kernel/linux.git/fs/fscache/page.c, branch linux-4.4.y

fscache: Fix hanging wait on page discarded by writeback

2018-05-30T05:49:08+00:00

[ Upstream commit 2c98425720233ae3e135add0c7e869b32913502f ] If the fscache asynchronous write operation elects to discard a page that's pending storage to the cache because the page would be over the store limit then it needs to wake the page as someone may be waiting on completion of the write. The problem is that the store limit may be updated by a different asynchronous operation - and so may miss the write - and that the store limit may not even get updated until later by the netfs. Fix the kernel hang by making fscache_write_op() mark as written any pages that are over the limit. Signed-off-by: David Howells Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

FS-Cache: Handle a write to the page immediately beyond the EOF marker

2015-11-11T07:11:02+00:00

Handle a write being requested to the page immediately beyond the EOF marker on a cache object. Currently this gets an assertion failure in CacheFiles because the EOF marker is used there to encode information about a partial page at the EOF - which could lead to an unknown blank spot in the file if we extend the file over it. The problem is actually in fscache where we check the index of the page being written against store_limit. store_limit is set to the number of pages that we're allowed to store by fscache_set_store_limit() - which means it's one more than the index of the last page we're allowed to store. The problem is that we permit writing to a page with an index _equal_ to the store limit - when we should reject that case. Whilst we're at it, change the triggered assertion in CacheFiles to just return -ENOBUFS instead. The assertion failure looks something like this: CacheFiles: Assertion failed 1000 < 7b1 is false ------------[ cut here ]------------ kernel BUG at fs/cachefiles/rdwr.c:962! ... RIP: 0010:[] [] cachefiles_write_page+0x273/0x2d0 [cachefiles] Cc: stable@vger.kernel.org # v2.6.31+; earlier - that + backport of a17754f (at least) Signed-off-by: David Howells Signed-off-by: Al Viro

mm, page_alloc: distinguish between being unable to sleep, unwilling to sleep and avoiding waking kswapd

2015-11-07T01:50:42+00:00

__GFP_WAIT has been used to identify atomic context in callers that hold spinlocks or are in interrupts. They are expected to be high priority and have access one of two watermarks lower than "min" which can be referred to as the "atomic reserve". __GFP_HIGH users get access to the first lower watermark and can be called the "high priority reserve". Over time, callers had a requirement to not block when fallback options were available. Some have abused __GFP_WAIT leading to a situation where an optimisitic allocation with a fallback option can access atomic reserves. This patch uses __GFP_ATOMIC to identify callers that are truely atomic, cannot sleep and have no alternative. High priority users continue to use __GFP_HIGH. __GFP_DIRECT_RECLAIM identifies callers that can sleep and are willing to enter direct reclaim. __GFP_KSWAPD_RECLAIM to identify callers that want to wake kswapd for background reclaim. __GFP_WAIT is redefined as a caller that is willing to enter direct reclaim and wake kswapd for background reclaim. This patch then converts a number of sites o __GFP_ATOMIC is used by callers that are high priority and have memory pools for those requests. GFP_ATOMIC uses this flag. o Callers that have a limited mempool to guarantee forward progress clear __GFP_DIRECT_RECLAIM but keep __GFP_KSWAPD_RECLAIM. bio allocations fall into this category where kswapd will still be woken but atomic reserves are not used as there is a one-entry mempool to guarantee progress. o Callers that are checking if they are non-blocking should use the helper gfpflags_allow_blocking() where possible. This is because checking for __GFP_WAIT as was done historically now can trigger false positives. Some exceptions like dm-crypt.c exist where the code intent is clearer if __GFP_DIRECT_RECLAIM is used instead of the helper due to flag manipulations. o Callers that built their own GFP flags instead of starting with GFP_KERNEL and friends now also need to specify __GFP_KSWAPD_RECLAIM. The first key hazard to watch out for is callers that removed __GFP_WAIT and was depending on access to atomic reserves for inconspicuous reasons. In some cases it may be appropriate for them to use __GFP_HIGH. The second key hazard is callers that assembled their own combination of GFP flags instead of starting with something like GFP_KERNEL. They may now wish to specify __GFP_KSWAPD_RECLAIM. It's almost certainly harmless if it's missed in most cases as other activity will wake kswapd. Signed-off-by: Mel Gorman Acked-by: Vlastimil Babka Acked-by: Michal Hocko Acked-by: Johannes Weiner Cc: Christoph Lameter Cc: David Rientjes Cc: Vitaly Wool Cc: Rik van Riel Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

FS-Cache: Retain the netfs context in the retrieval op earlier

2015-04-02T13:28:53+00:00

Now that the retrieval operation may be disposed of by fscache_put_operation() before we actually set the context, the retrieval-specific cleanup operation can produce a NULL-pointer dereference when it tries to unconditionally clean up the netfs context. Given that it is expected that we'll get at least as far as the place where we currently set the context pointer and it is unlikely we'll go through the error handling paths prior to that point, retain the context right from the point that the retrieval op is allocated. Concomitant to this, we need to retain the cookie pointer in the retrieval op also so that we can call the netfs to release its context in the release method. In addition, we might now get into fscache_release_retrieval_op() with the op only initialised. To this end, set the operation to DEAD only after the release method has been called and skip the n_pages test upon cleanup if the op is still in the INITIALISED state. Without these changes, the following oops might be seen: BUG: unable to handle kernel NULL pointer dereference at 00000000000000b8 ... RIP: 0010:[] fscache_release_retrieval_op+0xae/0x100 ... Call Trace: [] fscache_put_operation+0x117/0x2e0 [] __fscache_read_or_alloc_pages+0x351/0x3ac [] __nfs_readpages_from_fscache+0x59/0xbf [nfs] [] nfs_readpages+0x10c/0x185 [nfs] [] ? alloc_pages_current+0x119/0x13e [] ? __page_cache_alloc+0xfb/0x10a [] __do_page_cache_readahead+0x188/0x22c [] ondemand_readahead+0x29e/0x2af [] page_cache_sync_readahead+0x38/0x3a [] generic_file_read_iter+0x1a2/0x55a [] ? nfs_revalidate_mapping+0xd6/0x288 [nfs] [] nfs_file_read+0x49/0x70 [nfs] [] new_sync_read+0x78/0x9c [] __vfs_read+0x13/0x38 [] vfs_read+0x95/0x121 [] SyS_read+0x4c/0x8a [] system_call_fastpath+0x12/0x17 Signed-off-by: David Howells Reviewed-by: Steve Dickson Acked-by: Jeff Layton

FS-Cache: The operation cancellation method needs calling in more places

2015-04-02T13:28:53+00:00

Any time an incomplete operation is cancelled, the operation cancellation function needs to be called to clean up. This is currently being passed directly to some of the functions that might want to call it, but not all. Instead, pass the cancellation method pointer to the fscache_operation_init() and have that cache it in the operation struct. Further, plug in a dummy cancellation handler if the caller declines to set one as this allows us to call the function unconditionally (the extra overhead isn't worth bothering about as we don't expect to be calling this typically). The cancellation method must thence be called everywhere the CANCELLED state is set. Note that we call it *before* setting the CANCELLED state such that the method can use the old state value to guide its operation. fscache_do_cancel_retrieval() needs moving higher up in the sources so that the init function can use it now. Without this, the following oops may be seen: FS-Cache: Assertion failed FS-Cache: 3 == 0 is false ------------[ cut here ]------------ kernel BUG at ../fs/fscache/page.c:261! ... RIP: 0010:[] fscache_release_retrieval_op+0x77/0x100 [] fscache_put_operation+0x114/0x2da [] __fscache_read_or_alloc_pages+0x358/0x3b3 [] __nfs_readpages_from_fscache+0x59/0xbf [nfs] [] nfs_readpages+0x10c/0x185 [nfs] [] ? alloc_pages_current+0x119/0x13e [] ? __page_cache_alloc+0xfb/0x10a [] __do_page_cache_readahead+0x188/0x22c [] ondemand_readahead+0x29e/0x2af [] page_cache_sync_readahead+0x38/0x3a [] generic_file_read_iter+0x1a2/0x55a [] ? nfs_revalidate_mapping+0xd6/0x288 [nfs] [] nfs_file_read+0x49/0x70 [nfs] [] new_sync_read+0x78/0x9c [] __vfs_read+0x13/0x38 [] vfs_read+0x95/0x121 [] SyS_read+0x4c/0x8a [] system_call_fastpath+0x12/0x17 The assertion is showing that the remaining number of pages (n_pages) is not 0 when the operation is being released. Signed-off-by: David Howells Reviewed-by: Steve Dickson Acked-by: Jeff Layton

FS-Cache: Put an aborted initialised op so that it is accounted correctly

2015-04-02T13:28:53+00:00

Call fscache_put_operation() or a wrapper on any op that has gone through fscache_operation_init() so that the accounting shown in /proc is done correctly, specifically fscache_n_op_release. fscache_put_operation() therefore now allows an op in the INITIALISED state as well as in the CANCELLED and COMPLETE states. Note that this means that an operation can get put that doesn't have its ->object pointer filled in, so anything that depends on the object needs to be conditional in fscache_put_operation(). Signed-off-by: David Howells Reviewed-by: Steve Dickson Acked-by: Jeff Layton

FS-Cache: Permit fscache_cancel_op() to cancel in-progress operations too

2015-04-02T13:28:53+00:00

Currently, fscache_cancel_op() only cancels pending operations - attempts to cancel in-progress operations are ignored. This leads to a problem in fscache_wait_for_operation_activation() whereby the wait is terminated, but the object has been killed. The check at the end of the function now triggers because it's no longer contingent on the cache having produced an I/O error since the commit that fixed the logic error in fscache_object_is_dead(). The result of the check is that it tries to cancel the operation - but since the object may not be pending by this point, the cancellation request may be ignored - with the result that the the object is just put by the caller and fscache_put_operation has an assertion failure because the operation isn't in either the COMPLETE or the CANCELLED states. To fix this, we permit in-progress ops to be cancelled under some circumstances. The bug results in an oops that looks something like this: FS-Cache: fscache_wait_for_operation_activation() = -ENOBUFS [obj dead 3] FS-Cache: FS-Cache: Assertion failed FS-Cache: 3 == 5 is false ------------[ cut here ]------------ kernel BUG at ../fs/fscache/operation.c:432! ... RIP: 0010:[] fscache_put_operation+0xf2/0x2cd Call Trace: [] __fscache_read_or_alloc_pages+0x2ec/0x3b3 [] __nfs_readpages_from_fscache+0x59/0xbf [nfs] [] nfs_readpages+0x10c/0x185 [nfs] [] ? alloc_pages_current+0x119/0x13e [] ? __page_cache_alloc+0xfb/0x10a [] __do_page_cache_readahead+0x188/0x22c [] ondemand_readahead+0x29e/0x2af [] page_cache_sync_readahead+0x38/0x3a [] generic_file_read_iter+0x1a2/0x55a [] ? nfs_revalidate_mapping+0xd6/0x288 [nfs] [] nfs_file_read+0x49/0x70 [nfs] [] new_sync_read+0x78/0x9c [] __vfs_read+0x13/0x38 [] vfs_read+0x95/0x121 [] SyS_read+0x4c/0x8a [] system_call_fastpath+0x12/0x17 Signed-off-by: David Howells Reviewed-by: Steve Dickson Acked-by: Jeff Layton

FS-Cache: fscache_object_is_dead() has wrong logic, kill it

2015-04-02T13:28:53+00:00

fscache_object_is_dead() returns true only if the object is marked dead and the cache got an I/O error. This should be a logical OR instead. Since two of the callers got split up into handling for separate subcases, expand the other callers and kill the function. This is probably the right thing to do anyway since one of the subcases isn't about the object at all, but rather about the cache. Signed-off-by: David Howells Reviewed-by: Steve Dickson Acked-by: Jeff Layton

FS-Cache: refcount becomes corrupt under vma pressure.

2014-09-17T21:41:40+00:00

In rare cases under heavy VMA pressure the ref count for a fscache cookie becomes corrupt. In this case we decrement ref count even if we fail before incrementing the refcount. FS-Cache: Assertion failed bnode-eca5f9c6/syslog 0 > 0 is false ------------[ cut here ]------------ kernel BUG at fs/fscache/cookie.c:519! invalid opcode: 0000 [#1] SMP Call Trace: [] __fscache_relinquish_cookie+0x50/0x220 [fscache] [] ceph_fscache_unregister_inode_cookie+0x3e/0x50 [ceph] [] ceph_destroy_inode+0x33/0x200 [ceph] [] ? __fsnotify_inode_delete+0xe/0x10 [] destroy_inode+0x3c/0x70 [] evict+0x111/0x180 [] iput+0x103/0x190 [] __dentry_kill+0x1c8/0x220 [] shrink_dentry_list+0xf1/0x250 [] prune_dcache_sb+0x4c/0x60 [] super_cache_scan+0xff/0x170 [] shrink_slab_node+0x140/0x2c0 [] shrink_slab+0x8a/0x130 [] balance_pgdat+0x3e2/0x5d0 [] kswapd+0x16a/0x4a0 [] ? __wake_up_sync+0x20/0x20 [] ? balance_pgdat+0x5d0/0x5d0 [] kthread+0xc9/0xe0 [] ? ftrace_raw_event_xen_mmu_release_ptpage+0x70/0x90 [] ? flush_kthread_worker+0xb0/0xb0 [] ret_from_fork+0x7c/0xb0 [] ? flush_kthread_worker+0xb0/0xb0 RIP [] __fscache_disable_cookie+0x1db/0x210 [fscache] RSP ---[ end trace 254d0d7c74a01f25 ]--- Signed-off-by: Milosz Tanski Signed-off-by: David Howells

FS-Cache: Timeout for releasepage()

2014-08-27T14:24:06+00:00

This is meant to avoid a recusive hang caused by underlying filesystem trying to grab a free page and causing a write-out. INFO: task kworker/u30:7:28375 blocked for more than 120 seconds. Not tainted 3.15.0-virtual #74 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. kworker/u30:7 D 0000000000000000 0 28375 2 0x00000000 Workqueue: fscache_operation fscache_op_work_func [fscache] ffff88000b147148 0000000000000046 0000000000000000 ffff88000b1471c8 ffff8807aa031820 0000000000014040 ffff88000b147fd8 0000000000014040 ffff880f0c50c860 ffff8807aa031820 ffff88000b147158 ffff88007be59cd0 Call Trace: [] schedule+0x29/0x70 [] __fscache_wait_on_page_write+0x55/0x90 [fscache] [] ? __wake_up_sync+0x20/0x20 [] __fscache_maybe_release_page+0x65/0x1e0 [fscache] [] ceph_releasepage+0x83/0x100 [ceph] [] ? anon_vma_fork+0x130/0x130 [] try_to_release_page+0x32/0x50 [] shrink_page_list+0x7e6/0x9d0 [] ? isolate_lru_pages.isra.73+0x78/0x1e0 [] shrink_inactive_list+0x252/0x4c0 [] shrink_lruvec+0x3e1/0x670 [] shrink_zone+0x3f/0x110 [] do_try_to_free_pages+0x1d6/0x450 [] ? zone_statistics+0x99/0xc0 [] try_to_free_pages+0xc4/0x180 [] __alloc_pages_nodemask+0x6b2/0xa60 [] ? __find_get_block+0xbe/0x250 [] ? wake_up_bit+0x2e/0x40 [] alloc_pages_current+0xb3/0x180 [] __page_cache_alloc+0xb7/0xd0 [] grab_cache_page_write_begin+0x7c/0xe0 [] ? ext4_mark_inode_dirty+0x82/0x220 [] ext4_da_write_begin+0x89/0x2d0 [] generic_perform_write+0xbe/0x1d0 [] ? update_time+0x81/0xc0 [] ? mnt_clone_write+0x12/0x30 [] __generic_file_aio_write+0x1ce/0x3f0 [] generic_file_aio_write+0x5e/0xe0 [] ext4_file_write+0x9f/0x410 [] ? ext4_file_open+0x66/0x180 [] do_sync_write+0x5a/0x90 [] cachefiles_write_page+0x149/0x430 [cachefiles] [] ? radix_tree_gang_lookup_tag+0x89/0xd0 [] fscache_write_op+0x222/0x3b0 [fscache] [] fscache_op_work_func+0x3a/0x100 [fscache] [] process_one_work+0x179/0x4a0 [] worker_thread+0x11b/0x370 [] ? manage_workers.isra.21+0x2e0/0x2e0 [] kthread+0xc9/0xe0 [] ? ftrace_raw_event_xen_mmu_release_ptpage+0x70/0x90 [] ? flush_kthread_worker+0xb0/0xb0 [] ret_from_fork+0x7c/0xb0 [] ? flush_kthread_worker+0xb0/0xb0 Signed-off-by: Milosz Tanski Signed-off-by: David Howells