diff --git a/config/kernel-blk-queue.m4 b/config/kernel-blk-queue.m4 index 091952f5700e..b3899178381c 100644 --- a/config/kernel-blk-queue.m4 +++ b/config/kernel-blk-queue.m4 @@ -1,348 +1,370 @@ dnl # dnl # 2.6.39 API change, dnl # blk_start_plug() and blk_finish_plug() dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_PLUG], [ ZFS_LINUX_TEST_SRC([blk_plug], [ #include ],[ struct blk_plug plug __attribute__ ((unused)); blk_start_plug(&plug); blk_finish_plug(&plug); ]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_PLUG], [ AC_MSG_CHECKING([whether struct blk_plug is available]) ZFS_LINUX_TEST_RESULT([blk_plug], [ AC_MSG_RESULT(yes) ],[ ZFS_LINUX_TEST_ERROR([blk_plug]) ]) ]) dnl # dnl # 2.6.32 - 4.11: statically allocated bdi in request_queue dnl # 4.12: dynamically allocated bdi in request_queue dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_BDI], [ ZFS_LINUX_TEST_SRC([blk_queue_bdi], [ #include ],[ struct request_queue q; struct backing_dev_info bdi; q.backing_dev_info = &bdi; ]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_BDI], [ AC_MSG_CHECKING([whether blk_queue bdi is dynamic]) ZFS_LINUX_TEST_RESULT([blk_queue_bdi], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_BDI_DYNAMIC, 1, [blk queue backing_dev_info is dynamic]) ],[ AC_MSG_RESULT(no) ]) ]) dnl # dnl # 5.9: added blk_queue_update_readahead(), dnl # 5.15: renamed to disk_update_readahead() dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_UPDATE_READAHEAD], [ ZFS_LINUX_TEST_SRC([blk_queue_update_readahead], [ #include ],[ struct request_queue q; blk_queue_update_readahead(&q); ]) ZFS_LINUX_TEST_SRC([disk_update_readahead], [ #include ],[ struct gendisk disk; disk_update_readahead(&disk); ]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_UPDATE_READAHEAD], [ AC_MSG_CHECKING([whether blk_queue_update_readahead() exists]) ZFS_LINUX_TEST_RESULT([blk_queue_update_readahead], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_UPDATE_READAHEAD, 1, [blk_queue_update_readahead() exists]) ],[ + AC_MSG_RESULT(no) + AC_MSG_CHECKING([whether disk_update_readahead() exists]) ZFS_LINUX_TEST_RESULT([disk_update_readahead], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_DISK_UPDATE_READAHEAD, 1, [disk_update_readahead() exists]) ],[ AC_MSG_RESULT(no) ]) ]) ]) dnl # -dnl # 2.6.32 API, -dnl # blk_queue_discard() +dnl # 5.19: bdev_max_discard_sectors() available +dnl # 2.6.32: blk_queue_discard() available dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_DISCARD], [ + ZFS_LINUX_TEST_SRC([bdev_max_discard_sectors], [ + #include + ],[ + struct block_device *bdev __attribute__ ((unused)) = NULL; + unsigned int error __attribute__ ((unused)); + + error = bdev_max_discard_sectors(bdev); + ]) + ZFS_LINUX_TEST_SRC([blk_queue_discard], [ #include ],[ struct request_queue r; struct request_queue *q = &r; int value __attribute__ ((unused)); memset(q, 0, sizeof(r)); value = blk_queue_discard(q); ]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_DISCARD], [ - AC_MSG_CHECKING([whether blk_queue_discard() is available]) - ZFS_LINUX_TEST_RESULT([blk_queue_discard], [ + AC_MSG_CHECKING([whether bdev_max_discard_sectors() is available]) + ZFS_LINUX_TEST_RESULT([bdev_max_discard_sectors], [ AC_MSG_RESULT(yes) + AC_DEFINE(HAVE_BDEV_MAX_DISCARD_SECTORS, 1, + [bdev_max_discard_sectors() is available]) ],[ - ZFS_LINUX_TEST_ERROR([blk_queue_discard]) + AC_MSG_RESULT(no) + + AC_MSG_CHECKING([whether blk_queue_discard() is available]) + ZFS_LINUX_TEST_RESULT([blk_queue_discard], [ + AC_MSG_RESULT(yes) + AC_DEFINE(HAVE_BLK_QUEUE_DISCARD, 1, + [blk_queue_discard() is available]) + ],[ + ZFS_LINUX_TEST_ERROR([blk_queue_discard]) + ]) ]) ]) dnl # dnl # 4.8 API, dnl # blk_queue_secure_erase() dnl # dnl # 2.6.36 - 4.7 API, dnl # blk_queue_secdiscard() dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_SECURE_ERASE], [ ZFS_LINUX_TEST_SRC([blk_queue_secure_erase], [ #include ],[ struct request_queue r; struct request_queue *q = &r; int value __attribute__ ((unused)); memset(q, 0, sizeof(r)); value = blk_queue_secure_erase(q); ]) ZFS_LINUX_TEST_SRC([blk_queue_secdiscard], [ #include ],[ struct request_queue r; struct request_queue *q = &r; int value __attribute__ ((unused)); memset(q, 0, sizeof(r)); value = blk_queue_secdiscard(q); ]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_SECURE_ERASE], [ AC_MSG_CHECKING([whether blk_queue_secure_erase() is available]) ZFS_LINUX_TEST_RESULT([blk_queue_secure_erase], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_SECURE_ERASE, 1, [blk_queue_secure_erase() is available]) ],[ AC_MSG_RESULT(no) AC_MSG_CHECKING([whether blk_queue_secdiscard() is available]) ZFS_LINUX_TEST_RESULT([blk_queue_secdiscard], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_SECDISCARD, 1, [blk_queue_secdiscard() is available]) ],[ ZFS_LINUX_TEST_ERROR([blk_queue_secure_erase]) ]) ]) ]) dnl # dnl # 4.16 API change, dnl # Introduction of blk_queue_flag_set and blk_queue_flag_clear dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_FLAG_SET], [ ZFS_LINUX_TEST_SRC([blk_queue_flag_set], [ #include #include ],[ struct request_queue *q = NULL; blk_queue_flag_set(0, q); ]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_FLAG_SET], [ AC_MSG_CHECKING([whether blk_queue_flag_set() exists]) ZFS_LINUX_TEST_RESULT([blk_queue_flag_set], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_FLAG_SET, 1, [blk_queue_flag_set() exists]) ],[ AC_MSG_RESULT(no) ]) ]) AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_FLAG_CLEAR], [ ZFS_LINUX_TEST_SRC([blk_queue_flag_clear], [ #include #include ],[ struct request_queue *q = NULL; blk_queue_flag_clear(0, q); ]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_FLAG_CLEAR], [ AC_MSG_CHECKING([whether blk_queue_flag_clear() exists]) ZFS_LINUX_TEST_RESULT([blk_queue_flag_clear], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_FLAG_CLEAR, 1, [blk_queue_flag_clear() exists]) ],[ AC_MSG_RESULT(no) ]) ]) dnl # dnl # 2.6.36 API change, dnl # Added blk_queue_flush() interface, while the previous interface dnl # was available to all the new one is GPL-only. Thus in addition to dnl # detecting if this function is available we determine if it is dnl # GPL-only. If the GPL-only interface is there we implement our own dnl # compatibility function, otherwise we use the function. The hope dnl # is that long term this function will be opened up. dnl # dnl # 4.7 API change, dnl # Replace blk_queue_flush with blk_queue_write_cache dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_FLUSH], [ ZFS_LINUX_TEST_SRC([blk_queue_flush], [ #include ], [ struct request_queue *q = NULL; (void) blk_queue_flush(q, REQ_FLUSH); ], [$NO_UNUSED_BUT_SET_VARIABLE], [ZFS_META_LICENSE]) ZFS_LINUX_TEST_SRC([blk_queue_write_cache], [ #include #include ], [ struct request_queue *q = NULL; blk_queue_write_cache(q, true, true); ], [$NO_UNUSED_BUT_SET_VARIABLE], [ZFS_META_LICENSE]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_FLUSH], [ AC_MSG_CHECKING([whether blk_queue_flush() is available]) ZFS_LINUX_TEST_RESULT([blk_queue_flush], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_FLUSH, 1, [blk_queue_flush() is available]) AC_MSG_CHECKING([whether blk_queue_flush() is GPL-only]) ZFS_LINUX_TEST_RESULT([blk_queue_flush_license], [ AC_MSG_RESULT(no) ],[ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_FLUSH_GPL_ONLY, 1, [blk_queue_flush() is GPL-only]) ]) ],[ AC_MSG_RESULT(no) ]) dnl # dnl # 4.7 API change dnl # Replace blk_queue_flush with blk_queue_write_cache dnl # AC_MSG_CHECKING([whether blk_queue_write_cache() exists]) ZFS_LINUX_TEST_RESULT([blk_queue_write_cache], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_WRITE_CACHE, 1, [blk_queue_write_cache() exists]) AC_MSG_CHECKING([whether blk_queue_write_cache() is GPL-only]) ZFS_LINUX_TEST_RESULT([blk_queue_write_cache_license], [ AC_MSG_RESULT(no) ],[ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_BLK_QUEUE_WRITE_CACHE_GPL_ONLY, 1, [blk_queue_write_cache() is GPL-only]) ]) ],[ AC_MSG_RESULT(no) ]) ]) dnl # dnl # 2.6.34 API change dnl # blk_queue_max_hw_sectors() replaces blk_queue_max_sectors(). dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_MAX_HW_SECTORS], [ ZFS_LINUX_TEST_SRC([blk_queue_max_hw_sectors], [ #include ], [ struct request_queue *q = NULL; (void) blk_queue_max_hw_sectors(q, BLK_SAFE_MAX_SECTORS); ], [$NO_UNUSED_BUT_SET_VARIABLE]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_MAX_HW_SECTORS], [ AC_MSG_CHECKING([whether blk_queue_max_hw_sectors() is available]) ZFS_LINUX_TEST_RESULT([blk_queue_max_hw_sectors], [ AC_MSG_RESULT(yes) ],[ ZFS_LINUX_TEST_ERROR([blk_queue_max_hw_sectors]) ]) ]) dnl # dnl # 2.6.34 API change dnl # blk_queue_max_segments() consolidates blk_queue_max_hw_segments() dnl # and blk_queue_max_phys_segments(). dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE_MAX_SEGMENTS], [ ZFS_LINUX_TEST_SRC([blk_queue_max_segments], [ #include ], [ struct request_queue *q = NULL; (void) blk_queue_max_segments(q, BLK_MAX_SEGMENTS); ], [$NO_UNUSED_BUT_SET_VARIABLE]) ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE_MAX_SEGMENTS], [ AC_MSG_CHECKING([whether blk_queue_max_segments() is available]) ZFS_LINUX_TEST_RESULT([blk_queue_max_segments], [ AC_MSG_RESULT(yes) ], [ ZFS_LINUX_TEST_ERROR([blk_queue_max_segments]) ]) ]) AC_DEFUN([ZFS_AC_KERNEL_SRC_BLK_QUEUE], [ ZFS_AC_KERNEL_SRC_BLK_QUEUE_PLUG ZFS_AC_KERNEL_SRC_BLK_QUEUE_BDI ZFS_AC_KERNEL_SRC_BLK_QUEUE_UPDATE_READAHEAD ZFS_AC_KERNEL_SRC_BLK_QUEUE_DISCARD ZFS_AC_KERNEL_SRC_BLK_QUEUE_SECURE_ERASE ZFS_AC_KERNEL_SRC_BLK_QUEUE_FLAG_SET ZFS_AC_KERNEL_SRC_BLK_QUEUE_FLAG_CLEAR ZFS_AC_KERNEL_SRC_BLK_QUEUE_FLUSH ZFS_AC_KERNEL_SRC_BLK_QUEUE_MAX_HW_SECTORS ZFS_AC_KERNEL_SRC_BLK_QUEUE_MAX_SEGMENTS ]) AC_DEFUN([ZFS_AC_KERNEL_BLK_QUEUE], [ ZFS_AC_KERNEL_BLK_QUEUE_PLUG ZFS_AC_KERNEL_BLK_QUEUE_BDI ZFS_AC_KERNEL_BLK_QUEUE_UPDATE_READAHEAD ZFS_AC_KERNEL_BLK_QUEUE_DISCARD ZFS_AC_KERNEL_BLK_QUEUE_SECURE_ERASE ZFS_AC_KERNEL_BLK_QUEUE_FLAG_SET ZFS_AC_KERNEL_BLK_QUEUE_FLAG_CLEAR ZFS_AC_KERNEL_BLK_QUEUE_FLUSH ZFS_AC_KERNEL_BLK_QUEUE_MAX_HW_SECTORS ZFS_AC_KERNEL_BLK_QUEUE_MAX_SEGMENTS ]) diff --git a/config/kernel-vfs-iov_iter.m4 b/config/kernel-vfs-iov_iter.m4 index 57f78745a24b..e0617faab02c 100644 --- a/config/kernel-vfs-iov_iter.m4 +++ b/config/kernel-vfs-iov_iter.m4 @@ -1,202 +1,204 @@ dnl # dnl # Check for available iov_iter functionality. dnl # AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_IOV_ITER], [ ZFS_LINUX_TEST_SRC([iov_iter_types], [ #include #include ],[ int type __attribute__ ((unused)) = ITER_IOVEC | ITER_KVEC | ITER_BVEC | ITER_PIPE; ]) ZFS_LINUX_TEST_SRC([iov_iter_advance], [ #include #include ],[ struct iov_iter iter = { 0 }; size_t advance = 512; iov_iter_advance(&iter, advance); ]) ZFS_LINUX_TEST_SRC([iov_iter_revert], [ #include #include ],[ struct iov_iter iter = { 0 }; size_t revert = 512; iov_iter_revert(&iter, revert); ]) ZFS_LINUX_TEST_SRC([iov_iter_fault_in_readable], [ #include #include ],[ struct iov_iter iter = { 0 }; size_t size = 512; int error __attribute__ ((unused)); error = iov_iter_fault_in_readable(&iter, size); ]) ZFS_LINUX_TEST_SRC([fault_in_iov_iter_readable], [ #include #include ],[ struct iov_iter iter = { 0 }; size_t size = 512; int error __attribute__ ((unused)); error = fault_in_iov_iter_readable(&iter, size); ]) ZFS_LINUX_TEST_SRC([iov_iter_count], [ #include #include ],[ struct iov_iter iter = { 0 }; size_t bytes __attribute__ ((unused)); bytes = iov_iter_count(&iter); ]) ZFS_LINUX_TEST_SRC([copy_to_iter], [ #include #include ],[ struct iov_iter iter = { 0 }; char buf[512] = { 0 }; size_t size = 512; size_t bytes __attribute__ ((unused)); bytes = copy_to_iter((const void *)&buf, size, &iter); ]) ZFS_LINUX_TEST_SRC([copy_from_iter], [ #include #include ],[ struct iov_iter iter = { 0 }; char buf[512] = { 0 }; size_t size = 512; size_t bytes __attribute__ ((unused)); bytes = copy_from_iter((void *)&buf, size, &iter); ]) ZFS_LINUX_TEST_SRC([iov_iter_type], [ #include #include ],[ struct iov_iter iter = { 0 }; __attribute__((unused)) enum iter_type i = iov_iter_type(&iter); ]) ]) AC_DEFUN([ZFS_AC_KERNEL_VFS_IOV_ITER], [ enable_vfs_iov_iter="yes" AC_MSG_CHECKING([whether iov_iter types are available]) ZFS_LINUX_TEST_RESULT([iov_iter_types], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_IOV_ITER_TYPES, 1, [iov_iter types are available]) ],[ AC_MSG_RESULT(no) enable_vfs_iov_iter="no" ]) AC_MSG_CHECKING([whether iov_iter_advance() is available]) ZFS_LINUX_TEST_RESULT([iov_iter_advance], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_IOV_ITER_ADVANCE, 1, [iov_iter_advance() is available]) ],[ AC_MSG_RESULT(no) enable_vfs_iov_iter="no" ]) AC_MSG_CHECKING([whether iov_iter_revert() is available]) ZFS_LINUX_TEST_RESULT([iov_iter_revert], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_IOV_ITER_REVERT, 1, [iov_iter_revert() is available]) ],[ AC_MSG_RESULT(no) enable_vfs_iov_iter="no" ]) AC_MSG_CHECKING([whether iov_iter_fault_in_readable() is available]) ZFS_LINUX_TEST_RESULT([iov_iter_fault_in_readable], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_IOV_ITER_FAULT_IN_READABLE, 1, [iov_iter_fault_in_readable() is available]) ],[ + AC_MSG_RESULT(no) + AC_MSG_CHECKING([whether fault_in_iov_iter_readable() is available]) ZFS_LINUX_TEST_RESULT([fault_in_iov_iter_readable], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_FAULT_IN_IOV_ITER_READABLE, 1, [fault_in_iov_iter_readable() is available]) ],[ AC_MSG_RESULT(no) enable_vfs_iov_iter="no" ]) ]) AC_MSG_CHECKING([whether iov_iter_count() is available]) ZFS_LINUX_TEST_RESULT([iov_iter_count], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_IOV_ITER_COUNT, 1, [iov_iter_count() is available]) ],[ AC_MSG_RESULT(no) enable_vfs_iov_iter="no" ]) AC_MSG_CHECKING([whether copy_to_iter() is available]) ZFS_LINUX_TEST_RESULT([copy_to_iter], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_COPY_TO_ITER, 1, [copy_to_iter() is available]) ],[ AC_MSG_RESULT(no) enable_vfs_iov_iter="no" ]) AC_MSG_CHECKING([whether copy_from_iter() is available]) ZFS_LINUX_TEST_RESULT([copy_from_iter], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_COPY_FROM_ITER, 1, [copy_from_iter() is available]) ],[ AC_MSG_RESULT(no) enable_vfs_iov_iter="no" ]) dnl # dnl # This checks for iov_iter_type() in linux/uio.h. It is not dnl # required, however, and the module will compiled without it dnl # using direct access of the member attribute dnl # AC_MSG_CHECKING([whether iov_iter_type() is available]) ZFS_LINUX_TEST_RESULT([iov_iter_type], [ AC_MSG_RESULT(yes) AC_DEFINE(HAVE_IOV_ITER_TYPE, 1, [iov_iter_type() is available]) ],[ AC_MSG_RESULT(no) ]) dnl # dnl # As of the 4.9 kernel support is provided for iovecs, kvecs, dnl # bvecs and pipes in the iov_iter structure. As long as the dnl # other support interfaces are all available the iov_iter can dnl # be correctly used in the uio structure. dnl # AS_IF([test "x$enable_vfs_iov_iter" = "xyes"], [ AC_DEFINE(HAVE_VFS_IOV_ITER, 1, [All required iov_iter interfaces are available]) ]) ]) diff --git a/include/os/linux/kernel/linux/blkdev_compat.h b/include/os/linux/kernel/linux/blkdev_compat.h index 9fa8884bb7a1..084d48c876d5 100644 --- a/include/os/linux/kernel/linux/blkdev_compat.h +++ b/include/os/linux/kernel/linux/blkdev_compat.h @@ -1,582 +1,601 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (C) 2011 Lawrence Livermore National Security, LLC. * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER). * Written by Brian Behlendorf . * LLNL-CODE-403049. */ #ifndef _ZFS_BLKDEV_H #define _ZFS_BLKDEV_H #include #include #include #include #include /* for SECTOR_* */ #ifndef HAVE_BLK_QUEUE_FLAG_SET static inline void blk_queue_flag_set(unsigned int flag, struct request_queue *q) { queue_flag_set(flag, q); } #endif #ifndef HAVE_BLK_QUEUE_FLAG_CLEAR static inline void blk_queue_flag_clear(unsigned int flag, struct request_queue *q) { queue_flag_clear(flag, q); } #endif /* * 4.7 API, * The blk_queue_write_cache() interface has replaced blk_queue_flush() * interface. However, the new interface is GPL-only thus we implement * our own trivial wrapper when the GPL-only version is detected. * * 2.6.36 - 4.6 API, * The blk_queue_flush() interface has replaced blk_queue_ordered() * interface. However, while the old interface was available to all the * new one is GPL-only. Thus if the GPL-only version is detected we * implement our own trivial helper. */ static inline void blk_queue_set_write_cache(struct request_queue *q, bool wc, bool fua) { #if defined(HAVE_BLK_QUEUE_WRITE_CACHE_GPL_ONLY) if (wc) blk_queue_flag_set(QUEUE_FLAG_WC, q); else blk_queue_flag_clear(QUEUE_FLAG_WC, q); if (fua) blk_queue_flag_set(QUEUE_FLAG_FUA, q); else blk_queue_flag_clear(QUEUE_FLAG_FUA, q); #elif defined(HAVE_BLK_QUEUE_WRITE_CACHE) blk_queue_write_cache(q, wc, fua); #elif defined(HAVE_BLK_QUEUE_FLUSH_GPL_ONLY) if (wc) q->flush_flags |= REQ_FLUSH; if (fua) q->flush_flags |= REQ_FUA; #elif defined(HAVE_BLK_QUEUE_FLUSH) blk_queue_flush(q, (wc ? REQ_FLUSH : 0) | (fua ? REQ_FUA : 0)); #else #error "Unsupported kernel" #endif } static inline void blk_queue_set_read_ahead(struct request_queue *q, unsigned long ra_pages) { #if !defined(HAVE_BLK_QUEUE_UPDATE_READAHEAD) && \ !defined(HAVE_DISK_UPDATE_READAHEAD) #ifdef HAVE_BLK_QUEUE_BDI_DYNAMIC q->backing_dev_info->ra_pages = ra_pages; #else q->backing_dev_info.ra_pages = ra_pages; #endif #endif } #ifdef HAVE_BIO_BVEC_ITER #define BIO_BI_SECTOR(bio) (bio)->bi_iter.bi_sector #define BIO_BI_SIZE(bio) (bio)->bi_iter.bi_size #define BIO_BI_IDX(bio) (bio)->bi_iter.bi_idx #define BIO_BI_SKIP(bio) (bio)->bi_iter.bi_bvec_done #define bio_for_each_segment4(bv, bvp, b, i) \ bio_for_each_segment((bv), (b), (i)) typedef struct bvec_iter bvec_iterator_t; #else #define BIO_BI_SECTOR(bio) (bio)->bi_sector #define BIO_BI_SIZE(bio) (bio)->bi_size #define BIO_BI_IDX(bio) (bio)->bi_idx #define BIO_BI_SKIP(bio) (0) #define bio_for_each_segment4(bv, bvp, b, i) \ bio_for_each_segment((bvp), (b), (i)) typedef int bvec_iterator_t; #endif static inline void bio_set_flags_failfast(struct block_device *bdev, int *flags) { #ifdef CONFIG_BUG /* * Disable FAILFAST for loopback devices because of the * following incorrect BUG_ON() in loop_make_request(). * This support is also disabled for md devices because the * test suite layers md devices on top of loopback devices. * This may be removed when the loopback driver is fixed. * * BUG_ON(!lo || (rw != READ && rw != WRITE)); */ if ((MAJOR(bdev->bd_dev) == LOOP_MAJOR) || (MAJOR(bdev->bd_dev) == MD_MAJOR)) return; #ifdef BLOCK_EXT_MAJOR if (MAJOR(bdev->bd_dev) == BLOCK_EXT_MAJOR) return; #endif /* BLOCK_EXT_MAJOR */ #endif /* CONFIG_BUG */ *flags |= REQ_FAILFAST_MASK; } /* * Maximum disk label length, it may be undefined for some kernels. */ #if !defined(DISK_NAME_LEN) #define DISK_NAME_LEN 32 #endif /* DISK_NAME_LEN */ #ifdef HAVE_BIO_BI_STATUS static inline int bi_status_to_errno(blk_status_t status) { switch (status) { case BLK_STS_OK: return (0); case BLK_STS_NOTSUPP: return (EOPNOTSUPP); case BLK_STS_TIMEOUT: return (ETIMEDOUT); case BLK_STS_NOSPC: return (ENOSPC); case BLK_STS_TRANSPORT: return (ENOLINK); case BLK_STS_TARGET: return (EREMOTEIO); case BLK_STS_NEXUS: return (EBADE); case BLK_STS_MEDIUM: return (ENODATA); case BLK_STS_PROTECTION: return (EILSEQ); case BLK_STS_RESOURCE: return (ENOMEM); case BLK_STS_AGAIN: return (EAGAIN); case BLK_STS_IOERR: return (EIO); default: return (EIO); } } static inline blk_status_t errno_to_bi_status(int error) { switch (error) { case 0: return (BLK_STS_OK); case EOPNOTSUPP: return (BLK_STS_NOTSUPP); case ETIMEDOUT: return (BLK_STS_TIMEOUT); case ENOSPC: return (BLK_STS_NOSPC); case ENOLINK: return (BLK_STS_TRANSPORT); case EREMOTEIO: return (BLK_STS_TARGET); case EBADE: return (BLK_STS_NEXUS); case ENODATA: return (BLK_STS_MEDIUM); case EILSEQ: return (BLK_STS_PROTECTION); case ENOMEM: return (BLK_STS_RESOURCE); case EAGAIN: return (BLK_STS_AGAIN); case EIO: return (BLK_STS_IOERR); default: return (BLK_STS_IOERR); } } #endif /* HAVE_BIO_BI_STATUS */ /* * 4.3 API change * The bio_endio() prototype changed slightly. These are helper * macro's to ensure the prototype and invocation are handled. */ #ifdef HAVE_1ARG_BIO_END_IO_T #ifdef HAVE_BIO_BI_STATUS #define BIO_END_IO_ERROR(bio) bi_status_to_errno(bio->bi_status) #define BIO_END_IO_PROTO(fn, x, z) static void fn(struct bio *x) #define BIO_END_IO(bio, error) bio_set_bi_status(bio, error) static inline void bio_set_bi_status(struct bio *bio, int error) { ASSERT3S(error, <=, 0); bio->bi_status = errno_to_bi_status(-error); bio_endio(bio); } #else #define BIO_END_IO_ERROR(bio) (-(bio->bi_error)) #define BIO_END_IO_PROTO(fn, x, z) static void fn(struct bio *x) #define BIO_END_IO(bio, error) bio_set_bi_error(bio, error) static inline void bio_set_bi_error(struct bio *bio, int error) { ASSERT3S(error, <=, 0); bio->bi_error = error; bio_endio(bio); } #endif /* HAVE_BIO_BI_STATUS */ #else #define BIO_END_IO_PROTO(fn, x, z) static void fn(struct bio *x, int z) #define BIO_END_IO(bio, error) bio_endio(bio, error); #endif /* HAVE_1ARG_BIO_END_IO_T */ /* * 4.1 API, * 3.10.0 CentOS 7.x API, * blkdev_reread_part() * * For older kernels trigger a re-reading of the partition table by calling * check_disk_change() which calls flush_disk() to invalidate the device. * * For newer kernels (as of 5.10), bdev_check_media_change is used, in favor of * check_disk_change(), with the modification that invalidation is no longer * forced. */ #ifdef HAVE_CHECK_DISK_CHANGE #define zfs_check_media_change(bdev) check_disk_change(bdev) #ifdef HAVE_BLKDEV_REREAD_PART #define vdev_bdev_reread_part(bdev) blkdev_reread_part(bdev) #else #define vdev_bdev_reread_part(bdev) check_disk_change(bdev) #endif /* HAVE_BLKDEV_REREAD_PART */ #else #ifdef HAVE_BDEV_CHECK_MEDIA_CHANGE static inline int zfs_check_media_change(struct block_device *bdev) { #ifdef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK struct gendisk *gd = bdev->bd_disk; const struct block_device_operations *bdo = gd->fops; #endif if (!bdev_check_media_change(bdev)) return (0); #ifdef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK /* * Force revalidation, to mimic the old behavior of * check_disk_change() */ if (bdo->revalidate_disk) bdo->revalidate_disk(gd); #endif return (0); } #define vdev_bdev_reread_part(bdev) zfs_check_media_change(bdev) #else /* * This is encountered if check_disk_change() and bdev_check_media_change() * are not available in the kernel - likely due to an API change that needs * to be chased down. */ #error "Unsupported kernel: no usable disk change check" #endif /* HAVE_BDEV_CHECK_MEDIA_CHANGE */ #endif /* HAVE_CHECK_DISK_CHANGE */ /* * 2.6.27 API change * The function was exported for use, prior to this it existed but the * symbol was not exported. * * 4.4.0-6.21 API change for Ubuntu * lookup_bdev() gained a second argument, FMODE_*, to check inode permissions. * * 5.11 API change * Changed to take a dev_t argument which is set on success and return a * non-zero error code on failure. */ static inline int vdev_lookup_bdev(const char *path, dev_t *dev) { #if defined(HAVE_DEVT_LOOKUP_BDEV) return (lookup_bdev(path, dev)); #elif defined(HAVE_1ARG_LOOKUP_BDEV) struct block_device *bdev = lookup_bdev(path); if (IS_ERR(bdev)) return (PTR_ERR(bdev)); *dev = bdev->bd_dev; bdput(bdev); return (0); #elif defined(HAVE_MODE_LOOKUP_BDEV) struct block_device *bdev = lookup_bdev(path, FMODE_READ); if (IS_ERR(bdev)) return (PTR_ERR(bdev)); *dev = bdev->bd_dev; bdput(bdev); return (0); #else #error "Unsupported kernel" #endif } /* * Kernels without bio_set_op_attrs use bi_rw for the bio flags. */ #if !defined(HAVE_BIO_SET_OP_ATTRS) static inline void bio_set_op_attrs(struct bio *bio, unsigned rw, unsigned flags) { bio->bi_rw |= rw | flags; } #endif /* * bio_set_flush - Set the appropriate flags in a bio to guarantee * data are on non-volatile media on completion. * * 2.6.37 - 4.8 API, * Introduce WRITE_FLUSH, WRITE_FUA, and WRITE_FLUSH_FUA flags as a * replacement for WRITE_BARRIER to allow expressing richer semantics * to the block layer. It's up to the block layer to implement the * semantics correctly. Use the WRITE_FLUSH_FUA flag combination. * * 4.8 - 4.9 API, * REQ_FLUSH was renamed to REQ_PREFLUSH. For consistency with previous * OpenZFS releases, prefer the WRITE_FLUSH_FUA flag set if it's available. * * 4.10 API, * The read/write flags and their modifiers, including WRITE_FLUSH, * WRITE_FUA and WRITE_FLUSH_FUA were removed from fs.h in * torvalds/linux@70fd7614 and replaced by direct flag modification * of the REQ_ flags in bio->bi_opf. Use REQ_PREFLUSH. */ static inline void bio_set_flush(struct bio *bio) { #if defined(HAVE_REQ_PREFLUSH) /* >= 4.10 */ bio_set_op_attrs(bio, 0, REQ_PREFLUSH); #elif defined(WRITE_FLUSH_FUA) /* >= 2.6.37 and <= 4.9 */ bio_set_op_attrs(bio, 0, WRITE_FLUSH_FUA); #else #error "Allowing the build will cause bio_set_flush requests to be ignored." #endif } /* * 4.8 API, * REQ_OP_FLUSH * * 4.8-rc0 - 4.8-rc1, * REQ_PREFLUSH * * 2.6.36 - 4.7 API, * REQ_FLUSH * * in all cases but may have a performance impact for some kernels. It * has the advantage of minimizing kernel specific changes in the zvol code. * */ static inline boolean_t bio_is_flush(struct bio *bio) { #if defined(HAVE_REQ_OP_FLUSH) && defined(HAVE_BIO_BI_OPF) return ((bio_op(bio) == REQ_OP_FLUSH) || (bio->bi_opf & REQ_PREFLUSH)); #elif defined(HAVE_REQ_PREFLUSH) && defined(HAVE_BIO_BI_OPF) return (bio->bi_opf & REQ_PREFLUSH); #elif defined(HAVE_REQ_PREFLUSH) && !defined(HAVE_BIO_BI_OPF) return (bio->bi_rw & REQ_PREFLUSH); #elif defined(HAVE_REQ_FLUSH) return (bio->bi_rw & REQ_FLUSH); #else #error "Unsupported kernel" #endif } /* * 4.8 API, * REQ_FUA flag moved to bio->bi_opf * * 2.6.x - 4.7 API, * REQ_FUA */ static inline boolean_t bio_is_fua(struct bio *bio) { #if defined(HAVE_BIO_BI_OPF) return (bio->bi_opf & REQ_FUA); #elif defined(REQ_FUA) return (bio->bi_rw & REQ_FUA); #else #error "Allowing the build will cause fua requests to be ignored." #endif } /* * 4.8 API, * REQ_OP_DISCARD * * 2.6.36 - 4.7 API, * REQ_DISCARD * * In all cases the normal I/O path is used for discards. The only * difference is how the kernel tags individual I/Os as discards. */ static inline boolean_t bio_is_discard(struct bio *bio) { #if defined(HAVE_REQ_OP_DISCARD) return (bio_op(bio) == REQ_OP_DISCARD); #elif defined(HAVE_REQ_DISCARD) return (bio->bi_rw & REQ_DISCARD); #else #error "Unsupported kernel" #endif } /* * 4.8 API, * REQ_OP_SECURE_ERASE * * 2.6.36 - 4.7 API, * REQ_SECURE */ static inline boolean_t bio_is_secure_erase(struct bio *bio) { #if defined(HAVE_REQ_OP_SECURE_ERASE) return (bio_op(bio) == REQ_OP_SECURE_ERASE); #elif defined(REQ_SECURE) return (bio->bi_rw & REQ_SECURE); #else return (0); #endif } /* * 2.6.33 API change * Discard granularity and alignment restrictions may now be set. For * older kernels which do not support this it is safe to skip it. */ static inline void blk_queue_discard_granularity(struct request_queue *q, unsigned int dg) { q->limits.discard_granularity = dg; } +/* + * 5.19 API, + * bdev_max_discard_sectors() + * + * 2.6.32 API, + * blk_queue_discard() + */ +static inline boolean_t +bdev_discard_supported(struct block_device *bdev) +{ +#if defined(HAVE_BDEV_MAX_DISCARD_SECTORS) + return (!!bdev_max_discard_sectors(bdev)); +#elif defined(HAVE_BLK_QUEUE_DISCARD) + return (!!blk_queue_discard(bdev_get_queue(bdev))); +#else +#error "Unsupported kernel" +#endif +} + /* * 4.8 API, * blk_queue_secure_erase() * * 2.6.36 - 4.7 API, * blk_queue_secdiscard() */ static inline int blk_queue_discard_secure(struct request_queue *q) { #if defined(HAVE_BLK_QUEUE_SECURE_ERASE) return (blk_queue_secure_erase(q)); #elif defined(HAVE_BLK_QUEUE_SECDISCARD) return (blk_queue_secdiscard(q)); #else return (0); #endif } /* * A common holder for vdev_bdev_open() is used to relax the exclusive open * semantics slightly. Internal vdev disk callers may pass VDEV_HOLDER to * allow them to open the device multiple times. Other kernel callers and * user space processes which don't pass this value will get EBUSY. This is * currently required for the correct operation of hot spares. */ #define VDEV_HOLDER ((void *)0x2401de7) static inline unsigned long blk_generic_start_io_acct(struct request_queue *q __attribute__((unused)), struct gendisk *disk __attribute__((unused)), int rw __attribute__((unused)), struct bio *bio) { #if defined(HAVE_DISK_IO_ACCT) return (disk_start_io_acct(disk, bio_sectors(bio), bio_op(bio))); #elif defined(HAVE_BIO_IO_ACCT) return (bio_start_io_acct(bio)); #elif defined(HAVE_GENERIC_IO_ACCT_3ARG) unsigned long start_time = jiffies; generic_start_io_acct(rw, bio_sectors(bio), &disk->part0); return (start_time); #elif defined(HAVE_GENERIC_IO_ACCT_4ARG) unsigned long start_time = jiffies; generic_start_io_acct(q, rw, bio_sectors(bio), &disk->part0); return (start_time); #else /* Unsupported */ return (0); #endif } static inline void blk_generic_end_io_acct(struct request_queue *q __attribute__((unused)), struct gendisk *disk __attribute__((unused)), int rw __attribute__((unused)), struct bio *bio, unsigned long start_time) { #if defined(HAVE_DISK_IO_ACCT) disk_end_io_acct(disk, bio_op(bio), start_time); #elif defined(HAVE_BIO_IO_ACCT) bio_end_io_acct(bio, start_time); #elif defined(HAVE_GENERIC_IO_ACCT_3ARG) generic_end_io_acct(rw, &disk->part0, start_time); #elif defined(HAVE_GENERIC_IO_ACCT_4ARG) generic_end_io_acct(q, rw, &disk->part0, start_time); #endif } #ifndef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS static inline struct request_queue * blk_generic_alloc_queue(make_request_fn make_request, int node_id) { #if defined(HAVE_BLK_ALLOC_QUEUE_REQUEST_FN) return (blk_alloc_queue(make_request, node_id)); #elif defined(HAVE_BLK_ALLOC_QUEUE_REQUEST_FN_RH) return (blk_alloc_queue_rh(make_request, node_id)); #else struct request_queue *q = blk_alloc_queue(GFP_KERNEL); if (q != NULL) blk_queue_make_request(q, make_request); return (q); #endif } #endif /* !HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS */ #endif /* _ZFS_BLKDEV_H */ diff --git a/module/os/linux/zfs/vdev_disk.c b/module/os/linux/zfs/vdev_disk.c index d6f9311be72b..ed933656d210 100644 --- a/module/os/linux/zfs/vdev_disk.c +++ b/module/os/linux/zfs/vdev_disk.c @@ -1,995 +1,995 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (C) 2008-2010 Lawrence Livermore National Security, LLC. * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER). * Rewritten for Linux by Brian Behlendorf . * LLNL-CODE-403049. * Copyright (c) 2012, 2019 by Delphix. All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #ifdef HAVE_LINUX_BLK_CGROUP_HEADER #include #endif typedef struct vdev_disk { struct block_device *vd_bdev; krwlock_t vd_lock; } vdev_disk_t; /* * Unique identifier for the exclusive vdev holder. */ static void *zfs_vdev_holder = VDEV_HOLDER; /* * Wait up to zfs_vdev_open_timeout_ms milliseconds before determining the * device is missing. The missing path may be transient since the links * can be briefly removed and recreated in response to udev events. */ static unsigned zfs_vdev_open_timeout_ms = 1000; /* * Size of the "reserved" partition, in blocks. */ #define EFI_MIN_RESV_SIZE (16 * 1024) /* * Virtual device vector for disks. */ typedef struct dio_request { zio_t *dr_zio; /* Parent ZIO */ atomic_t dr_ref; /* References */ int dr_error; /* Bio error */ int dr_bio_count; /* Count of bio's */ struct bio *dr_bio[0]; /* Attached bio's */ } dio_request_t; static fmode_t vdev_bdev_mode(spa_mode_t spa_mode) { fmode_t mode = 0; if (spa_mode & SPA_MODE_READ) mode |= FMODE_READ; if (spa_mode & SPA_MODE_WRITE) mode |= FMODE_WRITE; return (mode); } /* * Returns the usable capacity (in bytes) for the partition or disk. */ static uint64_t bdev_capacity(struct block_device *bdev) { return (i_size_read(bdev->bd_inode)); } #if !defined(HAVE_BDEV_WHOLE) static inline struct block_device * bdev_whole(struct block_device *bdev) { return (bdev->bd_contains); } #endif /* * Returns the maximum expansion capacity of the block device (in bytes). * * It is possible to expand a vdev when it has been created as a wholedisk * and the containing block device has increased in capacity. Or when the * partition containing the pool has been manually increased in size. * * This function is only responsible for calculating the potential expansion * size so it can be reported by 'zpool list'. The efi_use_whole_disk() is * responsible for verifying the expected partition layout in the wholedisk * case, and updating the partition table if appropriate. Once the partition * size has been increased the additional capacity will be visible using * bdev_capacity(). * * The returned maximum expansion capacity is always expected to be larger, or * at the very least equal, to its usable capacity to prevent overestimating * the pool expandsize. */ static uint64_t bdev_max_capacity(struct block_device *bdev, uint64_t wholedisk) { uint64_t psize; int64_t available; if (wholedisk && bdev != bdev_whole(bdev)) { /* * When reporting maximum expansion capacity for a wholedisk * deduct any capacity which is expected to be lost due to * alignment restrictions. Over reporting this value isn't * harmful and would only result in slightly less capacity * than expected post expansion. * The estimated available space may be slightly smaller than * bdev_capacity() for devices where the number of sectors is * not a multiple of the alignment size and the partition layout * is keeping less than PARTITION_END_ALIGNMENT bytes after the * "reserved" EFI partition: in such cases return the device * usable capacity. */ available = i_size_read(bdev_whole(bdev)->bd_inode) - ((EFI_MIN_RESV_SIZE + NEW_START_BLOCK + PARTITION_END_ALIGNMENT) << SECTOR_BITS); psize = MAX(available, bdev_capacity(bdev)); } else { psize = bdev_capacity(bdev); } return (psize); } static void vdev_disk_error(zio_t *zio) { /* * This function can be called in interrupt context, for instance while * handling IRQs coming from a misbehaving disk device; use printk() * which is safe from any context. */ printk(KERN_WARNING "zio pool=%s vdev=%s error=%d type=%d " "offset=%llu size=%llu flags=%x\n", spa_name(zio->io_spa), zio->io_vd->vdev_path, zio->io_error, zio->io_type, (u_longlong_t)zio->io_offset, (u_longlong_t)zio->io_size, zio->io_flags); } static int vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize, uint64_t *logical_ashift, uint64_t *physical_ashift) { struct block_device *bdev; fmode_t mode = vdev_bdev_mode(spa_mode(v->vdev_spa)); hrtime_t timeout = MSEC2NSEC(zfs_vdev_open_timeout_ms); vdev_disk_t *vd; /* Must have a pathname and it must be absolute. */ if (v->vdev_path == NULL || v->vdev_path[0] != '/') { v->vdev_stat.vs_aux = VDEV_AUX_BAD_LABEL; vdev_dbgmsg(v, "invalid vdev_path"); return (SET_ERROR(EINVAL)); } /* * Reopen the device if it is currently open. When expanding a * partition force re-scanning the partition table if userland * did not take care of this already. We need to do this while closed * in order to get an accurate updated block device size. Then * since udev may need to recreate the device links increase the * open retry timeout before reporting the device as unavailable. */ vd = v->vdev_tsd; if (vd) { char disk_name[BDEVNAME_SIZE + 6] = "/dev/"; boolean_t reread_part = B_FALSE; rw_enter(&vd->vd_lock, RW_WRITER); bdev = vd->vd_bdev; vd->vd_bdev = NULL; if (bdev) { if (v->vdev_expanding && bdev != bdev_whole(bdev)) { bdevname(bdev_whole(bdev), disk_name + 5); /* * If userland has BLKPG_RESIZE_PARTITION, * then it should have updated the partition * table already. We can detect this by * comparing our current physical size * with that of the device. If they are * the same, then we must not have * BLKPG_RESIZE_PARTITION or it failed to * update the partition table online. We * fallback to rescanning the partition * table from the kernel below. However, * if the capacity already reflects the * updated partition, then we skip * rescanning the partition table here. */ if (v->vdev_psize == bdev_capacity(bdev)) reread_part = B_TRUE; } blkdev_put(bdev, mode | FMODE_EXCL); } if (reread_part) { bdev = blkdev_get_by_path(disk_name, mode | FMODE_EXCL, zfs_vdev_holder); if (!IS_ERR(bdev)) { int error = vdev_bdev_reread_part(bdev); blkdev_put(bdev, mode | FMODE_EXCL); if (error == 0) { timeout = MSEC2NSEC( zfs_vdev_open_timeout_ms * 2); } } } } else { vd = kmem_zalloc(sizeof (vdev_disk_t), KM_SLEEP); rw_init(&vd->vd_lock, NULL, RW_DEFAULT, NULL); rw_enter(&vd->vd_lock, RW_WRITER); } /* * Devices are always opened by the path provided at configuration * time. This means that if the provided path is a udev by-id path * then drives may be re-cabled without an issue. If the provided * path is a udev by-path path, then the physical location information * will be preserved. This can be critical for more complicated * configurations where drives are located in specific physical * locations to maximize the systems tolerance to component failure. * * Alternatively, you can provide your own udev rule to flexibly map * the drives as you see fit. It is not advised that you use the * /dev/[hd]d devices which may be reordered due to probing order. * Devices in the wrong locations will be detected by the higher * level vdev validation. * * The specified paths may be briefly removed and recreated in * response to udev events. This should be exceptionally unlikely * because the zpool command makes every effort to verify these paths * have already settled prior to reaching this point. Therefore, * a ENOENT failure at this point is highly likely to be transient * and it is reasonable to sleep and retry before giving up. In * practice delays have been observed to be on the order of 100ms. * * When ERESTARTSYS is returned it indicates the block device is * a zvol which could not be opened due to the deadlock detection * logic in zvol_open(). Extend the timeout and retry the open * subsequent attempts are expected to eventually succeed. */ hrtime_t start = gethrtime(); bdev = ERR_PTR(-ENXIO); while (IS_ERR(bdev) && ((gethrtime() - start) < timeout)) { bdev = blkdev_get_by_path(v->vdev_path, mode | FMODE_EXCL, zfs_vdev_holder); if (unlikely(PTR_ERR(bdev) == -ENOENT)) { schedule_timeout(MSEC_TO_TICK(10)); } else if (unlikely(PTR_ERR(bdev) == -ERESTARTSYS)) { timeout = MSEC2NSEC(zfs_vdev_open_timeout_ms * 10); continue; } else if (IS_ERR(bdev)) { break; } } if (IS_ERR(bdev)) { int error = -PTR_ERR(bdev); vdev_dbgmsg(v, "open error=%d timeout=%llu/%llu", error, (u_longlong_t)(gethrtime() - start), (u_longlong_t)timeout); vd->vd_bdev = NULL; v->vdev_tsd = vd; rw_exit(&vd->vd_lock); return (SET_ERROR(error)); } else { vd->vd_bdev = bdev; v->vdev_tsd = vd; rw_exit(&vd->vd_lock); } struct request_queue *q = bdev_get_queue(vd->vd_bdev); /* Determine the physical block size */ int physical_block_size = bdev_physical_block_size(vd->vd_bdev); /* Determine the logical block size */ int logical_block_size = bdev_logical_block_size(vd->vd_bdev); /* Clear the nowritecache bit, causes vdev_reopen() to try again. */ v->vdev_nowritecache = B_FALSE; /* Set when device reports it supports TRIM. */ - v->vdev_has_trim = !!blk_queue_discard(q); + v->vdev_has_trim = bdev_discard_supported(vd->vd_bdev); /* Set when device reports it supports secure TRIM. */ v->vdev_has_securetrim = !!blk_queue_discard_secure(q); /* Inform the ZIO pipeline that we are non-rotational */ v->vdev_nonrot = blk_queue_nonrot(q); /* Physical volume size in bytes for the partition */ *psize = bdev_capacity(vd->vd_bdev); /* Physical volume size in bytes including possible expansion space */ *max_psize = bdev_max_capacity(vd->vd_bdev, v->vdev_wholedisk); /* Based on the minimum sector size set the block size */ *physical_ashift = highbit64(MAX(physical_block_size, SPA_MINBLOCKSIZE)) - 1; *logical_ashift = highbit64(MAX(logical_block_size, SPA_MINBLOCKSIZE)) - 1; return (0); } static void vdev_disk_close(vdev_t *v) { vdev_disk_t *vd = v->vdev_tsd; if (v->vdev_reopening || vd == NULL) return; if (vd->vd_bdev != NULL) { blkdev_put(vd->vd_bdev, vdev_bdev_mode(spa_mode(v->vdev_spa)) | FMODE_EXCL); } rw_destroy(&vd->vd_lock); kmem_free(vd, sizeof (vdev_disk_t)); v->vdev_tsd = NULL; } static dio_request_t * vdev_disk_dio_alloc(int bio_count) { dio_request_t *dr = kmem_zalloc(sizeof (dio_request_t) + sizeof (struct bio *) * bio_count, KM_SLEEP); atomic_set(&dr->dr_ref, 0); dr->dr_bio_count = bio_count; dr->dr_error = 0; for (int i = 0; i < dr->dr_bio_count; i++) dr->dr_bio[i] = NULL; return (dr); } static void vdev_disk_dio_free(dio_request_t *dr) { int i; for (i = 0; i < dr->dr_bio_count; i++) if (dr->dr_bio[i]) bio_put(dr->dr_bio[i]); kmem_free(dr, sizeof (dio_request_t) + sizeof (struct bio *) * dr->dr_bio_count); } static void vdev_disk_dio_get(dio_request_t *dr) { atomic_inc(&dr->dr_ref); } static int vdev_disk_dio_put(dio_request_t *dr) { int rc = atomic_dec_return(&dr->dr_ref); /* * Free the dio_request when the last reference is dropped and * ensure zio_interpret is called only once with the correct zio */ if (rc == 0) { zio_t *zio = dr->dr_zio; int error = dr->dr_error; vdev_disk_dio_free(dr); if (zio) { zio->io_error = error; ASSERT3S(zio->io_error, >=, 0); if (zio->io_error) vdev_disk_error(zio); zio_delay_interrupt(zio); } } return (rc); } BIO_END_IO_PROTO(vdev_disk_physio_completion, bio, error) { dio_request_t *dr = bio->bi_private; int rc; if (dr->dr_error == 0) { #ifdef HAVE_1ARG_BIO_END_IO_T dr->dr_error = BIO_END_IO_ERROR(bio); #else if (error) dr->dr_error = -(error); else if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) dr->dr_error = EIO; #endif } /* Drop reference acquired by __vdev_disk_physio */ rc = vdev_disk_dio_put(dr); } static inline void vdev_submit_bio_impl(struct bio *bio) { #ifdef HAVE_1ARG_SUBMIT_BIO (void) submit_bio(bio); #else (void) submit_bio(bio_data_dir(bio), bio); #endif } /* * preempt_schedule_notrace is GPL-only which breaks the ZFS build, so * replace it with preempt_schedule under the following condition: */ #if defined(CONFIG_ARM64) && \ defined(CONFIG_PREEMPTION) && \ defined(CONFIG_BLK_CGROUP) #define preempt_schedule_notrace(x) preempt_schedule(x) #endif /* * As for the Linux 5.18 kernel bio_alloc() expects a block_device struct * as an argument removing the need to set it with bio_set_dev(). This * removes the need for all of the following compatibility code. */ #if !defined(HAVE_BIO_ALLOC_4ARG) #ifdef HAVE_BIO_SET_DEV #if defined(CONFIG_BLK_CGROUP) && defined(HAVE_BIO_SET_DEV_GPL_ONLY) /* * The Linux 5.5 kernel updated percpu_ref_tryget() which is inlined by * blkg_tryget() to use rcu_read_lock() instead of rcu_read_lock_sched(). * As a side effect the function was converted to GPL-only. Define our * own version when needed which uses rcu_read_lock_sched(). * * The Linux 5.17 kernel split linux/blk-cgroup.h into a private and a public * part, moving blkg_tryget into the private one. Define our own version. */ #if defined(HAVE_BLKG_TRYGET_GPL_ONLY) || !defined(HAVE_BLKG_TRYGET) static inline bool vdev_blkg_tryget(struct blkcg_gq *blkg) { struct percpu_ref *ref = &blkg->refcnt; unsigned long __percpu *count; bool rc; rcu_read_lock_sched(); if (__ref_is_percpu(ref, &count)) { this_cpu_inc(*count); rc = true; } else { #ifdef ZFS_PERCPU_REF_COUNT_IN_DATA rc = atomic_long_inc_not_zero(&ref->data->count); #else rc = atomic_long_inc_not_zero(&ref->count); #endif } rcu_read_unlock_sched(); return (rc); } #else #define vdev_blkg_tryget(bg) blkg_tryget(bg) #endif #ifdef HAVE_BIO_SET_DEV_MACRO /* * The Linux 5.0 kernel updated the bio_set_dev() macro so it calls the * GPL-only bio_associate_blkg() symbol thus inadvertently converting * the entire macro. Provide a minimal version which always assigns the * request queue's root_blkg to the bio. */ static inline void vdev_bio_associate_blkg(struct bio *bio) { #if defined(HAVE_BIO_BDEV_DISK) struct request_queue *q = bio->bi_bdev->bd_disk->queue; #else struct request_queue *q = bio->bi_disk->queue; #endif ASSERT3P(q, !=, NULL); ASSERT3P(bio->bi_blkg, ==, NULL); if (q->root_blkg && vdev_blkg_tryget(q->root_blkg)) bio->bi_blkg = q->root_blkg; } #define bio_associate_blkg vdev_bio_associate_blkg #else static inline void vdev_bio_set_dev(struct bio *bio, struct block_device *bdev) { #if defined(HAVE_BIO_BDEV_DISK) struct request_queue *q = bdev->bd_disk->queue; #else struct request_queue *q = bio->bi_disk->queue; #endif bio_clear_flag(bio, BIO_REMAPPED); if (bio->bi_bdev != bdev) bio_clear_flag(bio, BIO_THROTTLED); bio->bi_bdev = bdev; ASSERT3P(q, !=, NULL); ASSERT3P(bio->bi_blkg, ==, NULL); if (q->root_blkg && vdev_blkg_tryget(q->root_blkg)) bio->bi_blkg = q->root_blkg; } #define bio_set_dev vdev_bio_set_dev #endif #endif #else /* * Provide a bio_set_dev() helper macro for pre-Linux 4.14 kernels. */ static inline void bio_set_dev(struct bio *bio, struct block_device *bdev) { bio->bi_bdev = bdev; } #endif /* HAVE_BIO_SET_DEV */ #endif /* !HAVE_BIO_ALLOC_4ARG */ static inline void vdev_submit_bio(struct bio *bio) { struct bio_list *bio_list = current->bio_list; current->bio_list = NULL; vdev_submit_bio_impl(bio); current->bio_list = bio_list; } static inline struct bio * vdev_bio_alloc(struct block_device *bdev, gfp_t gfp_mask, unsigned short nr_vecs) { struct bio *bio; #ifdef HAVE_BIO_ALLOC_4ARG bio = bio_alloc(bdev, nr_vecs, 0, gfp_mask); #else bio = bio_alloc(gfp_mask, nr_vecs); if (likely(bio != NULL)) bio_set_dev(bio, bdev); #endif return (bio); } static inline unsigned int vdev_bio_max_segs(zio_t *zio, int bio_size, uint64_t abd_offset) { unsigned long nr_segs = abd_nr_pages_off(zio->io_abd, bio_size, abd_offset); #ifdef HAVE_BIO_MAX_SEGS return (bio_max_segs(nr_segs)); #else return (MIN(nr_segs, BIO_MAX_PAGES)); #endif } static int __vdev_disk_physio(struct block_device *bdev, zio_t *zio, size_t io_size, uint64_t io_offset, int rw, int flags) { dio_request_t *dr; uint64_t abd_offset; uint64_t bio_offset; int bio_size; int bio_count = 16; int error = 0; struct blk_plug plug; unsigned short nr_vecs; /* * Accessing outside the block device is never allowed. */ if (io_offset + io_size > bdev->bd_inode->i_size) { vdev_dbgmsg(zio->io_vd, "Illegal access %llu size %llu, device size %llu", io_offset, io_size, i_size_read(bdev->bd_inode)); return (SET_ERROR(EIO)); } retry: dr = vdev_disk_dio_alloc(bio_count); if (zio && !(zio->io_flags & (ZIO_FLAG_IO_RETRY | ZIO_FLAG_TRYHARD))) bio_set_flags_failfast(bdev, &flags); dr->dr_zio = zio; /* * Since bio's can have up to BIO_MAX_PAGES=256 iovec's, each of which * is at least 512 bytes and at most PAGESIZE (typically 4K), one bio * can cover at least 128KB and at most 1MB. When the required number * of iovec's exceeds this, we are forced to break the IO in multiple * bio's and wait for them all to complete. This is likely if the * recordsize property is increased beyond 1MB. The default * bio_count=16 should typically accommodate the maximum-size zio of * 16MB. */ abd_offset = 0; bio_offset = io_offset; bio_size = io_size; for (int i = 0; i <= dr->dr_bio_count; i++) { /* Finished constructing bio's for given buffer */ if (bio_size <= 0) break; /* * If additional bio's are required, we have to retry, but * this should be rare - see the comment above. */ if (dr->dr_bio_count == i) { vdev_disk_dio_free(dr); bio_count *= 2; goto retry; } nr_vecs = vdev_bio_max_segs(zio, bio_size, abd_offset); dr->dr_bio[i] = vdev_bio_alloc(bdev, GFP_NOIO, nr_vecs); if (unlikely(dr->dr_bio[i] == NULL)) { vdev_disk_dio_free(dr); return (SET_ERROR(ENOMEM)); } /* Matching put called by vdev_disk_physio_completion */ vdev_disk_dio_get(dr); BIO_BI_SECTOR(dr->dr_bio[i]) = bio_offset >> 9; dr->dr_bio[i]->bi_end_io = vdev_disk_physio_completion; dr->dr_bio[i]->bi_private = dr; bio_set_op_attrs(dr->dr_bio[i], rw, flags); /* Remaining size is returned to become the new size */ bio_size = abd_bio_map_off(dr->dr_bio[i], zio->io_abd, bio_size, abd_offset); /* Advance in buffer and construct another bio if needed */ abd_offset += BIO_BI_SIZE(dr->dr_bio[i]); bio_offset += BIO_BI_SIZE(dr->dr_bio[i]); } /* Extra reference to protect dio_request during vdev_submit_bio */ vdev_disk_dio_get(dr); if (dr->dr_bio_count > 1) blk_start_plug(&plug); /* Submit all bio's associated with this dio */ for (int i = 0; i < dr->dr_bio_count; i++) { if (dr->dr_bio[i]) vdev_submit_bio(dr->dr_bio[i]); } if (dr->dr_bio_count > 1) blk_finish_plug(&plug); (void) vdev_disk_dio_put(dr); return (error); } BIO_END_IO_PROTO(vdev_disk_io_flush_completion, bio, error) { zio_t *zio = bio->bi_private; #ifdef HAVE_1ARG_BIO_END_IO_T zio->io_error = BIO_END_IO_ERROR(bio); #else zio->io_error = -error; #endif if (zio->io_error && (zio->io_error == EOPNOTSUPP)) zio->io_vd->vdev_nowritecache = B_TRUE; bio_put(bio); ASSERT3S(zio->io_error, >=, 0); if (zio->io_error) vdev_disk_error(zio); zio_interrupt(zio); } static int vdev_disk_io_flush(struct block_device *bdev, zio_t *zio) { struct request_queue *q; struct bio *bio; q = bdev_get_queue(bdev); if (!q) return (SET_ERROR(ENXIO)); bio = vdev_bio_alloc(bdev, GFP_NOIO, 0); if (unlikely(bio == NULL)) return (SET_ERROR(ENOMEM)); bio->bi_end_io = vdev_disk_io_flush_completion; bio->bi_private = zio; bio_set_flush(bio); vdev_submit_bio(bio); invalidate_bdev(bdev); return (0); } static void vdev_disk_io_start(zio_t *zio) { vdev_t *v = zio->io_vd; vdev_disk_t *vd = v->vdev_tsd; unsigned long trim_flags = 0; int rw, error; /* * If the vdev is closed, it's likely in the REMOVED or FAULTED state. * Nothing to be done here but return failure. */ if (vd == NULL) { zio->io_error = ENXIO; zio_interrupt(zio); return; } rw_enter(&vd->vd_lock, RW_READER); /* * If the vdev is closed, it's likely due to a failed reopen and is * in the UNAVAIL state. Nothing to be done here but return failure. */ if (vd->vd_bdev == NULL) { rw_exit(&vd->vd_lock); zio->io_error = ENXIO; zio_interrupt(zio); return; } switch (zio->io_type) { case ZIO_TYPE_IOCTL: if (!vdev_readable(v)) { rw_exit(&vd->vd_lock); zio->io_error = SET_ERROR(ENXIO); zio_interrupt(zio); return; } switch (zio->io_cmd) { case DKIOCFLUSHWRITECACHE: if (zfs_nocacheflush) break; if (v->vdev_nowritecache) { zio->io_error = SET_ERROR(ENOTSUP); break; } error = vdev_disk_io_flush(vd->vd_bdev, zio); if (error == 0) { rw_exit(&vd->vd_lock); return; } zio->io_error = error; break; default: zio->io_error = SET_ERROR(ENOTSUP); } rw_exit(&vd->vd_lock); zio_execute(zio); return; case ZIO_TYPE_WRITE: rw = WRITE; break; case ZIO_TYPE_READ: rw = READ; break; case ZIO_TYPE_TRIM: #if defined(BLKDEV_DISCARD_SECURE) if (zio->io_trim_flags & ZIO_TRIM_SECURE) trim_flags |= BLKDEV_DISCARD_SECURE; #endif zio->io_error = -blkdev_issue_discard(vd->vd_bdev, zio->io_offset >> 9, zio->io_size >> 9, GFP_NOFS, trim_flags); rw_exit(&vd->vd_lock); zio_interrupt(zio); return; default: rw_exit(&vd->vd_lock); zio->io_error = SET_ERROR(ENOTSUP); zio_interrupt(zio); return; } zio->io_target_timestamp = zio_handle_io_delay(zio); error = __vdev_disk_physio(vd->vd_bdev, zio, zio->io_size, zio->io_offset, rw, 0); rw_exit(&vd->vd_lock); if (error) { zio->io_error = error; zio_interrupt(zio); return; } } static void vdev_disk_io_done(zio_t *zio) { /* * If the device returned EIO, we revalidate the media. If it is * determined the media has changed this triggers the asynchronous * removal of the device from the configuration. */ if (zio->io_error == EIO) { vdev_t *v = zio->io_vd; vdev_disk_t *vd = v->vdev_tsd; if (zfs_check_media_change(vd->vd_bdev)) { invalidate_bdev(vd->vd_bdev); v->vdev_remove_wanted = B_TRUE; spa_async_request(zio->io_spa, SPA_ASYNC_REMOVE); } } } static void vdev_disk_hold(vdev_t *vd) { ASSERT(spa_config_held(vd->vdev_spa, SCL_STATE, RW_WRITER)); /* We must have a pathname, and it must be absolute. */ if (vd->vdev_path == NULL || vd->vdev_path[0] != '/') return; /* * Only prefetch path and devid info if the device has * never been opened. */ if (vd->vdev_tsd != NULL) return; } static void vdev_disk_rele(vdev_t *vd) { ASSERT(spa_config_held(vd->vdev_spa, SCL_STATE, RW_WRITER)); /* XXX: Implement me as a vnode rele for the device */ } vdev_ops_t vdev_disk_ops = { .vdev_op_init = NULL, .vdev_op_fini = NULL, .vdev_op_open = vdev_disk_open, .vdev_op_close = vdev_disk_close, .vdev_op_asize = vdev_default_asize, .vdev_op_min_asize = vdev_default_min_asize, .vdev_op_min_alloc = NULL, .vdev_op_io_start = vdev_disk_io_start, .vdev_op_io_done = vdev_disk_io_done, .vdev_op_state_change = NULL, .vdev_op_need_resilver = NULL, .vdev_op_hold = vdev_disk_hold, .vdev_op_rele = vdev_disk_rele, .vdev_op_remap = NULL, .vdev_op_xlate = vdev_default_xlate, .vdev_op_rebuild_asize = NULL, .vdev_op_metaslab_init = NULL, .vdev_op_config_generate = NULL, .vdev_op_nparity = NULL, .vdev_op_ndisks = NULL, .vdev_op_type = VDEV_TYPE_DISK, /* name of this vdev type */ .vdev_op_leaf = B_TRUE /* leaf vdev */ }; /* * The zfs_vdev_scheduler module option has been deprecated. Setting this * value no longer has any effect. It has not yet been entirely removed * to allow the module to be loaded if this option is specified in the * /etc/modprobe.d/zfs.conf file. The following warning will be logged. */ static int param_set_vdev_scheduler(const char *val, zfs_kernel_param_t *kp) { int error = param_set_charp(val, kp); if (error == 0) { printk(KERN_INFO "The 'zfs_vdev_scheduler' module option " "is not supported.\n"); } return (error); } char *zfs_vdev_scheduler = "unused"; module_param_call(zfs_vdev_scheduler, param_set_vdev_scheduler, param_get_charp, &zfs_vdev_scheduler, 0644); MODULE_PARM_DESC(zfs_vdev_scheduler, "I/O scheduler"); int param_set_min_auto_ashift(const char *buf, zfs_kernel_param_t *kp) { uint64_t val; int error; error = kstrtoull(buf, 0, &val); if (error < 0) return (SET_ERROR(error)); if (val < ASHIFT_MIN || val > zfs_vdev_max_auto_ashift) return (SET_ERROR(-EINVAL)); error = param_set_ulong(buf, kp); if (error < 0) return (SET_ERROR(error)); return (0); } int param_set_max_auto_ashift(const char *buf, zfs_kernel_param_t *kp) { uint64_t val; int error; error = kstrtoull(buf, 0, &val); if (error < 0) return (SET_ERROR(error)); if (val > ASHIFT_MAX || val < zfs_vdev_min_auto_ashift) return (SET_ERROR(-EINVAL)); error = param_set_ulong(buf, kp); if (error < 0) return (SET_ERROR(error)); return (0); } diff --git a/module/os/linux/zfs/zvol_os.c b/module/os/linux/zfs/zvol_os.c index 5c509a06a4fb..bb92c65ca810 100644 --- a/module/os/linux/zfs/zvol_os.c +++ b/module/os/linux/zfs/zvol_os.c @@ -1,1225 +1,1227 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2012, 2020 by Delphix. All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include unsigned int zvol_major = ZVOL_MAJOR; unsigned int zvol_request_sync = 0; unsigned int zvol_prefetch_bytes = (128 * 1024); unsigned long zvol_max_discard_blocks = 16384; unsigned int zvol_threads = 32; unsigned int zvol_open_timeout_ms = 1000; struct zvol_state_os { struct gendisk *zvo_disk; /* generic disk */ struct request_queue *zvo_queue; /* request queue */ dev_t zvo_dev; /* device id */ }; taskq_t *zvol_taskq; static struct ida zvol_ida; typedef struct zv_request_stack { zvol_state_t *zv; struct bio *bio; } zv_request_t; typedef struct zv_request_task { zv_request_t zvr; taskq_ent_t ent; } zv_request_task_t; static zv_request_task_t * zv_request_task_create(zv_request_t zvr) { zv_request_task_t *task; task = kmem_alloc(sizeof (zv_request_task_t), KM_SLEEP); taskq_init_ent(&task->ent); task->zvr = zvr; return (task); } static void zv_request_task_free(zv_request_task_t *task) { kmem_free(task, sizeof (*task)); } /* * Given a path, return TRUE if path is a ZVOL. */ static boolean_t zvol_is_zvol_impl(const char *path) { dev_t dev = 0; if (vdev_lookup_bdev(path, &dev) != 0) return (B_FALSE); if (MAJOR(dev) == zvol_major) return (B_TRUE); return (B_FALSE); } static void zvol_write(zv_request_t *zvr) { struct bio *bio = zvr->bio; int error = 0; zfs_uio_t uio; zfs_uio_bvec_init(&uio, bio); zvol_state_t *zv = zvr->zv; ASSERT3P(zv, !=, NULL); ASSERT3U(zv->zv_open_count, >, 0); ASSERT3P(zv->zv_zilog, !=, NULL); /* bio marked as FLUSH need to flush before write */ if (bio_is_flush(bio)) zil_commit(zv->zv_zilog, ZVOL_OBJ); /* Some requests are just for flush and nothing else. */ if (uio.uio_resid == 0) { rw_exit(&zv->zv_suspend_lock); BIO_END_IO(bio, 0); return; } struct request_queue *q = zv->zv_zso->zvo_queue; struct gendisk *disk = zv->zv_zso->zvo_disk; ssize_t start_resid = uio.uio_resid; unsigned long start_time; boolean_t acct = blk_queue_io_stat(q); if (acct) start_time = blk_generic_start_io_acct(q, disk, WRITE, bio); boolean_t sync = bio_is_fua(bio) || zv->zv_objset->os_sync == ZFS_SYNC_ALWAYS; zfs_locked_range_t *lr = zfs_rangelock_enter(&zv->zv_rangelock, uio.uio_loffset, uio.uio_resid, RL_WRITER); uint64_t volsize = zv->zv_volsize; while (uio.uio_resid > 0 && uio.uio_loffset < volsize) { uint64_t bytes = MIN(uio.uio_resid, DMU_MAX_ACCESS >> 1); uint64_t off = uio.uio_loffset; dmu_tx_t *tx = dmu_tx_create(zv->zv_objset); if (bytes > volsize - off) /* don't write past the end */ bytes = volsize - off; dmu_tx_hold_write_by_dnode(tx, zv->zv_dn, off, bytes); /* This will only fail for ENOSPC */ error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); break; } error = dmu_write_uio_dnode(zv->zv_dn, &uio, bytes, tx); if (error == 0) { zvol_log_write(zv, tx, off, bytes, sync); } dmu_tx_commit(tx); if (error) break; } zfs_rangelock_exit(lr); int64_t nwritten = start_resid - uio.uio_resid; dataset_kstats_update_write_kstats(&zv->zv_kstat, nwritten); task_io_account_write(nwritten); if (sync) zil_commit(zv->zv_zilog, ZVOL_OBJ); rw_exit(&zv->zv_suspend_lock); if (acct) blk_generic_end_io_acct(q, disk, WRITE, bio, start_time); BIO_END_IO(bio, -error); } static void zvol_write_task(void *arg) { zv_request_task_t *task = arg; zvol_write(&task->zvr); zv_request_task_free(task); } static void zvol_discard(zv_request_t *zvr) { struct bio *bio = zvr->bio; zvol_state_t *zv = zvr->zv; uint64_t start = BIO_BI_SECTOR(bio) << 9; uint64_t size = BIO_BI_SIZE(bio); uint64_t end = start + size; boolean_t sync; int error = 0; dmu_tx_t *tx; ASSERT3P(zv, !=, NULL); ASSERT3U(zv->zv_open_count, >, 0); ASSERT3P(zv->zv_zilog, !=, NULL); struct request_queue *q = zv->zv_zso->zvo_queue; struct gendisk *disk = zv->zv_zso->zvo_disk; unsigned long start_time; boolean_t acct = blk_queue_io_stat(q); if (acct) start_time = blk_generic_start_io_acct(q, disk, WRITE, bio); sync = bio_is_fua(bio) || zv->zv_objset->os_sync == ZFS_SYNC_ALWAYS; if (end > zv->zv_volsize) { error = SET_ERROR(EIO); goto unlock; } /* * Align the request to volume block boundaries when a secure erase is * not required. This will prevent dnode_free_range() from zeroing out * the unaligned parts which is slow (read-modify-write) and useless * since we are not freeing any space by doing so. */ if (!bio_is_secure_erase(bio)) { start = P2ROUNDUP(start, zv->zv_volblocksize); end = P2ALIGN(end, zv->zv_volblocksize); size = end - start; } if (start >= end) goto unlock; zfs_locked_range_t *lr = zfs_rangelock_enter(&zv->zv_rangelock, start, size, RL_WRITER); tx = dmu_tx_create(zv->zv_objset); dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, TXG_WAIT); if (error != 0) { dmu_tx_abort(tx); } else { zvol_log_truncate(zv, tx, start, size, B_TRUE); dmu_tx_commit(tx); error = dmu_free_long_range(zv->zv_objset, ZVOL_OBJ, start, size); } zfs_rangelock_exit(lr); if (error == 0 && sync) zil_commit(zv->zv_zilog, ZVOL_OBJ); unlock: rw_exit(&zv->zv_suspend_lock); if (acct) blk_generic_end_io_acct(q, disk, WRITE, bio, start_time); BIO_END_IO(bio, -error); } static void zvol_discard_task(void *arg) { zv_request_task_t *task = arg; zvol_discard(&task->zvr); zv_request_task_free(task); } static void zvol_read(zv_request_t *zvr) { struct bio *bio = zvr->bio; int error = 0; zfs_uio_t uio; zfs_uio_bvec_init(&uio, bio); zvol_state_t *zv = zvr->zv; ASSERT3P(zv, !=, NULL); ASSERT3U(zv->zv_open_count, >, 0); struct request_queue *q = zv->zv_zso->zvo_queue; struct gendisk *disk = zv->zv_zso->zvo_disk; ssize_t start_resid = uio.uio_resid; unsigned long start_time; boolean_t acct = blk_queue_io_stat(q); if (acct) start_time = blk_generic_start_io_acct(q, disk, READ, bio); zfs_locked_range_t *lr = zfs_rangelock_enter(&zv->zv_rangelock, uio.uio_loffset, uio.uio_resid, RL_READER); uint64_t volsize = zv->zv_volsize; while (uio.uio_resid > 0 && uio.uio_loffset < volsize) { uint64_t bytes = MIN(uio.uio_resid, DMU_MAX_ACCESS >> 1); /* don't read past the end */ if (bytes > volsize - uio.uio_loffset) bytes = volsize - uio.uio_loffset; error = dmu_read_uio_dnode(zv->zv_dn, &uio, bytes); if (error) { /* convert checksum errors into IO errors */ if (error == ECKSUM) error = SET_ERROR(EIO); break; } } zfs_rangelock_exit(lr); int64_t nread = start_resid - uio.uio_resid; dataset_kstats_update_read_kstats(&zv->zv_kstat, nread); task_io_account_read(nread); rw_exit(&zv->zv_suspend_lock); if (acct) blk_generic_end_io_acct(q, disk, READ, bio, start_time); BIO_END_IO(bio, -error); } static void zvol_read_task(void *arg) { zv_request_task_t *task = arg; zvol_read(&task->zvr); zv_request_task_free(task); } #ifdef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS #ifdef HAVE_BDEV_SUBMIT_BIO_RETURNS_VOID static void zvol_submit_bio(struct bio *bio) #else static blk_qc_t zvol_submit_bio(struct bio *bio) #endif #else static MAKE_REQUEST_FN_RET zvol_request(struct request_queue *q, struct bio *bio) #endif { #ifdef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS #if defined(HAVE_BIO_BDEV_DISK) struct request_queue *q = bio->bi_bdev->bd_disk->queue; #else struct request_queue *q = bio->bi_disk->queue; #endif #endif zvol_state_t *zv = q->queuedata; fstrans_cookie_t cookie = spl_fstrans_mark(); uint64_t offset = BIO_BI_SECTOR(bio) << 9; uint64_t size = BIO_BI_SIZE(bio); int rw = bio_data_dir(bio); if (bio_has_data(bio) && offset + size > zv->zv_volsize) { printk(KERN_INFO "%s: bad access: offset=%llu, size=%lu\n", zv->zv_zso->zvo_disk->disk_name, (long long unsigned)offset, (long unsigned)size); BIO_END_IO(bio, -SET_ERROR(EIO)); goto out; } zv_request_t zvr = { .zv = zv, .bio = bio, }; zv_request_task_t *task; if (rw == WRITE) { if (unlikely(zv->zv_flags & ZVOL_RDONLY)) { BIO_END_IO(bio, -SET_ERROR(EROFS)); goto out; } /* * Prevents the zvol from being suspended, or the ZIL being * concurrently opened. Will be released after the i/o * completes. */ rw_enter(&zv->zv_suspend_lock, RW_READER); /* * Open a ZIL if this is the first time we have written to this * zvol. We protect zv->zv_zilog with zv_suspend_lock rather * than zv_state_lock so that we don't need to acquire an * additional lock in this path. */ if (zv->zv_zilog == NULL) { rw_exit(&zv->zv_suspend_lock); rw_enter(&zv->zv_suspend_lock, RW_WRITER); if (zv->zv_zilog == NULL) { zv->zv_zilog = zil_open(zv->zv_objset, zvol_get_data); zv->zv_flags |= ZVOL_WRITTEN_TO; /* replay / destroy done in zvol_create_minor */ VERIFY0((zv->zv_zilog->zl_header->zh_flags & ZIL_REPLAY_NEEDED)); } rw_downgrade(&zv->zv_suspend_lock); } /* * We don't want this thread to be blocked waiting for i/o to * complete, so we instead wait from a taskq callback. The * i/o may be a ZIL write (via zil_commit()), or a read of an * indirect block, or a read of a data block (if this is a * partial-block write). We will indicate that the i/o is * complete by calling BIO_END_IO() from the taskq callback. * * This design allows the calling thread to continue and * initiate more concurrent operations by calling * zvol_request() again. There are typically only a small * number of threads available to call zvol_request() (e.g. * one per iSCSI target), so keeping the latency of * zvol_request() low is important for performance. * * The zvol_request_sync module parameter allows this * behavior to be altered, for performance evaluation * purposes. If the callback blocks, setting * zvol_request_sync=1 will result in much worse performance. * * We can have up to zvol_threads concurrent i/o's being * processed for all zvols on the system. This is typically * a vast improvement over the zvol_request_sync=1 behavior * of one i/o at a time per zvol. However, an even better * design would be for zvol_request() to initiate the zio * directly, and then be notified by the zio_done callback, * which would call BIO_END_IO(). Unfortunately, the DMU/ZIL * interfaces lack this functionality (they block waiting for * the i/o to complete). */ if (bio_is_discard(bio) || bio_is_secure_erase(bio)) { if (zvol_request_sync) { zvol_discard(&zvr); } else { task = zv_request_task_create(zvr); taskq_dispatch_ent(zvol_taskq, zvol_discard_task, task, 0, &task->ent); } } else { if (zvol_request_sync) { zvol_write(&zvr); } else { task = zv_request_task_create(zvr); taskq_dispatch_ent(zvol_taskq, zvol_write_task, task, 0, &task->ent); } } } else { /* * The SCST driver, and possibly others, may issue READ I/Os * with a length of zero bytes. These empty I/Os contain no * data and require no additional handling. */ if (size == 0) { BIO_END_IO(bio, 0); goto out; } rw_enter(&zv->zv_suspend_lock, RW_READER); /* See comment in WRITE case above. */ if (zvol_request_sync) { zvol_read(&zvr); } else { task = zv_request_task_create(zvr); taskq_dispatch_ent(zvol_taskq, zvol_read_task, task, 0, &task->ent); } } out: spl_fstrans_unmark(cookie); #if (defined(HAVE_MAKE_REQUEST_FN_RET_QC) || \ defined(HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS)) && \ !defined(HAVE_BDEV_SUBMIT_BIO_RETURNS_VOID) return (BLK_QC_T_NONE); #endif } static int zvol_open(struct block_device *bdev, fmode_t flag) { zvol_state_t *zv; int error = 0; boolean_t drop_suspend = B_FALSE; #ifndef HAVE_BLKDEV_GET_ERESTARTSYS hrtime_t timeout = MSEC2NSEC(zvol_open_timeout_ms); hrtime_t start = gethrtime(); retry: #endif rw_enter(&zvol_state_lock, RW_READER); /* * Obtain a copy of private_data under the zvol_state_lock to make * sure that either the result of zvol free code path setting * bdev->bd_disk->private_data to NULL is observed, or zvol_free() * is not called on this zv because of the positive zv_open_count. */ zv = bdev->bd_disk->private_data; if (zv == NULL) { rw_exit(&zvol_state_lock); return (SET_ERROR(-ENXIO)); } mutex_enter(&zv->zv_state_lock); /* * Make sure zvol is not suspended during first open * (hold zv_suspend_lock) and respect proper lock acquisition * ordering - zv_suspend_lock before zv_state_lock */ if (zv->zv_open_count == 0) { if (!rw_tryenter(&zv->zv_suspend_lock, RW_READER)) { mutex_exit(&zv->zv_state_lock); rw_enter(&zv->zv_suspend_lock, RW_READER); mutex_enter(&zv->zv_state_lock); /* check to see if zv_suspend_lock is needed */ if (zv->zv_open_count != 0) { rw_exit(&zv->zv_suspend_lock); } else { drop_suspend = B_TRUE; } } else { drop_suspend = B_TRUE; } } rw_exit(&zvol_state_lock); ASSERT(MUTEX_HELD(&zv->zv_state_lock)); if (zv->zv_open_count == 0) { boolean_t drop_namespace = B_FALSE; ASSERT(RW_READ_HELD(&zv->zv_suspend_lock)); /* * In all other call paths the spa_namespace_lock is taken * before the bdev->bd_mutex lock. However, on open(2) * the __blkdev_get() function calls fops->open() with the * bdev->bd_mutex lock held. This can result in a deadlock * when zvols from one pool are used as vdevs in another. * * To prevent a lock inversion deadlock we preemptively * take the spa_namespace_lock. Normally the lock will not * be contended and this is safe because spa_open_common() * handles the case where the caller already holds the * spa_namespace_lock. * * When the lock cannot be aquired after multiple retries * this must be the vdev on zvol deadlock case and we have * no choice but to return an error. For 5.12 and older * kernels returning -ERESTARTSYS will result in the * bdev->bd_mutex being dropped, then reacquired, and * fops->open() being called again. This process can be * repeated safely until both locks are acquired. For 5.13 * and newer the -ERESTARTSYS retry logic was removed from * the kernel so the only option is to return the error for * the caller to handle it. */ if (!mutex_owned(&spa_namespace_lock)) { if (!mutex_tryenter(&spa_namespace_lock)) { mutex_exit(&zv->zv_state_lock); rw_exit(&zv->zv_suspend_lock); #ifdef HAVE_BLKDEV_GET_ERESTARTSYS schedule(); return (SET_ERROR(-ERESTARTSYS)); #else if ((gethrtime() - start) > timeout) return (SET_ERROR(-ERESTARTSYS)); schedule_timeout(MSEC_TO_TICK(10)); goto retry; #endif } else { drop_namespace = B_TRUE; } } error = -zvol_first_open(zv, !(flag & FMODE_WRITE)); if (drop_namespace) mutex_exit(&spa_namespace_lock); } if (error == 0) { if ((flag & FMODE_WRITE) && (zv->zv_flags & ZVOL_RDONLY)) { if (zv->zv_open_count == 0) zvol_last_close(zv); error = SET_ERROR(-EROFS); } else { zv->zv_open_count++; } } mutex_exit(&zv->zv_state_lock); if (drop_suspend) rw_exit(&zv->zv_suspend_lock); if (error == 0) zfs_check_media_change(bdev); return (error); } static void zvol_release(struct gendisk *disk, fmode_t mode) { zvol_state_t *zv; boolean_t drop_suspend = B_TRUE; rw_enter(&zvol_state_lock, RW_READER); zv = disk->private_data; mutex_enter(&zv->zv_state_lock); ASSERT3U(zv->zv_open_count, >, 0); /* * make sure zvol is not suspended during last close * (hold zv_suspend_lock) and respect proper lock acquisition * ordering - zv_suspend_lock before zv_state_lock */ if (zv->zv_open_count == 1) { if (!rw_tryenter(&zv->zv_suspend_lock, RW_READER)) { mutex_exit(&zv->zv_state_lock); rw_enter(&zv->zv_suspend_lock, RW_READER); mutex_enter(&zv->zv_state_lock); /* check to see if zv_suspend_lock is needed */ if (zv->zv_open_count != 1) { rw_exit(&zv->zv_suspend_lock); drop_suspend = B_FALSE; } } } else { drop_suspend = B_FALSE; } rw_exit(&zvol_state_lock); ASSERT(MUTEX_HELD(&zv->zv_state_lock)); zv->zv_open_count--; if (zv->zv_open_count == 0) { ASSERT(RW_READ_HELD(&zv->zv_suspend_lock)); zvol_last_close(zv); } mutex_exit(&zv->zv_state_lock); if (drop_suspend) rw_exit(&zv->zv_suspend_lock); } static int zvol_ioctl(struct block_device *bdev, fmode_t mode, unsigned int cmd, unsigned long arg) { zvol_state_t *zv = bdev->bd_disk->private_data; int error = 0; ASSERT3U(zv->zv_open_count, >, 0); switch (cmd) { case BLKFLSBUF: fsync_bdev(bdev); invalidate_bdev(bdev); rw_enter(&zv->zv_suspend_lock, RW_READER); if (!(zv->zv_flags & ZVOL_RDONLY)) txg_wait_synced(dmu_objset_pool(zv->zv_objset), 0); rw_exit(&zv->zv_suspend_lock); break; case BLKZNAME: mutex_enter(&zv->zv_state_lock); error = copy_to_user((void *)arg, zv->zv_name, MAXNAMELEN); mutex_exit(&zv->zv_state_lock); break; default: error = -ENOTTY; break; } return (SET_ERROR(error)); } #ifdef CONFIG_COMPAT static int zvol_compat_ioctl(struct block_device *bdev, fmode_t mode, unsigned cmd, unsigned long arg) { return (zvol_ioctl(bdev, mode, cmd, arg)); } #else #define zvol_compat_ioctl NULL #endif static unsigned int zvol_check_events(struct gendisk *disk, unsigned int clearing) { unsigned int mask = 0; rw_enter(&zvol_state_lock, RW_READER); zvol_state_t *zv = disk->private_data; if (zv != NULL) { mutex_enter(&zv->zv_state_lock); mask = zv->zv_changed ? DISK_EVENT_MEDIA_CHANGE : 0; zv->zv_changed = 0; mutex_exit(&zv->zv_state_lock); } rw_exit(&zvol_state_lock); return (mask); } static int zvol_revalidate_disk(struct gendisk *disk) { rw_enter(&zvol_state_lock, RW_READER); zvol_state_t *zv = disk->private_data; if (zv != NULL) { mutex_enter(&zv->zv_state_lock); set_capacity(zv->zv_zso->zvo_disk, zv->zv_volsize >> SECTOR_BITS); mutex_exit(&zv->zv_state_lock); } rw_exit(&zvol_state_lock); return (0); } static int zvol_update_volsize(zvol_state_t *zv, uint64_t volsize) { struct gendisk *disk = zv->zv_zso->zvo_disk; #if defined(HAVE_REVALIDATE_DISK_SIZE) revalidate_disk_size(disk, zvol_revalidate_disk(disk) == 0); #elif defined(HAVE_REVALIDATE_DISK) revalidate_disk(disk); #else zvol_revalidate_disk(disk); #endif return (0); } static void zvol_clear_private(zvol_state_t *zv) { /* * Cleared while holding zvol_state_lock as a writer * which will prevent zvol_open() from opening it. */ zv->zv_zso->zvo_disk->private_data = NULL; } /* * Provide a simple virtual geometry for legacy compatibility. For devices * smaller than 1 MiB a small head and sector count is used to allow very * tiny devices. For devices over 1 Mib a standard head and sector count * is used to keep the cylinders count reasonable. */ static int zvol_getgeo(struct block_device *bdev, struct hd_geometry *geo) { zvol_state_t *zv = bdev->bd_disk->private_data; sector_t sectors; ASSERT3U(zv->zv_open_count, >, 0); sectors = get_capacity(zv->zv_zso->zvo_disk); if (sectors > 2048) { geo->heads = 16; geo->sectors = 63; } else { geo->heads = 2; geo->sectors = 4; } geo->start = 0; geo->cylinders = sectors / (geo->heads * geo->sectors); return (0); } static struct block_device_operations zvol_ops = { .open = zvol_open, .release = zvol_release, .ioctl = zvol_ioctl, .compat_ioctl = zvol_compat_ioctl, .check_events = zvol_check_events, #ifdef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK .revalidate_disk = zvol_revalidate_disk, #endif .getgeo = zvol_getgeo, .owner = THIS_MODULE, #ifdef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS .submit_bio = zvol_submit_bio, #endif }; /* * Allocate memory for a new zvol_state_t and setup the required * request queue and generic disk structures for the block device. */ static zvol_state_t * zvol_alloc(dev_t dev, const char *name) { zvol_state_t *zv; struct zvol_state_os *zso; uint64_t volmode; if (dsl_prop_get_integer(name, "volmode", &volmode, NULL) != 0) return (NULL); if (volmode == ZFS_VOLMODE_DEFAULT) volmode = zvol_volmode; if (volmode == ZFS_VOLMODE_NONE) return (NULL); zv = kmem_zalloc(sizeof (zvol_state_t), KM_SLEEP); zso = kmem_zalloc(sizeof (struct zvol_state_os), KM_SLEEP); zv->zv_zso = zso; zv->zv_volmode = volmode; list_link_init(&zv->zv_next); mutex_init(&zv->zv_state_lock, NULL, MUTEX_DEFAULT, NULL); #ifdef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS #ifdef HAVE_BLK_ALLOC_DISK zso->zvo_disk = blk_alloc_disk(NUMA_NO_NODE); if (zso->zvo_disk == NULL) goto out_kmem; zso->zvo_disk->minors = ZVOL_MINORS; zso->zvo_queue = zso->zvo_disk->queue; #else zso->zvo_queue = blk_alloc_queue(NUMA_NO_NODE); if (zso->zvo_queue == NULL) goto out_kmem; zso->zvo_disk = alloc_disk(ZVOL_MINORS); if (zso->zvo_disk == NULL) { blk_cleanup_queue(zso->zvo_queue); goto out_kmem; } zso->zvo_disk->queue = zso->zvo_queue; #endif /* HAVE_BLK_ALLOC_DISK */ #else zso->zvo_queue = blk_generic_alloc_queue(zvol_request, NUMA_NO_NODE); if (zso->zvo_queue == NULL) goto out_kmem; zso->zvo_disk = alloc_disk(ZVOL_MINORS); if (zso->zvo_disk == NULL) { blk_cleanup_queue(zso->zvo_queue); goto out_kmem; } zso->zvo_disk->queue = zso->zvo_queue; #endif /* HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS */ blk_queue_set_write_cache(zso->zvo_queue, B_TRUE, B_TRUE); /* Limit read-ahead to a single page to prevent over-prefetching. */ blk_queue_set_read_ahead(zso->zvo_queue, 1); /* Disable write merging in favor of the ZIO pipeline. */ blk_queue_flag_set(QUEUE_FLAG_NOMERGES, zso->zvo_queue); /* Enable /proc/diskstats */ blk_queue_flag_set(QUEUE_FLAG_IO_STAT, zso->zvo_queue); zso->zvo_queue->queuedata = zv; zso->zvo_dev = dev; zv->zv_open_count = 0; strlcpy(zv->zv_name, name, MAXNAMELEN); zfs_rangelock_init(&zv->zv_rangelock, NULL, NULL); rw_init(&zv->zv_suspend_lock, NULL, RW_DEFAULT, NULL); zso->zvo_disk->major = zvol_major; zso->zvo_disk->events = DISK_EVENT_MEDIA_CHANGE; /* * Setting ZFS_VOLMODE_DEV disables partitioning on ZVOL devices. * This is accomplished by limiting the number of minors for the * device to one and explicitly disabling partition scanning. */ if (volmode == ZFS_VOLMODE_DEV) { zso->zvo_disk->minors = 1; zso->zvo_disk->flags &= ~ZFS_GENHD_FL_EXT_DEVT; zso->zvo_disk->flags |= ZFS_GENHD_FL_NO_PART; } zso->zvo_disk->first_minor = (dev & MINORMASK); zso->zvo_disk->fops = &zvol_ops; zso->zvo_disk->private_data = zv; snprintf(zso->zvo_disk->disk_name, DISK_NAME_LEN, "%s%d", ZVOL_DEV_NAME, (dev & MINORMASK)); return (zv); out_kmem: kmem_free(zso, sizeof (struct zvol_state_os)); kmem_free(zv, sizeof (zvol_state_t)); return (NULL); } /* * Cleanup then free a zvol_state_t which was created by zvol_alloc(). * At this time, the structure is not opened by anyone, is taken off * the zvol_state_list, and has its private data set to NULL. * The zvol_state_lock is dropped. * * This function may take many milliseconds to complete (e.g. we've seen * it take over 256ms), due to the calls to "blk_cleanup_queue" and * "del_gendisk". Thus, consumers need to be careful to account for this * latency when calling this function. */ static void zvol_free(zvol_state_t *zv) { ASSERT(!RW_LOCK_HELD(&zv->zv_suspend_lock)); ASSERT(!MUTEX_HELD(&zv->zv_state_lock)); ASSERT0(zv->zv_open_count); ASSERT3P(zv->zv_zso->zvo_disk->private_data, ==, NULL); rw_destroy(&zv->zv_suspend_lock); zfs_rangelock_fini(&zv->zv_rangelock); del_gendisk(zv->zv_zso->zvo_disk); #if defined(HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS) && \ defined(HAVE_BLK_ALLOC_DISK) blk_cleanup_disk(zv->zv_zso->zvo_disk); #else blk_cleanup_queue(zv->zv_zso->zvo_queue); put_disk(zv->zv_zso->zvo_disk); #endif ida_simple_remove(&zvol_ida, MINOR(zv->zv_zso->zvo_dev) >> ZVOL_MINOR_BITS); mutex_destroy(&zv->zv_state_lock); dataset_kstats_destroy(&zv->zv_kstat); kmem_free(zv->zv_zso, sizeof (struct zvol_state_os)); kmem_free(zv, sizeof (zvol_state_t)); } void zvol_wait_close(zvol_state_t *zv) { } /* * Create a block device minor node and setup the linkage between it * and the specified volume. Once this function returns the block * device is live and ready for use. */ static int zvol_os_create_minor(const char *name) { zvol_state_t *zv; objset_t *os; dmu_object_info_t *doi; uint64_t volsize; uint64_t len; unsigned minor = 0; int error = 0; int idx; uint64_t hash = zvol_name_hash(name); if (zvol_inhibit_dev) return (0); idx = ida_simple_get(&zvol_ida, 0, 0, kmem_flags_convert(KM_SLEEP)); if (idx < 0) return (SET_ERROR(-idx)); minor = idx << ZVOL_MINOR_BITS; zv = zvol_find_by_name_hash(name, hash, RW_NONE); if (zv) { ASSERT(MUTEX_HELD(&zv->zv_state_lock)); mutex_exit(&zv->zv_state_lock); ida_simple_remove(&zvol_ida, idx); return (SET_ERROR(EEXIST)); } doi = kmem_alloc(sizeof (dmu_object_info_t), KM_SLEEP); error = dmu_objset_own(name, DMU_OST_ZVOL, B_TRUE, B_TRUE, FTAG, &os); if (error) goto out_doi; error = dmu_object_info(os, ZVOL_OBJ, doi); if (error) goto out_dmu_objset_disown; error = zap_lookup(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize); if (error) goto out_dmu_objset_disown; zv = zvol_alloc(MKDEV(zvol_major, minor), name); if (zv == NULL) { error = SET_ERROR(EAGAIN); goto out_dmu_objset_disown; } zv->zv_hash = hash; if (dmu_objset_is_snapshot(os)) zv->zv_flags |= ZVOL_RDONLY; zv->zv_volblocksize = doi->doi_data_block_size; zv->zv_volsize = volsize; zv->zv_objset = os; set_capacity(zv->zv_zso->zvo_disk, zv->zv_volsize >> 9); blk_queue_max_hw_sectors(zv->zv_zso->zvo_queue, (DMU_MAX_ACCESS / 4) >> 9); blk_queue_max_segments(zv->zv_zso->zvo_queue, UINT16_MAX); blk_queue_max_segment_size(zv->zv_zso->zvo_queue, UINT_MAX); blk_queue_physical_block_size(zv->zv_zso->zvo_queue, zv->zv_volblocksize); blk_queue_io_opt(zv->zv_zso->zvo_queue, zv->zv_volblocksize); blk_queue_max_discard_sectors(zv->zv_zso->zvo_queue, (zvol_max_discard_blocks * zv->zv_volblocksize) >> 9); blk_queue_discard_granularity(zv->zv_zso->zvo_queue, zv->zv_volblocksize); +#ifdef QUEUE_FLAG_DISCARD blk_queue_flag_set(QUEUE_FLAG_DISCARD, zv->zv_zso->zvo_queue); +#endif #ifdef QUEUE_FLAG_NONROT blk_queue_flag_set(QUEUE_FLAG_NONROT, zv->zv_zso->zvo_queue); #endif #ifdef QUEUE_FLAG_ADD_RANDOM blk_queue_flag_clear(QUEUE_FLAG_ADD_RANDOM, zv->zv_zso->zvo_queue); #endif /* This flag was introduced in kernel version 4.12. */ #ifdef QUEUE_FLAG_SCSI_PASSTHROUGH blk_queue_flag_set(QUEUE_FLAG_SCSI_PASSTHROUGH, zv->zv_zso->zvo_queue); #endif ASSERT3P(zv->zv_zilog, ==, NULL); zv->zv_zilog = zil_open(os, zvol_get_data); if (spa_writeable(dmu_objset_spa(os))) { if (zil_replay_disable) zil_destroy(zv->zv_zilog, B_FALSE); else zil_replay(os, zv, zvol_replay_vector); } zil_close(zv->zv_zilog); zv->zv_zilog = NULL; ASSERT3P(zv->zv_kstat.dk_kstats, ==, NULL); dataset_kstats_create(&zv->zv_kstat, zv->zv_objset); /* * When udev detects the addition of the device it will immediately * invoke blkid(8) to determine the type of content on the device. * Prefetching the blocks commonly scanned by blkid(8) will speed * up this process. */ len = MIN(MAX(zvol_prefetch_bytes, 0), SPA_MAXBLOCKSIZE); if (len > 0) { dmu_prefetch(os, ZVOL_OBJ, 0, 0, len, ZIO_PRIORITY_SYNC_READ); dmu_prefetch(os, ZVOL_OBJ, 0, volsize - len, len, ZIO_PRIORITY_SYNC_READ); } zv->zv_objset = NULL; out_dmu_objset_disown: dmu_objset_disown(os, B_TRUE, FTAG); out_doi: kmem_free(doi, sizeof (dmu_object_info_t)); /* * Keep in mind that once add_disk() is called, the zvol is * announced to the world, and zvol_open()/zvol_release() can * be called at any time. Incidentally, add_disk() itself calls * zvol_open()->zvol_first_open() and zvol_release()->zvol_last_close() * directly as well. */ if (error == 0) { rw_enter(&zvol_state_lock, RW_WRITER); zvol_insert(zv); rw_exit(&zvol_state_lock); #ifdef HAVE_ADD_DISK_RET error = add_disk(zv->zv_zso->zvo_disk); #else add_disk(zv->zv_zso->zvo_disk); #endif } else { ida_simple_remove(&zvol_ida, idx); } return (error); } static void zvol_rename_minor(zvol_state_t *zv, const char *newname) { int readonly = get_disk_ro(zv->zv_zso->zvo_disk); ASSERT(RW_LOCK_HELD(&zvol_state_lock)); ASSERT(MUTEX_HELD(&zv->zv_state_lock)); strlcpy(zv->zv_name, newname, sizeof (zv->zv_name)); /* move to new hashtable entry */ zv->zv_hash = zvol_name_hash(zv->zv_name); hlist_del(&zv->zv_hlink); hlist_add_head(&zv->zv_hlink, ZVOL_HT_HEAD(zv->zv_hash)); /* * The block device's read-only state is briefly changed causing * a KOBJ_CHANGE uevent to be issued. This ensures udev detects * the name change and fixes the symlinks. This does not change * ZVOL_RDONLY in zv->zv_flags so the actual read-only state never * changes. This would normally be done using kobject_uevent() but * that is a GPL-only symbol which is why we need this workaround. */ set_disk_ro(zv->zv_zso->zvo_disk, !readonly); set_disk_ro(zv->zv_zso->zvo_disk, readonly); } static void zvol_set_disk_ro_impl(zvol_state_t *zv, int flags) { set_disk_ro(zv->zv_zso->zvo_disk, flags); } static void zvol_set_capacity_impl(zvol_state_t *zv, uint64_t capacity) { set_capacity(zv->zv_zso->zvo_disk, capacity); } const static zvol_platform_ops_t zvol_linux_ops = { .zv_free = zvol_free, .zv_rename_minor = zvol_rename_minor, .zv_create_minor = zvol_os_create_minor, .zv_update_volsize = zvol_update_volsize, .zv_clear_private = zvol_clear_private, .zv_is_zvol = zvol_is_zvol_impl, .zv_set_disk_ro = zvol_set_disk_ro_impl, .zv_set_capacity = zvol_set_capacity_impl, }; int zvol_init(void) { int error; int threads = MIN(MAX(zvol_threads, 1), 1024); error = register_blkdev(zvol_major, ZVOL_DRIVER); if (error) { printk(KERN_INFO "ZFS: register_blkdev() failed %d\n", error); return (error); } zvol_taskq = taskq_create(ZVOL_DRIVER, threads, maxclsyspri, threads * 2, INT_MAX, TASKQ_PREPOPULATE | TASKQ_DYNAMIC); if (zvol_taskq == NULL) { unregister_blkdev(zvol_major, ZVOL_DRIVER); return (-ENOMEM); } zvol_init_impl(); ida_init(&zvol_ida); zvol_register_ops(&zvol_linux_ops); return (0); } void zvol_fini(void) { zvol_fini_impl(); unregister_blkdev(zvol_major, ZVOL_DRIVER); taskq_destroy(zvol_taskq); ida_destroy(&zvol_ida); } /* BEGIN CSTYLED */ module_param(zvol_inhibit_dev, uint, 0644); MODULE_PARM_DESC(zvol_inhibit_dev, "Do not create zvol device nodes"); module_param(zvol_major, uint, 0444); MODULE_PARM_DESC(zvol_major, "Major number for zvol device"); module_param(zvol_threads, uint, 0444); MODULE_PARM_DESC(zvol_threads, "Max number of threads to handle I/O requests"); module_param(zvol_request_sync, uint, 0644); MODULE_PARM_DESC(zvol_request_sync, "Synchronously handle bio requests"); module_param(zvol_max_discard_blocks, ulong, 0444); MODULE_PARM_DESC(zvol_max_discard_blocks, "Max number of blocks to discard"); module_param(zvol_prefetch_bytes, uint, 0644); MODULE_PARM_DESC(zvol_prefetch_bytes, "Prefetch N bytes at zvol start+end"); module_param(zvol_volmode, uint, 0644); MODULE_PARM_DESC(zvol_volmode, "Default volmode property value"); /* END CSTYLED */