From d635711daa98be86d4c7fd01499c34f566b54ccb Mon Sep 17 00:00:00 2001 From: André Fabian Silva Delgado Date: Fri, 10 Jun 2016 05:30:17 -0300 Subject: Linux-libre 4.6.2-gnu --- fs/nfs/blocklayout/blocklayout.c | 96 +++++++++++++++++++------- fs/nfs/blocklayout/blocklayout.h | 18 +++-- fs/nfs/blocklayout/dev.c | 144 ++++++++++++++++++++++++++++++++++++++- fs/nfs/blocklayout/extent_tree.c | 44 ++++++++---- fs/nfs/blocklayout/rpc_pipefs.c | 2 +- 5 files changed, 259 insertions(+), 45 deletions(-) (limited to 'fs/nfs/blocklayout') diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c index ddd0138f4..17a42e4eb 100644 --- a/fs/nfs/blocklayout/blocklayout.c +++ b/fs/nfs/blocklayout/blocklayout.c @@ -231,7 +231,7 @@ bl_read_pagelist(struct nfs_pgio_header *header) size_t bytes_left = header->args.count; unsigned int pg_offset = header->args.pgbase, pg_len; struct page **pages = header->args.pages; - int pg_index = header->args.pgbase >> PAGE_CACHE_SHIFT; + int pg_index = header->args.pgbase >> PAGE_SHIFT; const bool is_dio = (header->dreq != NULL); struct blk_plug plug; int i; @@ -263,13 +263,13 @@ bl_read_pagelist(struct nfs_pgio_header *header) } if (is_dio) { - if (pg_offset + bytes_left > PAGE_CACHE_SIZE) - pg_len = PAGE_CACHE_SIZE - pg_offset; + if (pg_offset + bytes_left > PAGE_SIZE) + pg_len = PAGE_SIZE - pg_offset; else pg_len = bytes_left; } else { BUG_ON(pg_offset != 0); - pg_len = PAGE_CACHE_SIZE; + pg_len = PAGE_SIZE; } if (is_hole(&be)) { @@ -339,9 +339,9 @@ static void bl_write_cleanup(struct work_struct *work) if (likely(!hdr->pnfs_error)) { struct pnfs_block_layout *bl = BLK_LSEG2EXT(hdr->lseg); - u64 start = hdr->args.offset & (loff_t)PAGE_CACHE_MASK; + u64 start = hdr->args.offset & (loff_t)PAGE_MASK; u64 end = (hdr->args.offset + hdr->args.count + - PAGE_CACHE_SIZE - 1) & (loff_t)PAGE_CACHE_MASK; + PAGE_SIZE - 1) & (loff_t)PAGE_MASK; ext_tree_mark_written(bl, start >> SECTOR_SHIFT, (end - start) >> SECTOR_SHIFT); @@ -373,7 +373,7 @@ bl_write_pagelist(struct nfs_pgio_header *header, int sync) loff_t offset = header->args.offset; size_t count = header->args.count; struct page **pages = header->args.pages; - int pg_index = header->args.pgbase >> PAGE_CACHE_SHIFT; + int pg_index = header->args.pgbase >> PAGE_SHIFT; unsigned int pg_len; struct blk_plug plug; int i; @@ -392,7 +392,7 @@ bl_write_pagelist(struct nfs_pgio_header *header, int sync) blk_start_plug(&plug); /* we always write out the whole page */ - offset = offset & (loff_t)PAGE_CACHE_MASK; + offset = offset & (loff_t)PAGE_MASK; isect = offset >> SECTOR_SHIFT; for (i = pg_index; i < header->page_array.npages; i++) { @@ -408,7 +408,7 @@ bl_write_pagelist(struct nfs_pgio_header *header, int sync) extent_length = be.be_length - (isect - be.be_f_offset); } - pg_len = PAGE_CACHE_SIZE; + pg_len = PAGE_SIZE; bio = do_add_page_to_bio(bio, header->page_array.npages - i, WRITE, isect, pages[i], &map, &be, bl_end_io_write, par, @@ -446,8 +446,8 @@ static void bl_free_layout_hdr(struct pnfs_layout_hdr *lo) kfree(bl); } -static struct pnfs_layout_hdr *bl_alloc_layout_hdr(struct inode *inode, - gfp_t gfp_flags) +static struct pnfs_layout_hdr *__bl_alloc_layout_hdr(struct inode *inode, + gfp_t gfp_flags, bool is_scsi_layout) { struct pnfs_block_layout *bl; @@ -460,9 +460,22 @@ static struct pnfs_layout_hdr *bl_alloc_layout_hdr(struct inode *inode, bl->bl_ext_ro = RB_ROOT; spin_lock_init(&bl->bl_ext_lock); + bl->bl_scsi_layout = is_scsi_layout; return &bl->bl_layout; } +static struct pnfs_layout_hdr *bl_alloc_layout_hdr(struct inode *inode, + gfp_t gfp_flags) +{ + return __bl_alloc_layout_hdr(inode, gfp_flags, false); +} + +static struct pnfs_layout_hdr *sl_alloc_layout_hdr(struct inode *inode, + gfp_t gfp_flags) +{ + return __bl_alloc_layout_hdr(inode, gfp_flags, true); +} + static void bl_free_lseg(struct pnfs_layout_segment *lseg) { dprintk("%s enter\n", __func__); @@ -743,7 +756,7 @@ bl_set_layoutdriver(struct nfs_server *server, const struct nfs_fh *fh) static bool is_aligned_req(struct nfs_pageio_descriptor *pgio, - struct nfs_page *req, unsigned int alignment) + struct nfs_page *req, unsigned int alignment, bool is_write) { /* * Always accept buffered writes, higher layers take care of the @@ -758,7 +771,8 @@ is_aligned_req(struct nfs_pageio_descriptor *pgio, if (IS_ALIGNED(req->wb_bytes, alignment)) return true; - if (req_offset(req) + req->wb_bytes == i_size_read(pgio->pg_inode)) { + if (is_write && + (req_offset(req) + req->wb_bytes == i_size_read(pgio->pg_inode))) { /* * If the write goes up to the inode size, just write * the full page. Data past the inode size is @@ -775,7 +789,7 @@ is_aligned_req(struct nfs_pageio_descriptor *pgio, static void bl_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req) { - if (!is_aligned_req(pgio, req, SECTOR_SIZE)) { + if (!is_aligned_req(pgio, req, SECTOR_SIZE, false)) { nfs_pageio_reset_read_mds(pgio); return; } @@ -791,7 +805,7 @@ static size_t bl_pg_test_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req) { - if (!is_aligned_req(pgio, req, SECTOR_SIZE)) + if (!is_aligned_req(pgio, req, SECTOR_SIZE, false)) return 0; return pnfs_generic_pg_test(pgio, prev, req); } @@ -806,7 +820,7 @@ static u64 pnfs_num_cont_bytes(struct inode *inode, pgoff_t idx) pgoff_t end; /* Optimize common case that writes from 0 to end of file */ - end = DIV_ROUND_UP(i_size_read(inode), PAGE_CACHE_SIZE); + end = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE); if (end != inode->i_mapping->nrpages) { rcu_read_lock(); end = page_cache_next_hole(mapping, idx + 1, ULONG_MAX); @@ -814,9 +828,9 @@ static u64 pnfs_num_cont_bytes(struct inode *inode, pgoff_t idx) } if (!end) - return i_size_read(inode) - (idx << PAGE_CACHE_SHIFT); + return i_size_read(inode) - (idx << PAGE_SHIFT); else - return (end - idx) << PAGE_CACHE_SHIFT; + return (end - idx) << PAGE_SHIFT; } static void @@ -824,7 +838,7 @@ bl_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req) { u64 wb_size; - if (!is_aligned_req(pgio, req, PAGE_SIZE)) { + if (!is_aligned_req(pgio, req, PAGE_SIZE, true)) { nfs_pageio_reset_write_mds(pgio); return; } @@ -846,7 +860,7 @@ static size_t bl_pg_test_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req) { - if (!is_aligned_req(pgio, req, PAGE_SIZE)) + if (!is_aligned_req(pgio, req, PAGE_SIZE, true)) return 0; return pnfs_generic_pg_test(pgio, prev, req); } @@ -888,22 +902,53 @@ static struct pnfs_layoutdriver_type blocklayout_type = { .sync = pnfs_generic_sync, }; +static struct pnfs_layoutdriver_type scsilayout_type = { + .id = LAYOUT_SCSI, + .name = "LAYOUT_SCSI", + .owner = THIS_MODULE, + .flags = PNFS_LAYOUTRET_ON_SETATTR | + PNFS_READ_WHOLE_PAGE, + .read_pagelist = bl_read_pagelist, + .write_pagelist = bl_write_pagelist, + .alloc_layout_hdr = sl_alloc_layout_hdr, + .free_layout_hdr = bl_free_layout_hdr, + .alloc_lseg = bl_alloc_lseg, + .free_lseg = bl_free_lseg, + .return_range = bl_return_range, + .prepare_layoutcommit = bl_prepare_layoutcommit, + .cleanup_layoutcommit = bl_cleanup_layoutcommit, + .set_layoutdriver = bl_set_layoutdriver, + .alloc_deviceid_node = bl_alloc_deviceid_node, + .free_deviceid_node = bl_free_deviceid_node, + .pg_read_ops = &bl_pg_read_ops, + .pg_write_ops = &bl_pg_write_ops, + .sync = pnfs_generic_sync, +}; + + static int __init nfs4blocklayout_init(void) { int ret; dprintk("%s: NFSv4 Block Layout Driver Registering...\n", __func__); - ret = pnfs_register_layoutdriver(&blocklayout_type); + ret = bl_init_pipefs(); if (ret) goto out; - ret = bl_init_pipefs(); + + ret = pnfs_register_layoutdriver(&blocklayout_type); if (ret) - goto out_unregister; + goto out_cleanup_pipe; + + ret = pnfs_register_layoutdriver(&scsilayout_type); + if (ret) + goto out_unregister_block; return 0; -out_unregister: +out_unregister_block: pnfs_unregister_layoutdriver(&blocklayout_type); +out_cleanup_pipe: + bl_cleanup_pipefs(); out: return ret; } @@ -913,8 +958,9 @@ static void __exit nfs4blocklayout_exit(void) dprintk("%s: NFSv4 Block Layout Driver Unregistering...\n", __func__); - bl_cleanup_pipefs(); + pnfs_unregister_layoutdriver(&scsilayout_type); pnfs_unregister_layoutdriver(&blocklayout_type); + bl_cleanup_pipefs(); } MODULE_ALIAS("nfs-layouttype4-3"); diff --git a/fs/nfs/blocklayout/blocklayout.h b/fs/nfs/blocklayout/blocklayout.h index c556640dc..18e6fd0b9 100644 --- a/fs/nfs/blocklayout/blocklayout.h +++ b/fs/nfs/blocklayout/blocklayout.h @@ -40,8 +40,8 @@ #include "../pnfs.h" #include "../netns.h" -#define PAGE_CACHE_SECTORS (PAGE_CACHE_SIZE >> SECTOR_SHIFT) -#define PAGE_CACHE_SECTOR_SHIFT (PAGE_CACHE_SHIFT - SECTOR_SHIFT) +#define PAGE_CACHE_SECTORS (PAGE_SIZE >> SECTOR_SHIFT) +#define PAGE_CACHE_SECTOR_SHIFT (PAGE_SHIFT - SECTOR_SHIFT) #define SECTOR_SIZE (1 << SECTOR_SHIFT) struct pnfs_block_dev; @@ -55,7 +55,6 @@ struct pnfs_block_dev; */ #define PNFS_BLOCK_UUID_LEN 128 - struct pnfs_block_volume { enum pnfs_block_volume_type type; union { @@ -82,6 +81,13 @@ struct pnfs_block_volume { u32 volumes_count; u32 volumes[PNFS_BLOCK_MAX_DEVICES]; } stripe; + struct { + enum scsi_code_set code_set; + enum scsi_designator_type designator_type; + int designator_len; + u8 designator[256]; + u64 pr_key; + } scsi; }; }; @@ -106,6 +112,9 @@ struct pnfs_block_dev { struct block_device *bdev; u64 disk_offset; + u64 pr_key; + bool pr_registered; + bool (*map)(struct pnfs_block_dev *dev, u64 offset, struct pnfs_block_dev_map *map); }; @@ -131,6 +140,7 @@ struct pnfs_block_layout { struct rb_root bl_ext_rw; struct rb_root bl_ext_ro; spinlock_t bl_ext_lock; /* Protects list manipulation */ + bool bl_scsi_layout; }; static inline struct pnfs_block_layout * @@ -182,6 +192,6 @@ void ext_tree_mark_committed(struct nfs4_layoutcommit_args *arg, int status); dev_t bl_resolve_deviceid(struct nfs_server *server, struct pnfs_block_volume *b, gfp_t gfp_mask); int __init bl_init_pipefs(void); -void __exit bl_cleanup_pipefs(void); +void bl_cleanup_pipefs(void); #endif /* FS_NFS_NFS4BLOCKLAYOUT_H */ diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index a861bbdfe..e5b896752 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -1,11 +1,12 @@ /* - * Copyright (c) 2014 Christoph Hellwig. + * Copyright (c) 2014-2016 Christoph Hellwig. */ #include #include #include #include #include +#include #include "blocklayout.h" @@ -21,6 +22,17 @@ bl_free_device(struct pnfs_block_dev *dev) bl_free_device(&dev->children[i]); kfree(dev->children); } else { + if (dev->pr_registered) { + const struct pr_ops *ops = + dev->bdev->bd_disk->fops->pr_ops; + int error; + + error = ops->pr_register(dev->bdev, dev->pr_key, 0, + false); + if (error) + pr_err("failed to unregister PR key.\n"); + } + if (dev->bdev) blkdev_put(dev->bdev, FMODE_READ | FMODE_WRITE); } @@ -113,6 +125,24 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) for (i = 0; i < b->stripe.volumes_count; i++) b->stripe.volumes[i] = be32_to_cpup(p++); break; + case PNFS_BLOCK_VOLUME_SCSI: + p = xdr_inline_decode(xdr, 4 + 4 + 4); + if (!p) + return -EIO; + b->scsi.code_set = be32_to_cpup(p++); + b->scsi.designator_type = be32_to_cpup(p++); + b->scsi.designator_len = be32_to_cpup(p++); + p = xdr_inline_decode(xdr, b->scsi.designator_len); + if (!p) + return -EIO; + if (b->scsi.designator_len > 256) + return -EIO; + memcpy(&b->scsi.designator, p, b->scsi.designator_len); + p = xdr_inline_decode(xdr, 8); + if (!p) + return -EIO; + p = xdr_decode_hyper(p, &b->scsi.pr_key); + break; default: dprintk("unknown volume type!\n"); return -EIO; @@ -216,6 +246,116 @@ bl_parse_simple(struct nfs_server *server, struct pnfs_block_dev *d, return 0; } +static bool +bl_validate_designator(struct pnfs_block_volume *v) +{ + switch (v->scsi.designator_type) { + case PS_DESIGNATOR_EUI64: + if (v->scsi.code_set != PS_CODE_SET_BINARY) + return false; + + if (v->scsi.designator_len != 8 && + v->scsi.designator_len != 10 && + v->scsi.designator_len != 16) + return false; + + return true; + case PS_DESIGNATOR_NAA: + if (v->scsi.code_set != PS_CODE_SET_BINARY) + return false; + + if (v->scsi.designator_len != 8 && + v->scsi.designator_len != 16) + return false; + + return true; + case PS_DESIGNATOR_T10: + case PS_DESIGNATOR_NAME: + pr_err("pNFS: unsupported designator " + "(code set %d, type %d, len %d.\n", + v->scsi.code_set, + v->scsi.designator_type, + v->scsi.designator_len); + return false; + default: + pr_err("pNFS: invalid designator " + "(code set %d, type %d, len %d.\n", + v->scsi.code_set, + v->scsi.designator_type, + v->scsi.designator_len); + return false; + } +} + +static int +bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, + struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) +{ + struct pnfs_block_volume *v = &volumes[idx]; + const struct pr_ops *ops; + const char *devname; + int error; + + if (!bl_validate_designator(v)) + return -EINVAL; + + switch (v->scsi.designator_len) { + case 8: + devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%8phN", + v->scsi.designator); + break; + case 12: + devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%12phN", + v->scsi.designator); + break; + case 16: + devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%16phN", + v->scsi.designator); + break; + default: + return -EINVAL; + } + + d->bdev = blkdev_get_by_path(devname, FMODE_READ, NULL); + if (IS_ERR(d->bdev)) { + pr_warn("pNFS: failed to open device %s (%ld)\n", + devname, PTR_ERR(d->bdev)); + kfree(devname); + return PTR_ERR(d->bdev); + } + + kfree(devname); + + d->len = i_size_read(d->bdev->bd_inode); + d->map = bl_map_simple; + d->pr_key = v->scsi.pr_key; + + pr_info("pNFS: using block device %s (reservation key 0x%llx)\n", + d->bdev->bd_disk->disk_name, d->pr_key); + + ops = d->bdev->bd_disk->fops->pr_ops; + if (!ops) { + pr_err("pNFS: block device %s does not support reservations.", + d->bdev->bd_disk->disk_name); + error = -EINVAL; + goto out_blkdev_put; + } + + error = ops->pr_register(d->bdev, 0, d->pr_key, true); + if (error) { + pr_err("pNFS: failed to register key for block device %s.", + d->bdev->bd_disk->disk_name); + goto out_blkdev_put; + } + + d->pr_registered = true; + return 0; + +out_blkdev_put: + blkdev_put(d->bdev, FMODE_READ); + return error; +} + static int bl_parse_slice(struct nfs_server *server, struct pnfs_block_dev *d, struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) @@ -303,6 +443,8 @@ bl_parse_deviceid(struct nfs_server *server, struct pnfs_block_dev *d, return bl_parse_concat(server, d, volumes, idx, gfp_mask); case PNFS_BLOCK_VOLUME_STRIPE: return bl_parse_stripe(server, d, volumes, idx, gfp_mask); + case PNFS_BLOCK_VOLUME_SCSI: + return bl_parse_scsi(server, d, volumes, idx, gfp_mask); default: dprintk("unsupported volume type: %d\n", volumes[idx].type); return -EIO; diff --git a/fs/nfs/blocklayout/extent_tree.c b/fs/nfs/blocklayout/extent_tree.c index 35ab51c04..720b3ff55 100644 --- a/fs/nfs/blocklayout/extent_tree.c +++ b/fs/nfs/blocklayout/extent_tree.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2014 Christoph Hellwig. + * Copyright (c) 2014-2016 Christoph Hellwig. */ #include @@ -462,10 +462,12 @@ out: return err; } -static size_t ext_tree_layoutupdate_size(size_t count) +static size_t ext_tree_layoutupdate_size(struct pnfs_block_layout *bl, size_t count) { - return sizeof(__be32) /* number of entries */ + - PNFS_BLOCK_EXTENT_SIZE * count; + if (bl->bl_scsi_layout) + return sizeof(__be32) + PNFS_SCSI_RANGE_SIZE * count; + else + return sizeof(__be32) + PNFS_BLOCK_EXTENT_SIZE * count; } static void ext_tree_free_commitdata(struct nfs4_layoutcommit_args *arg, @@ -483,6 +485,23 @@ static void ext_tree_free_commitdata(struct nfs4_layoutcommit_args *arg, } } +static __be32 *encode_block_extent(struct pnfs_block_extent *be, __be32 *p) +{ + p = xdr_encode_opaque_fixed(p, be->be_device->deviceid.data, + NFS4_DEVICEID4_SIZE); + p = xdr_encode_hyper(p, be->be_f_offset << SECTOR_SHIFT); + p = xdr_encode_hyper(p, be->be_length << SECTOR_SHIFT); + p = xdr_encode_hyper(p, 0LL); + *p++ = cpu_to_be32(PNFS_BLOCK_READWRITE_DATA); + return p; +} + +static __be32 *encode_scsi_range(struct pnfs_block_extent *be, __be32 *p) +{ + p = xdr_encode_hyper(p, be->be_f_offset << SECTOR_SHIFT); + return xdr_encode_hyper(p, be->be_length << SECTOR_SHIFT); +} + static int ext_tree_encode_commit(struct pnfs_block_layout *bl, __be32 *p, size_t buffer_size, size_t *count) { @@ -496,19 +515,16 @@ static int ext_tree_encode_commit(struct pnfs_block_layout *bl, __be32 *p, continue; (*count)++; - if (ext_tree_layoutupdate_size(*count) > buffer_size) { + if (ext_tree_layoutupdate_size(bl, *count) > buffer_size) { /* keep counting.. */ ret = -ENOSPC; continue; } - p = xdr_encode_opaque_fixed(p, be->be_device->deviceid.data, - NFS4_DEVICEID4_SIZE); - p = xdr_encode_hyper(p, be->be_f_offset << SECTOR_SHIFT); - p = xdr_encode_hyper(p, be->be_length << SECTOR_SHIFT); - p = xdr_encode_hyper(p, 0LL); - *p++ = cpu_to_be32(PNFS_BLOCK_READWRITE_DATA); - + if (bl->bl_scsi_layout) + p = encode_scsi_range(be, p); + else + p = encode_block_extent(be, p); be->be_tag = EXTENT_COMMITTING; } spin_unlock(&bl->bl_ext_lock); @@ -537,7 +553,7 @@ retry: if (unlikely(ret)) { ext_tree_free_commitdata(arg, buffer_size); - buffer_size = ext_tree_layoutupdate_size(count); + buffer_size = ext_tree_layoutupdate_size(bl, count); count = 0; arg->layoutupdate_pages = @@ -556,7 +572,7 @@ retry: } *start_p = cpu_to_be32(count); - arg->layoutupdate_len = ext_tree_layoutupdate_size(count); + arg->layoutupdate_len = ext_tree_layoutupdate_size(bl, count); if (unlikely(arg->layoutupdate_pages != &arg->layoutupdate_page)) { void *p = start_p, *end = p + arg->layoutupdate_len; diff --git a/fs/nfs/blocklayout/rpc_pipefs.c b/fs/nfs/blocklayout/rpc_pipefs.c index dbe5839cd..9fb067a6f 100644 --- a/fs/nfs/blocklayout/rpc_pipefs.c +++ b/fs/nfs/blocklayout/rpc_pipefs.c @@ -281,7 +281,7 @@ out: return ret; } -void __exit bl_cleanup_pipefs(void) +void bl_cleanup_pipefs(void) { rpc_pipefs_notifier_unregister(&nfs4blocklayout_block); unregister_pernet_subsys(&nfs4blocklayout_net_ops); -- cgit v1.2.3-54-g00ecf