERNEL); + r = kmalloc_flex(*r, data, a->length); if (r == NULL) return -ENOMEM; - if (copy_from_user(r->data, u64_to_uptr(a->data), a->length * 4)) { + if (copy_from_user(r->data, u64_to_uptr(a->data), + flex_array_size(r, data, a->length))) { ret = -EFAULT; goto failed; } @@ -1001,7 +1006,7 @@ static void iso_mc_callback(struct fw_iso_context *context, struct client *client = data; struct iso_interrupt_mc_event *e; - e = kmalloc(sizeof(*e), GFP_KERNEL); + e = kmalloc_obj(*e); if (e == NULL) return; @@ -1021,25 +1026,10 @@ static enum dma_data_direction iso_dma_direction(struct fw_iso_context *context) return DMA_FROM_DEVICE; } -static struct fw_iso_context *fw_iso_mc_context_create(struct fw_card *card, - fw_iso_mc_callback_t callback, - void *callback_data) -{ - struct fw_iso_context *ctx; - - ctx = fw_iso_context_create(card, FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL, - 0, 0, 0, NULL, callback_data); - if (!IS_ERR(ctx)) - ctx->callback.mc = callback; - - return ctx; -} - static int ioctl_create_iso_context(struct client *client, union ioctl_arg *arg) { struct fw_cdev_create_iso_context *a = &arg->create_iso_context; struct fw_iso_context *context; - union fw_iso_callback cb; int ret; BUILD_BUG_ON(FW_CDEV_ISO_CONTEXT_TRANSMIT != FW_ISO_CONTEXT_TRANSMIT || @@ -1051,20 +1041,15 @@ static int ioctl_create_iso_context(struct client *client, union ioctl_arg *arg) case FW_ISO_CONTEXT_TRANSMIT: if (a->speed > SCODE_3200 || a->channel > 63) return -EINVAL; - - cb.sc = iso_callback; break; case FW_ISO_CONTEXT_RECEIVE: if (a->header_size < 4 || (a->header_size & 3) || a->channel > 63) return -EINVAL; - - cb.sc = iso_callback; break; case FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL: - cb.mc = iso_mc_callback; break; default: @@ -1072,38 +1057,36 @@ static int ioctl_create_iso_context(struct client *client, union ioctl_arg *arg) } if (a->type == FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL) - context = fw_iso_mc_context_create(client->device->card, cb.mc, - client); + context = fw_iso_mc_context_create(client->device->card, iso_mc_callback, client); else - context = fw_iso_context_create(client->device->card, a->type, - a->channel, a->speed, - a->header_size, cb.sc, client); + context = fw_iso_context_create(client->device->card, a->type, a->channel, a->speed, + a->header_size, iso_callback, client); if (IS_ERR(context)) return PTR_ERR(context); if (client->version < FW_CDEV_VERSION_AUTO_FLUSH_ISO_OVERFLOW) - context->drop_overflow_headers = true; + context->flags |= FW_ISO_CONTEXT_FLAG_DROP_OVERFLOW_HEADERS; // We only support one context at this time. - guard(spinlock_irq)(&client->lock); - - if (client->iso_context != NULL) { - fw_iso_context_destroy(context); - - return -EBUSY; - } - if (!client->buffer_is_mapped) { - ret = fw_iso_buffer_map_dma(&client->buffer, - client->device->card, - iso_dma_direction(context)); - if (ret < 0) { + scoped_guard(mutex, &client->iso_context_mutex) { + if (client->iso_context != NULL) { fw_iso_context_destroy(context); - return ret; + return -EBUSY; + } + // The DMA mapping operation is available if the buffer is already allocated by + // mmap(2) system call. If not, it is delegated to the system call. + if (client->buffer.pages && !client->buffer.dma_addrs) { + ret = fw_iso_buffer_map_dma(&client->buffer, client->device->card, + iso_dma_direction(context)); + if (ret < 0) { + fw_iso_context_destroy(context); + + return ret; + } } - client->buffer_is_mapped = true; + client->iso_closure = a->closure; + client->iso_context = context; } - client->iso_closure = a->closure; - client->iso_context = context; a->handle = 0; @@ -1137,10 +1120,7 @@ static int ioctl_queue_iso(struct client *client, union ioctl_arg *arg) unsigned long payload, buffer_end, transmit_header_bytes = 0; u32 control; int count; - struct { - struct fw_iso_packet packet; - u8 header[256]; - } u; + DEFINE_RAW_FLEX(struct fw_iso_packet, u, header, 64); if (ctx == NULL || a->handle != 0) return -EINVAL; @@ -1172,29 +1152,29 @@ static int ioctl_queue_iso(struct client *client, union ioctl_arg *arg) while (p < end) { if (get_user(control, &p->control)) return -EFAULT; - u.packet.payload_length = GET_PAYLOAD_LENGTH(control); - u.packet.interrupt = GET_INTERRUPT(control); - u.packet.skip = GET_SKIP(control); - u.packet.tag = GET_TAG(control); - u.packet.sy = GET_SY(control); - u.packet.header_length = GET_HEADER_LENGTH(control); + u->payload_length = GET_PAYLOAD_LENGTH(control); + u->interrupt = GET_INTERRUPT(control); + u->skip = GET_SKIP(control); + u->tag = GET_TAG(control); + u->sy = GET_SY(control); + u->header_length = GET_HEADER_LENGTH(control); switch (ctx->type) { case FW_ISO_CONTEXT_TRANSMIT: - if (u.packet.header_length & 3) + if (u->header_length & 3) return -EINVAL; - transmit_header_bytes = u.packet.header_length; + transmit_header_bytes = u->header_length; break; case FW_ISO_CONTEXT_RECEIVE: - if (u.packet.header_length == 0 || - u.packet.header_length % ctx->header_size != 0) + if (u->header_length == 0 || + u->header_length % ctx->header_size != 0) return -EINVAL; break; case FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL: - if (u.packet.payload_length == 0 || - u.packet.payload_length & 3) + if (u->payload_length == 0 || + u->payload_length & 3) return -EINVAL; break; } @@ -1204,20 +1184,19 @@ static int ioctl_queue_iso(struct client *client, union ioctl_arg *arg) if (next > end) return -EINVAL; if (copy_from_user - (u.packet.header, p->header, transmit_header_bytes)) + (u->header, p->header, transmit_header_bytes)) return -EFAULT; - if (u.packet.skip && ctx->type == FW_ISO_CONTEXT_TRANSMIT && - u.packet.header_length + u.packet.payload_length > 0) + if (u->skip && ctx->type == FW_ISO_CONTEXT_TRANSMIT && + u->header_length + u->payload_length > 0) return -EINVAL; - if (payload + u.packet.payload_length > buffer_end) + if (payload + u->payload_length > buffer_end) return -EINVAL; - if (fw_iso_context_queue(ctx, &u.packet, - &client->buffer, payload)) + if (fw_iso_context_queue(ctx, u, &client->buffer, payload)) break; p = next; - payload += u.packet.payload_length; + payload += u->payload_length; count++; } fw_iso_context_queue_flush(ctx); @@ -1317,8 +1296,7 @@ static int ioctl_get_cycle_timer(struct client *client, union ioctl_arg *arg) static void iso_resource_work(struct work_struct *work) { struct iso_resource_event *e; - struct iso_resource *r = - container_of(work, struct iso_resource, work.work); + struct iso_resource *r = from_work(r, work, work.work); struct client *client = r->client; unsigned long index = r->resource.handle; int generation, channel, bandwidth, todo; @@ -1329,8 +1307,8 @@ static void iso_resource_work(struct work_struct *work) todo = r->todo; // Allow 1000ms grace period for other reallocations. if (todo == ISO_RES_ALLOC && - time_before64(get_jiffies_64(), client->device->card->reset_jiffies + HZ)) { - schedule_iso_resource(r, DIV_ROUND_UP(HZ, 3)); + time_is_after_jiffies64(client->device->card->reset_jiffies + secs_to_jiffies(1))) { + schedule_iso_resource(r, msecs_to_jiffies(333)); skip = true; } else { // We could be called twice within the same generation. @@ -1431,9 +1409,9 @@ static int init_iso_resource(struct client *client, request->bandwidth > BANDWIDTH_AVAILABLE_INITIAL) return -EINVAL; - r = kmalloc(sizeof(*r), GFP_KERNEL); - e1 = kmalloc(sizeof(*e1), GFP_KERNEL); - e2 = kmalloc(sizeof(*e2), GFP_KERNEL); + r = kmalloc_obj(*r); + e1 = kmalloc_obj(*e1); + e2 = kmalloc_obj(*e2); if (r == NULL || e1 == NULL || e2 == NULL) { ret = -ENOMEM; goto fail; @@ -1674,15 +1652,16 @@ static int ioctl_send_phy_packet(struct client *client, union ioctl_arg *arg) static int ioctl_receive_phy_packets(struct client *client, union ioctl_arg *arg) { struct fw_cdev_receive_phy_packets *a = &arg->receive_phy_packets; - struct fw_card *card = client->device->card; /* Access policy: Allow this ioctl only on local nodes' device files. */ if (!client->device->is_local) return -ENOSYS; - guard(spinlock_irq)(&card->lock); + // NOTE: This can be without irq when we can guarantee that __fw_send_request() for local + // destination never runs in any type of IRQ context. + scoped_guard(spinlock_irq, &phy_receiver_list_lock) + list_move_tail(&client->phy_receiver_link, &phy_receiver_list); - list_move_tail(&client->phy_receiver_link, &card->phy_receiver_list); client->phy_receiver_closure = a->closure; return 0; @@ -1692,10 +1671,17 @@ void fw_cdev_handle_phy_packet(struct fw_card *card, struct fw_packet *p) { struct client *client; - guard(spinlock_irqsave)(&card->lock); + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for local + // destination never runs in any type of IRQ context. + guard(spinlock_irqsave)(&phy_receiver_list_lock); + + list_for_each_entry(client, &phy_receiver_list, phy_receiver_link) { + struct inbound_phy_packet_event *e; - list_for_each_entry(client, &card->phy_receiver_list, phy_receiver_link) { - struct inbound_phy_packet_event *e = kmalloc(sizeof(*e) + 8, GFP_ATOMIC); + if (client->device->card != card) + continue; + + e = kmalloc(sizeof(*e) + 8, GFP_ATOMIC); if (e == NULL) break; @@ -1819,13 +1805,14 @@ static int fw_device_op_mmap(struct file *file, struct vm_area_struct *vma) if (ret < 0) return ret; - scoped_guard(spinlock_irq, &client->lock) { + scoped_guard(mutex, &client->iso_context_mutex) { + // The direction of DMA can be determined if the isochronous context is already + // allocated. If not, the DMA mapping operation is postponed after the allocation. if (client->iso_context) { ret = fw_iso_buffer_map_dma(&client->buffer, client->device->card, iso_dma_direction(client->iso_context)); if (ret < 0) goto fail; - client->buffer_is_mapped = true; } } @@ -1862,7 +1849,9 @@ static int fw_device_op_release(struct inode *inode, struct file *file) struct client_resource *resource; unsigned long index; - scoped_guard(spinlock_irq, &client->device->card->lock) + // NOTE: This can be without irq when we can guarantee that __fw_send_request() for local + // destination never runs in any type of IRQ context. + scoped_guard(spinlock_irq, &phy_receiver_list_lock) list_del(&client->phy_receiver_link); scoped_guard(mutex, &client->device->client_list_mutex) @@ -1870,6 +1859,7 @@ static int fw_device_op_release(struct inode *inode, struct file *file) if (client->iso_context) fw_iso_context_destroy(client->iso_context); + mutex_destroy(&client->iso_context_mutex); if (client->buffer.pages) fw_iso_buffer_destroy(&client->buffer, client->device->card); diff --git a/drivers/firewire/core-device.c b/drivers/firewire/core-device.c index a99fe35f1f0d..c0f17da27a22 100644 --- a/drivers/firewire/core-device.c +++ b/drivers/firewire/core-device.c @@ -542,8 +542,83 @@ static struct device_attribute fw_device_attributes[] = { __ATTR_NULL, }; -static int read_rom(struct fw_device *device, - int generation, int index, u32 *data) +#define CANON_OUI 0x000085 + +static int detect_quirks_by_bus_information_block(const u32 *bus_information_block) +{ + int quirks = 0; + + if ((bus_information_block[2] & 0x000000f0) == 0) + quirks |= FW_DEVICE_QUIRK_IRM_IS_1394_1995_ONLY; + + if ((bus_information_block[3] >> 8) == CANON_OUI) + quirks |= FW_DEVICE_QUIRK_IRM_IGNORES_BUS_MANAGER; + + return quirks; +} + +struct entry_match { + unsigned int index; + u32 value; +}; + +static const struct entry_match motu_audio_express_matches[] = { + { 1, 0x030001f2 }, + { 3, 0xd1000002 }, + { 4, 0x8d000005 }, + { 6, 0x120001f2 }, + { 7, 0x13000033 }, + { 8, 0x17104800 }, +}; + +static const struct entry_match tascam_fw_series_matches[] = { + { 1, 0x0300022e }, + { 3, 0x8d000006 }, + { 4, 0xd1000001 }, + { 6, 0x1200022e }, + { 8, 0xd4000004 }, +}; + +static int detect_quirks_by_root_directory(const u32 *root_directory, unsigned int length) +{ + static const struct { + enum fw_device_quirk quirk; + const struct entry_match *matches; + unsigned int match_count; + } *entry, entries[] = { + { + .quirk = FW_DEVICE_QUIRK_ACK_PACKET_WITH_INVALID_PENDING_CODE, + .matches = motu_audio_express_matches, + .match_count = ARRAY_SIZE(motu_audio_express_matches), + }, + { + .quirk = FW_DEVICE_QUIRK_UNSTABLE_AT_S400, + .matches = tascam_fw_series_matches, + .match_count = ARRAY_SIZE(tascam_fw_series_matches), + }, + }; + int quirks = 0; + int i; + + for (i = 0; i < ARRAY_SIZE(entries); ++i) { + int j; + + entry = entries + i; + for (j = 0; j < entry->match_count; ++j) { + unsigned int index = entry->matches[j].index; + unsigned int value = entry->matches[j].value; + + if ((length < index) || (root_directory[index] != value)) + break; + } + if (j == entry->match_count) + quirks |= entry->quirk; + } + + return quirks; +} + +static int read_rom(struct fw_device *device, int generation, int speed, int index, u32 *data) { u64 offset = (CSR_REGISTER_BASE | CSR_CONFIG_ROM) + index * 4; int i, rcode; @@ -554,7 +629,7 @@ static int read_rom(struct fw_device *device, for (i = 10; i < 100; i += 10) { rcode = fw_run_transaction(device->card, TCODE_READ_QUADLET_REQUEST, device->node_id, - generation, device->max_speed, offset, data, 4); + generation, speed, offset, data, 4); if (rcode != RCODE_BUSY) break; msleep(i); @@ -578,10 +653,11 @@ static int read_rom(struct fw_device *device, static int read_config_rom(struct fw_device *device, int generation) { struct fw_card *card = device->card; - const u32 *old_rom, *new_rom; - u32 *rom, *stack; + const u32 *new_rom, *old_rom __free(kfree) = NULL; + u32 *stack, *rom __free(kfree) = NULL; u32 sp, key; - int i, end, length, ret; + int i, end, length, ret, speed; + int quirks; rom = kmalloc(sizeof(*rom) * MAX_CONFIG_ROM_SIZE + sizeof(*stack) * MAX_CONFIG_ROM_SIZE, GFP_KERNEL); @@ -591,13 +667,13 @@ static int read_config_rom(struct fw_device *device, int generation) stack = &rom[MAX_CONFIG_ROM_SIZE]; memset(rom, 0, sizeof(*rom) * MAX_CONFIG_ROM_SIZE); - device->max_speed = SCODE_100; + speed = SCODE_100; /* First read the bus info block. */ for (i = 0; i < 5; i++) { - ret = read_rom(device, generation, i, &rom[i]); + ret = read_rom(device, generation, speed, i, &rom[i]); if (ret != RCODE_COMPLETE) - goto out; + return ret; /* * As per IEEE1212 7.2, during initialization, devices can * reply with a 0 for the first quadlet of the config @@ -606,39 +682,14 @@ static int read_config_rom(struct fw_device *device, int generation) * harddisk). In that case we just fail, and the * retry mechanism will try again later. */ - if (i == 0 && rom[i] == 0) { - ret = RCODE_BUSY; - goto out; - } + if (i == 0 && rom[i] == 0) + return RCODE_BUSY; } - device->max_speed = device->node->max_speed; - - /* - * Determine the speed of - * - devices with link speed less than PHY speed, - * - devices with 1394b PHY (unless only connected to 1394a PHYs), - * - all devices if there are 1394b repeaters. - * Note, we cannot use the bus info block's link_spd as starting point - * because some buggy firmwares set it lower than necessary and because - * 1394-1995 nodes do not have the field. - */ - if ((rom[2] & 0x7) < device->max_speed || - device->max_speed == SCODE_BETA || - card->beta_repeaters_present) { - u32 dummy; - - /* for S1600 and S3200 */ - if (device->max_speed == SCODE_BETA) - device->max_speed = card->link_speed; + quirks = detect_quirks_by_bus_information_block(rom); - while (device->max_speed > SCODE_100) { - if (read_rom(device, generation, 0, &dummy) == - RCODE_COMPLETE) - break; - device->max_speed--; - } - } + // Just prevent from torn writing/reading. + WRITE_ONCE(device->quirks, quirks); /* * Now parse the config rom. The config rom is a recursive @@ -659,15 +710,13 @@ static int read_config_rom(struct fw_device *device, int generation) */ key = stack[--sp]; i = key & 0xffffff; - if (WARN_ON(i >= MAX_CONFIG_ROM_SIZE)) { - ret = -ENXIO; - goto out; - } + if (WARN_ON(i >= MAX_CONFIG_ROM_SIZE)) + return -ENXIO; /* Read header quadlet for the block to get the length. */ - ret = read_rom(device, generation, i, &rom[i]); + ret = read_rom(device, generation, speed, i, &rom[i]); if (ret != RCODE_COMPLETE) - goto out; + return ret; end = i + (rom[i] >> 16) + 1; if (end > MAX_CONFIG_ROM_SIZE) { /* @@ -689,9 +738,9 @@ static int read_config_rom(struct fw_device *device, int generation) * it references another block, and push it in that case. */ for (; i < end; i++) { - ret = read_rom(device, generation, i, &rom[i]); + ret = read_rom(device, generation, speed, i, &rom[i]); if (ret != RCODE_COMPLETE) - goto out; + return ret; if ((key >> 30) != 3 || (rom[i] >> 30) < 2) continue; @@ -716,27 +765,54 @@ static int read_config_rom(struct fw_device *device, int generation) length = i; } + quirks |= detect_quirks_by_root_directory(rom + ROOT_DIR_OFFSET, length - ROOT_DIR_OFFSET); + + // Just prevent from torn writing/reading. + WRITE_ONCE(device->quirks, quirks); + + if (unlikely(quirks & FW_DEVICE_QUIRK_UNSTABLE_AT_S400)) + speed = SCODE_200; + else + speed = device->node->max_speed; + + // Determine the speed of + // - devices with link speed less than PHY speed, + // - devices with 1394b PHY (unless only connected to 1394a PHYs), + // - all devices if there are 1394b repeaters. + // Note, we cannot use the bus info block's link_spd as starting point because some buggy + // firmwares set it lower than necessary and because 1394-1995 nodes do not have the field. + if ((rom[2] & 0x7) < speed || speed == SCODE_BETA || card->beta_repeaters_present) { + u32 dummy; + + // for S1600 and S3200. + if (speed == SCODE_BETA) + speed = card->link_speed; + + while (speed > SCODE_100) { + if (read_rom(device, generation, speed, 0, &dummy) == + RCODE_COMPLETE) + break; + --speed; + } + } + + device->max_speed = speed; + old_rom = device->config_rom; new_rom = kmemdup(rom, length * 4, GFP_KERNEL); - if (new_rom == NULL) { - ret = -ENOMEM; - goto out; - } + if (new_rom == NULL) + return -ENOMEM; scoped_guard(rwsem_write, &fw_device_rwsem) { device->config_rom = new_rom; device->config_rom_length = length; } - kfree(old_rom); - ret = RCODE_COMPLETE; device->max_rec = rom[2] >> 12 & 0xf; device->cmc = rom[2] >> 30 & 1; device->irmc = rom[2] >> 31 & 1; - out: - kfree(rom); - return ret; + return RCODE_COMPLETE; } static void fw_unit_release(struct device *dev) @@ -773,7 +849,7 @@ static void create_units(struct fw_device *device) * Get the address of the unit directory and try to * match the drivers id_tables against it. */ - unit = kzalloc(sizeof(*unit), GFP_KERNEL); + unit = kzalloc_obj(*unit); if (unit == NULL) continue; @@ -847,17 +923,15 @@ static void fw_schedule_device_work(struct fw_device *device, */ #define MAX_RETRIES 10 -#define RETRY_DELAY (3 * HZ) -#define INITIAL_DELAY (HZ / 2) -#define SHUTDOWN_DELAY (2 * HZ) +#define RETRY_DELAY secs_to_jiffies(3) +#define INITIAL_DELAY msecs_to_jiffies(500) +#define SHUTDOWN_DELAY secs_to_jiffies(2) static void fw_device_shutdown(struct work_struct *work) { - struct fw_device *device = - container_of(work, struct fw_device, work.work); + struct fw_device *device = from_work(device, work, work.work); - if (time_before64(get_jiffies_64(), - device->card->reset_jiffies + SHUTDOWN_DELAY) + if (time_is_after_jiffies64(device->card->reset_jiffies + SHUTDOWN_DELAY) && !list_empty(&device->card->link)) { fw_schedule_device_work(device, SHUTDOWN_DELAY); return; @@ -888,7 +962,7 @@ static void fw_device_release(struct device *dev) * bus manager work looks at this node. */ scoped_guard(spinlock_irqsave, &card->lock) - device->node->data = NULL; + fw_node_set_device(device->node, NULL); fw_node_put(device->node); kfree(device->config_rom); @@ -921,8 +995,7 @@ static int update_unit(struct device *dev, void *data) static void fw_device_update(struct work_struct *work) { - struct fw_device *device = - container_of(work, struct fw_device, work.work); + struct fw_device *device = from_work(device, work, work.work); fw_device_cdev_update(device); device_for_each_child(&device->device, NULL, update_unit); @@ -988,7 +1061,7 @@ int fw_device_set_broadcast_channel(struct device *dev, void *gen) return 0; } -static int compare_configuration_rom(struct device *dev, void *data) +static int compare_configuration_rom(struct device *dev, const void *data) { const struct fw_device *old = fw_device(dev); const u32 *config_rom = data; @@ -1002,15 +1075,14 @@ static int compare_configuration_rom(struct device *dev, void *data) static void fw_device_init(struct work_struct *work) { - struct fw_device *device = - container_of(work, struct fw_device, work.work); + struct fw_device *device = from_work(device, work, work.work); struct fw_card *card = device->card; struct device *found; u32 minor; int ret; /* - * All failure paths here set node->data to NULL, so that we + * All failure paths here call fw_node_set_device(node, NULL), so that we * don't try to do device_for_each_child() on a kfree()'d * device. */ @@ -1039,7 +1111,7 @@ static void fw_device_init(struct work_struct *work) // // serialize config_rom access. scoped_guard(rwsem_read, &fw_device_rwsem) { - found = device_find_child(card->device, (void *)device->config_rom, + found = device_find_child(card->device, device->config_rom, compare_configuration_rom); } if (found) { @@ -1054,9 +1126,9 @@ static void fw_device_init(struct work_struct *work) struct fw_node *obsolete_node = reused->node; device->node = obsolete_node; - device->node->data = device; + fw_node_set_device(device->node, device); reused->node = current_node; - reused->node->data = reused; + fw_node_set_device(reused->node, reused); reused->max_speed = device->max_speed; reused->node_id = current_node->node_id; @@ -1126,10 +1198,10 @@ static void fw_device_init(struct work_struct *work) device->workfn = fw_device_shutdown; fw_schedule_device_work(device, SHUTDOWN_DELAY); } else { - fw_notice(card, "created device %s: GUID %08x%08x, S%d00\n", + fw_notice(card, "created device %s: GUID %08x%08x, S%d00, quirks %08x\n", dev_name(&device->device), device->config_rom[3], device->config_rom[4], - 1 << device->max_speed); + 1 << device->max_speed, device->quirks); device->config_rom_retries = 0; set_broadcast_channel(device, device->generation); @@ -1164,7 +1236,7 @@ static int reread_config_rom(struct fw_device *device, int generation, int i, rcode; for (i = 0; i < 6; i++) { - rcode = read_rom(device, generation, i, &q); + rcode = read_rom(device, generation, device->max_speed, i, &q); if (rcode != RCODE_COMPLETE) return rcode; @@ -1184,8 +1256,7 @@ static int reread_config_rom(struct fw_device *device, int generation, static void fw_device_refresh(struct work_struct *work) { - struct fw_device *device = - container_of(work, struct fw_device, work.work); + struct fw_device *device = from_work(device, work, work.work); struct fw_card *card = device->card; int ret, node_id = device->node_id; bool changed; @@ -1251,8 +1322,7 @@ static void fw_device_refresh(struct work_struct *work) static void fw_device_workfn(struct work_struct *work) { - struct fw_device *device = container_of(to_delayed_work(work), - struct fw_device, work); + struct fw_device *device = from_work(device, to_delayed_work(work), work); device->workfn(work); } @@ -1269,7 +1339,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event) * without actually having a link. */ create: - device = kzalloc(sizeof(*device), GFP_ATOMIC); + device = kzalloc_obj(*device, GFP_ATOMIC); if (device == NULL) break; @@ -1297,7 +1367,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event) * FW_NODE_UPDATED callbacks can update the node_id * and generation for the device. */ - node->data = device; + fw_node_set_device(node, device); /* * Many devices are slow to respond after bus resets, @@ -1312,7 +1382,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event) case FW_NODE_INITIATED_RESET: case FW_NODE_LINK_ON: - device = node->data; + device = fw_node_get_device(node); if (device == NULL) goto create; @@ -1329,7 +1399,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event) break; case FW_NODE_UPDATED: - device = node->data; + device = fw_node_get_device(node); if (device == NULL) break; @@ -1344,7 +1414,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event) case FW_NODE_DESTROYED: case FW_NODE_LINK_OFF: - if (!node->data) + if (!fw_node_get_device(node)) break; /* @@ -1359,7 +1429,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event) * the device in shutdown state to have that code fail * to create the device. */ - device = node->data; + device = fw_node_get_device(node); if (atomic_xchg(&device->state, FW_DEVICE_GONE) == FW_DEVICE_RUNNING) { device->workfn = fw_device_shutdown; diff --git a/drivers/firewire/core-iso.c b/drivers/firewire/core-iso.c index a67493862c85..d2b40a1a565e 100644 --- a/drivers/firewire/core-iso.c +++ b/drivers/firewire/core-iso.c @@ -30,48 +30,58 @@ int fw_iso_buffer_alloc(struct fw_iso_buffer *buffer, int page_count) { - int i; + struct page **page_array __free(kfree) = kzalloc_objs(page_array[0], + page_count); - buffer->page_count = 0; - buffer->page_count_mapped = 0; - buffer->pages = kmalloc_array(page_count, sizeof(buffer->pages[0]), - GFP_KERNEL); - if (buffer->pages == NULL) + if (!page_array) return -ENOMEM; - for (i = 0; i < page_count; i++) { - buffer->pages[i] = alloc_page(GFP_KERNEL | GFP_DMA32 | __GFP_ZERO); - if (buffer->pages[i] == NULL) - break; - } - buffer->page_count = i; - if (i < page_count) { - fw_iso_buffer_destroy(buffer, NULL); + // Retrieve noncontiguous pages. The descriptors for 1394 OHCI isochronous DMA contexts + // have a set of address and length per each, while the reason to use pages is the + // convenience to map them into virtual address space of user process. + unsigned long nr_populated = alloc_pages_bulk(GFP_KERNEL | GFP_DMA32 | __GFP_ZERO, + page_count, page_array); + if (nr_populated != page_count) { + // Assuming the above call fills page_array sequentially from the beginning. + release_pages(page_array, nr_populated); return -ENOMEM; } + buffer->page_count = page_count; + buffer->pages = no_free_ptr(page_array); + return 0; } int fw_iso_buffer_map_dma(struct fw_iso_buffer *buffer, struct fw_card *card, enum dma_data_direction direction) { - dma_addr_t address; + dma_addr_t *dma_addrs __free(kfree) = kzalloc_objs(dma_addrs[0], + buffer->page_count); int i; - buffer->direction = direction; + if (!dma_addrs) + return -ENOMEM; + // Retrieve DMA mapping addresses for the pages. They are not contiguous. Maintain the cache + // coherency for the pages by hand. for (i = 0; i < buffer->page_count; i++) { - address = dma_map_page(card->device, buffer->pages[i], - 0, PAGE_SIZE, direction); - if (dma_mapping_error(card->device, address)) + // The dma_map_phys() with a physical address per page is available here, instead. + dma_addr_t dma_addr = dma_map_page(card->device, buffer->pages[i], 0, PAGE_SIZE, + direction); + if (dma_mapping_error(card->device, dma_addr)) break; - set_page_private(buffer->pages[i], address); + dma_addrs[i] = dma_addr; } - buffer->page_count_mapped = i; - if (i < buffer->page_count) + if (i < buffer->page_count) { + while (i-- > 0) + dma_unmap_page(card->device, dma_addrs[i], PAGE_SIZE, buffer->direction); return -ENOMEM; + } + + buffer->direction = direction; + buffer->dma_addrs = no_free_ptr(dma_addrs); return 0; } @@ -96,34 +106,31 @@ EXPORT_SYMBOL(fw_iso_buffer_init); void fw_iso_buffer_destroy(struct fw_iso_buffer *buffer, struct fw_card *card) { - int i; - dma_addr_t address; + if (buffer->dma_addrs) { + for (int i = 0; i < buffer->page_count; ++i) { + dma_addr_t dma_addr = buffer->dma_addrs[i]; + dma_unmap_page(card->device, dma_addr, PAGE_SIZE, buffer->direction); + } + kfree(buffer->dma_addrs); + buffer->dma_addrs = NULL; + } - for (i = 0; i < buffer->page_count_mapped; i++) { - address = page_private(buffer->pages[i]); - dma_unmap_page(card->device, address, - PAGE_SIZE, buffer->direction); + if (buffer->pages) { + release_pages(buffer->pages, buffer->page_count); + kfree(buffer->pages); + buffer->pages = NULL; } - for (i = 0; i < buffer->page_count; i++) - __free_page(buffer->pages[i]); - kfree(buffer->pages); - buffer->pages = NULL; buffer->page_count = 0; - buffer->page_count_mapped = 0; } EXPORT_SYMBOL(fw_iso_buffer_destroy); /* Convert DMA address to offset into virtually contiguous buffer. */ size_t fw_iso_buffer_lookup(struct fw_iso_buffer *buffer, dma_addr_t completed) { - size_t i; - dma_addr_t address; - ssize_t offset; - - for (i = 0; i < buffer->page_count; i++) { - address = page_private(buffer->pages[i]); - offset = (ssize_t)completed - (ssize_t)address; + for (int i = 0; i < buffer->page_count; i++) { + dma_addr_t dma_addr = buffer->dma_addrs[i]; + ssize_t offset = (ssize_t)completed - (ssize_t)dma_addr; if (offset > 0 && offset <= PAGE_SIZE) return (i << PAGE_SHIFT) + offset; } @@ -131,14 +138,14 @@ size_t fw_iso_buffer_lookup(struct fw_iso_buffer *buffer, dma_addr_t completed) return 0; } -struct fw_iso_context *fw_iso_context_create(struct fw_card *card, - int type, int channel, int speed, size_t header_size, - fw_iso_callback_t callback, void *callback_data) +struct fw_iso_context *__fw_iso_context_create(struct fw_card *card, int type, int channel, + int speed, size_t header_size, size_t header_storage_size, + union fw_iso_callback callback, void *callback_data) { struct fw_iso_context *ctx; - ctx = card->driver->allocate_iso_context(card, - type, channel, header_size); + ctx = card->driver->allocate_iso_context(card, type, channel, header_size, + header_storage_size); if (IS_ERR(ctx)) return ctx; @@ -146,8 +153,10 @@ struct fw_iso_context *fw_iso_context_create(struct fw_card *card, ctx->type = type; ctx->channel = channel; ctx->speed = speed; + ctx->flags = 0; ctx->header_size = header_size; - ctx->callback.sc = callback; + ctx->header_storage_size = header_storage_size; + ctx->callback = callback; ctx->callback_data = callback_data; trace_isoc_outbound_allocate(ctx, channel, speed); @@ -156,7 +165,7 @@ struct fw_iso_context *fw_iso_context_create(struct fw_card *card, return ctx; } -EXPORT_SYMBOL(fw_iso_context_create); +EXPORT_SYMBOL(__fw_iso_context_create); void fw_iso_context_destroy(struct fw_iso_context *ctx) { diff --git a/drivers/firewire/core-topology.c b/drivers/firewire/core-topology.c index 74a6aa7d8cc9..bb2d2db30795 100644 --- a/drivers/firewire/core-topology.c +++ b/drivers/firewire/core-topology.c @@ -27,7 +27,7 @@ static struct fw_node *fw_node_create(u32 sid, int port_count, int color) { struct fw_node *node; - node = kzalloc(struct_size(node, ports, port_count), GFP_ATOMIC); + node = kzalloc_flex(*node, ports, port_count, GFP_ATOMIC); if (node == NULL) return NULL; @@ -241,7 +241,7 @@ static struct fw_node *build_tree(struct fw_card *card, const u32 *sid, int self // If PHYs report different gap counts, set an invalid count which will force a gap // count reconfiguration and a reset. if (phy_packet_self_id_zero_get_gap_count(self_id_sequence[0]) != gap_count) - gap_count = 0; + gap_count = GAP_COUNT_MISMATCHED; update_hop_count(node); @@ -325,9 +325,11 @@ static void report_found_node(struct fw_card *card, card->bm_retries = 0; } -/* Must be called with card->lock held */ void fw_destroy_nodes(struct fw_card *card) +__must_hold(&card->lock) { + lockdep_assert_held(&card->lock); + card->color++; if (card->local_node != NULL) for_each_fw_node(card, card->local_node, report_lost_node); @@ -435,20 +437,23 @@ static void update_tree(struct fw_card *card, struct fw_node *root) } } -static void update_topology_map(struct fw_card *card, - u32 *self_ids, int self_id_count) +static void update_topology_map(__be32 *buffer, size_t buffer_size, int root_node_id, + const u32 *self_ids, int self_id_count) { - int node_count = (card->root_node->node_id & 0x3f) + 1; - __be32 *map = card->topology_map; + __be32 *map = buffer; + u32 next_generation = be32_to_cpu(buffer[1]) + 1; + int node_count = (root_node_id & 0x3f) + 1; + + memset(map, 0, buffer_size); *map++ = cpu_to_be32((self_id_count + 2) << 16); - *map++ = cpu_to_be32(be32_to_cpu(card->topology_map[1]) + 1); + *map++ = cpu_to_be32(next_generation); *map++ = cpu_to_be32((node_count << 16) | self_id_count); while (self_id_count--) *map++ = cpu_to_be32p(self_ids++); - fw_compute_block_crc(card->topology_map); + fw_compute_block_crc(buffer); } void fw_core_handle_bus_reset(struct fw_card *card, int node_id, int generation, @@ -458,46 +463,45 @@ void fw_core_handle_bus_reset(struct fw_card *card, int node_id, int generation, trace_bus_reset_handle(card->index, generation, node_id, bm_abdicate, self_ids, self_id_count); - guard(spinlock_irqsave)(&card->lock); - - /* - * If the selfID buffer is not the immediate successor of the - * previously processed one, we cannot reliably compare the - * old and new topologies. - */ - if (!is_next_generation(generation, card->generation) && - card->local_node != NULL) { - fw_destroy_nodes(card); - card->bm_retries = 0; + scoped_guard(spinlock, &card->lock) { + // If the selfID buffer is not the immediate successor of the + // previously processed one, we cannot reliably compare the + // old and new topologies. + if (!is_next_generation(generation, card->generation) && card->local_node != NULL) { + fw_destroy_nodes(card); + card->bm_retries = 0; + } + card->broadcast_channel_allocated = card->broadcast_channel_auto_allocated; + card->node_id = node_id; + // Update node_id before generation to prevent anybody from using + // a stale node_id together with a current generation. + smp_wmb(); + card->generation = generation; + card->reset_jiffies = get_jiffies_64(); + card->bm_node_id = 0xffff; + card->bm_abdicate = bm_abdicate; + + local_node = build_tree(card, self_ids, self_id_count, generation); + + card->color++; + + if (local_node == NULL) { + fw_err(card, "topology build failed\n"); + // FIXME: We need to issue a bus reset in this case. + } else if (card->local_node == NULL) { + card->local_node = local_node; + for_each_fw_node(card, local_node, report_found_node); + } else { + update_tree(card, local_node); + } } - card->broadcast_channel_allocated = card->broadcast_channel_auto_allocated; - card->node_id = node_id; - /* - * Update node_id before generation to prevent anybody from using - * a stale node_id together with a current generation. - */ - smp_wmb(); - card->generation = generation; - card->reset_jiffies = get_jiffies_64(); - card->bm_node_id = 0xffff; - card->bm_abdicate = bm_abdicate; fw_schedule_bm_work(card, 0); - local_node = build_tree(card, self_ids, self_id_count, generation); - - update_topology_map(card, self_ids, self_id_count); - - card->color++; - - if (local_node == NULL) { - fw_err(card, "topology build failed\n"); - /* FIXME: We need to issue a bus reset in this case. */ - } else if (card->local_node == NULL) { - card->local_node = local_node; - for_each_fw_node(card, local_node, report_found_node); - } else { - update_tree(card, local_node); + // Just used by transaction layer. + scoped_guard(spinlock, &card->topology_map.lock) { + update_topology_map(card->topology_map.buffer, sizeof(card->topology_map.buffer), + card->root_node->node_id, self_ids, self_id_count); } } EXPORT_SYMBOL(fw_core_handle_bus_reset); diff --git a/drivers/firewire/core-transaction.c b/drivers/firewire/core-transaction.c index e141d24a7644..22ae387ae03c 100644 --- a/drivers/firewire/core-transaction.c +++ b/drivers/firewire/core-transaction.c @@ -39,31 +39,73 @@ static int try_cancel_split_timeout(struct fw_transaction *t) { if (t->is_split_transaction) - return del_timer(&t->split_timeout_timer); + return timer_delete(&t->split_timeout_timer); else return 1; } -static int close_transaction(struct fw_transaction *transaction, struct fw_card *card, int rcode, - u32 response_tstamp) +// card->transactions.lock must be acquired in advance. +static void remove_transaction_entry(struct fw_card *card, struct fw_transaction *entry) { - struct fw_transaction *t = NULL, *iter; + list_del_init(&entry->link); + card->transactions.tlabel_mask &= ~(1ULL << entry->tlabel); +} - scoped_guard(spinlock_irqsave, &card->lock) { - list_for_each_entry(iter, &card->transaction_list, link) { - if (iter == transaction) { - if (try_cancel_split_timeout(iter)) { - list_del_init(&iter->link); - card->tlabel_mask &= ~(1ULL << iter->tlabel); - t = iter; - } - break; - } +// Must be called without holding card->transactions.lock. +void fw_cancel_pending_transactions(struct fw_card *card) +{ + struct fw_transaction *t, *tmp; + LIST_HEAD(pending_list); + + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->transactions.lock) { + list_for_each_entry_safe(t, tmp, &card->transactions.list, link) { + if (try_cancel_split_timeout(t)) + list_move(&t->link, &pending_list); + } + } + + list_for_each_entry_safe(t, tmp, &pending_list, link) { + list_del(&t->link); + + if (!t->with_tstamp) { + t->callback.without_tstamp(card, RCODE_CANCELLED, NULL, 0, + t->callback_data); + } else { + t->callback.with_tstamp(card, RCODE_CANCELLED, t->packet.timestamp, 0, + NULL, 0, t->callback_data); } } +} - if (!t) - return -ENOENT; +// card->transactions.lock must be acquired in advance. +#define find_and_pop_transaction_entry(card, condition) \ +({ \ + struct fw_transaction *iter, *t = NULL; \ + list_for_each_entry(iter, &card->transactions.list, link) { \ + if (condition) { \ + t = iter; \ + break; \ + } \ + } \ + if (t && try_cancel_split_timeout(t)) \ + remove_transaction_entry(card, t); \ + t; \ +}) + +static int close_transaction(struct fw_transaction *transaction, struct fw_card *card, int rcode, + u32 response_tstamp) +{ + struct fw_transaction *t; + + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->transactions.lock) { + t = find_and_pop_transaction_entry(card, iter == transaction); + if (!t) + return -ENOENT; + } if (!t->with_tstamp) { t->callback.without_tstamp(card, rcode, NULL, 0, t->callback_data); @@ -114,14 +156,13 @@ EXPORT_SYMBOL(fw_cancel_transaction); static void split_transaction_timeout_callback(struct timer_list *timer) { - struct fw_transaction *t = from_timer(t, timer, split_timeout_timer); + struct fw_transaction *t = timer_container_of(t, timer, split_timeout_timer); struct fw_card *card = t->card; - scoped_guard(spinlock_irqsave, &card->lock) { + scoped_guard(spinlock_irqsave, &card->transactions.lock) { if (list_empty(&t->link)) return; - list_del(&t->link); - card->tlabel_mask &= ~(1ULL << t->tlabel); + remove_transaction_entry(card, t); } if (!t->with_tstamp) { @@ -132,17 +173,15 @@ static void split_transaction_timeout_callback(struct timer_list *timer) } } -static void start_split_transaction_timeout(struct fw_transaction *t, - struct fw_card *card) +// card->transactions.lock should be acquired in advance for the linked list. +static void start_split_transaction_timeout(struct fw_transaction *t, unsigned int delta) { - guard(spinlock_irqsave)(&card->lock); - if (list_empty(&t->link) || WARN_ON(t->is_split_transaction)) return; t->is_split_transaction = true; - mod_timer(&t->split_timeout_timer, - jiffies + card->split_timeout_jiffies); + + mod_timer(&t->split_timeout_timer, jiffies + delta); } static u32 compute_split_timeout_timestamp(struct fw_card *card, u32 request_timestamp); @@ -162,9 +201,20 @@ static void transmit_complete_callback(struct fw_packet *packet, break; case ACK_PENDING: { - t->split_timeout_cycle = - compute_split_timeout_timestamp(card, packet->timestamp) & 0xffff; - start_split_transaction_timeout(t, card); + unsigned int delta; + + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->split_timeout.lock) { + t->split_timeout_cycle = + compute_split_timeout_timestamp(card, packet->timestamp) & 0xffff; + delta = card->split_timeout.jiffies; + } + + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->transactions.lock) + start_split_transaction_timeout(t, delta); break; } case ACK_BUSY_X: @@ -259,18 +309,21 @@ static void fw_fill_request(struct fw_packet *packet, int tcode, int tlabel, } static int allocate_tlabel(struct fw_card *card) +__must_hold(&card->transactions.lock) { int tlabel; - tlabel = card->current_tlabel; - while (card->tlabel_mask & (1ULL << tlabel)) { + lockdep_assert_held(&card->transactions.lock); + + tlabel = card->transactions.current_tlabel; + while (card->transactions.tlabel_mask & (1ULL << tlabel)) { tlabel = (tlabel + 1) & 0x3f; - if (tlabel == card->current_tlabel) + if (tlabel == card->transactions.current_tlabel) return -EBUSY; } - card->current_tlabel = (tlabel + 1) & 0x3f; - card->tlabel_mask |= 1ULL << tlabel; + card->transactions.current_tlabel = (tlabel + 1) & 0x3f; + card->transactions.tlabel_mask |= 1ULL << tlabel; return tlabel; } @@ -331,7 +384,6 @@ void __fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode void *payload, size_t length, union fw_transaction_callback callback, bool with_tstamp, void *callback_data) { - unsigned long flags; int tlabel; /* @@ -339,11 +391,11 @@ void __fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode * the list while holding the card spinlock. */ - spin_lock_irqsave(&card->lock, flags); - - tlabel = allocate_tlabel(card); + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->transactions.lock) + tlabel = allocate_tlabel(card); if (tlabel < 0) { - spin_unlock_irqrestore(&card->lock, flags); if (!with_tstamp) { callback.without_tstamp(card, RCODE_SEND_ERROR, NULL, 0, callback_data); } else { @@ -368,15 +420,22 @@ void __fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode t->callback = callback; t->with_tstamp = with_tstamp; t->callback_data = callback_data; - - fw_fill_request(&t->packet, tcode, t->tlabel, destination_id, card->node_id, generation, - speed, offset, payload, length); t->packet.callback = transmit_complete_callback; - list_add_tail(&t->link, &card->transaction_list); + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->lock) { + // The node_id field of fw_card can be updated when handling SelfIDComplete. + fw_fill_request(&t->packet, tcode, t->tlabel, destination_id, card->node_id, + generation, speed, offset, payload, length); + } - spin_unlock_irqrestore(&card->lock, flags); + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->transactions.lock) + list_add_tail(&t->link, &card->transactions.list); + // Safe with no lock, since the index field of fw_card is immutable once assigned. trace_async_request_outbound_initiate((uintptr_t)t, card->index, generation, speed, t->packet.header, payload, tcode_is_read_request(tcode) ? 0 : length / 4); @@ -431,7 +490,7 @@ int fw_run_transaction(struct fw_card *card, int tcode, int destination_id, fw_send_request(card, &t, tcode, destination_id, generation, speed, offset, payload, length, transaction_callback, &d); wait_for_completion(&d.done); - destroy_timer_on_stack(&t.split_timeout_timer); + timer_destroy_on_stack(&t.split_timeout_timer); return d.rcode; } @@ -458,7 +517,7 @@ static struct fw_packet phy_config_packet = { void fw_send_phy_config(struct fw_card *card, int node_id, int generation, int gap_count) { - long timeout = DIV_ROUND_UP(HZ, 10); + long timeout = msecs_to_jiffies(100); u32 data = 0; phy_packet_set_packet_identifier(&data, PHY_PACKET_PACKET_IDENTIFIER_PHY_CONFIG); @@ -550,6 +609,23 @@ const struct fw_address_region fw_unit_space_region = { .start = 0xfffff0000900ULL, .end = 0x1000000000000ULL, }; #endif /* 0 */ +static void complete_address_handler(struct kref *kref) +{ + struct fw_address_handler *handler = container_of(kref, struct fw_address_handler, kref); + + complete(&handler->done); +} + +static void get_address_handler(struct fw_address_handler *handler) +{ + kref_get(&handler->kref); +} + +static int put_address_handler(struct fw_address_handler *handler) +{ + return kref_put(&handler->kref, complete_address_handler); +} + /** * fw_core_add_address_handler() - register for incoming requests * @handler: callback @@ -557,9 +633,10 @@ const struct fw_address_region fw_unit_space_region = * * region->start, ->end, and handler->length have to be quadlet-aligned. * - * When a request is received that falls within the specified address range, - * the specified callback is invoked. The parameters passed to the callback - * give the details of the particular request. + * When a request is received that falls within the specified address range, the specified callback + * is invoked. The parameters passed to the callback give the details of the particular request. + * The callback is invoked in the workqueue context in most cases. However, if the request is + * initiated by the local node, the callback is invoked in the initiator's context. * * To be called in process context. * Return value: 0 on success, non-zero otherwise. @@ -595,6 +672,8 @@ int fw_core_add_address_handler(struct fw_address_handler *handler, if (other != NULL) { handler->offset += other->length; } else { + init_completion(&handler->done); + kref_init(&handler->kref); list_add_tail_rcu(&handler->link, &address_handler_list); ret = 0; break; @@ -620,6 +699,9 @@ void fw_core_remove_address_handler(struct fw_address_handler *handler) list_del_rcu(&handler->link); synchronize_rcu(); + + if (!put_address_handler(handler)) + wait_for_completion(&handler->done); } EXPORT_SYMBOL(fw_core_remove_address_handler); @@ -756,11 +838,14 @@ EXPORT_SYMBOL(fw_fill_response); static u32 compute_split_timeout_timestamp(struct fw_card *card, u32 request_timestamp) +__must_hold(&card->split_timeout.lock) { unsigned int cycles; u32 timestamp; - cycles = card->split_timeout_cycles; + lockdep_assert_held(&card->split_timeout.lock); + + cycles = card->split_timeout.cycles; cycles += request_timestamp & 0x1fff; timestamp = request_timestamp & ~0x1fff; @@ -811,9 +896,12 @@ static struct fw_request *allocate_request(struct fw_card *card, return NULL; kref_init(&request->kref); + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->split_timeout.lock) + request->response.timestamp = compute_split_timeout_timestamp(card, p->timestamp); + request->response.speed = p->speed; - request->response.timestamp = - compute_split_timeout_timestamp(card, p->timestamp); request->response.generation = p->generation; request->response.ack = 0; request->response.callback = free_response_callback; @@ -913,22 +1001,31 @@ static void handle_exclusive_region_request(struct fw_card *card, handler = lookup_enclosing_address_handler(&address_handler_list, offset, request->length); if (handler) - handler->address_callback(card, request, tcode, destination, source, - p->generation, offset, request->data, - request->length, handler->callback_data); + get_address_handler(handler); } - if (!handler) + if (!handler) { fw_send_response(card, request, RCODE_ADDRESS_ERROR); + return; + } + + // Outside the RCU read-side critical section. Without spinlock. With reference count. + handler->address_callback(card, request, tcode, destination, source, p->generation, offset, + request->data, request->length, handler->callback_data); + put_address_handler(handler); } +// To use kmalloc allocator efficiently, this should be power of two. +#define BUFFER_ON_KERNEL_STACK_SIZE 4 + static void handle_fcp_region_request(struct fw_card *card, struct fw_packet *p, struct fw_request *request, unsigned long long offset) { - struct fw_address_handler *handler; - int tcode, destination, source; + struct fw_address_handler *buffer_on_kernel_stack[BUFFER_ON_KERNEL_STACK_SIZE]; + struct fw_address_handler *handler, **handlers; + int tcode, destination, source, i, count, buffer_size; if ((offset != (CSR_REGISTER_BASE | CSR_FCP_COMMAND) && offset != (CSR_REGISTER_BASE | CSR_FCP_RESPONSE)) || @@ -949,15 +1046,55 @@ static void handle_fcp_region_request(struct fw_card *card, return; } + count = 0; + handlers = buffer_on_kernel_stack; + buffer_size = ARRAY_SIZE(buffer_on_kernel_stack); scoped_guard(rcu) { list_for_each_entry_rcu(handler, &address_handler_list, link) { - if (is_enclosing_handler(handler, offset, request->length)) - handler->address_callback(card, request, tcode, destination, source, - p->generation, offset, request->data, - request->length, handler->callback_data); + if (is_enclosing_handler(handler, offset, request->length)) { + if (count >= buffer_size) { + int next_size = buffer_size * 2; + struct fw_address_handler **buffer_on_kernel_heap; + + if (handlers == buffer_on_kernel_stack) + buffer_on_kernel_heap = NULL; + else + buffer_on_kernel_heap = handlers; + + buffer_on_kernel_heap = + krealloc_array(buffer_on_kernel_heap, next_size, + sizeof(*buffer_on_kernel_heap), GFP_ATOMIC); + // FCP is used for purposes unrelated to significant system + // resources (e.g. storage or networking), so allocation + // failures are not considered so critical. + if (!buffer_on_kernel_heap) + break; + + if (handlers == buffer_on_kernel_stack) { + memcpy(buffer_on_kernel_heap, buffer_on_kernel_stack, + sizeof(buffer_on_kernel_stack)); + } + + handlers = buffer_on_kernel_heap; + buffer_size = next_size; + } + get_address_handler(handler); + handlers[count++] = handler; + } } } + for (i = 0; i < count; ++i) { + handler = handlers[i]; + handler->address_callback(card, request, tcode, destination, source, + p->generation, offset, request->data, + request->length, handler->callback_data); + put_address_handler(handler); + } + + if (handlers != buffer_on_kernel_stack) + kfree(handlers); + fw_send_response(card, request, RCODE_COMPLETE); } @@ -1000,7 +1137,7 @@ EXPORT_SYMBOL(fw_core_handle_request); void fw_core_handle_response(struct fw_card *card, struct fw_packet *p) { - struct fw_transaction *t = NULL, *iter; + struct fw_transaction *t = NULL; u32 *data; size_t data_length; int tcode, tlabel, source, rcode; @@ -1039,17 +1176,11 @@ void fw_core_handle_response(struct fw_card *card, struct fw_packet *p) break; } - scoped_guard(spinlock_irqsave, &card->lock) { - list_for_each_entry(iter, &card->transaction_list, link) { - if (iter->node_id == source && iter->tlabel == tlabel) { - if (try_cancel_split_timeout(iter)) { - list_del_init(&iter->link); - card->tlabel_mask &= ~(1ULL << iter->tlabel); - t = iter; - } - break; - } - } + // NOTE: This can be without irqsave when we can guarantee that __fw_send_request() for + // local destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->transactions.lock) { + t = find_and_pop_transaction_entry(card, + iter->node_id == source && iter->tlabel == tlabel); } trace_async_response_inbound((uintptr_t)t, card->index, p->generation, p->speed, p->ack, @@ -1124,7 +1255,11 @@ static void handle_topology_map(struct fw_card *card, struct fw_request *request } start = (offset - topology_map_region.start) / 4; - memcpy(payload, &card->topology_map[start], length); + + // NOTE: This can be without irqsave when we can guarantee that fw_send_request() for local + // destination never runs in any type of IRQ context. + scoped_guard(spinlock_irqsave, &card->topology_map.lock) + memcpy(payload, &card->topology_map.buffer[start], length); fw_send_response(card, request, RCODE_COMPLETE); } @@ -1139,16 +1274,17 @@ static const struct fw_address_region registers_region = .end = CSR_REGISTER_BASE | CSR_CONFIG_ROM, }; static void update_split_timeout(struct fw_card *card) +__must_hold(&card->split_timeout.lock) { unsigned int cycles; - cycles = card->split_timeout_hi * 8000 + (card->split_timeout_lo >> 19); + cycles = card->split_timeout.hi * 8000 + (card->split_timeout.lo >> 19); /* minimum per IEEE 1394, maximum which doesn't overflow OHCI */ cycles = clamp(cycles, 800u, 3u * 8000u); - card->split_timeout_cycles = cycles; - card->split_timeout_jiffies = DIV_ROUND_UP(cycles * HZ, 8000); + card->split_timeout.cycles = cycles; + card->split_timeout.jiffies = isoc_cycles_to_jiffies(cycles); } static void handle_registers(struct fw_card *card, struct fw_request *request, @@ -1198,12 +1334,15 @@ static void handle_registers(struct fw_card *card, struct fw_request *request, case CSR_SPLIT_TIMEOUT_HI: if (tcode == TCODE_READ_QUADLET_REQUEST) { - *data = cpu_to_be32(card->split_timeout_hi); + *data = cpu_to_be32(card->split_timeout.hi); } else if (tcode == TCODE_WRITE_QUADLET_REQUEST) { - guard(spinlock_irqsave)(&card->lock); - - card->split_timeout_hi = be32_to_cpu(*data) & 7; - update_split_timeout(card); + // NOTE: This can be without irqsave when we can guarantee that + // __fw_send_request() for local destination never runs in any type of IRQ + // context. + scoped_guard(spinlock_irqsave, &card->split_timeout.lock) { + card->split_timeout.hi = be32_to_cpu(*data) & 7; + update_split_timeout(card); + } } else { rcode = RCODE_TYPE_ERROR; } @@ -1211,12 +1350,15 @@ static void handle_registers(struct fw_card *card, struct fw_request *request, case CSR_SPLIT_TIMEOUT_LO: if (tcode == TCODE_READ_QUADLET_REQUEST) { - *data = cpu_to_be32(card->split_timeout_lo); + *data = cpu_to_be32(card->split_timeout.lo); } else if (tcode == TCODE_WRITE_QUADLET_REQUEST) { - guard(spinlock_irqsave)(&card->lock); - - card->split_timeout_lo = be32_to_cpu(*data) & 0xfff80000; - update_split_timeout(card); + // NOTE: This can be without irqsave when we can guarantee that + // __fw_send_request() for local destination never runs in any type of IRQ + // context. + scoped_guard(spinlock_irqsave, &card->split_timeout.lock) { + card->split_timeout.lo = be32_to_cpu(*data) & 0xfff80000; + update_split_timeout(card); + } } else { rcode = RCODE_TYPE_ERROR; } @@ -1327,7 +1469,8 @@ static int __init fw_core_init(void) { int ret; - fw_workqueue = alloc_workqueue("firewire", WQ_MEM_RECLAIM, 0); + fw_workqueue = alloc_workqueue("firewire", WQ_MEM_RECLAIM | WQ_UNBOUND, + 0); if (!fw_workqueue) return -ENOMEM; diff --git a/drivers/firewire/core.h b/drivers/firewire/core.h index 9b298af1cac0..8b49d7480c37 100644 --- a/drivers/firewire/core.h +++ b/drivers/firewire/core.h @@ -27,6 +27,11 @@ struct fw_packet; /* -card */ +// This is the arbitrary value we use to indicate a mismatched gap count. +#define GAP_COUNT_MISMATCHED 0 + +#define isoc_cycles_to_jiffies(cycles) usecs_to_jiffies((u32)div_u64((u64)cycles * USEC_PER_SEC, 8000)) + extern __printf(2, 3) void fw_err(const struct fw_card *card, const char *fmt, ...); extern __printf(2, 3) @@ -60,6 +65,9 @@ struct fw_card_driver { int (*enable)(struct fw_card *card, const __be32 *config_rom, size_t length); + // After returning the call, any function is no longer triggered to handle hardware event. + void (*disable)(struct fw_card *card); + int (*read_phy_reg)(struct fw_card *card, int address); int (*update_phy_reg)(struct fw_card *card, int address, int clear_bits, int set_bits); @@ -92,8 +100,8 @@ struct fw_card_driver { void (*write_csr)(struct fw_card *card, int csr_offset, u32 value); struct fw_iso_context * - (*allocate_iso_context)(struct fw_card *card, - int type, int channel, size_t header_size); + (*allocate_iso_context)(struct fw_card *card, int type, int channel, size_t header_size, + size_t header_storage_size); void (*free_iso_context)(struct fw_iso_context *ctx); int (*start_iso)(struct fw_iso_context *ctx, @@ -158,15 +166,28 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event); int fw_iso_buffer_alloc(struct fw_iso_buffer *buffer, int page_count); int fw_iso_buffer_map_dma(struct fw_iso_buffer *buffer, struct fw_card *card, enum dma_data_direction direction); +size_t fw_iso_buffer_lookup(struct fw_iso_buffer *buffer, dma_addr_t completed); static inline void fw_iso_context_init_work(struct fw_iso_context *ctx, work_func_t func) { INIT_WORK(&ctx->work, func); } +static inline struct fw_iso_context *fw_iso_mc_context_create(struct fw_card *card, + fw_iso_mc_callback_t callback, void *callback_data) +{ + union fw_iso_callback cb = { .mc = callback }; + + return __fw_iso_context_create(card, FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL, 0, 0, 0, 0, cb, + callback_data); +} + /* -topology */ +// The initial value of BUS_MANAGER_ID register, to express nothing registered. +#define BUS_MANAGER_ID_NOT_REGISTERED 0x3f + enum { FW_NODE_CREATED, FW_NODE_UPDATED, @@ -194,8 +215,8 @@ struct fw_node { /* For serializing node topology into a list. */ struct list_head link; - /* Upper layer specific data. */ - void *data; + // The device when already associated, else NULL. + struct fw_device *device; struct fw_node *ports[] __counted_by(port_count); }; @@ -219,6 +240,16 @@ static inline void fw_node_put(struct fw_node *node) kref_put(&node->kref, release_node); } +static inline struct fw_device *fw_node_get_device(struct fw_node *node) +{ + return node->device; +} + +static inline void fw_node_set_device(struct fw_node *node, struct fw_device *device) +{ + node->device = device; +} + void fw_core_handle_bus_reset(struct fw_card *card, int node_id, int generation, int self_id_count, u32 *self_ids, bool bm_abdicate); void fw_destroy_nodes(struct fw_card *card); @@ -266,6 +297,8 @@ void fw_fill_response(struct fw_packet *response, u32 *request_header, void fw_request_get(struct fw_request *request); void fw_request_put(struct fw_request *request); +void fw_cancel_pending_transactions(struct fw_card *card); + // Convert the value of IEEE 1394 CYCLE_TIME register to the format of timeStamp field in // descriptors of 1394 OHCI. static inline u32 cycle_time_to_ohci_tstamp(u32 tstamp) diff --git a/drivers/firewire/device-attribute-test.c b/drivers/firewire/device-attribute-test.c index 2f123c6b0a16..97478a96d1c9 100644 --- a/drivers/firewire/device-attribute-test.c +++ b/drivers/firewire/device-attribute-test.c @@ -99,6 +99,7 @@ static void device_attr_simple_avc(struct kunit *test) struct device *unit0_dev = (struct device *)&unit0.device; static const int unit0_expected_ids[] = {0x00ffffff, 0x00ffffff, 0x0000a02d, 0x00010001}; char *buf = kunit_kzalloc(test, PAGE_SIZE, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, buf); int ids[4] = {0, 0, 0, 0}; // Ensure associations for node and unit devices. @@ -180,6 +181,7 @@ static void device_attr_legacy_avc(struct kunit *test) struct device *unit0_dev = (struct device *)&unit0.device; static const int unit0_expected_ids[] = {0x00012345, 0x00fedcba, 0x00abcdef, 0x00543210}; char *buf = kunit_kzalloc(test, PAGE_SIZE, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, buf); int ids[4] = {0, 0, 0, 0}; // Ensure associations for node and unit devices. diff --git a/drivers/firewire/init_ohci1394_dma.c b/drivers/firewire/init_ohci1394_dma.c index 48b879e9e831..121f0c2f6401 100644 --- a/drivers/firewire/init_ohci1394_dma.c +++ b/drivers/firewire/init_ohci1394_dma.c @@ -167,6 +167,7 @@ static inline void __init init_ohci1394_initialize(struct ohci *ohci) /** * init_ohci1394_wait_for_busresets - wait until bus resets are completed + * @ohci: Pointer to the OHCI-1394 controller structure * * OHCI1394 initialization itself and any device going on- or offline * and any cable issue cause a IEEE1394 bus reset. The OHCI1394 spec @@ -189,6 +190,8 @@ static inline void __init init_ohci1394_wait_for_busresets(struct ohci *ohci) /** * init_ohci1394_enable_physical_dma - Enable physical DMA for remote debugging + * @ohci: Pointer to the OHCI-1394 controller structure + * * This enables remote DMA access over IEEE1394 from every host for the low * 4GB of address space. DMA accesses above 4GB are not available currently. */ @@ -201,6 +204,8 @@ static inline void __init init_ohci1394_enable_physical_dma(struct ohci *ohci) /** * init_ohci1394_reset_and_init_dma - init controller and enable DMA + * @ohci: Pointer to the OHCI-1394 controller structure + * * This initializes the given controller and enables physical DMA engine in it. */ static inline void __init init_ohci1394_reset_and_init_dma(struct ohci *ohci) @@ -230,6 +235,10 @@ static inline void __init init_ohci1394_reset_and_init_dma(struct ohci *ohci) /** * init_ohci1394_controller - Map the registers of the controller and init DMA + * @num: PCI bus number + * @slot: PCI device number + * @func: PCI function number + * * This maps the registers of the specified controller and initializes it */ static inline void __init init_ohci1394_controller(int num, int slot, int func) @@ -284,6 +293,7 @@ void __init init_ohci1394_dma_on_all_controllers(void) /** * setup_ohci1394_dma - enables early OHCI1394 DMA initialization + * @opt: Kernel command line parameter string */ static int __init setup_ohci1394_dma(char *opt) { diff --git a/drivers/firewire/net.c b/drivers/firewire/net.c index 1bf0e15c1540..82b3b6d9ed2d 100644 --- a/drivers/firewire/net.c +++ b/drivers/firewire/net.c @@ -257,9 +257,10 @@ static void fwnet_header_cache_update(struct hh_cache *hh, memcpy((u8 *)hh->hh_data + HH_DATA_OFF(FWNET_HLEN), haddr, net->addr_len); } -static int fwnet_header_parse(const struct sk_buff *skb, unsigned char *haddr) +static int fwnet_header_parse(const struct sk_buff *skb, const struct net_device *dev, + unsigned char *haddr) { - memcpy(haddr, skb->dev->dev_addr, FWNET_ALEN); + memcpy(haddr, dev->dev_addr, FWNET_ALEN); return FWNET_ALEN; } @@ -338,7 +339,7 @@ static struct fwnet_fragment_info *fwnet_frag_new( } } - new = kmalloc(sizeof(*new), GFP_ATOMIC); + new = kmalloc_obj(*new, GFP_ATOMIC); if (!new) return NULL; @@ -356,7 +357,7 @@ static struct fwnet_partial_datagram *fwnet_pd_new(struct net_device *net, struct fwnet_partial_datagram *new; struct fwnet_fragment_info *fi; - new = kmalloc(sizeof(*new), GFP_ATOMIC); + new = kmalloc_obj(*new, GFP_ATOMIC); if (!new) goto fail; @@ -1007,7 +1008,7 @@ static int fwnet_send_packet(struct fwnet_packet_task *ptask) spin_lock_irqsave(&dev->lock, flags); - /* If the AT tasklet already ran, we may be last user. */ + /* If the AT work item already ran, we may be last user. */ free = (ptask->outstanding_pkts == 0 && !ptask->enqueued); if (!free) ptask->enqueued = true; @@ -1026,7 +1027,7 @@ static int fwnet_send_packet(struct fwnet_packet_task *ptask) spin_lock_irqsave(&dev->lock, flags); - /* If the AT tasklet already ran, we may be last user. */ + /* If the AT work item already ran, we may be last user. */ free = (ptask->outstanding_pkts == 0 && !ptask->enqueued); if (!free) ptask->enqueued = true; @@ -1402,7 +1403,7 @@ static int fwnet_add_peer(struct fwnet_device *dev, { struct fwnet_peer *peer; - peer = kmalloc(sizeof(*peer), GFP_KERNEL); + peer = kmalloc_obj(*peer); if (!peer) return -ENOMEM; diff --git a/drivers/firewire/nosy.c b/drivers/firewire/nosy.c index ea31ac7ac1ca..900b7d904b41 100644 --- a/drivers/firewire/nosy.c +++ b/drivers/firewire/nosy.c @@ -36,6 +36,8 @@ static char driver_name[] = KBUILD_MODNAME; +#define RCV_BUFFER_SIZE (16 * 1024) + /* this is the physical layout of a PCL, its size is 128 bytes */ struct pcl { __le32 next; @@ -279,7 +281,7 @@ nosy_open(struct inode *inode, struct file *file) if (lynx == NULL) return -ENODEV; - client = kmalloc(sizeof *client, GFP_KERNEL); + client = kmalloc_obj(*client); if (client == NULL) goto fail; @@ -517,16 +519,14 @@ remove_card(struct pci_dev *dev) lynx->rcv_start_pcl, lynx->rcv_start_pcl_bus); dma_free_coherent(&lynx->pci_device->dev, sizeof(struct pcl), lynx->rcv_pcl, lynx->rcv_pcl_bus); - dma_free_coherent(&lynx->pci_device->dev, PAGE_SIZE, lynx->rcv_buffer, - lynx->rcv_buffer_bus); + dma_free_coherent(&lynx->pci_device->dev, RCV_BUFFER_SIZE, + lynx->rcv_buffer, lynx->rcv_buffer_bus); iounmap(lynx->registers); pci_disable_device(dev); lynx_put(lynx); } -#define RCV_BUFFER_SIZE (16 * 1024) - static int add_card(struct pci_dev *dev, const struct pci_device_id *unused) { @@ -545,7 +545,7 @@ add_card(struct pci_dev *dev, const struct pci_device_id *unused) } pci_set_master(dev); - lynx = kzalloc(sizeof *lynx, GFP_KERNEL); + lynx = kzalloc_obj(*lynx); if (lynx == NULL) { dev_err(&dev->dev, "Failed to allocate control structure\n"); ret = -ENOMEM; @@ -680,7 +680,7 @@ fail_deallocate_buffers: dma_free_coherent(&lynx->pci_device->dev, sizeof(struct pcl), lynx->rcv_pcl, lynx->rcv_pcl_bus); if (lynx->rcv_buffer) - dma_free_coherent(&lynx->pci_device->dev, PAGE_SIZE, + dma_free_coherent(&lynx->pci_device->dev, RCV_BUFFER_SIZE, lynx->rcv_buffer, lynx->rcv_buffer_bus); iounmap(lynx->registers); diff --git a/drivers/firewire/ohci.c b/drivers/firewire/ohci.c index c02aed11b590..8153d62c58f0 100644 --- a/drivers/firewire/ohci.c +++ b/drivers/firewire/ohci.c @@ -86,7 +86,7 @@ struct descriptor { #define AR_BUFFER_SIZE (32*1024) #define AR_BUFFERS_MIN DIV_ROUND_UP(AR_BUFFER_SIZE, PAGE_SIZE) /* we need at least two pages for proper list management */ -#define AR_BUFFERS (AR_BUFFERS_MIN >= 2 ? AR_BUFFERS_MIN : 2) +#define AR_BUFFERS MAX(2, AR_BUFFERS_MIN) #define MAX_ASYNC_PAYLOAD 4096 #define MAX_AR_PACKET_SIZE (16 + MAX_ASYNC_PAYLOAD + 4) @@ -96,12 +96,13 @@ struct ar_context { struct fw_ohci *ohci; struct page *pages[AR_BUFFERS]; void *buffer; + dma_addr_t dma_addrs[AR_BUFFERS]; struct descriptor *descriptors; dma_addr_t descriptors_bus; void *pointer; unsigned int last_buffer_index; u32 regs; - struct tasklet_struct tasklet; + struct work_struct work; }; struct context; @@ -128,7 +129,6 @@ struct context { int total_allocation; u32 current_bus; bool running; - bool flushing; /* * List of page-sized buffers for storing DMA descriptors. @@ -157,21 +157,31 @@ struct context { int prev_z; descriptor_callback_t callback; +}; - struct tasklet_struct tasklet; +struct at_context { + struct context context; + struct work_struct work; + bool flushing; }; struct iso_context { struct fw_iso_context base; struct context context; - void *header; - size_t header_length; unsigned long flushing_completions; - u32 mc_buffer_bus; - u16 mc_completed; - u16 last_timestamp; u8 sync; u8 tags; + union { + struct { + u16 last_timestamp; + size_t header_length; + void *header; + } sc; + struct { + u32 buffer_bus; + u16 completed; + } mc; + }; }; #define CONFIG_ROM_SIZE (CSR_CONFIG_ROM_END - CSR_CONFIG_ROM) @@ -204,8 +214,8 @@ struct fw_ohci { struct ar_context ar_request_ctx; struct ar_context ar_response_ctx; - struct context at_request_ctx; - struct context at_response_ctx; + struct at_context at_request_ctx; + struct at_context at_response_ctx; u32 it_context_support; u32 it_context_mask; /* unoccupied IT contexts */ @@ -225,13 +235,10 @@ struct fw_ohci { __le32 *self_id; dma_addr_t self_id_bus; - struct work_struct bus_reset_work; u32 self_id_buffer[512]; }; -static struct workqueue_struct *selfid_workqueue; - static inline struct fw_ohci *fw_ohci(struct fw_card *card) { return container_of(card, struct fw_ohci, card); @@ -390,225 +397,10 @@ MODULE_PARM_DESC(quirks, "Chip quirks (default = 0" ", IR wake unreliable = " __stringify(QUIRK_IR_WAKE) ")"); -#define OHCI_PARAM_DEBUG_AT_AR 1 -#define OHCI_PARAM_DEBUG_SELFIDS 2 -#define OHCI_PARAM_DEBUG_IRQS 4 - -static int param_debug; -module_param_named(debug, param_debug, int, 0644); -MODULE_PARM_DESC(debug, "Verbose logging, deprecated in v6.11 kernel or later. (default = 0" - ", AT/AR events = " __stringify(OHCI_PARAM_DEBUG_AT_AR) - ", self-IDs = " __stringify(OHCI_PARAM_DEBUG_SELFIDS) - ", IRQs = " __stringify(OHCI_PARAM_DEBUG_IRQS) - ", or a combination, or all = -1)"); - static bool param_remote_dma; module_param_named(remote_dma, param_remote_dma, bool, 0444); MODULE_PARM_DESC(remote_dma, "Enable unfiltered remote DMA (default = N)"); -static void log_irqs(struct fw_ohci *ohci, u32 evt) -{ - if (likely(!(param_debug & OHCI_PARAM_DEBUG_IRQS))) - return; - - ohci_notice(ohci, "IRQ %08x%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", evt, - evt & OHCI1394_selfIDComplete ? " selfID" : "", - evt & OHCI1394_RQPkt ? " AR_req" : "", - evt & OHCI1394_RSPkt ? " AR_resp" : "", - evt & OHCI1394_reqTxComplete ? " AT_req" : "", - evt & OHCI1394_respTxComplete ? " AT_resp" : "", - evt & OHCI1394_isochRx ? " IR" : "", - evt & OHCI1394_isochTx ? " IT" : "", - evt & OHCI1394_postedWriteErr ? " postedWriteErr" : "", - evt & OHCI1394_cycleTooLong ? " cycleTooLong" : "", - evt & OHCI1394_cycle64Seconds ? " cycle64Seconds" : "", - evt & OHCI1394_cycleInconsistent ? " cycleInconsistent" : "", - evt & OHCI1394_regAccessFail ? " regAccessFail" : "", - evt & OHCI1394_unrecoverableError ? " unrecoverableError" : "", - evt & OHCI1394_busReset ? " busReset" : "", - evt & ~(OHCI1394_selfIDComplete | OHCI1394_RQPkt | - OHCI1394_RSPkt | OHCI1394_reqTxComplete | - OHCI1394_respTxComplete | OHCI1394_isochRx | - OHCI1394_isochTx | OHCI1394_postedWriteErr | - OHCI1394_cycleTooLong | OHCI1394_cycle64Seconds | - OHCI1394_cycleInconsistent | - OHCI1394_regAccessFail | OHCI1394_busReset) - ? " ?" : ""); -} - -static void log_selfids(struct fw_ohci *ohci, int generation, int self_id_count) -{ - static const char *const speed[] = { - [0] = "S100", [1] = "S200", [2] = "S400", [3] = "beta", - }; - static const char *const power[] = { - [0] = "+0W", [1] = "+15W", [2] = "+30W", [3] = "+45W", - [4] = "-3W", [5] = " ?W", [6] = "-3..-6W", [7] = "-3..-10W", - }; - static const char port[] = { - [PHY_PACKET_SELF_ID_PORT_STATUS_NONE] = '.', - [PHY_PACKET_SELF_ID_PORT_STATUS_NCONN] = '-', - [PHY_PACKET_SELF_ID_PORT_STATUS_PARENT] = 'p', - [PHY_PACKET_SELF_ID_PORT_STATUS_CHILD] = 'c', - }; - struct self_id_sequence_enumerator enumerator = { - .cursor = ohci->self_id_buffer, - .quadlet_count = self_id_count, - }; - - if (likely(!(param_debug & OHCI_PARAM_DEBUG_SELFIDS))) - return; - - ohci_notice(ohci, "%d selfIDs, generation %d, local node ID %04x\n", - self_id_count, generation, ohci->node_id); - - while (enumerator.quadlet_count > 0) { - unsigned int quadlet_count; - unsigned int port_index; - const u32 *s; - int i; - - s = self_id_sequence_enumerator_next(&enumerator, &quadlet_count); - if (IS_ERR(s)) - break; - - ohci_notice(ohci, - "selfID 0: %08x, phy %d [%c%c%c] %s gc=%d %s %s%s%s\n", - *s, - phy_packet_self_id_get_phy_id(*s), - port[self_id_sequence_get_port_status(s, quadlet_count, 0)], - port[self_id_sequence_get_port_status(s, quadlet_count, 1)], - port[self_id_sequence_get_port_status(s, quadlet_count, 2)], - speed[*s >> 14 & 3], *s >> 16 & 63, - power[*s >> 8 & 7], *s >> 22 & 1 ? "L" : "", - *s >> 11 & 1 ? "c" : "", *s & 2 ? "i" : ""); - - port_index = 3; - for (i = 1; i < quadlet_count; ++i) { - ohci_notice(ohci, - "selfID n: %08x, phy %d [%c%c%c%c%c%c%c%c]\n", - s[i], - phy_packet_self_id_get_phy_id(s[i]), - port[self_id_sequence_get_port_status(s, quadlet_count, port_index)], - port[self_id_sequence_get_port_status(s, quadlet_count, port_index + 1)], - port[self_id_sequence_get_port_status(s, quadlet_count, port_index + 2)], - port[self_id_sequence_get_port_status(s, quadlet_count, port_index + 3)], - port[self_id_sequence_get_port_status(s, quadlet_count, port_index + 4)], - port[self_id_sequence_get_port_status(s, quadlet_count, port_index + 5)], - port[self_id_sequence_get_port_status(s, quadlet_count, port_index + 6)], - port[self_id_sequence_get_port_status(s, quadlet_count, port_index + 7)] - ); - - port_index += 8; - } - } -} - -static const char *evts[] = { - [0x00] = "evt_no_status", [0x01] = "-reserved-", - [0x02] = "evt_long_packet", [0x03] = "evt_missing_ack", - [0x04] = "evt_underrun", [0x05] = "evt_overrun", - [0x06] = "evt_descriptor_read", [0x07] = "evt_data_read", - [0x08] = "evt_data_write", [0x09] = "evt_bus_reset", - [0x0a] = "evt_timeout", [0x0b] = "evt_tcode_err", - [0x0c] = "-reserved-", [0x0d] = "-reserved-", - [0x0e] = "evt_unknown", [0x0f] = "evt_flushed", - [0x10] = "-reserved-", [0x11] = "ack_complete", - [0x12] = "ack_pending ", [0x13] = "-reserved-", - [0x14] = "ack_busy_X", [0x15] = "ack_busy_A", - [0x16] = "ack_busy_B", [0x17] = "-reserved-", - [0x18] = "-reserved-", [0x19] = "-reserved-", - [0x1a] = "-reserved-", [0x1b] = "ack_tardy", - [0x1c] = "-reserved-", [0x1d] = "ack_data_error", - [0x1e] = "ack_type_error", [0x1f] = "-reserved-", - [0x20] = "pending/cancelled", -}; - -static void log_ar_at_event(struct fw_ohci *ohci, - char dir, int speed, u32 *header, int evt) -{ - static const char *const tcodes[] = { - [TCODE_WRITE_QUADLET_REQUEST] = "QW req", - [TCODE_WRITE_BLOCK_REQUEST] = "BW req", - [TCODE_WRITE_RESPONSE] = "W resp", - [0x3] = "-reserved-", - [TCODE_READ_QUADLET_REQUEST] = "QR req", - [TCODE_READ_BLOCK_REQUEST] = "BR req", - [TCODE_READ_QUADLET_RESPONSE] = "QR resp", - [TCODE_READ_BLOCK_RESPONSE] = "BR resp", - [TCODE_CYCLE_START] = "cycle start", - [TCODE_LOCK_REQUEST] = "Lk req", - [TCODE_STREAM_DATA] = "async stream packet", - [TCODE_LOCK_RESPONSE] = "Lk resp", - [0xc] = "-reserved-", - [0xd] = "-reserved-", - [TCODE_LINK_INTERNAL] = "link internal", - [0xf] = "-reserved-", - }; - int tcode = async_header_get_tcode(header); - char specific[12]; - - if (likely(!(param_debug & OHCI_PARAM_DEBUG_AT_AR))) - return; - - if (unlikely(evt >= ARRAY_SIZE(evts))) - evt = 0x1f; - - if (evt == OHCI1394_evt_bus_reset) { - ohci_notice(ohci, "A%c evt_bus_reset, generation %d\n", - dir, (header[2] >> 16) & 0xff); - return; - } - - switch (tcode) { - case TCODE_WRITE_QUADLET_REQUEST: - case TCODE_READ_QUADLET_RESPONSE: - case TCODE_CYCLE_START: - snprintf(specific, sizeof(specific), " = %08x", - be32_to_cpu((__force __be32)header[3])); - break; - case TCODE_WRITE_BLOCK_REQUEST: - case TCODE_READ_BLOCK_REQUEST: - case TCODE_READ_BLOCK_RESPONSE: - case TCODE_LOCK_REQUEST: - case TCODE_LOCK_RESPONSE: - snprintf(specific, sizeof(specific), " %x,%x", - async_header_get_data_length(header), - async_header_get_extended_tcode(header)); - break; - default: - specific[0] = '\0'; - } - - switch (tcode) { - case TCODE_STREAM_DATA: - ohci_notice(ohci, "A%c %s, %s\n", - dir, evts[evt], tcodes[tcode]); - break; - case TCODE_LINK_INTERNAL: - ohci_notice(ohci, "A%c %s, PHY %08x %08x\n", - dir, evts[evt], header[1], header[2]); - break; - case TCODE_WRITE_QUADLET_REQUEST: - case TCODE_WRITE_BLOCK_REQUEST: - case TCODE_READ_QUADLET_REQUEST: - case TCODE_READ_BLOCK_REQUEST: - case TCODE_LOCK_REQUEST: - ohci_notice(ohci, - "A%c spd %x tl %02x, %04x -> %04x, %s, %s, %012llx%s\n", - dir, speed, async_header_get_tlabel(header), - async_header_get_source(header), async_header_get_destination(header), - evts[evt], tcodes[tcode], async_header_get_offset(header), specific); - break; - default: - ohci_notice(ohci, - "A%c spd %x tl %02x, %04x -> %04x, %s, %s%s\n", - dir, speed, async_header_get_tlabel(header), - async_header_get_source(header), async_header_get_destination(header), - evts[evt], tcodes[tcode], specific); - } -} - static inline void reg_write(const struct fw_ohci *ohci, int offset, u32 data) { writel(data, ohci->registers + offset); @@ -728,11 +520,6 @@ static int ohci_update_phy_reg(struct fw_card *card, int addr, return update_phy_reg(ohci, addr, clear_bits, set_bits); } -static inline dma_addr_t ar_buffer_bus(struct ar_context *ctx, unsigned int i) -{ - return page_private(ctx->pages[i]); -} - static void ar_context_link_page(struct ar_context *ctx, unsigned int index) { struct descriptor *d; @@ -754,18 +541,22 @@ static void ar_context_link_page(struct ar_context *ctx, unsigned int index) static void ar_context_release(struct ar_context *ctx) { struct device *dev = ctx->ohci->card.device; - unsigned int i; if (!ctx->buffer) return; + for (int i = 0; i < AR_BUFFERS; ++i) { + dma_addr_t dma_addr = ctx->dma_addrs[i]; + if (dma_addr) + dma_unmap_page(dev, dma_addr, PAGE_SIZE, DMA_FROM_DEVICE); + } + memset(ctx->dma_addrs, 0, sizeof(ctx->dma_addrs)); + vunmap(ctx->buffer); + ctx->buffer = NULL; - for (i = 0; i < AR_BUFFERS; i++) { - if (ctx->pages[i]) - dma_free_pages(dev, PAGE_SIZE, ctx->pages[i], - ar_buffer_bus(ctx, i), DMA_FROM_DEVICE); - } + release_pages(ctx->pages, AR_BUFFERS); + memset(ctx->pages, 0, sizeof(ctx->pages)); } static void ar_context_abort(struct ar_context *ctx, const char *error_msg) @@ -858,14 +649,12 @@ static void ar_sync_buffers_for_cpu(struct ar_context *ctx, i = ar_first_buffer_index(ctx); while (i != end_buffer_index) { - dma_sync_single_for_cpu(ctx->ohci->card.device, - ar_buffer_bus(ctx, i), - PAGE_SIZE, DMA_FROM_DEVICE); + dma_sync_single_for_cpu(ctx->ohci->card.device, ctx->dma_addrs[i], PAGE_SIZE, + DMA_FROM_DEVICE); i = ar_next_buffer_index(i); } if (end_buffer_offset > 0) - dma_sync_single_for_cpu(ctx->ohci->card.device, - ar_buffer_bus(ctx, i), + dma_sync_single_for_cpu(ctx->ohci->card.device, ctx->dma_addrs[i], end_buffer_offset, DMA_FROM_DEVICE); } @@ -954,8 +743,6 @@ static __le32 *handle_ar_packet(struct ar_context *ctx, __le32 *buffer) p.timestamp = status & 0xffff; p.generation = ohci->request_generation; - log_ar_at_event(ohci, 'R', p.speed, p.header, evt); - /* * Several controllers, notably from NEC and VIA, forget to * write ack_complete status at PHY packet reception. @@ -974,7 +761,7 @@ static __le32 *handle_ar_packet(struct ar_context *ctx, __le32 *buffer) * * Alas some chips sometimes emit bus reset packets with a * wrong generation. We set the correct generation for these - * at a slightly incorrect time (in bus_reset_work). + * at a slightly incorrect time (in handle_selfid_complete_event). */ if (evt == OHCI1394_evt_bus_reset) { if (!(ohci->quirks & QUIRK_RESET_PACKET)) @@ -1008,17 +795,16 @@ static void ar_recycle_buffers(struct ar_context *ctx, unsigned int end_buffer) i = ar_first_buffer_index(ctx); while (i != end_buffer) { - dma_sync_single_for_device(ctx->ohci->card.device, - ar_buffer_bus(ctx, i), - PAGE_SIZE, DMA_FROM_DEVICE); + dma_sync_single_for_device(ctx->ohci->card.device, ctx->dma_addrs[i], PAGE_SIZE, + DMA_FROM_DEVICE); ar_context_link_page(ctx, i); i = ar_next_buffer_index(i); } } -static void ar_context_tasklet(unsigned long data) +static void ohci_ar_context_work(struct work_struct *work) { - struct ar_context *ctx = (struct ar_context *)data; + struct ar_context *ctx = from_work(ctx, work, work); unsigned int end_buffer_index, end_buffer_offset; void *p, *end; @@ -1026,23 +812,19 @@ static void ar_context_tasklet(unsigned long data) if (!p) return; - end_buffer_index = ar_search_last_active_buffer(ctx, - &end_buffer_offset); + end_buffer_index = ar_search_last_active_buffer(ctx, &end_buffer_offset); ar_sync_buffers_for_cpu(ctx, end_buffer_index, end_buffer_offset); end = ctx->buffer + end_buffer_index * PAGE_SIZE + end_buffer_offset; if (end_buffer_index < ar_first_buffer_index(ctx)) { - /* - * The filled part of the overall buffer wraps around; handle - * all packets up to the buffer end here. If the last packet - * wraps around, its tail will be visible after the buffer end - * because the buffer start pages are mapped there again. - */ + // The filled part of the overall buffer wraps around; handle all packets up to the + // buffer end here. If the last packet wraps around, its tail will be visible after + // the buffer end because the buffer start pages are mapped there again. void *buffer_end = ctx->buffer + AR_BUFFERS * PAGE_SIZE; p = handle_ar_packets(ctx, p, buffer_end); if (p < buffer_end) goto error; - /* adjust p to point back into the actual buffer */ + // adjust p to point back into the actual buffer p -= AR_BUFFERS * PAGE_SIZE; } @@ -1057,7 +839,6 @@ static void ar_context_tasklet(unsigned long data) ar_recycle_buffers(ctx, end_buffer_index); return; - error: ctx->pointer = NULL; } @@ -1067,31 +848,57 @@ static int ar_context_init(struct ar_context *ctx, struct fw_ohci *ohci, { struct device *dev = ohci->card.device; unsigned int i; - dma_addr_t dma_addr; - struct page *pages[AR_BUFFERS + AR_WRAPAROUND_PAGES]; + struct page *pages[AR_BUFFERS + AR_WRAPAROUND_PAGES] = { NULL }; + dma_addr_t dma_addrs[AR_BUFFERS]; + void *vaddr; struct descriptor *d; ctx->regs = regs; ctx->ohci = ohci; - tasklet_init(&ctx->tasklet, ar_context_tasklet, (unsigned long)ctx); + INIT_WORK(&ctx->work, ohci_ar_context_work); - for (i = 0; i < AR_BUFFERS; i++) { - ctx->pages[i] = dma_alloc_pages(dev, PAGE_SIZE, &dma_addr, - DMA_FROM_DEVICE, GFP_KERNEL); - if (!ctx->pages[i]) - goto out_of_memory; - set_page_private(ctx->pages[i], dma_addr); - dma_sync_single_for_device(dev, dma_addr, PAGE_SIZE, - DMA_FROM_DEVICE); + // Retrieve noncontiguous pages. The descriptors for 1394 OHCI AR DMA contexts have a set + // of address and length per each. The reason to use pages is to construct contiguous + // address range in kernel virtual address space. + unsigned long nr_populated = alloc_pages_bulk(GFP_KERNEL | GFP_DMA32, AR_BUFFERS, pages); + + if (nr_populated != AR_BUFFERS) { + release_pages(pages, nr_populated); + return -ENOMEM; } - for (i = 0; i < AR_BUFFERS; i++) - pages[i] = ctx->pages[i]; + // Map the pages into contiguous kernel virtual addresses so that the packet data + // across the pages can be referred as being contiguous, especially across the last + // and first pages. for (i = 0; i < AR_WRAPAROUND_PAGES; i++) - pages[AR_BUFFERS + i] = ctx->pages[i]; - ctx->buffer = vmap(pages, ARRAY_SIZE(pages), VM_MAP, PAGE_KERNEL); - if (!ctx->buffer) - goto out_of_memory; + pages[AR_BUFFERS + i] = pages[i]; + vaddr = vmap(pages, ARRAY_SIZE(pages), VM_MAP, PAGE_KERNEL); + if (!vaddr) { + release_pages(pages, nr_populated); + return -ENOMEM; + } + + // Retrieve DMA mapping addresses for the pages. They are not contiguous. Maintain the cache + // coherency for the pages by hand. + for (i = 0; i < AR_BUFFERS; i++) { + // The dma_map_phys() with a physical address per page is available here, instead. + dma_addr_t dma_addr = dma_map_page(dev, pages[i], 0, PAGE_SIZE, DMA_FROM_DEVICE); + if (dma_mapping_error(dev, dma_addr)) + break; + dma_addrs[i] = dma_addr; + dma_sync_single_for_device(dev, dma_addr, PAGE_SIZE, DMA_FROM_DEVICE); + } + if (i < AR_BUFFERS) { + while (i-- > 0) + dma_unmap_page(dev, dma_addrs[i], PAGE_SIZE, DMA_FROM_DEVICE); + vunmap(vaddr); + release_pages(pages, nr_populated); + return -ENOMEM; + } + + memcpy(ctx->dma_addrs, dma_addrs, sizeof(ctx->dma_addrs)); + ctx->buffer = vaddr; + memcpy(ctx->pages, pages, sizeof(ctx->pages)); ctx->descriptors = ohci->misc_buffer + descriptors_offset; ctx->descriptors_bus = ohci->misc_buffer_bus + descriptors_offset; @@ -1102,17 +909,12 @@ static int ar_context_init(struct ar_context *ctx, struct fw_ohci *ohci, d->control = cpu_to_le16(DESCRIPTOR_INPUT_MORE | DESCRIPTOR_STATUS | DESCRIPTOR_BRANCH_ALWAYS); - d->data_address = cpu_to_le32(ar_buffer_bus(ctx, i)); + d->data_address = cpu_to_le32(ctx->dma_addrs[i]); d->branch_address = cpu_to_le32(ctx->descriptors_bus + ar_next_buffer_index(i) * sizeof(struct descriptor)); } return 0; - -out_of_memory: - ar_context_release(ctx); - - return -ENOMEM; } static void ar_context_run(struct ar_context *ctx) @@ -1181,16 +983,16 @@ static void context_retire_descriptors(struct context *ctx) } } -static void context_tasklet(unsigned long data) +static void ohci_at_context_work(struct work_struct *work) { - struct context *ctx = (struct context *) data; + struct at_context *ctx = from_work(ctx, work, work); - context_retire_descriptors(ctx); + context_retire_descriptors(&ctx->context); } static void ohci_isoc_context_work(struct work_struct *work) { - struct fw_iso_context *base = container_of(work, struct fw_iso_context, work); + struct fw_iso_context *base = from_work(base, work, work); struct iso_context *isoc_ctx = container_of(base, struct iso_context, base); context_retire_descriptors(&isoc_ctx->context); @@ -1248,7 +1050,6 @@ static int context_init(struct context *ctx, struct fw_ohci *ohci, ctx->buffer_tail = list_entry(ctx->buffer_list.next, struct descriptor_buffer, list); - tasklet_init(&ctx->tasklet, context_tasklet, (unsigned long)ctx); ctx->callback = callback; /* @@ -1388,17 +1189,17 @@ struct driver_data { * Must always be called with the ochi->lock held to ensure proper * generation handling and locking around packet queue manipulation. */ -static int at_context_queue_packet(struct context *ctx, - struct fw_packet *packet) +static int at_context_queue_packet(struct at_context *ctx, struct fw_packet *packet) { - struct fw_ohci *ohci = ctx->ohci; + struct context *context = &ctx->context; + struct fw_ohci *ohci = context->ohci; dma_addr_t d_bus, payload_bus; struct driver_data *driver_data; struct descriptor *d, *last; __le32 *header; int z, tcode; - d = context_get_descriptors(ctx, 4, &d_bus); + d = context_get_descriptors(context, 4, &d_bus); if (d == NULL) { packet->ack = RCODE_SEND_ERROR; return -1; @@ -1428,7 +1229,7 @@ static int at_context_queue_packet(struct context *ctx, ohci1394_at_data_set_destination_id(header, async_header_get_destination(packet->header)); - if (ctx == &ctx->ohci->at_response_ctx) { + if (ctx == &ohci->at_response_ctx) { ohci1394_at_data_set_rcode(header, async_header_get_rcode(packet->header)); } else { ohci1394_at_data_set_destination_offset(header, @@ -1517,37 +1318,50 @@ static int at_context_queue_packet(struct context *ctx, return -1; } - context_append(ctx, d, z, 4 - z); + context_append(context, d, z, 4 - z); - if (ctx->running) - reg_write(ohci, CONTROL_SET(ctx->regs), CONTEXT_WAKE); + if (context->running) + reg_write(ohci, CONTROL_SET(context->regs), CONTEXT_WAKE); else - context_run(ctx, 0); + context_run(context, 0); return 0; } -static void at_context_flush(struct context *ctx) +static void at_context_flush(struct at_context *ctx) { - tasklet_disable(&ctx->tasklet); + // Avoid dead lock due to programming mistake. + if (WARN_ON_ONCE(current_work() == &ctx->work)) + return; + + disable_work_sync(&ctx->work); - ctx->flushing = true; - context_tasklet((unsigned long)ctx); - ctx->flushing = false; + WRITE_ONCE(ctx->flushing, true); + ohci_at_context_work(&ctx->work); + WRITE_ONCE(ctx->flushing, false); - tasklet_enable(&ctx->tasklet); + enable_work(&ctx->work); +} + +static int find_fw_device(struct device *dev, const void *data) +{ + struct fw_device *device = fw_device(dev); + const u32 *params = data; + + return (device->generation == params[0]) && (device->node_id == params[1]); } static int handle_at_packet(struct context *context, struct descriptor *d, struct descriptor *last) { + struct at_context *ctx = container_of(context, struct at_context, context); + struct fw_ohci *ohci = ctx->context.ohci; struct driver_data *driver_data; struct fw_packet *packet; - struct fw_ohci *ohci = context->ohci; int evt; - if (last->transfer_status == 0 && !context->flushing) + if (last->transfer_status == 0 && !READ_ONCE(ctx->flushing)) /* This descriptor isn't done yet, stop iteration. */ return 0; @@ -1564,8 +1378,6 @@ static int handle_at_packet(struct context *context, evt = le16_to_cpu(last->transfer_status) & 0x1f; packet->timestamp = le16_to_cpu(last->res_count); - log_ar_at_event(ohci, 'T', packet->speed, packet->header, evt); - switch (evt) { case OHCI1394_evt_timeout: /* Async response transmit timed out. */ @@ -1581,7 +1393,7 @@ static int handle_at_packet(struct context *context, break; case OHCI1394_evt_missing_ack: - if (context->flushing) + if (READ_ONCE(ctx->flushing)) packet->ack = RCODE_GENERATION; else { /* @@ -1603,13 +1415,34 @@ static int handle_at_packet(struct context *context, break; case OHCI1394_evt_no_status: - if (context->flushing) { + if (READ_ONCE(ctx->flushing)) { packet->ack = RCODE_GENERATION; break; } fallthrough; default: + if (unlikely(evt == 0x10)) { + u32 params[2] = { + packet->generation, + async_header_get_destination(packet->header), + }; + struct device *dev; + + fw_card_get(&ohci->card); + dev = device_find_child(ohci->card.device, (const void *)params, find_fw_device); + fw_card_put(&ohci->card); + if (dev) { + struct fw_device *device = fw_device(dev); + int quirks = READ_ONCE(device->quirks); + + put_device(dev); + if (quirks & FW_DEVICE_QUIRK_ACK_PACKET_WITH_INVALID_PENDING_CODE) { + packet->ack = ACK_PENDING; + break; + } + } + } packet->ack = RCODE_SEND_ERROR; break; } @@ -1700,13 +1533,14 @@ static void handle_local_lock(struct fw_ohci *ohci, fw_core_handle_response(&ohci->card, &response); } -static void handle_local_request(struct context *ctx, struct fw_packet *packet) +static void handle_local_request(struct at_context *ctx, struct fw_packet *packet) { + struct fw_ohci *ohci = ctx->context.ohci; u64 offset, csr; - if (ctx == &ctx->ohci->at_request_ctx) { + if (ctx == &ohci->at_request_ctx) { packet->ack = ACK_PENDING; - packet->callback(packet, &ctx->ohci->card, packet->ack); + packet->callback(packet, &ohci->card, packet->ack); } offset = async_header_get_offset(packet->header); @@ -1714,60 +1548,80 @@ static void handle_local_request(struct context *ctx, struct fw_packet *packet) /* Handle config rom reads. */ if (csr >= CSR_CONFIG_ROM && csr < CSR_CONFIG_ROM_END) - handle_local_rom(ctx->ohci, packet, csr); + handle_local_rom(ohci, packet, csr); else switch (csr) { case CSR_BUS_MANAGER_ID: case CSR_BANDWIDTH_AVAILABLE: case CSR_CHANNELS_AVAILABLE_HI: case CSR_CHANNELS_AVAILABLE_LO: - handle_local_lock(ctx->ohci, packet, csr); + handle_local_lock(ohci, packet, csr); break; default: - if (ctx == &ctx->ohci->at_request_ctx) - fw_core_handle_request(&ctx->ohci->card, packet); + if (ctx == &ohci->at_request_ctx) + fw_core_handle_request(&ohci->card, packet); else - fw_core_handle_response(&ctx->ohci->card, packet); + fw_core_handle_response(&ohci->card, packet); break; } - if (ctx == &ctx->ohci->at_response_ctx) { + if (ctx == &ohci->at_response_ctx) { packet->ack = ACK_COMPLETE; - packet->callback(packet, &ctx->ohci->card, packet->ack); + packet->callback(packet, &ohci->card, packet->ack); } } -static void at_context_transmit(struct context *ctx, struct fw_packet *packet) +static void at_context_transmit(struct at_context *ctx, struct fw_packet *packet) { + struct fw_ohci *ohci = ctx->context.ohci; unsigned long flags; int ret; - spin_lock_irqsave(&ctx->ohci->lock, flags); + spin_lock_irqsave(&ohci->lock, flags); - if (async_header_get_destination(packet->header) == ctx->ohci->node_id && - ctx->ohci->generation == packet->generation) { - spin_unlock_irqrestore(&ctx->ohci->lock, flags); + if (async_header_get_destination(packet->header) == ohci->node_id && + ohci->generation == packet->generation) { + spin_unlock_irqrestore(&ohci->lock, flags); // Timestamping on behalf of the hardware. - packet->timestamp = cycle_time_to_ohci_tstamp(get_cycle_time(ctx->ohci)); + packet->timestamp = cycle_time_to_ohci_tstamp(get_cycle_time(ohci)); handle_local_request(ctx, packet); return; } ret = at_context_queue_packet(ctx, packet); - spin_unlock_irqrestore(&ctx->ohci->lock, flags); + spin_unlock_irqrestore(&ohci->lock, flags); if (ret < 0) { // Timestamping on behalf of the hardware. - packet->timestamp = cycle_time_to_ohci_tstamp(get_cycle_time(ctx->ohci)); + packet->timestamp = cycle_time_to_ohci_tstamp(get_cycle_time(ohci)); - packet->callback(packet, &ctx->ohci->card, packet->ack); + packet->callback(packet, &ohci->card, packet->ack); } } static void detect_dead_context(struct fw_ohci *ohci, const char *name, unsigned int regs) { + static const char *const evts[] = { + [0x00] = "evt_no_status", [0x01] = "-reserved-", + [0x02] = "evt_long_packet", [0x03] = "evt_missing_ack", + [0x04] = "evt_underrun", [0x05] = "evt_overrun", + [0x06] = "evt_descriptor_read", [0x07] = "evt_data_read", + [0x08] = "evt_data_write", [0x09] = "evt_bus_reset", + [0x0a] = "evt_timeout", [0x0b] = "evt_tcode_err", + [0x0c] = "-reserved-", [0x0d] = "-reserved-", + [0x0e] = "evt_unknown", [0x0f] = "evt_flushed", + [0x10] = "-reserved-", [0x11] = "ack_complete", + [0x12] = "ack_pending ", [0x13] = "-reserved-", + [0x14] = "ack_busy_X", [0x15] = "ack_busy_A", + [0x16] = "ack_busy_B", [0x17] = "-reserved-", + [0x18] = "-reserved-", [0x19] = "-reserved-", + [0x1a] = "-reserved-", [0x1b] = "ack_tardy", + [0x1c] = "-reserved-", [0x1d] = "ack_data_error", + [0x1e] = "ack_type_error", [0x1f] = "-reserved-", + [0x20] = "pending/cancelled", + }; u32 ctl; ctl = reg_read(ohci, CONTROL_SET(regs)); @@ -2026,10 +1880,9 @@ static int find_and_insert_self_id(struct fw_ohci *ohci, int self_id_count) return self_id_count; } -static void bus_reset_work(struct work_struct *work) +static irqreturn_t handle_selfid_complete_event(int irq, void *data) { - struct fw_ohci *ohci = - container_of(work, struct fw_ohci, bus_reset_work); + struct fw_ohci *ohci = data; int self_id_count, generation, new_generation, i, j; u32 reg, quadlet; void *free_rom = NULL; @@ -2040,11 +1893,11 @@ static void bus_reset_work(struct work_struct *work) if (!(reg & OHCI1394_NodeID_idValid)) { ohci_notice(ohci, "node ID not valid, new bus reset in progress\n"); - return; + goto end; } if ((reg & OHCI1394_NodeID_nodeNumber) == 63) { ohci_notice(ohci, "malconfigured bus\n"); - return; + goto end; } ohci->node_id = reg & (OHCI1394_NodeID_busNumber | OHCI1394_NodeID_nodeNumber); @@ -2058,8 +1911,11 @@ static void bus_reset_work(struct work_struct *work) reg = reg_read(ohci, OHCI1394_SelfIDCount); if (ohci1394_self_id_count_is_error(reg)) { ohci_notice(ohci, "self ID receive error\n"); - return; + goto end; } + + trace_self_id_complete(ohci->card.index, reg, ohci->self_id, has_be_header_quirk(ohci)); + /* * The count in the SelfIDCount register is the number of * bytes in the self ID receive buffer. Since we also receive @@ -2070,7 +1926,7 @@ static void bus_reset_work(struct work_struct *work) if (self_id_count > 252) { ohci_notice(ohci, "bad selfIDSize (%08x)\n", reg); - return; + goto end; } quadlet = cond_le32_to_cpu(ohci->self_id[0], has_be_header_quirk(ohci)); @@ -2097,7 +1953,7 @@ static void bus_reset_work(struct work_struct *work) ohci_notice(ohci, "bad self ID %d/%d (%08x != ~%08x)\n", j, self_id_count, id, id2); - return; + goto end; } ohci->self_id_buffer[j] = id; } @@ -2107,13 +1963,13 @@ static void bus_reset_work(struct work_struct *work) if (self_id_count < 0) { ohci_notice(ohci, "could not construct local self ID\n"); - return; + goto end; } } if (self_id_count == 0) { ohci_notice(ohci, "no self IDs\n"); - return; + goto end; } rmb(); @@ -2135,14 +1991,14 @@ static void bus_reset_work(struct work_struct *work) new_generation = ohci1394_self_id_count_get_generation(reg); if (new_generation != generation) { ohci_notice(ohci, "new bus reset, discarding self ids\n"); - return; + goto end; } // FIXME: Document how the locking works. scoped_guard(spinlock_irq, &ohci->lock) { ohci->generation = -1; // prevent AT packet queueing - context_stop(&ohci->at_request_ctx); - context_stop(&ohci->at_response_ctx); + context_stop(&ohci->at_request_ctx.context); + context_stop(&ohci->at_response_ctx.context); } /* @@ -2192,12 +2048,12 @@ static void bus_reset_work(struct work_struct *work) if (free_rom) dmam_free_coherent(ohci->card.device, CONFIG_ROM_SIZE, free_rom, free_rom_bus); - log_selfids(ohci, generation, self_id_count); - fw_core_handle_bus_reset(&ohci->card, ohci->node_id, generation, self_id_count, ohci->self_id_buffer, ohci->csr_state_setclear_abdicate); ohci->csr_state_setclear_abdicate = false; +end: + return IRQ_HANDLED; } static irqreturn_t irq_handler(int irq, void *data) @@ -2211,11 +2067,6 @@ static irqreturn_t irq_handler(int irq, void *data) if (!event || !~event) return IRQ_NONE; - if (unlikely(param_debug > 0)) { - dev_notice_ratelimited(ohci->card.device, - "The debug parameter is superseded by tracepoints events, and deprecated."); - } - /* * busReset and postedWriteErr events must not be cleared yet * (OHCI 1.1 clauses 7.2.3.2 and 13.2.8.1) @@ -2223,32 +2074,22 @@ static irqreturn_t irq_handler(int irq, void *data) reg_write(ohci, OHCI1394_IntEventClear, event & ~(OHCI1394_busReset | OHCI1394_postedWriteErr)); trace_irqs(ohci->card.index, event); - log_irqs(ohci, event); - // The flag is masked again at bus_reset_work() scheduled by selfID event. + + // The flag is masked again at handle_selfid_complete_event() scheduled by selfID event. if (event & OHCI1394_busReset) reg_write(ohci, OHCI1394_IntMaskClear, OHCI1394_busReset); - if (event & OHCI1394_selfIDComplete) { - if (trace_self_id_complete_enabled()) { - u32 reg = reg_read(ohci, OHCI1394_SelfIDCount); - - trace_self_id_complete(ohci->card.index, reg, ohci->self_id, - has_be_header_quirk(ohci)); - } - queue_work(selfid_workqueue, &ohci->bus_reset_work); - } - if (event & OHCI1394_RQPkt) - tasklet_schedule(&ohci->ar_request_ctx.tasklet); + queue_work(ohci->card.async_wq, &ohci->ar_request_ctx.work); if (event & OHCI1394_RSPkt) - tasklet_schedule(&ohci->ar_response_ctx.tasklet); + queue_work(ohci->card.async_wq, &ohci->ar_response_ctx.work); if (event & OHCI1394_reqTxComplete) - tasklet_schedule(&ohci->at_request_ctx.tasklet); + queue_work(ohci->card.async_wq, &ohci->at_request_ctx.work); if (event & OHCI1394_respTxComplete) - tasklet_schedule(&ohci->at_response_ctx.tasklet); + queue_work(ohci->card.async_wq, &ohci->at_response_ctx.work); if (event & OHCI1394_isochRx) { iso_event = reg_read(ohci, OHCI1394_IsoRecvIntEventClear); @@ -2308,7 +2149,10 @@ static irqreturn_t irq_handler(int irq, void *data) } else flush_writes(ohci); - return IRQ_HANDLED; + if (event & OHCI1394_selfIDComplete) + return IRQ_WAKE_THREAD; + else + return IRQ_HANDLED; } static int software_reset(struct fw_ohci *ohci) @@ -2528,7 +2372,7 @@ static int ohci_enable(struct fw_card *card, * They shouldn't do that in this initial case where the link * isn't enabled. This means we have to use the same * workaround here, setting the bus header to 0 and then write - * the right values in the bus reset tasklet. + * the right values in the bus reset work item. */ if (config_rom) { @@ -2588,6 +2432,41 @@ static int ohci_enable(struct fw_card *card, return 0; } +static void ohci_disable(struct fw_card *card) +{ + struct pci_dev *pdev = to_pci_dev(card->device); + struct fw_ohci *ohci = pci_get_drvdata(pdev); + int i, irq = pci_irq_vector(pdev, 0); + + // If the removal is happening from the suspend state, LPS won't be enabled and host + // registers (eg., IntMaskClear) won't be accessible. + if (!(reg_read(ohci, OHCI1394_HCControlSet) & OHCI1394_HCControl_LPS)) + return; + + reg_write(ohci, OHCI1394_IntMaskClear, ~0); + flush_writes(ohci); + + if (irq >= 0) + synchronize_irq(irq); + + flush_work(&ohci->ar_request_ctx.work); + flush_work(&ohci->ar_response_ctx.work); + flush_work(&ohci->at_request_ctx.work); + flush_work(&ohci->at_response_ctx.work); + + for (i = 0; i < ohci->n_ir; ++i) { + if (!(ohci->ir_context_mask & BIT(i))) + flush_work(&ohci->ir_context_list[i].base.work); + } + for (i = 0; i < ohci->n_it; ++i) { + if (!(ohci->it_context_mask & BIT(i))) + flush_work(&ohci->it_context_list[i].base.work); + } + + at_context_flush(&ohci->at_request_ctx); + at_context_flush(&ohci->at_response_ctx); +} + static int ohci_set_config_rom(struct fw_card *card, const __be32 *config_rom, size_t length) { @@ -2617,11 +2496,11 @@ static int ohci_set_config_rom(struct fw_card *card, * during the atomic update, even on little endian * architectures. The workaround we use is to put a 0 in the * header quadlet; 0 is endian agnostic and means that the - * config rom isn't ready yet. In the bus reset tasklet we + * config rom isn't ready yet. In the bus reset work item we * then set up the real values for the two registers. * * We use ohci->lock to avoid racing with the code that sets - * ohci->next_config_rom to NULL (see bus_reset_work). + * ohci->next_config_rom to NULL (see handle_selfid_complete_event). */ next_config_rom = dmam_alloc_coherent(ohci->card.device, CONFIG_ROM_SIZE, @@ -2659,7 +2538,7 @@ static int ohci_set_config_rom(struct fw_card *card, /* * Now initiate a bus reset to have the changes take * effect. We clean up the old config rom memory and DMA - * mappings in the bus reset tasklet, since the OHCI + * mappings in the bus reset work item, since the OHCI * controller could need to access it before the bus reset * takes effect. */ @@ -2686,11 +2565,14 @@ static void ohci_send_response(struct fw_card *card, struct fw_packet *packet) static int ohci_cancel_packet(struct fw_card *card, struct fw_packet *packet) { struct fw_ohci *ohci = fw_ohci(card); - struct context *ctx = &ohci->at_request_ctx; + struct at_context *ctx = &ohci->at_request_ctx; struct driver_data *driver_data = packet->driver_data; int ret = -ENOENT; - tasklet_disable_in_atomic(&ctx->tasklet); + // Avoid dead lock due to programming mistake. + if (WARN_ON_ONCE(current_work() == &ctx->work)) + return 0; + disable_work_sync(&ctx->work); if (packet->ack != 0) goto out; @@ -2699,7 +2581,6 @@ static int ohci_cancel_packet(struct fw_card *card, struct fw_packet *packet) dma_unmap_single(ohci->card.device, packet->payload_bus, packet->payload_length, DMA_TO_DEVICE); - log_ar_at_event(ohci, 'T', packet->speed, packet->header, 0x20); driver_data->packet = NULL; packet->ack = RCODE_CANCELLED; @@ -2709,7 +2590,7 @@ static int ohci_cancel_packet(struct fw_card *card, struct fw_packet *packet) packet->callback(packet, &ohci->card, packet->ack); ret = 0; out: - tasklet_enable(&ctx->tasklet); + enable_work(&ctx->work); return ret; } @@ -2860,29 +2741,28 @@ static void ohci_write_csr(struct fw_card *card, int csr_offset, u32 value) static void flush_iso_completions(struct iso_context *ctx, enum fw_iso_context_completions_cause cause) { - trace_isoc_inbound_single_completions(&ctx->base, ctx->last_timestamp, cause, ctx->header, - ctx->header_length); - trace_isoc_outbound_completions(&ctx->base, ctx->last_timestamp, cause, ctx->header, - ctx->header_length); + trace_isoc_inbound_single_completions(&ctx->base, ctx->sc.last_timestamp, cause, + ctx->sc.header, ctx->sc.header_length); + trace_isoc_outbound_completions(&ctx->base, ctx->sc.last_timestamp, cause, ctx->sc.header, + ctx->sc.header_length); - ctx->base.callback.sc(&ctx->base, ctx->last_timestamp, - ctx->header_length, ctx->header, - ctx->base.callback_data); - ctx->header_length = 0; + ctx->base.callback.sc(&ctx->base, ctx->sc.last_timestamp, ctx->sc.header_length, + ctx->sc.header, ctx->base.callback_data); + ctx->sc.header_length = 0; } static void copy_iso_headers(struct iso_context *ctx, const u32 *dma_hdr) { u32 *ctx_hdr; - if (ctx->header_length + ctx->base.header_size > PAGE_SIZE) { - if (ctx->base.drop_overflow_headers) + if (ctx->sc.header_length + ctx->base.header_size > ctx->base.header_storage_size) { + if (ctx->base.flags & FW_ISO_CONTEXT_FLAG_DROP_OVERFLOW_HEADERS) return; flush_iso_completions(ctx, FW_ISO_CONTEXT_COMPLETIONS_CAUSE_HEADER_OVERFLOW); } - ctx_hdr = ctx->header + ctx->header_length; - ctx->last_timestamp = (u16)le32_to_cpu((__force __le32)dma_hdr[0]); + ctx_hdr = ctx->sc.header + ctx->sc.header_length; + ctx->sc.last_timestamp = (u16)le32_to_cpu((__force __le32)dma_hdr[0]); /* * The two iso header quadlets are byteswapped to little @@ -2895,7 +2775,7 @@ static void copy_iso_headers(struct iso_context *ctx, const u32 *dma_hdr) ctx_hdr[1] = swab32(dma_hdr[0]); /* timestamp */ if (ctx->base.header_size > 8) memcpy(&ctx_hdr[2], &dma_hdr[2], ctx->base.header_size - 8); - ctx->header_length += ctx->base.header_size; + ctx->sc.header_length += ctx->base.header_size; } static int handle_ir_packet_per_buffer(struct context *context, @@ -2948,8 +2828,8 @@ static int handle_ir_buffer_fill(struct context *context, buffer_dma = le32_to_cpu(last->data_address); if (completed > 0) { - ctx->mc_buffer_bus = buffer_dma; - ctx->mc_completed = completed; + ctx->mc.buffer_bus = buffer_dma; + ctx->mc.completed = completed; } if (res_count != 0) @@ -2968,7 +2848,7 @@ static int handle_ir_buffer_fill(struct context *context, ctx->base.callback.mc(&ctx->base, buffer_dma + completed, ctx->base.callback_data); - ctx->mc_completed = 0; + ctx->mc.completed = 0; } return 1; @@ -2977,17 +2857,16 @@ static int handle_ir_buffer_fill(struct context *context, static void flush_ir_buffer_fill(struct iso_context *ctx) { dma_sync_single_range_for_cpu(ctx->context.ohci->card.device, - ctx->mc_buffer_bus & PAGE_MASK, - ctx->mc_buffer_bus & ~PAGE_MASK, - ctx->mc_completed, DMA_FROM_DEVICE); + ctx->mc.buffer_bus & PAGE_MASK, + ctx->mc.buffer_bus & ~PAGE_MASK, + ctx->mc.completed, DMA_FROM_DEVICE); - trace_isoc_inbound_multiple_completions(&ctx->base, ctx->mc_completed, + trace_isoc_inbound_multiple_completions(&ctx->base, ctx->mc.completed, FW_ISO_CONTEXT_COMPLETIONS_CAUSE_FLUSH); - ctx->base.callback.mc(&ctx->base, - ctx->mc_buffer_bus + ctx->mc_completed, + ctx->base.callback.mc(&ctx->base, ctx->mc.buffer_bus + ctx->mc.completed, ctx->base.callback_data); - ctx->mc_completed = 0; + ctx->mc.completed = 0; } static inline void sync_it_packet_for_cpu(struct context *context, @@ -3045,18 +2924,18 @@ static int handle_it_packet(struct context *context, sync_it_packet_for_cpu(context, d); - if (ctx->header_length + 4 > PAGE_SIZE) { - if (ctx->base.drop_overflow_headers) + if (ctx->sc.header_length + 4 > ctx->base.header_storage_size) { + if (ctx->base.flags & FW_ISO_CONTEXT_FLAG_DROP_OVERFLOW_HEADERS) return 1; flush_iso_completions(ctx, FW_ISO_CONTEXT_COMPLETIONS_CAUSE_HEADER_OVERFLOW); } - ctx_hdr = ctx->header + ctx->header_length; - ctx->last_timestamp = le16_to_cpu(last->res_count); + ctx_hdr = ctx->sc.header + ctx->sc.header_length; + ctx->sc.last_timestamp = le16_to_cpu(last->res_count); /* Present this value as big-endian to match the receive code */ *ctx_hdr = cpu_to_be32((le16_to_cpu(pd->transfer_status) << 16) | le16_to_cpu(pd->res_count)); - ctx->header_length += 4; + ctx->sc.header_length += 4; if (last->control & cpu_to_le16(DESCRIPTOR_IRQ_ALWAYS)) flush_iso_completions(ctx, FW_ISO_CONTEXT_COMPLETIONS_CAUSE_INTERRUPT); @@ -3075,10 +2954,11 @@ static void set_multichannel_mask(struct fw_ohci *ohci, u64 channels) ohci->mc_channels = channels; } -static struct fw_iso_context *ohci_allocate_iso_context(struct fw_card *card, - int type, int channel, size_t header_size) +static struct fw_iso_context *ohci_allocate_iso_context(struct fw_card *card, int type, int channel, + size_t header_size, size_t header_storage_size) { struct fw_ohci *ohci = fw_ohci(card); + void *header __free(kvfree) = NULL; struct iso_context *ctx; descriptor_callback_t callback; u64 *channels; @@ -3133,26 +3013,29 @@ static struct fw_iso_context *ohci_allocate_iso_context(struct fw_card *card, } memset(ctx, 0, sizeof(*ctx)); - ctx->header_length = 0; - ctx->header = (void *) __get_free_page(GFP_KERNEL); - if (ctx->header == NULL) { - ret = -ENOMEM; - goto out; + + if (type != FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL) { + ctx->sc.header_length = 0; + header = kvmalloc(header_storage_size, GFP_KERNEL); + if (!header) { + ret = -ENOMEM; + goto out; + } } + ret = context_init(&ctx->context, ohci, regs, callback); if (ret < 0) - goto out_with_header; + goto out; fw_iso_context_init_work(&ctx->base, ohci_isoc_context_work); - if (type == FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL) { + if (type != FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL) { + ctx->sc.header = no_free_ptr(header); + } else { set_multichannel_mask(ohci, 0); - ctx->mc_completed = 0; + ctx->mc.completed = 0; } return &ctx->base; - - out_with_header: - free_page((unsigned long)ctx->header); out: scoped_guard(spinlock_irq, &ohci->lock) { switch (type) { @@ -3252,7 +3135,11 @@ static void ohci_free_iso_context(struct fw_iso_context *base) ohci_stop_iso(base); context_release(&ctx->context); - free_page((unsigned long)ctx->header); + + if (base->type != FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL) { + kvfree(ctx->sc.header); + ctx->sc.header = NULL; + } guard(spinlock_irqsave)(&ohci->lock); @@ -3301,8 +3188,7 @@ static int ohci_set_iso_channels(struct fw_iso_context *base, u64 *channels) } } -#ifdef CONFIG_PM -static void ohci_resume_iso_dma(struct fw_ohci *ohci) +static void __maybe_unused ohci_resume_iso_dma(struct fw_ohci *ohci) { int i; struct iso_context *ctx; @@ -3319,7 +3205,6 @@ static void ohci_resume_iso_dma(struct fw_ohci *ohci) ohci_start_iso(&ctx->base, 0, ctx->sync, ctx->tags); } } -#endif static int queue_iso_transmit(struct iso_context *ctx, struct fw_iso_packet *packet, @@ -3329,7 +3214,7 @@ static int queue_iso_transmit(struct iso_context *ctx, struct descriptor *d, *last, *pd; struct fw_iso_packet *p; __le32 *header; - dma_addr_t d_bus, page_bus; + dma_addr_t d_bus; u32 z, header_z, payload_z, irq; u32 payload_index, payload_end_index, next_page_index; int page, end_page, i, length, offset; @@ -3399,11 +3284,11 @@ static int queue_iso_transmit(struct iso_context *ctx, min(next_page_index, payload_end_index) - payload_index; pd[i].req_count = cpu_to_le16(length); - page_bus = page_private(buffer->pages[page]); - pd[i].data_address = cpu_to_le32(page_bus + offset); + dma_addr_t dma_addr = buffer->dma_addrs[page]; + pd[i].data_address = cpu_to_le32(dma_addr + offset); dma_sync_single_range_for_device(ctx->context.ohci->card.device, - page_bus, offset, length, + dma_addr, offset, length, DMA_TO_DEVICE); payload_index += length; @@ -3432,7 +3317,7 @@ static int queue_iso_packet_per_buffer(struct iso_context *ctx, { struct device *device = ctx->context.ohci->card.device; struct descriptor *d, *pd; - dma_addr_t d_bus, page_bus; + dma_addr_t d_bus; u32 z, header_z, rest; int i, j, length; int page, offset, packet_count, header_size, payload_per_buffer; @@ -3482,10 +3367,10 @@ static int queue_iso_packet_per_buffer(struct iso_context *ctx, pd->res_count = pd->req_count; pd->transfer_status = 0; - page_bus = page_private(buffer->pages[page]); - pd->data_address = cpu_to_le32(page_bus + offset); + dma_addr_t dma_addr = buffer->dma_addrs[page]; + pd->data_address = cpu_to_le32(dma_addr + offset); - dma_sync_single_range_for_device(device, page_bus, + dma_sync_single_range_for_device(device, dma_addr, offset, length, DMA_FROM_DEVICE); @@ -3512,7 +3397,7 @@ static int queue_iso_buffer_fill(struct iso_context *ctx, unsigned long payload) { struct descriptor *d; - dma_addr_t d_bus, page_bus; + dma_addr_t d_bus; int page, offset, rest, z, i, length; page = payload >> PAGE_SHIFT; @@ -3545,11 +3430,11 @@ static int queue_iso_buffer_fill(struct iso_context *ctx, d->res_count = d->req_count; d->transfer_status = 0; - page_bus = page_private(buffer->pages[page]); - d->data_address = cpu_to_le32(page_bus + offset); + dma_addr_t dma_addr = buffer->dma_addrs[page]; + d->data_address = cpu_to_le32(dma_addr + offset); dma_sync_single_range_for_device(ctx->context.ohci->card.device, - page_bus, offset, length, + dma_addr, offset, length, DMA_FROM_DEVICE); rest -= length; @@ -3602,11 +3487,11 @@ static int ohci_flush_iso_completions(struct fw_iso_context *base) switch (base->type) { case FW_ISO_CONTEXT_TRANSMIT: case FW_ISO_CONTEXT_RECEIVE: - if (ctx->header_length != 0) + if (ctx->sc.header_length != 0) flush_iso_completions(ctx, FW_ISO_CONTEXT_COMPLETIONS_CAUSE_FLUSH); break; case FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL: - if (ctx->mc_completed != 0) + if (ctx->mc.completed != 0) flush_ir_buffer_fill(ctx); break; default: @@ -3622,6 +3507,7 @@ static int ohci_flush_iso_completions(struct fw_iso_context *base) static const struct fw_card_driver ohci_driver = { .enable = ohci_enable, + .disable = ohci_disable, .read_phy_reg = ohci_read_phy_reg, .update_phy_reg = ohci_update_phy_reg, .set_config_rom = ohci_set_config_rom, @@ -3691,7 +3577,6 @@ static int pci_probe(struct pci_dev *dev, u32 bus_options, max_receive, link_speed, version; u64 guid; int i, flags, irq, err; - size_t size; if (dev->vendor == PCI_VENDOR_ID_PINNACLE_SYSTEMS) { dev_err(&dev->dev, "Pinnacle MovieBoard is not yet supported\n"); @@ -3718,8 +3603,6 @@ static int pci_probe(struct pci_dev *dev, spin_lock_init(&ohci->lock); mutex_init(&ohci->phy_reg_mutex); - INIT_WORK(&ohci->bus_reset_work, bus_reset_work); - if (!(pci_resource_flags(dev, 0) & IORESOURCE_MEM) || pci_resource_len(dev, 0) < OHCI1394_REGISTER_SIZE) { ohci_err(ohci, "invalid MMIO resource\n"); @@ -3769,15 +3652,17 @@ static int pci_probe(struct pci_dev *dev, if (err < 0) return err; - err = context_init(&ohci->at_request_ctx, ohci, + err = context_init(&ohci->at_request_ctx.context, ohci, OHCI1394_AsReqTrContextControlSet, handle_at_packet); if (err < 0) return err; + INIT_WORK(&ohci->at_request_ctx.work, ohci_at_context_work); - err = context_init(&ohci->at_response_ctx, ohci, + err = context_init(&ohci->at_response_ctx.context, ohci, OHCI1394_AsRspTrContextControlSet, handle_at_packet); if (err < 0) return err; + INIT_WORK(&ohci->at_response_ctx.work, ohci_at_context_work); reg_write(ohci, OHCI1394_IsoRecvIntMaskSet, ~0); ohci->ir_context_channels = ~0ULL; @@ -3785,8 +3670,7 @@ static int pci_probe(struct pci_dev *dev, reg_write(ohci, OHCI1394_IsoRecvIntMaskClear, ~0); ohci->ir_context_mask = ohci->ir_context_support; ohci->n_ir = hweight32(ohci->ir_context_mask); - size = sizeof(struct iso_context) * ohci->n_ir; - ohci->ir_context_list = devm_kzalloc(&dev->dev, size, GFP_KERNEL); + ohci->ir_context_list = devm_kcalloc(&dev->dev, ohci->n_ir, sizeof(struct iso_context), GFP_KERNEL); if (!ohci->ir_context_list) return -ENOMEM; @@ -3800,8 +3684,7 @@ static int pci_probe(struct pci_dev *dev, reg_write(ohci, OHCI1394_IsoXmitIntMaskClear, ~0); ohci->it_context_mask = ohci->it_context_support; ohci->n_it = hweight32(ohci->it_context_mask); - size = sizeof(struct iso_context) * ohci->n_it; - ohci->it_context_list = devm_kzalloc(&dev->dev, size, GFP_KERNEL); + ohci->it_context_list = devm_kcalloc(&dev->dev, ohci->n_it, sizeof(struct iso_context), GFP_KERNEL); if (!ohci->it_context_list) return -ENOMEM; @@ -3826,7 +3709,9 @@ static int pci_probe(struct pci_dev *dev, goto fail_msi; } - err = request_threaded_irq(irq, irq_handler, NULL, + // IRQF_ONESHOT is not applied so that any events are handled in the hardIRQ handler during + // invoking the threaded IRQ handler for SelfIDComplete event. + err = request_threaded_irq(irq, irq_handler, handle_selfid_complete_event, pci_dev_msi_enabled(dev) ? 0 : IRQF_SHARED, ohci_driver_name, ohci); if (err < 0) { @@ -3862,22 +3747,8 @@ static void pci_remove(struct pci_dev *dev) struct fw_ohci *ohci = pci_get_drvdata(dev); int irq; - /* - * If the removal is happening from the suspend state, LPS won't be - * enabled and host registers (eg., IntMaskClear) won't be accessible. - */ - if (reg_read(ohci, OHCI1394_HCControlSet) & OHCI1394_HCControl_LPS) { - reg_write(ohci, OHCI1394_IntMaskClear, ~0); - flush_writes(ohci); - } - cancel_work_sync(&ohci->bus_reset_work); fw_core_remove_card(&ohci->card); - /* - * FIXME: Fail all pending packets here, now that the upper - * layers can't queue any more. - */ - software_reset(ohci); irq = pci_irq_vector(dev, 0); @@ -3888,39 +3759,25 @@ static void pci_remove(struct pci_dev *dev) dev_notice(&dev->dev, "removing fw-ohci device\n"); } -#ifdef CONFIG_PM -static int pci_suspend(struct pci_dev *dev, pm_message_t state) +static int __maybe_unused pci_suspend(struct device *dev) { - struct fw_ohci *ohci = pci_get_drvdata(dev); - int err; + struct pci_dev *pdev = to_pci_dev(dev); + struct fw_ohci *ohci = pci_get_drvdata(pdev); software_reset(ohci); - err = pci_save_state(dev); - if (err) { - ohci_err(ohci, "pci_save_state failed\n"); - return err; - } - err = pci_set_power_state(dev, pci_choose_state(dev, state)); - if (err) - ohci_err(ohci, "pci_set_power_state failed with %d\n", err); - pmac_ohci_off(dev); + pmac_ohci_off(pdev); return 0; } -static int pci_resume(struct pci_dev *dev) + +static int __maybe_unused pci_resume(struct device *dev) { - struct fw_ohci *ohci = pci_get_drvdata(dev); + struct pci_dev *pdev = to_pci_dev(dev); + struct fw_ohci *ohci = pci_get_drvdata(pdev); int err; - pmac_ohci_on(dev); - pci_set_power_state(dev, PCI_D0); - pci_restore_state(dev); - err = pci_enable_device(dev); - if (err) { - ohci_err(ohci, "pci_enable_device failed\n"); - return err; - } + pmac_ohci_on(pdev); /* Some systems don't setup GUID register on resume from ram */ if (!reg_read(ohci, OHCI1394_GUIDLo) && @@ -3937,7 +3794,6 @@ static int pci_resume(struct pci_dev *dev) return 0; } -#endif static const struct pci_device_id pci_table[] = { { PCI_DEVICE_CLASS(PCI_CLASS_SERIAL_FIREWIRE_OHCI, ~0) }, @@ -3946,30 +3802,24 @@ static const struct pci_device_id pci_table[] = { MODULE_DEVICE_TABLE(pci, pci_table); +static SIMPLE_DEV_PM_OPS(pci_pm_ops, pci_suspend, pci_resume); + static struct pci_driver fw_ohci_pci_driver = { .name = ohci_driver_name, .id_table = pci_table, .probe = pci_probe, .remove = pci_remove, -#ifdef CONFIG_PM - .resume = pci_resume, - .suspend = pci_suspend, -#endif + .driver.pm = &pci_pm_ops, }; static int __init fw_ohci_init(void) { - selfid_workqueue = alloc_workqueue(KBUILD_MODNAME, WQ_MEM_RECLAIM, 0); - if (!selfid_workqueue) - return -ENOMEM; - return pci_register_driver(&fw_ohci_pci_driver); } static void __exit fw_ohci_cleanup(void) { pci_unregister_driver(&fw_ohci_pci_driver); - destroy_workqueue(selfid_workqueue); } module_init(fw_ohci_init); diff --git a/drivers/firewire/sbp2.c b/drivers/firewire/sbp2.c index 827dee0f57dd..021b8f698e34 100644 --- a/drivers/firewire/sbp2.c +++ b/drivers/firewire/sbp2.c @@ -558,7 +558,7 @@ static int sbp2_send_management_orb(struct sbp2_logical_unit *lu, int node_id, if (function == SBP2_LOGOUT_REQUEST && fw_device_is_shutdown(device)) return 0; - orb = kzalloc(sizeof(*orb), GFP_NOIO); + orb = kzalloc_obj(*orb, GFP_NOIO); if (orb == NULL) return -ENOMEM; @@ -667,7 +667,7 @@ static void sbp2_agent_reset_no_wait(struct sbp2_logical_unit *lu) struct fw_transaction *t; static __be32 d; - t = kmalloc(sizeof(*t), GFP_ATOMIC); + t = kmalloc_obj(*t, GFP_ATOMIC); if (t == NULL) return; @@ -966,7 +966,7 @@ static int sbp2_add_logical_unit(struct sbp2_target *tgt, int lun_entry) { struct sbp2_logical_unit *lu; - lu = kmalloc(sizeof(*lu), GFP_KERNEL); + lu = kmalloc_obj(*lu); if (!lu) return -ENOMEM; @@ -1440,15 +1440,16 @@ static int sbp2_map_scatterlist(struct sbp2_command_orb *orb, /* SCSI stack integration */ -static int sbp2_scsi_queuecommand(struct Scsi_Host *shost, - struct scsi_cmnd *cmd) +static enum scsi_qc_status sbp2_scsi_queuecommand(struct Scsi_Host *shost, + struct scsi_cmnd *cmd) { struct sbp2_logical_unit *lu = cmd->device->hostdata; struct fw_device *device = target_parent_device(lu->tgt); + enum scsi_qc_status retval = SCSI_MLQUEUE_HOST_BUSY; struct sbp2_command_orb *orb; - int generation, retval = SCSI_MLQUEUE_HOST_BUSY; + int generation; - orb = kzalloc(sizeof(*orb), GFP_ATOMIC); + orb = kzalloc_obj(*orb, GFP_ATOMIC); if (orb == NULL) return SCSI_MLQUEUE_HOST_BUSY; @@ -1490,7 +1491,7 @@ static int sbp2_scsi_queuecommand(struct Scsi_Host *shost, return retval; } -static int sbp2_scsi_slave_alloc(struct scsi_device *sdev) +static int sbp2_scsi_sdev_init(struct scsi_device *sdev) { struct sbp2_logical_unit *lu = sdev->hostdata; @@ -1506,8 +1507,8 @@ static int sbp2_scsi_slave_alloc(struct scsi_device *sdev) return 0; } -static int sbp2_scsi_device_configure(struct scsi_device *sdev, - struct queue_limits *lim) +static int sbp2_scsi_sdev_configure(struct scsi_device *sdev, + struct queue_limits *lim) { struct sbp2_logical_unit *lu = sdev->hostdata; @@ -1590,8 +1591,8 @@ static const struct scsi_host_template scsi_driver_template = { .name = "SBP-2 IEEE-1394", .proc_name = "sbp2", .queuecommand = sbp2_scsi_queuecommand, - .slave_alloc = sbp2_scsi_slave_alloc, - .device_configure = sbp2_scsi_device_configure, + .sdev_init = sbp2_scsi_sdev_init, + .sdev_configure = sbp2_scsi_sdev_configure, .eh_abort_handler = sbp2_scsi_abort, .this_id = -1, .sg_tablesize = SG_ALL, |