From eec70897d81bb2913e22c6792ea2739faf59c3e1 Mon Sep 17 00:00:00 2001 From: Nicholas Bellinger Date: Thu, 20 Feb 2014 00:52:01 +0000 Subject: bio-integrity: Drop bio_integrity_verify BUG_ON in post bip->bip_iter world Given that bip->bip_iter.bi_size is decremented after bio_advance() -> bio_integrity_advance() is called, the BUG_ON() in bio_integrity_verify() ends up tripping in v3.14-rc1 code with the advent of immutable biovecs in: commit d57a5f7c6605f15f3b5134837e68b448a7cea88e Author: Kent Overstreet Date: Sat Nov 23 17:20:16 2013 -0800 bio-integrity: Convert to bvec_iter Given that there is no easy way to ascertain the original bi_size value, go ahead and drop this BUG_ON(). Reported-by: Sagi Grimberg Reported-by: Akinobu Mita Acked-by: Martin K. Petersen Cc: Kent Overstreet Signed-off-by: Nicholas Bellinger Signed-off-by: Jens Axboe --- fs/bio-integrity.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/bio-integrity.c b/fs/bio-integrity.c index 0129b78a690..4f70f383132 100644 --- a/fs/bio-integrity.c +++ b/fs/bio-integrity.c @@ -458,11 +458,10 @@ static int bio_integrity_verify(struct bio *bio) struct blk_integrity_exchg bix; struct bio_vec *bv; sector_t sector = bio->bi_integrity->bip_iter.bi_sector; - unsigned int sectors, total, ret; + unsigned int sectors, ret = 0; void *prot_buf = bio->bi_integrity->bip_buf; int i; - ret = total = 0; bix.disk_name = bio->bi_bdev->bd_disk->disk_name; bix.sector_size = bi->sector_size; @@ -484,8 +483,6 @@ static int bio_integrity_verify(struct bio *bio) sectors = bv->bv_len / bi->sector_size; sector += sectors; prot_buf += sectors * bi->tuple_size; - total += sectors * bi->tuple_size; - BUG_ON(total > bio->bi_integrity->bip_iter.bi_size); kunmap_atomic(kaddr); } -- cgit v1.2.3-70-g09d2 From b7e63a1079b266866a732cf699d8c4d61391bbda Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 26 Feb 2014 11:19:14 -0800 Subject: NFSv4: Fix another nfs4_sequence corruptor nfs4_release_lockowner needs to set the rpc_message reply to point to the nfs4_sequence_res in order to avoid another Oopsable situation in nfs41_assign_slot. Fixes: fbd4bfd1d9d21 (NFS: Add nfs4_sequence calls for RELEASE_LOCKOWNER) Cc: stable@vger.kernel.org # 3.12+ Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 10 +++++----- include/linux/nfs_xdr.h | 5 +++++ 2 files changed, 10 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 2da6a698b8f..44e088dc357 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -5828,8 +5828,7 @@ struct nfs_release_lockowner_data { struct nfs4_lock_state *lsp; struct nfs_server *server; struct nfs_release_lockowner_args args; - struct nfs4_sequence_args seq_args; - struct nfs4_sequence_res seq_res; + struct nfs_release_lockowner_res res; unsigned long timestamp; }; @@ -5837,7 +5836,7 @@ static void nfs4_release_lockowner_prepare(struct rpc_task *task, void *calldata { struct nfs_release_lockowner_data *data = calldata; nfs40_setup_sequence(data->server, - &data->seq_args, &data->seq_res, task); + &data->args.seq_args, &data->res.seq_res, task); data->timestamp = jiffies; } @@ -5846,7 +5845,7 @@ static void nfs4_release_lockowner_done(struct rpc_task *task, void *calldata) struct nfs_release_lockowner_data *data = calldata; struct nfs_server *server = data->server; - nfs40_sequence_done(task, &data->seq_res); + nfs40_sequence_done(task, &data->res.seq_res); switch (task->tk_status) { case 0: @@ -5887,7 +5886,6 @@ static int nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_st data = kmalloc(sizeof(*data), GFP_NOFS); if (!data) return -ENOMEM; - nfs4_init_sequence(&data->seq_args, &data->seq_res, 0); data->lsp = lsp; data->server = server; data->args.lock_owner.clientid = server->nfs_client->cl_clientid; @@ -5895,6 +5893,8 @@ static int nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_st data->args.lock_owner.s_dev = server->s_dev; msg.rpc_argp = &data->args; + msg.rpc_resp = &data->res; + nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0); rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, data); return 0; } diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index b2fb167b2e6..5624e4e2763 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -467,9 +467,14 @@ struct nfs_lockt_res { }; struct nfs_release_lockowner_args { + struct nfs4_sequence_args seq_args; struct nfs_lowner lock_owner; }; +struct nfs_release_lockowner_res { + struct nfs4_sequence_res seq_res; +}; + struct nfs4_delegreturnargs { struct nfs4_sequence_args seq_args; const struct nfs_fh *fhandle; -- cgit v1.2.3-70-g09d2 From 755a48a7a4eb05b9c8424e3017d947b2961a60e0 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 2 Mar 2014 22:03:12 -0500 Subject: NFS: Fix a delegation callback race The clean-up in commit 36281caa839f ended up removing a NULL pointer check that is needed in order to prevent an Oops in nfs_async_inode_return_delegation(). Reported-by: "Yan, Zheng" Link: http://lkml.kernel.org/r/5313E9F6.2020405@intel.com Fixes: 36281caa839f (NFSv4: Further clean-ups of delegation stateid validation) Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Trond Myklebust --- fs/nfs/delegation.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index ef792f29f83..5d8ccecf5f5 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -659,16 +659,19 @@ int nfs_async_inode_return_delegation(struct inode *inode, rcu_read_lock(); delegation = rcu_dereference(NFS_I(inode)->delegation); + if (delegation == NULL) + goto out_enoent; - if (!clp->cl_mvops->match_stateid(&delegation->stateid, stateid)) { - rcu_read_unlock(); - return -ENOENT; - } + if (!clp->cl_mvops->match_stateid(&delegation->stateid, stateid)) + goto out_enoent; nfs_mark_return_delegation(server, delegation); rcu_read_unlock(); nfs_delegation_run_state_manager(clp); return 0; +out_enoent: + rcu_read_unlock(); + return -ENOENT; } static struct inode * -- cgit v1.2.3-70-g09d2 From 668f9abbd4334e6c29fa8acd71635c4f9101caa7 Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Mon, 3 Mar 2014 15:38:18 -0800 Subject: mm: close PageTail race Commit bf6bddf1924e ("mm: introduce compaction and migration for ballooned pages") introduces page_count(page) into memory compaction which dereferences page->first_page if PageTail(page). This results in a very rare NULL pointer dereference on the aforementioned page_count(page). Indeed, anything that does compound_head(), including page_count() is susceptible to racing with prep_compound_page() and seeing a NULL or dangling page->first_page pointer. This patch uses Andrea's implementation of compound_trans_head() that deals with such a race and makes it the default compound_head() implementation. This includes a read memory barrier that ensures that if PageTail(head) is true that we return a head page that is neither NULL nor dangling. The patch then adds a store memory barrier to prep_compound_page() to ensure page->first_page is set. This is the safest way to ensure we see the head page that we are expecting, PageTail(page) is already in the unlikely() path and the memory barriers are unfortunately required. Hugetlbfs is the exception, we don't enforce a store memory barrier during init since no race is possible. Signed-off-by: David Rientjes Cc: Holger Kiehl Cc: Christoph Lameter Cc: Rafael Aquini Cc: Vlastimil Babka Cc: Michal Hocko Cc: Mel Gorman Cc: Andrea Arcangeli Cc: Rik van Riel Cc: "Kirill A. Shutemov" Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/block/aoe/aoecmd.c | 4 ++-- drivers/vfio/vfio_iommu_type1.c | 4 ++-- fs/proc/page.c | 5 ++--- include/linux/huge_mm.h | 41 ----------------------------------------- include/linux/mm.h | 14 ++++++++++++-- mm/ksm.c | 2 +- mm/memory-failure.c | 2 +- mm/page_alloc.c | 4 +++- mm/swap.c | 4 ++-- 9 files changed, 25 insertions(+), 55 deletions(-) (limited to 'fs') diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c index 8184451b57c..422b7d84f68 100644 --- a/drivers/block/aoe/aoecmd.c +++ b/drivers/block/aoe/aoecmd.c @@ -874,7 +874,7 @@ bio_pageinc(struct bio *bio) /* Non-zero page count for non-head members of * compound pages is no longer allowed by the kernel. */ - page = compound_trans_head(bv.bv_page); + page = compound_head(bv.bv_page); atomic_inc(&page->_count); } } @@ -887,7 +887,7 @@ bio_pagedec(struct bio *bio) struct bvec_iter iter; bio_for_each_segment(bv, bio, iter) { - page = compound_trans_head(bv.bv_page); + page = compound_head(bv.bv_page); atomic_dec(&page->_count); } } diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 4fb7a8f83c8..54af4e93369 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -186,12 +186,12 @@ static bool is_invalid_reserved_pfn(unsigned long pfn) if (pfn_valid(pfn)) { bool reserved; struct page *tail = pfn_to_page(pfn); - struct page *head = compound_trans_head(tail); + struct page *head = compound_head(tail); reserved = !!(PageReserved(head)); if (head != tail) { /* * "head" is not a dangling pointer - * (compound_trans_head takes care of that) + * (compound_head takes care of that) * but the hugepage may have been split * from under us (and we may not hold a * reference count on the head page so it can diff --git a/fs/proc/page.c b/fs/proc/page.c index 02174a61031..e647c55275d 100644 --- a/fs/proc/page.c +++ b/fs/proc/page.c @@ -121,9 +121,8 @@ u64 stable_page_flags(struct page *page) * just checks PG_head/PG_tail, so we need to check PageLRU/PageAnon * to make sure a given page is a thp, not a non-huge compound page. */ - else if (PageTransCompound(page) && - (PageLRU(compound_trans_head(page)) || - PageAnon(compound_trans_head(page)))) + else if (PageTransCompound(page) && (PageLRU(compound_head(page)) || + PageAnon(compound_head(page)))) u |= 1 << KPF_THP; /* diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index db512014e06..b826239bdce 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -157,46 +157,6 @@ static inline int hpage_nr_pages(struct page *page) return HPAGE_PMD_NR; return 1; } -/* - * compound_trans_head() should be used instead of compound_head(), - * whenever the "page" passed as parameter could be the tail of a - * transparent hugepage that could be undergoing a - * __split_huge_page_refcount(). The page structure layout often - * changes across releases and it makes extensive use of unions. So if - * the page structure layout will change in a way that - * page->first_page gets clobbered by __split_huge_page_refcount, the - * implementation making use of smp_rmb() will be required. - * - * Currently we define compound_trans_head as compound_head, because - * page->private is in the same union with page->first_page, and - * page->private isn't clobbered. However this also means we're - * currently leaving dirt into the page->private field of anonymous - * pages resulting from a THP split, instead of setting page->private - * to zero like for every other page that has PG_private not set. But - * anonymous pages don't use page->private so this is not a problem. - */ -#if 0 -/* This will be needed if page->private will be clobbered in split_huge_page */ -static inline struct page *compound_trans_head(struct page *page) -{ - if (PageTail(page)) { - struct page *head; - head = page->first_page; - smp_rmb(); - /* - * head may be a dangling pointer. - * __split_huge_page_refcount clears PageTail before - * overwriting first_page, so if PageTail is still - * there it means the head pointer isn't dangling. - */ - if (PageTail(page)) - return head; - } - return page; -} -#else -#define compound_trans_head(page) compound_head(page) -#endif extern int do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, pmd_t pmd, pmd_t *pmdp); @@ -226,7 +186,6 @@ static inline int split_huge_page(struct page *page) do { } while (0) #define split_huge_page_pmd_mm(__mm, __address, __pmd) \ do { } while (0) -#define compound_trans_head(page) compound_head(page) static inline int hugepage_madvise(struct vm_area_struct *vma, unsigned long *vm_flags, int advice) { diff --git a/include/linux/mm.h b/include/linux/mm.h index f28f46eade6..03ab3e58f51 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -399,8 +399,18 @@ static inline void compound_unlock_irqrestore(struct page *page, static inline struct page *compound_head(struct page *page) { - if (unlikely(PageTail(page))) - return page->first_page; + if (unlikely(PageTail(page))) { + struct page *head = page->first_page; + + /* + * page->first_page may be a dangling pointer to an old + * compound page, so recheck that it is still a tail + * page before returning. + */ + smp_rmb(); + if (likely(PageTail(page))) + return head; + } return page; } diff --git a/mm/ksm.c b/mm/ksm.c index aa4c7c7250c..68710e80994 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -444,7 +444,7 @@ static void break_cow(struct rmap_item *rmap_item) static struct page *page_trans_compound_anon(struct page *page) { if (PageTransCompound(page)) { - struct page *head = compound_trans_head(page); + struct page *head = compound_head(page); /* * head may actually be splitted and freed from under * us but it's ok here. diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 2f2f34a4e77..90002ea4363 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -1651,7 +1651,7 @@ int soft_offline_page(struct page *page, int flags) { int ret; unsigned long pfn = page_to_pfn(page); - struct page *hpage = compound_trans_head(page); + struct page *hpage = compound_head(page); if (PageHWPoison(page)) { pr_info("soft offline: %#lx page already poisoned\n", pfn); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index e3758a09a00..3d1bf889465 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -369,9 +369,11 @@ void prep_compound_page(struct page *page, unsigned long order) __SetPageHead(page); for (i = 1; i < nr_pages; i++) { struct page *p = page + i; - __SetPageTail(p); set_page_count(p, 0); p->first_page = page; + /* Make sure p->first_page is always valid for PageTail() */ + smp_wmb(); + __SetPageTail(p); } } diff --git a/mm/swap.c b/mm/swap.c index b31ba67d440..0092097b3f4 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -98,7 +98,7 @@ static void put_compound_page(struct page *page) } /* __split_huge_page_refcount can run under us */ - page_head = compound_trans_head(page); + page_head = compound_head(page); /* * THP can not break up slab pages so avoid taking @@ -253,7 +253,7 @@ bool __get_page_tail(struct page *page) */ unsigned long flags; bool got; - struct page *page_head = compound_trans_head(page); + struct page *page_head = compound_head(page); /* Ref to put_compound_page() comment. */ if (!__compound_tail_refcounted(page_head)) { -- cgit v1.2.3-70-g09d2 From 15c34a760630ca2c803848fba90ca0646a9907dd Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 3 Mar 2014 15:38:32 -0800 Subject: ocfs2: fix quota file corruption Global quota files are accessed from different nodes. Thus we cannot cache offset of quota structure in the quota file after we drop our node reference count to it because after that moment quota structure may be freed and reallocated elsewhere by a different node resulting in corruption of quota file. Fix the problem by clearing dq_off when we are releasing dquot structure. We also remove the DB_READ_B handling because it is useless - DQ_ACTIVE_B is set iff DQ_READ_B is set. Signed-off-by: Jan Kara Cc: Goldwyn Rodrigues Cc: Joel Becker Reviewed-by: Mark Fasheh Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/quota_global.c | 27 +++++++++++++++++---------- fs/ocfs2/quota_local.c | 4 ---- 2 files changed, 17 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index aaa50611ec6..d7b5108789e 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -717,6 +717,12 @@ static int ocfs2_release_dquot(struct dquot *dquot) */ if (status < 0) mlog_errno(status); + /* + * Clear dq_off so that we search for the structure in quota file next + * time we acquire it. The structure might be deleted and reallocated + * elsewhere by another node while our dquot structure is on freelist. + */ + dquot->dq_off = 0; clear_bit(DQ_ACTIVE_B, &dquot->dq_flags); out_trans: ocfs2_commit_trans(osb, handle); @@ -756,16 +762,17 @@ static int ocfs2_acquire_dquot(struct dquot *dquot) status = ocfs2_lock_global_qf(info, 1); if (status < 0) goto out; - if (!test_bit(DQ_READ_B, &dquot->dq_flags)) { - status = ocfs2_qinfo_lock(info, 0); - if (status < 0) - goto out_dq; - status = qtree_read_dquot(&info->dqi_gi, dquot); - ocfs2_qinfo_unlock(info, 0); - if (status < 0) - goto out_dq; - } - set_bit(DQ_READ_B, &dquot->dq_flags); + status = ocfs2_qinfo_lock(info, 0); + if (status < 0) + goto out_dq; + /* + * We always want to read dquot structure from disk because we don't + * know what happened with it while it was on freelist. + */ + status = qtree_read_dquot(&info->dqi_gi, dquot); + ocfs2_qinfo_unlock(info, 0); + if (status < 0) + goto out_dq; OCFS2_DQUOT(dquot)->dq_use_count++; OCFS2_DQUOT(dquot)->dq_origspace = dquot->dq_dqb.dqb_curspace; diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index 2e4344be3b9..2001862bf2b 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -1303,10 +1303,6 @@ int ocfs2_local_release_dquot(handle_t *handle, struct dquot *dquot) ocfs2_journal_dirty(handle, od->dq_chunk->qc_headerbh); out: - /* Clear the read bit so that next time someone uses this - * dquot he reads fresh info from disk and allocates local - * dquot structure */ - clear_bit(DQ_READ_B, &dquot->dq_flags); return status; } -- cgit v1.2.3-70-g09d2 From bd2c00353286d63542a8a0896a8c747f7c880edd Mon Sep 17 00:00:00 2001 From: Vyacheslav Dubeyko Date: Mon, 3 Mar 2014 15:38:35 -0800 Subject: hfsplus: fix remount issue Current implementation of HFS+ driver has small issue with remount option. Namely, for example, you are unable to remount from RO mode into RW mode by means of command "mount -o remount,rw /dev/loop0 /mnt/hfsplus". Trying to execute sequence of commands results in an error message: mount /dev/loop0 /mnt/hfsplus mount -o remount,ro /dev/loop0 /mnt/hfsplus mount -o remount,rw /dev/loop0 /mnt/hfsplus mount: you must specify the filesystem type mount -t hfsplus -o remount,rw /dev/loop0 /mnt/hfsplus mount: /mnt/hfsplus not mounted or bad option The reason of such issue is failure of mount syscall: mount("/dev/loop0", "/mnt/hfsplus", 0x2282a60, MS_MGC_VAL|MS_REMOUNT, NULL) = -1 EINVAL (Invalid argument) Namely, hfsplus_parse_options_remount() method receives empty "input" argument and return false in such case. As a result, hfsplus_remount() returns -EINVAL error code. This patch fixes the issue by means of return true for the case of empty "input" argument in hfsplus_parse_options_remount() method. Signed-off-by: Vyacheslav Dubeyko Cc: Al Viro Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hfsplus/options.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/hfsplus/options.c b/fs/hfsplus/options.c index 968eab5bc1f..68537e8b7a0 100644 --- a/fs/hfsplus/options.c +++ b/fs/hfsplus/options.c @@ -75,7 +75,7 @@ int hfsplus_parse_options_remount(char *input, int *force) int token; if (!input) - return 0; + return 1; while ((p = strsep(&input, ",")) != NULL) { if (!*p) -- cgit v1.2.3-70-g09d2 From e1253be0ece1a95a02c7f5843194877471af8179 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 5 Mar 2014 08:44:23 -0500 Subject: NFSv4: nfs4_stateid_is_current should return 'true' for an invalid stateid When nfs4_set_rw_stateid() can fails by returning EIO to indicate that the stateid is completely invalid, then it makes no sense to have it trigger a retry of the READ or WRITE operation. Instead, we should just have it fall through and attempt a recovery. This fixes an infinite loop in which the client keeps replaying the same bad stateid back to the server. Reported-by: Andy Adamson Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com Cc: stable@vger.kernel.org # 3.10+ Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 44e088dc357..4ae8141452c 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -4011,8 +4011,9 @@ static bool nfs4_stateid_is_current(nfs4_stateid *stateid, { nfs4_stateid current_stateid; - if (nfs4_set_rw_stateid(¤t_stateid, ctx, l_ctx, fmode)) - return false; + /* If the current stateid represents a lost lock, then exit */ + if (nfs4_set_rw_stateid(¤t_stateid, ctx, l_ctx, fmode) == -EIO) + return true; return nfs4_stateid_match(stateid, ¤t_stateid); } -- cgit v1.2.3-70-g09d2 From 927864cd92aaad1d6285e3bb16e503caf3d6e27e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 4 Mar 2014 13:12:03 -0500 Subject: NFSv4: Fix the return value of nfs4_select_rw_stateid In commit 5521abfdcf4d6 (NFSv4: Resend the READ/WRITE RPC call if a stateid change causes an error), we overloaded the return value of nfs4_select_rw_stateid() to cause it to return -EWOULDBLOCK if an RPC call is outstanding that would cause the NFSv4 lock or open stateid to change. That is all redundant when we actually copy the stateid used in the read/write RPC call that failed, and check that against the current stateid. It is doubly so, when we consider that in the NFSv4.1 case, we also set the stateid's seqid to the special value '0', which means 'match the current valid stateid'. Reported-by: Andy Adamson Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com Signed-off-by: Trond Myklebust --- fs/nfs/nfs4state.c | 14 +++----------- 1 file changed, 3 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index e1a47217c05..0deb32105cc 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -974,9 +974,6 @@ static int nfs4_copy_lock_stateid(nfs4_stateid *dst, else if (lsp != NULL && test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0) { nfs4_stateid_copy(dst, &lsp->ls_stateid); ret = 0; - smp_rmb(); - if (!list_empty(&lsp->ls_seqid.list)) - ret = -EWOULDBLOCK; } spin_unlock(&state->state_lock); nfs4_put_lock_state(lsp); @@ -984,10 +981,9 @@ out: return ret; } -static int nfs4_copy_open_stateid(nfs4_stateid *dst, struct nfs4_state *state) +static void nfs4_copy_open_stateid(nfs4_stateid *dst, struct nfs4_state *state) { const nfs4_stateid *src; - int ret; int seq; do { @@ -996,12 +992,7 @@ static int nfs4_copy_open_stateid(nfs4_stateid *dst, struct nfs4_state *state) if (test_bit(NFS_OPEN_STATE, &state->flags)) src = &state->open_stateid; nfs4_stateid_copy(dst, src); - ret = 0; - smp_rmb(); - if (!list_empty(&state->owner->so_seqid.list)) - ret = -EWOULDBLOCK; } while (read_seqretry(&state->seqlock, seq)); - return ret; } /* @@ -1026,7 +1017,8 @@ int nfs4_select_rw_stateid(nfs4_stateid *dst, struct nfs4_state *state, * choose to use. */ goto out; - ret = nfs4_copy_open_stateid(dst, state); + nfs4_copy_open_stateid(dst, state); + ret = 0; out: if (nfs_server_capable(state->inode, NFS_CAP_STATEID_NFSV41)) dst->seqid = 0; -- cgit v1.2.3-70-g09d2 From 869a9d375dca601dde1dee1344f3d1d665505f19 Mon Sep 17 00:00:00 2001 From: Andy Adamson Date: Tue, 4 Mar 2014 12:31:09 -0500 Subject: NFSv4.1 Fail data server I/O if stateid represents a lost lock Signed-off-by: Andy Adamson Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com Signed-off-by: Trond Myklebust --- fs/nfs/nfs4filelayout.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 12c8132ad40..b9a35c05b60 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -324,8 +324,9 @@ static void filelayout_read_prepare(struct rpc_task *task, void *data) &rdata->res.seq_res, task)) return; - nfs4_set_rw_stateid(&rdata->args.stateid, rdata->args.context, - rdata->args.lock_context, FMODE_READ); + if (nfs4_set_rw_stateid(&rdata->args.stateid, rdata->args.context, + rdata->args.lock_context, FMODE_READ) == -EIO) + rpc_exit(task, -EIO); /* lost lock, terminate I/O */ } static void filelayout_read_call_done(struct rpc_task *task, void *data) @@ -435,8 +436,9 @@ static void filelayout_write_prepare(struct rpc_task *task, void *data) &wdata->res.seq_res, task)) return; - nfs4_set_rw_stateid(&wdata->args.stateid, wdata->args.context, - wdata->args.lock_context, FMODE_WRITE); + if (nfs4_set_rw_stateid(&wdata->args.stateid, wdata->args.context, + wdata->args.lock_context, FMODE_WRITE) == -EIO) + rpc_exit(task, -EIO); /* lost lock, terminate I/O */ } static void filelayout_write_call_done(struct rpc_task *task, void *data) -- cgit v1.2.3-70-g09d2 From 0418dae1056d6091e9527b7158a3763f7aa92353 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 4 Mar 2014 13:48:16 -0500 Subject: NFSv4: Fail the truncate() if the lock/open stateid is invalid If the open stateid could not be recovered, or the file locks were lost, then we should fail the truncate() operation altogether. Reported-by: Andy Adamson Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 4ae8141452c..450bfedbe2f 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2398,13 +2398,16 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, if (nfs4_copy_delegation_stateid(&arg.stateid, inode, fmode)) { /* Use that stateid */ - } else if (truncate && state != NULL && nfs4_valid_open_stateid(state)) { + } else if (truncate && state != NULL) { struct nfs_lockowner lockowner = { .l_owner = current->files, .l_pid = current->tgid, }; - nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, - &lockowner); + if (!nfs4_valid_open_stateid(state)) + return -EBADF; + if (nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE, + &lockowner) == -EIO) + return -EBADF; } else nfs4_stateid_copy(&arg.stateid, &zero_stateid); -- cgit v1.2.3-70-g09d2