From d635711daa98be86d4c7fd01499c34f566b54ccb Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Andr=C3=A9=20Fabian=20Silva=20Delgado?= 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 +- fs/nfs/callback.h | 3 +- fs/nfs/callback_proc.c | 71 ++++++++------- fs/nfs/callback_xdr.c | 12 +-- fs/nfs/client.c | 8 +- fs/nfs/dir.c | 16 ++-- fs/nfs/direct.c | 8 +- fs/nfs/file.c | 32 +++---- fs/nfs/flexfilelayout/flexfilelayoutdev.c | 2 + fs/nfs/inode.c | 2 +- fs/nfs/internal.h | 12 +-- fs/nfs/nfs4file.c | 33 +------ fs/nfs/nfs4proc.c | 76 +++++++++++++--- fs/nfs/nfs4session.c | 54 ++++++++--- fs/nfs/nfs4session.h | 8 ++ fs/nfs/nfs4xdr.c | 2 +- fs/nfs/objlayout/objio_osd.c | 2 +- fs/nfs/pagelist.c | 6 +- fs/nfs/pnfs.c | 6 +- fs/nfs/pnfs_nfs.c | 44 ++++++++- fs/nfs/read.c | 16 ++-- fs/nfs/write.c | 8 +- 26 files changed, 525 insertions(+), 200 deletions(-) (limited to 'fs/nfs') 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); diff --git a/fs/nfs/callback.h b/fs/nfs/callback.h index ff8195bd7..5fe1cecbf 100644 --- a/fs/nfs/callback.h +++ b/fs/nfs/callback.h @@ -37,10 +37,11 @@ enum nfs4_callback_opnum { OP_CB_ILLEGAL = 10044, }; +struct nfs4_slot; struct cb_process_state { __be32 drc_status; struct nfs_client *clp; - u32 slotid; + struct nfs4_slot *slot; u32 minorversion; struct net *net; }; diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index f0939d097..7c9fbf504 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -354,47 +354,38 @@ out: * a single outstanding callback request at a time. */ static __be32 -validate_seqid(struct nfs4_slot_table *tbl, struct cb_sequenceargs * args) +validate_seqid(const struct nfs4_slot_table *tbl, const struct nfs4_slot *slot, + const struct cb_sequenceargs * args) { - struct nfs4_slot *slot; - - dprintk("%s enter. slotid %u seqid %u\n", - __func__, args->csa_slotid, args->csa_sequenceid); + dprintk("%s enter. slotid %u seqid %u, slot table seqid: %u\n", + __func__, args->csa_slotid, args->csa_sequenceid, slot->seq_nr); - if (args->csa_slotid >= NFS41_BC_MAX_CALLBACKS) + if (args->csa_slotid > tbl->server_highest_slotid) return htonl(NFS4ERR_BADSLOT); - slot = tbl->slots + args->csa_slotid; - dprintk("%s slot table seqid: %u\n", __func__, slot->seq_nr); - - /* Normal */ - if (likely(args->csa_sequenceid == slot->seq_nr + 1)) - goto out_ok; - /* Replay */ if (args->csa_sequenceid == slot->seq_nr) { dprintk("%s seqid %u is a replay\n", __func__, args->csa_sequenceid); + if (nfs4_test_locked_slot(tbl, slot->slot_nr)) + return htonl(NFS4ERR_DELAY); /* Signal process_op to set this error on next op */ if (args->csa_cachethis == 0) return htonl(NFS4ERR_RETRY_UNCACHED_REP); - /* The ca_maxresponsesize_cached is 0 with no DRC */ - else if (args->csa_cachethis == 1) - return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); + /* Liar! We never allowed you to set csa_cachethis != 0 */ + return htonl(NFS4ERR_SEQ_FALSE_RETRY); } /* Wraparound */ - if (args->csa_sequenceid == 1 && (slot->seq_nr + 1) == 0) { - slot->seq_nr = 1; - goto out_ok; - } + if (unlikely(slot->seq_nr == 0xFFFFFFFFU)) { + if (args->csa_sequenceid == 1) + return htonl(NFS4_OK); + } else if (likely(args->csa_sequenceid == slot->seq_nr + 1)) + return htonl(NFS4_OK); /* Misordered request */ return htonl(NFS4ERR_SEQ_MISORDERED); -out_ok: - tbl->highest_used_slotid = args->csa_slotid; - return htonl(NFS4_OK); } /* @@ -473,6 +464,12 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, tbl = &clp->cl_session->bc_slot_table; slot = tbl->slots + args->csa_slotid; + /* Set up res before grabbing the spinlock */ + memcpy(&res->csr_sessionid, &args->csa_sessionid, + sizeof(res->csr_sessionid)); + res->csr_sequenceid = args->csa_sequenceid; + res->csr_slotid = args->csa_slotid; + spin_lock(&tbl->slot_tbl_lock); /* state manager is resetting the session */ if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state)) { @@ -485,18 +482,28 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, goto out_unlock; } - memcpy(&res->csr_sessionid, &args->csa_sessionid, - sizeof(res->csr_sessionid)); - res->csr_sequenceid = args->csa_sequenceid; - res->csr_slotid = args->csa_slotid; - res->csr_highestslotid = NFS41_BC_MAX_CALLBACKS - 1; - res->csr_target_highestslotid = NFS41_BC_MAX_CALLBACKS - 1; + status = htonl(NFS4ERR_BADSLOT); + slot = nfs4_lookup_slot(tbl, args->csa_slotid); + if (IS_ERR(slot)) + goto out_unlock; + + res->csr_highestslotid = tbl->server_highest_slotid; + res->csr_target_highestslotid = tbl->target_highest_slotid; - status = validate_seqid(tbl, args); + status = validate_seqid(tbl, slot, args); if (status) goto out_unlock; + if (!nfs4_try_to_lock_slot(tbl, slot)) { + status = htonl(NFS4ERR_DELAY); + goto out_unlock; + } + cps->slot = slot; - cps->slotid = args->csa_slotid; + /* The ca_maxresponsesize_cached is 0 with no DRC */ + if (args->csa_cachethis != 0) { + status = htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE); + goto out_unlock; + } /* * Check for pending referring calls. If a match is found, a @@ -513,7 +520,7 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args, * If CB_SEQUENCE returns an error, then the state of the slot * (sequence ID, cached reply) MUST NOT change. */ - slot->seq_nr++; + slot->seq_nr = args->csa_sequenceid; out_unlock: spin_unlock(&tbl->slot_tbl_lock); diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c index 646cdac73..976c90608 100644 --- a/fs/nfs/callback_xdr.c +++ b/fs/nfs/callback_xdr.c @@ -752,7 +752,8 @@ preprocess_nfs41_op(int nop, unsigned int op_nr, struct callback_op **op) return htonl(NFS_OK); } -static void nfs4_callback_free_slot(struct nfs4_session *session) +static void nfs4_callback_free_slot(struct nfs4_session *session, + struct nfs4_slot *slot) { struct nfs4_slot_table *tbl = &session->bc_slot_table; @@ -761,15 +762,17 @@ static void nfs4_callback_free_slot(struct nfs4_session *session) * Let the state manager know callback processing done. * A single slot, so highest used slotid is either 0 or -1 */ - tbl->highest_used_slotid = NFS4_NO_SLOT; + nfs4_free_slot(tbl, slot); nfs4_slot_tbl_drain_complete(tbl); spin_unlock(&tbl->slot_tbl_lock); } static void nfs4_cb_free_slot(struct cb_process_state *cps) { - if (cps->slotid != NFS4_NO_SLOT) - nfs4_callback_free_slot(cps->clp->cl_session); + if (cps->slot) { + nfs4_callback_free_slot(cps->clp->cl_session, cps->slot); + cps->slot = NULL; + } } #else /* CONFIG_NFS_V4_1 */ @@ -893,7 +896,6 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r struct cb_process_state cps = { .drc_status = 0, .clp = NULL, - .slotid = NFS4_NO_SLOT, .net = SVC_NET(rqstp), }; unsigned int nops = 0; diff --git a/fs/nfs/client.c b/fs/nfs/client.c index d6d5d2a48..0c96528db 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -736,7 +736,7 @@ static void nfs_server_set_fsinfo(struct nfs_server *server, server->rsize = max_rpc_payload; if (server->rsize > NFS_MAX_FILE_IO_SIZE) server->rsize = NFS_MAX_FILE_IO_SIZE; - server->rpages = (server->rsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + server->rpages = (server->rsize + PAGE_SIZE - 1) >> PAGE_SHIFT; server->backing_dev_info.name = "nfs"; server->backing_dev_info.ra_pages = server->rpages * NFS_MAX_READAHEAD; @@ -745,13 +745,13 @@ static void nfs_server_set_fsinfo(struct nfs_server *server, server->wsize = max_rpc_payload; if (server->wsize > NFS_MAX_FILE_IO_SIZE) server->wsize = NFS_MAX_FILE_IO_SIZE; - server->wpages = (server->wsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + server->wpages = (server->wsize + PAGE_SIZE - 1) >> PAGE_SHIFT; server->wtmult = nfs_block_bits(fsinfo->wtmult, NULL); server->dtsize = nfs_block_size(fsinfo->dtpref, NULL); - if (server->dtsize > PAGE_CACHE_SIZE * NFS_MAX_READDIR_PAGES) - server->dtsize = PAGE_CACHE_SIZE * NFS_MAX_READDIR_PAGES; + if (server->dtsize > PAGE_SIZE * NFS_MAX_READDIR_PAGES) + server->dtsize = PAGE_SIZE * NFS_MAX_READDIR_PAGES; if (server->dtsize > server->rsize) server->dtsize = server->rsize; diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 7ded17764..33eb81738 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -707,7 +707,7 @@ void cache_page_release(nfs_readdir_descriptor_t *desc) { if (!desc->page->mapping) nfs_readdir_clear_array(desc->page); - page_cache_release(desc->page); + put_page(desc->page); desc->page = NULL; } @@ -1360,19 +1360,15 @@ struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned in dfprintk(VFS, "NFS: lookup(%pd2)\n", dentry); nfs_inc_stats(dir, NFSIOS_VFSLOOKUP); - res = ERR_PTR(-ENAMETOOLONG); - if (dentry->d_name.len > NFS_SERVER(dir)->namelen) - goto out; + if (unlikely(dentry->d_name.len > NFS_SERVER(dir)->namelen)) + return ERR_PTR(-ENAMETOOLONG); /* * If we're doing an exclusive create, optimize away the lookup * but don't hash the dentry. */ - if (nfs_is_exclusive_create(dir, flags)) { - d_instantiate(dentry, NULL); - res = NULL; - goto out; - } + if (nfs_is_exclusive_create(dir, flags)) + return NULL; res = ERR_PTR(-ENOMEM); fhandle = nfs_alloc_fhandle(); @@ -1927,7 +1923,7 @@ int nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname) * add_to_page_cache_lru() grabs an extra page refcount. * Drop it here to avoid leaking this page later. */ - page_cache_release(page); + put_page(page); } else __free_page(page); diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 7a0cfd326..c93826e4a 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -269,7 +269,7 @@ static void nfs_direct_release_pages(struct page **pages, unsigned int npages) { unsigned int i; for (i = 0; i < npages; i++) - page_cache_release(pages[i]); + put_page(pages[i]); } void nfs_init_cinfo_from_dreq(struct nfs_commit_info *cinfo, @@ -1003,7 +1003,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) iov_iter_count(iter)); pos = iocb->ki_pos; - end = (pos + iov_iter_count(iter) - 1) >> PAGE_CACHE_SHIFT; + end = (pos + iov_iter_count(iter) - 1) >> PAGE_SHIFT; inode_lock(inode); @@ -1013,7 +1013,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) if (mapping->nrpages) { result = invalidate_inode_pages2_range(mapping, - pos >> PAGE_CACHE_SHIFT, end); + pos >> PAGE_SHIFT, end); if (result) goto out_unlock; } @@ -1042,7 +1042,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) if (mapping->nrpages) { invalidate_inode_pages2_range(mapping, - pos >> PAGE_CACHE_SHIFT, end); + pos >> PAGE_SHIFT, end); } inode_unlock(inode); diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 748bb813b..be01095b9 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -233,7 +233,7 @@ EXPORT_SYMBOL_GPL(nfs_file_mmap); * nfs_file_write() that a write error occurred, and hence cause it to * fall back to doing a synchronous write. */ -int +static int nfs_file_fsync_commit(struct file *file, loff_t start, loff_t end, int datasync) { struct nfs_open_context *ctx = nfs_file_open_context(file); @@ -263,9 +263,8 @@ nfs_file_fsync_commit(struct file *file, loff_t start, loff_t end, int datasync) out: return ret; } -EXPORT_SYMBOL_GPL(nfs_file_fsync_commit); -static int +int nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) { int ret; @@ -273,13 +272,15 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) trace_nfs_fsync_enter(inode); - nfs_inode_dio_wait(inode); + inode_dio_wait(inode); do { ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret != 0) break; inode_lock(inode); ret = nfs_file_fsync_commit(file, start, end, datasync); + if (!ret) + ret = pnfs_sync_inode(inode, !!datasync); inode_unlock(inode); /* * If nfs_file_fsync_commit detected a server reboot, then @@ -293,6 +294,7 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) trace_nfs_fsync_exit(inode, ret); return ret; } +EXPORT_SYMBOL_GPL(nfs_file_fsync); /* * Decide whether a read/modify/write cycle may be more efficient @@ -318,7 +320,7 @@ static int nfs_want_read_modify_write(struct file *file, struct page *page, loff_t pos, unsigned len) { unsigned int pglen = nfs_page_length(page); - unsigned int offset = pos & (PAGE_CACHE_SIZE - 1); + unsigned int offset = pos & (PAGE_SIZE - 1); unsigned int end = offset + len; if (pnfs_ld_read_whole_page(file->f_mapping->host)) { @@ -349,7 +351,7 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping, struct page **pagep, void **fsdata) { int ret; - pgoff_t index = pos >> PAGE_CACHE_SHIFT; + pgoff_t index = pos >> PAGE_SHIFT; struct page *page; int once_thru = 0; @@ -368,7 +370,7 @@ start: /* * Wait for O_DIRECT to complete */ - nfs_inode_dio_wait(mapping->host); + inode_dio_wait(mapping->host); page = grab_cache_page_write_begin(mapping, index, flags); if (!page) @@ -378,12 +380,12 @@ start: ret = nfs_flush_incompatible(file, page); if (ret) { unlock_page(page); - page_cache_release(page); + put_page(page); } else if (!once_thru && nfs_want_read_modify_write(file, page, pos, len)) { once_thru = 1; ret = nfs_readpage(file, page); - page_cache_release(page); + put_page(page); if (!ret) goto start; } @@ -394,7 +396,7 @@ static int nfs_write_end(struct file *file, struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata) { - unsigned offset = pos & (PAGE_CACHE_SIZE - 1); + unsigned offset = pos & (PAGE_SIZE - 1); struct nfs_open_context *ctx = nfs_file_open_context(file); int status; @@ -411,20 +413,20 @@ static int nfs_write_end(struct file *file, struct address_space *mapping, if (pglen == 0) { zero_user_segments(page, 0, offset, - end, PAGE_CACHE_SIZE); + end, PAGE_SIZE); SetPageUptodate(page); } else if (end >= pglen) { - zero_user_segment(page, end, PAGE_CACHE_SIZE); + zero_user_segment(page, end, PAGE_SIZE); if (offset == 0) SetPageUptodate(page); } else - zero_user_segment(page, pglen, PAGE_CACHE_SIZE); + zero_user_segment(page, pglen, PAGE_SIZE); } status = nfs_updatepage(file, page, offset, copied); unlock_page(page); - page_cache_release(page); + put_page(page); if (status < 0) return status; @@ -452,7 +454,7 @@ static void nfs_invalidate_page(struct page *page, unsigned int offset, dfprintk(PAGECACHE, "NFS: invalidate_page(%p, %u, %u)\n", page, offset, length); - if (offset != 0 || length < PAGE_CACHE_SIZE) + if (offset != 0 || length < PAGE_SIZE) return; /* Cancel any unstarted writes on this page */ nfs_wb_page_cancel(page_file_mapping(page)->host, page); diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c index eb370460c..add0e5a70 100644 --- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c +++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c @@ -418,6 +418,8 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx, pnfs_error_mark_layout_for_return(ino, lseg); } else pnfs_error_mark_layout_for_return(ino, lseg); + ds = NULL; + goto out; } out_update_creds: if (ff_layout_update_mirror_cred(mirror, ds)) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 847b678af..738c84a42 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -141,7 +141,7 @@ void nfs_evict_inode(struct inode *inode) int nfs_sync_inode(struct inode *inode) { - nfs_inode_dio_wait(inode); + inode_dio_wait(inode); return nfs_wb_all(inode); } EXPORT_SYMBOL_GPL(nfs_sync_inode); diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 9a547aa3e..f1d1d2c47 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -358,7 +358,7 @@ int nfs_mknod(struct inode *, struct dentry *, umode_t, dev_t); int nfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); /* file.c */ -int nfs_file_fsync_commit(struct file *, loff_t, loff_t, int); +int nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync); loff_t nfs_file_llseek(struct file *, loff_t, int); ssize_t nfs_file_read(struct kiocb *, struct iov_iter *); ssize_t nfs_file_splice_read(struct file *, loff_t *, struct pipe_inode_info *, @@ -515,10 +515,6 @@ extern int nfs_sillyrename(struct inode *dir, struct dentry *dentry); /* direct.c */ void nfs_init_cinfo_from_dreq(struct nfs_commit_info *cinfo, struct nfs_direct_req *dreq); -static inline void nfs_inode_dio_wait(struct inode *inode) -{ - inode_dio_wait(inode); -} extern ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq); /* nfs4proc.c */ @@ -642,11 +638,11 @@ unsigned int nfs_page_length(struct page *page) if (i_size > 0) { pgoff_t page_index = page_file_index(page); - pgoff_t end_index = (i_size - 1) >> PAGE_CACHE_SHIFT; + pgoff_t end_index = (i_size - 1) >> PAGE_SHIFT; if (page_index < end_index) - return PAGE_CACHE_SIZE; + return PAGE_SIZE; if (page_index == end_index) - return ((i_size - 1) & ~PAGE_CACHE_MASK) + 1; + return ((i_size - 1) & ~PAGE_MASK) + 1; } return 0; } diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c index 2a9ff14cf..d03905164 100644 --- a/fs/nfs/nfs4file.c +++ b/fs/nfs/nfs4file.c @@ -128,37 +128,6 @@ nfs4_file_flush(struct file *file, fl_owner_t id) return vfs_fsync(file, 0); } -static int -nfs4_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) -{ - int ret; - struct inode *inode = file_inode(file); - - trace_nfs_fsync_enter(inode); - - nfs_inode_dio_wait(inode); - do { - ret = filemap_write_and_wait_range(inode->i_mapping, start, end); - if (ret != 0) - break; - inode_lock(inode); - ret = nfs_file_fsync_commit(file, start, end, datasync); - if (!ret) - ret = pnfs_sync_inode(inode, !!datasync); - inode_unlock(inode); - /* - * If nfs_file_fsync_commit detected a server reboot, then - * resend all dirty pages that might have been covered by - * the NFS_CONTEXT_RESEND_WRITES flag - */ - start = 0; - end = LLONG_MAX; - } while (ret == -EAGAIN); - - trace_nfs_fsync_exit(inode, ret); - return ret; -} - #ifdef CONFIG_NFS_V4_2 static loff_t nfs4_file_llseek(struct file *filep, loff_t offset, int whence) { @@ -266,7 +235,7 @@ const struct file_operations nfs4_file_operations = { .open = nfs4_file_open, .flush = nfs4_file_flush, .release = nfs_file_release, - .fsync = nfs4_file_fsync, + .fsync = nfs_file_fsync, .lock = nfs_lock, .flock = nfs_flock, .splice_read = nfs_file_splice_read, diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 14881594d..327b8c34d 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2461,14 +2461,15 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata, dentry = opendata->dentry; if (d_really_is_negative(dentry)) { - /* FIXME: Is this d_drop() ever needed? */ + struct dentry *alias; d_drop(dentry); - dentry = d_add_unique(dentry, igrab(state->inode)); - if (dentry == NULL) { - dentry = opendata->dentry; - } else { + alias = d_exact_alias(dentry, state->inode); + if (!alias) + alias = d_splice_alias(igrab(state->inode), dentry); + /* d_splice_alias() can't fail here - it's a non-directory */ + if (alias) { dput(ctx->dentry); - ctx->dentry = dentry; + ctx->dentry = dentry = alias; } nfs_set_verifier(dentry, nfs_save_change_attribute(d_inode(opendata->dir))); @@ -6782,13 +6783,26 @@ nfs41_same_server_scope(struct nfs41_server_scope *a, return false; } +static void +nfs4_bind_one_conn_to_session_done(struct rpc_task *task, void *calldata) +{ +} + +static const struct rpc_call_ops nfs4_bind_one_conn_to_session_ops = { + .rpc_call_done = &nfs4_bind_one_conn_to_session_done, +}; + /* - * nfs4_proc_bind_conn_to_session() + * nfs4_proc_bind_one_conn_to_session() * * The 4.1 client currently uses the same TCP connection for the * fore and backchannel. */ -int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred) +static +int nfs4_proc_bind_one_conn_to_session(struct rpc_clnt *clnt, + struct rpc_xprt *xprt, + struct nfs_client *clp, + struct rpc_cred *cred) { int status; struct nfs41_bind_conn_to_session_args args = { @@ -6803,6 +6817,14 @@ int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred .rpc_resp = &res, .rpc_cred = cred, }; + struct rpc_task_setup task_setup_data = { + .rpc_client = clnt, + .rpc_xprt = xprt, + .callback_ops = &nfs4_bind_one_conn_to_session_ops, + .rpc_message = &msg, + .flags = RPC_TASK_TIMEOUT, + }; + struct rpc_task *task; dprintk("--> %s\n", __func__); @@ -6810,7 +6832,16 @@ int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred if (!(clp->cl_session->flags & SESSION4_BACK_CHAN)) args.dir = NFS4_CDFC4_FORE; - status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT); + /* Do not set the backchannel flag unless this is clnt->cl_xprt */ + if (xprt != rcu_access_pointer(clnt->cl_xprt)) + args.dir = NFS4_CDFC4_FORE; + + task = rpc_run_task(&task_setup_data); + if (!IS_ERR(task)) { + status = task->tk_status; + rpc_put_task(task); + } else + status = PTR_ERR(task); trace_nfs4_bind_conn_to_session(clp, status); if (status == 0) { if (memcmp(res.sessionid.data, @@ -6837,6 +6868,31 @@ out: return status; } +struct rpc_bind_conn_calldata { + struct nfs_client *clp; + struct rpc_cred *cred; +}; + +static int +nfs4_proc_bind_conn_to_session_callback(struct rpc_clnt *clnt, + struct rpc_xprt *xprt, + void *calldata) +{ + struct rpc_bind_conn_calldata *p = calldata; + + return nfs4_proc_bind_one_conn_to_session(clnt, xprt, p->clp, p->cred); +} + +int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred) +{ + struct rpc_bind_conn_calldata data = { + .clp = clp, + .cred = cred, + }; + return rpc_clnt_iterate_for_each_xprt(clp->cl_rpcclient, + nfs4_proc_bind_conn_to_session_callback, &data); +} + /* * Minimum set of SP4_MACH_CRED operations from RFC 5661 in the enforce map * and operations we'd like to see to enable certain features in the allow map @@ -7319,7 +7375,7 @@ static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args) args->bc_attrs.max_resp_sz = PAGE_SIZE; args->bc_attrs.max_resp_sz_cached = 0; args->bc_attrs.max_ops = NFS4_MAX_BACK_CHANNEL_OPS; - args->bc_attrs.max_reqs = 1; + args->bc_attrs.max_reqs = NFS41_BC_MAX_CALLBACKS; dprintk("%s: Back Channel : max_rqst_sz=%u max_resp_sz=%u " "max_resp_sz_cached=%u max_ops=%u max_reqs=%u\n", diff --git a/fs/nfs/nfs4session.c b/fs/nfs/nfs4session.c index e23366eff..332d06e64 100644 --- a/fs/nfs/nfs4session.c +++ b/fs/nfs/nfs4session.c @@ -135,6 +135,43 @@ static struct nfs4_slot *nfs4_find_or_create_slot(struct nfs4_slot_table *tbl, return ERR_PTR(-ENOMEM); } +static void nfs4_lock_slot(struct nfs4_slot_table *tbl, + struct nfs4_slot *slot) +{ + u32 slotid = slot->slot_nr; + + __set_bit(slotid, tbl->used_slots); + if (slotid > tbl->highest_used_slotid || + tbl->highest_used_slotid == NFS4_NO_SLOT) + tbl->highest_used_slotid = slotid; + slot->generation = tbl->generation; +} + +/* + * nfs4_try_to_lock_slot - Given a slot try to allocate it + * + * Note: must be called with the slot_tbl_lock held. + */ +bool nfs4_try_to_lock_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot) +{ + if (nfs4_test_locked_slot(tbl, slot->slot_nr)) + return false; + nfs4_lock_slot(tbl, slot); + return true; +} + +/* + * nfs4_lookup_slot - Find a slot but don't allocate it + * + * Note: must be called with the slot_tbl_lock held. + */ +struct nfs4_slot *nfs4_lookup_slot(struct nfs4_slot_table *tbl, u32 slotid) +{ + if (slotid <= tbl->max_slotid) + return nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT); + return ERR_PTR(-E2BIG); +} + /* * nfs4_alloc_slot - efficiently look for a free slot * @@ -153,18 +190,11 @@ struct nfs4_slot *nfs4_alloc_slot(struct nfs4_slot_table *tbl) __func__, tbl->used_slots[0], tbl->highest_used_slotid, tbl->max_slotid + 1); slotid = find_first_zero_bit(tbl->used_slots, tbl->max_slotid + 1); - if (slotid > tbl->max_slotid) - goto out; - ret = nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT); - if (IS_ERR(ret)) - goto out; - __set_bit(slotid, tbl->used_slots); - if (slotid > tbl->highest_used_slotid || - tbl->highest_used_slotid == NFS4_NO_SLOT) - tbl->highest_used_slotid = slotid; - ret->generation = tbl->generation; - -out: + if (slotid <= tbl->max_slotid) { + ret = nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT); + if (!IS_ERR(ret)) + nfs4_lock_slot(tbl, ret); + } dprintk("<-- %s used_slots=%04lx highest_used=%u slotid=%u\n", __func__, tbl->used_slots[0], tbl->highest_used_slotid, !IS_ERR(ret) ? ret->slot_nr : NFS4_NO_SLOT); diff --git a/fs/nfs/nfs4session.h b/fs/nfs/nfs4session.h index e3ea2c532..5b51298d1 100644 --- a/fs/nfs/nfs4session.h +++ b/fs/nfs/nfs4session.h @@ -77,6 +77,8 @@ extern int nfs4_setup_slot_table(struct nfs4_slot_table *tbl, unsigned int max_reqs, const char *queue); extern void nfs4_shutdown_slot_table(struct nfs4_slot_table *tbl); extern struct nfs4_slot *nfs4_alloc_slot(struct nfs4_slot_table *tbl); +extern struct nfs4_slot *nfs4_lookup_slot(struct nfs4_slot_table *tbl, u32 slotid); +extern bool nfs4_try_to_lock_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot); extern void nfs4_free_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot); extern void nfs4_slot_tbl_drain_complete(struct nfs4_slot_table *tbl); bool nfs41_wake_and_assign_slot(struct nfs4_slot_table *tbl, @@ -88,6 +90,12 @@ static inline bool nfs4_slot_tbl_draining(struct nfs4_slot_table *tbl) return !!test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state); } +static inline bool nfs4_test_locked_slot(const struct nfs4_slot_table *tbl, + u32 slotid) +{ + return !!test_bit(slotid, tbl->used_slots); +} + #if defined(CONFIG_NFS_V4_1) extern void nfs41_set_target_slotid(struct nfs4_slot_table *tbl, u32 target_highest_slotid); diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 4e4441216..88474a4fc 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -5001,7 +5001,7 @@ static int decode_space_limit(struct xdr_stream *xdr, blocksize = be32_to_cpup(p); maxsize = (uint64_t)nblocks * (uint64_t)blocksize; } - maxsize >>= PAGE_CACHE_SHIFT; + maxsize >>= PAGE_SHIFT; *pagemod_limit = min_t(u64, maxsize, ULONG_MAX); return 0; out_overflow: diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c index 9aebffb40..049c1b1f2 100644 --- a/fs/nfs/objlayout/objio_osd.c +++ b/fs/nfs/objlayout/objio_osd.c @@ -486,7 +486,7 @@ static void __r4w_put_page(void *priv, struct page *page) dprintk("%s: index=0x%lx\n", __func__, (page == ZERO_PAGE(0)) ? -1UL : page->index); if (ZERO_PAGE(0) != page) - page_cache_release(page); + put_page(page); return; } diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c index 8ce4f61cb..1f6db4231 100644 --- a/fs/nfs/pagelist.c +++ b/fs/nfs/pagelist.c @@ -342,7 +342,7 @@ nfs_create_request(struct nfs_open_context *ctx, struct page *page, * update_nfs_request below if the region is not locked. */ req->wb_page = page; req->wb_index = page_file_index(page); - page_cache_get(page); + get_page(page); req->wb_offset = offset; req->wb_pgbase = offset; req->wb_bytes = count; @@ -392,7 +392,7 @@ static void nfs_clear_request(struct nfs_page *req) struct nfs_lock_context *l_ctx = req->wb_lock_context; if (page != NULL) { - page_cache_release(page); + put_page(page); req->wb_page = NULL; } if (l_ctx != NULL) { @@ -904,7 +904,7 @@ static bool nfs_can_coalesce_requests(struct nfs_page *prev, return false; } else { if (req->wb_pgbase != 0 || - prev->wb_pgbase + prev->wb_bytes != PAGE_CACHE_SIZE) + prev->wb_pgbase + prev->wb_bytes != PAGE_SIZE) return false; } } diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 2fa483e6d..89a5ef4df 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -841,7 +841,7 @@ send_layoutget(struct pnfs_layout_hdr *lo, i_size = i_size_read(ino); - lgp->args.minlength = PAGE_CACHE_SIZE; + lgp->args.minlength = PAGE_SIZE; if (lgp->args.minlength > range->length) lgp->args.minlength = range->length; if (range->iomode == IOMODE_READ) { @@ -1618,13 +1618,13 @@ lookup_again: spin_unlock(&clp->cl_lock); } - pg_offset = arg.offset & ~PAGE_CACHE_MASK; + pg_offset = arg.offset & ~PAGE_MASK; if (pg_offset) { arg.offset -= pg_offset; arg.length += pg_offset; } if (arg.length != NFS4_MAX_UINT64) - arg.length = PAGE_CACHE_ALIGN(arg.length); + arg.length = PAGE_ALIGN(arg.length); lseg = send_layoutget(lo, ctx, &arg, gfp_flags); atomic_dec(&lo->plh_outstanding); diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index 81ac6480f..776dccbc3 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -246,6 +246,23 @@ void pnfs_fetch_commit_bucket_list(struct list_head *pages, } +/* Helper function for pnfs_generic_commit_pagelist to catch an empty + * page list. This can happen when two commits race. */ +static bool +pnfs_generic_commit_cancel_empty_pagelist(struct list_head *pages, + struct nfs_commit_data *data, + struct nfs_commit_info *cinfo) +{ + if (list_empty(pages)) { + if (atomic_dec_and_test(&cinfo->mds->rpcs_out)) + wake_up_atomic_t(&cinfo->mds->rpcs_out); + nfs_commitdata_release(data); + return true; + } + + return false; +} + /* This follows nfs_commit_list pretty closely */ int pnfs_generic_commit_pagelist(struct inode *inode, struct list_head *mds_pages, @@ -280,6 +297,11 @@ pnfs_generic_commit_pagelist(struct inode *inode, struct list_head *mds_pages, list_for_each_entry_safe(data, tmp, &list, pages) { list_del_init(&data->pages); if (data->ds_commit_index < 0) { + /* another commit raced with us */ + if (pnfs_generic_commit_cancel_empty_pagelist(mds_pages, + data, cinfo)) + continue; + nfs_init_commit(data, mds_pages, NULL, cinfo); nfs_initiate_commit(NFS_CLIENT(inode), data, NFS_PROTO(data->inode), @@ -288,6 +310,12 @@ pnfs_generic_commit_pagelist(struct inode *inode, struct list_head *mds_pages, LIST_HEAD(pages); pnfs_fetch_commit_bucket_list(&pages, data, cinfo); + + /* another commit raced with us */ + if (pnfs_generic_commit_cancel_empty_pagelist(&pages, + data, cinfo)) + continue; + nfs_init_commit(data, &pages, data->lseg, cinfo); initiate_commit(data, how); } @@ -606,12 +634,22 @@ static int _nfs4_pnfs_v3_ds_connect(struct nfs_server *mds_srv, dprintk("%s: DS %s: trying address %s\n", __func__, ds->ds_remotestr, da->da_remotestr); - clp = get_v3_ds_connect(mds_srv->nfs_client, + if (!IS_ERR(clp)) { + struct xprt_create xprt_args = { + .ident = XPRT_TRANSPORT_TCP, + .net = clp->cl_net, + .dstaddr = (struct sockaddr *)&da->da_addr, + .addrlen = da->da_addrlen, + .servername = clp->cl_hostname, + }; + /* Add this address as an alias */ + rpc_clnt_add_xprt(clp->cl_rpcclient, &xprt_args, + rpc_clnt_test_and_add_xprt, NULL); + } else + clp = get_v3_ds_connect(mds_srv->nfs_client, (struct sockaddr *)&da->da_addr, da->da_addrlen, IPPROTO_TCP, timeo, retrans, au_flavor); - if (!IS_ERR(clp)) - break; } if (IS_ERR(clp)) { diff --git a/fs/nfs/read.c b/fs/nfs/read.c index eb31e23e7..6776d7a78 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -46,7 +46,7 @@ static void nfs_readhdr_free(struct nfs_pgio_header *rhdr) static int nfs_return_empty_page(struct page *page) { - zero_user(page, 0, PAGE_CACHE_SIZE); + zero_user(page, 0, PAGE_SIZE); SetPageUptodate(page); unlock_page(page); return 0; @@ -118,8 +118,8 @@ int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode, unlock_page(page); return PTR_ERR(new); } - if (len < PAGE_CACHE_SIZE) - zero_user_segment(page, len, PAGE_CACHE_SIZE); + if (len < PAGE_SIZE) + zero_user_segment(page, len, PAGE_SIZE); nfs_pageio_init_read(&pgio, inode, false, &nfs_async_read_completion_ops); @@ -295,7 +295,7 @@ int nfs_readpage(struct file *file, struct page *page) int error; dprintk("NFS: nfs_readpage (%p %ld@%lu)\n", - page, PAGE_CACHE_SIZE, page_file_index(page)); + page, PAGE_SIZE, page_file_index(page)); nfs_inc_stats(inode, NFSIOS_VFSREADPAGE); nfs_add_stats(inode, NFSIOS_READPAGES, 1); @@ -361,8 +361,8 @@ readpage_async_filler(void *data, struct page *page) if (IS_ERR(new)) goto out_error; - if (len < PAGE_CACHE_SIZE) - zero_user_segment(page, len, PAGE_CACHE_SIZE); + if (len < PAGE_SIZE) + zero_user_segment(page, len, PAGE_SIZE); if (!nfs_pageio_add_request(desc->pgio, new)) { nfs_list_remove_request(new); nfs_readpage_release(new); @@ -424,8 +424,8 @@ int nfs_readpages(struct file *filp, struct address_space *mapping, pgm = &pgio.pg_mirrors[0]; NFS_I(inode)->read_io += pgm->pg_bytes_written; - npages = (pgm->pg_bytes_written + PAGE_CACHE_SIZE - 1) >> - PAGE_CACHE_SHIFT; + npages = (pgm->pg_bytes_written + PAGE_SIZE - 1) >> + PAGE_SHIFT; nfs_add_stats(inode, NFSIOS_READPAGES, npages); read_complete: put_nfs_open_context(desc.ctx); diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 5754835a2..f5e613395 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -150,7 +150,7 @@ static void nfs_grow_file(struct page *page, unsigned int offset, unsigned int c spin_lock(&inode->i_lock); i_size = i_size_read(inode); - end_index = (i_size - 1) >> PAGE_CACHE_SHIFT; + end_index = (i_size - 1) >> PAGE_SHIFT; if (i_size > 0 && page_file_index(page) < end_index) goto out; end = page_file_offset(page) + ((loff_t)offset+count); @@ -1709,6 +1709,10 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how, { struct nfs_commit_data *data; + /* another commit raced with us */ + if (list_empty(head)) + return 0; + data = nfs_commitdata_alloc(); if (!data) @@ -1942,7 +1946,7 @@ int nfs_wb_page_cancel(struct inode *inode, struct page *page) int nfs_wb_single_page(struct inode *inode, struct page *page, bool launder) { loff_t range_start = page_file_offset(page); - loff_t range_end = range_start + (loff_t)(PAGE_CACHE_SIZE - 1); + loff_t range_end = range_start + (loff_t)(PAGE_SIZE - 1); struct writeback_control wbc = { .sync_mode = WB_SYNC_ALL, .nr_to_write = 0, -- cgit v1.2.3