Pull block fixes from Jens Axboe: "A few fixes for the current series. This contains: - Two fixes for NVMe: One fixes a reset race that can be triggered by repeated insert/removal of the module. The other fixes an issue on some platforms, where we get probe timeouts since legacy interrupts isn't working. This used not to be a problem since we had the worker thread poll for completions, but since that was killed off, it means those poor souls can't successfully probe their NVMe device. Use a proper IRQ check and probe (msi-x -> msi ->legacy), like most other drivers to work around this. Both from Keith. - A loop corruption issue with offset in iters, from Ming Lei. - A fix for not having the partition stat per cpu ref count initialized before sending out the KOBJ_ADD, which could cause user space to access the counter prior to initialization. Also from Ming Lei. - A fix for using the wrong congestion state, from Kaixu Xia" * 'for-linus' of git://git.kernel.dk/linux-block: block: loop: fix filesystem corruption in case of aio/dio NVMe: Always use MSI/MSI-x interrupts NVMe: Fix reset/remove race writeback: fix the wrong congested state variable definition block: partition: initialize percpuref before sending out KOBJ_ADD
Unified · Split
| @@ -361,15 +361,20 @@ struct hd_struct *add_partition(struct gendisk *disk, int partno, · 360 unmodified lines | |||
| 361 | goto out_del; | 361 | goto out_del; |
| 362 | } | 362 | } |
| 363 | 363 | ||
| 364 | err = hd_ref_init(p); | ||
| 365 | if (err) { | ||
| 366 | if (flags & ADDPART_FLAG_WHOLEDISK) | ||
| 367 | goto out_remove_file; | ||
| 368 | goto out_del; | ||
| 369 | } | ||
| 370 | |||
| 364 | /* everything is up and running, commence */ | 371 | /* everything is up and running, commence */ |
| 365 | rcu_assign_pointer(ptbl->part[partno], p); | 372 | rcu_assign_pointer(ptbl->part[partno], p); |
| 366 | 373 | ||
| 367 | /* suppress uevent if the disk suppresses it */ | 374 | /* suppress uevent if the disk suppresses it */ |
| 368 | if (!dev_get_uevent_suppress(ddev)) | 375 | if (!dev_get_uevent_suppress(ddev)) |
| 369 | kobject_uevent(&pdev->kobj, KOBJ_ADD); | 376 | kobject_uevent(&pdev->kobj, KOBJ_ADD); |
| 370 | 377 | return p; | |
| 371 | if (!hd_ref_init(p)) | ||
| 372 | return p; | ||
| 373 | 378 | ||
| 374 | out_free_info: | 379 | out_free_info: |
| 375 | free_part_info(p); | 380 | free_part_info(p); |
| @@ -378,6 +383,8 @@ out_free_stats: · 2 unmodified lines | |||
| 378 | out_free: | 383 | out_free: |
| 379 | kfree(p); | 384 | kfree(p); |
| 380 | return ERR_PTR(err); | 385 | return ERR_PTR(err); |
| 386 | out_remove_file: | ||
| 387 | device_remove_file(pdev, &dev_attr_whole_disk); | ||
| 381 | out_del: | 388 | out_del: |
| 382 | kobject_put(p->holder_dir); | 389 | kobject_put(p->holder_dir); |
| 383 | device_del(pdev); | 390 | device_del(pdev); |
| @@ -488,6 +488,12 @@ static int lo_rw_aio(struct loop_device *lo, struct loop_cmd *cmd, · 487 unmodified lines | |||
| 488 | bvec = __bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter); | 488 | bvec = __bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter); |
| 489 | iov_iter_bvec(&iter, ITER_BVEC | rw, bvec, | 489 | iov_iter_bvec(&iter, ITER_BVEC | rw, bvec, |
| 490 | bio_segments(bio), blk_rq_bytes(cmd->rq)); | 490 | bio_segments(bio), blk_rq_bytes(cmd->rq)); |
| 491 | /* | ||
| 492 | * This bio may be started from the middle of the 'bvec' | ||
| 493 | * because of bio splitting, so offset from the bvec must | ||
| 494 | * be passed to iov iterator | ||
| 495 | */ | ||
| 496 | iter.iov_offset = bio->bi_iter.bi_bvec_done; | ||
| 491 | 497 | ||
| 492 | cmd->iocb.ki_pos = pos; | 498 | cmd->iocb.ki_pos = pos; |
| 493 | cmd->iocb.ki_filp = file; | 499 | cmd->iocb.ki_filp = file; |
| @@ -1478,8 +1478,7 @@ static int nvme_setup_io_queues(struct nvme_dev *dev) · 1477 unmodified lines | |||
| 1478 | if (result > 0) { | 1478 | if (result > 0) { |
| 1479 | dev_err(dev->ctrl.device, | 1479 | dev_err(dev->ctrl.device, |
| 1480 | "Could not set queue count (%d)\n", result); | 1480 | "Could not set queue count (%d)\n", result); |
| 1481 | nr_io_queues = 0; | 1481 | return 0; |
| 1482 | result = 0; | ||
| 1483 | } | 1482 | } |
| 1484 | 1483 | ||
| 1485 | if (dev->cmb && NVME_CMB_SQS(dev->cmbsz)) { | 1484 | if (dev->cmb && NVME_CMB_SQS(dev->cmbsz)) { |
| @@ -1513,7 +1512,9 @@ static int nvme_setup_io_queues(struct nvme_dev *dev) · 27 unmodified lines | |||
| 1513 | * If we enable msix early due to not intx, disable it again before | 1512 | * If we enable msix early due to not intx, disable it again before |
| 1514 | * setting up the full range we need. | 1513 | * setting up the full range we need. |
| 1515 | */ | 1514 | */ |
| 1516 | if (!pdev->irq) | 1515 | if (pdev->msi_enabled) |
| 1516 | pci_disable_msi(pdev); | ||
| 1517 | else if (pdev->msix_enabled) | ||
| 1517 | pci_disable_msix(pdev); | 1518 | pci_disable_msix(pdev); |
| 1518 | 1519 | ||
| 1519 | for (i = 0; i < nr_io_queues; i++) | 1520 | for (i = 0; i < nr_io_queues; i++) |
| @@ -1696,7 +1697,6 @@ static int nvme_pci_enable(struct nvme_dev *dev) · 176 unmodified lines | |||
| 1696 | if (pci_enable_device_mem(pdev)) | 1697 | if (pci_enable_device_mem(pdev)) |
| 1697 | return result; | 1698 | return result; |
| 1698 | 1699 | ||
| 1699 | dev->entry[0].vector = pdev->irq; | ||
| 1700 | pci_set_master(pdev); | 1700 | pci_set_master(pdev); |
| 1701 | 1701 | ||
| 1702 | if (dma_set_mask_and_coherent(dev->dev, DMA_BIT_MASK(64)) && | 1702 | if (dma_set_mask_and_coherent(dev->dev, DMA_BIT_MASK(64)) && |
| @@ -1709,13 +1709,18 @@ static int nvme_pci_enable(struct nvme_dev *dev) · 6 unmodified lines | |||
| 1709 | } | 1709 | } |
| 1710 | 1710 | ||
| 1711 | /* | 1711 | /* |
| 1712 | * Some devices don't advertse INTx interrupts, pre-enable a single | 1712 | * Some devices and/or platforms don't advertise or work with INTx |
| 1713 | * MSIX vec for setup. We'll adjust this later. | 1713 | * interrupts. Pre-enable a single MSIX or MSI vec for setup. We'll |
| 1714 | * adjust this later. | ||
| 1714 | */ | 1715 | */ |
| 1715 | if (!pdev->irq) { | 1716 | if (pci_enable_msix(pdev, dev->entry, 1)) { |
| 1716 | result = pci_enable_msix(pdev, dev->entry, 1); | 1717 | pci_enable_msi(pdev); |
| 1717 | if (result < 0) | 1718 | dev->entry[0].vector = pdev->irq; |
| 1718 | goto disable; | 1719 | } |
| 1720 | |||
| 1721 | if (!dev->entry[0].vector) { | ||
| 1722 | result = -ENODEV; | ||
| 1723 | goto disable; | ||
| 1719 | } | 1724 | } |
| 1720 | 1725 | ||
| 1721 | cap = lo_hi_readq(dev->bar + NVME_REG_CAP); | 1726 | cap = lo_hi_readq(dev->bar + NVME_REG_CAP); |
| @@ -1859,6 +1864,9 @@ static void nvme_reset_work(struct work_struct *work) · 137 unmodified lines | |||
| 1859 | if (dev->ctrl.ctrl_config & NVME_CC_ENABLE) | 1864 | if (dev->ctrl.ctrl_config & NVME_CC_ENABLE) |
| 1860 | nvme_dev_disable(dev, false); | 1865 | nvme_dev_disable(dev, false); |
| 1861 | 1866 | ||
| 1867 | if (test_bit(NVME_CTRL_REMOVING, &dev->flags)) | ||
| 1868 | goto out; | ||
| 1869 | |||
| 1862 | set_bit(NVME_CTRL_RESETTING, &dev->flags); | 1870 | set_bit(NVME_CTRL_RESETTING, &dev->flags); |
| 1863 | 1871 | ||
| 1864 | result = nvme_pci_enable(dev); | 1872 | result = nvme_pci_enable(dev); |
| @@ -2078,11 +2086,10 @@ static void nvme_remove(struct pci_dev *pdev) · 213 unmodified lines | |||
| 2078 | { | 2086 | { |
| 2079 | struct nvme_dev *dev = pci_get_drvdata(pdev); | 2087 | struct nvme_dev *dev = pci_get_drvdata(pdev); |
| 2080 | 2088 | ||
| 2081 | del_timer_sync(&dev->watchdog_timer); | ||
| 2082 | |||
| 2083 | set_bit(NVME_CTRL_REMOVING, &dev->flags); | 2089 | set_bit(NVME_CTRL_REMOVING, &dev->flags); |
| 2084 | pci_set_drvdata(pdev, NULL); | 2090 | pci_set_drvdata(pdev, NULL); |
| 2085 | flush_work(&dev->async_work); | 2091 | flush_work(&dev->async_work); |
| 2092 | flush_work(&dev->reset_work); | ||
| 2086 | flush_work(&dev->scan_work); | 2093 | flush_work(&dev->scan_work); |
| 2087 | nvme_remove_namespaces(&dev->ctrl); | 2094 | nvme_remove_namespaces(&dev->ctrl); |
| 2088 | nvme_uninit_ctrl(&dev->ctrl); | 2095 | nvme_uninit_ctrl(&dev->ctrl); |
| @@ -898,7 +898,7 @@ static atomic_t nr_wb_congested[2]; · 897 unmodified lines | |||
| 898 | void clear_wb_congested(struct bdi_writeback_congested *congested, int sync) | 898 | void clear_wb_congested(struct bdi_writeback_congested *congested, int sync) |
| 899 | { | 899 | { |
| 900 | wait_queue_head_t *wqh = &congestion_wqh[sync]; | 900 | wait_queue_head_t *wqh = &congestion_wqh[sync]; |
| 901 | enum wb_state bit; | 901 | enum wb_congested_state bit; |
| 902 | 902 | ||
| 903 | bit = sync ? WB_sync_congested : WB_async_congested; | 903 | bit = sync ? WB_sync_congested : WB_async_congested; |
| 904 | if (test_and_clear_bit(bit, &congested->state)) | 904 | if (test_and_clear_bit(bit, &congested->state)) |
| @@ -911,7 +911,7 @@ EXPORT_SYMBOL(clear_wb_congested); · 6 unmodified lines | |||
| 911 | 911 | ||
| 912 | void set_wb_congested(struct bdi_writeback_congested *congested, int sync) | 912 | void set_wb_congested(struct bdi_writeback_congested *congested, int sync) |
| 913 | { | 913 | { |
| 914 | enum wb_state bit; | 914 | enum wb_congested_state bit; |
| 915 | 915 | ||
| 916 | bit = sync ? WB_sync_congested : WB_async_congested; | 916 | bit = sync ? WB_sync_congested : WB_async_congested; |
| 917 | if (!test_and_set_bit(bit, &congested->state)) | 917 | if (!test_and_set_bit(bit, &congested->state)) |