kAFS can be given certain network errors (EADDRNOTAVAIL, EHOSTDOWN and
ERFKILL) that it doesn't handle in its server/address rotation algorithms.
They cause the probing and rotation to abort immediately rather than
rotating.
Fix this by:
(1) Abstracting out the error prioritisation from the VL and FS rotation
algorithms into a common function and expand usage into the server
probing code.
When multiple errors are available, this code selects the one we'd
prefer to return.
(2) Add handling for EADDRNOTAVAIL, EHOSTDOWN and ERFKILL.
Fixes: 0fafdc9f88 ("afs: Fix file locking")
Fixes: 0338747d8454 ("afs: Probe multiple fileservers simultaneously")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Send probes to all the unprobed fileservers in a fileserver list on all
addresses simultaneously in an attempt to find out the fastest route whilst
not getting stuck for 20s on any server or address that we don't get a
reply from.
This alleviates the problem whereby attempting to access a new server can
take a long time because the rotation algorithm ends up rotating through
all servers and addresses until it finds one that responds.
Signed-off-by: David Howells <dhowells@redhat.com>
Provide an option to allow the file or volume location server cursor to be
dumped if the rotation routine falls off the end without managing to
contact a server.
Signed-off-by: David Howells <dhowells@redhat.com>
Increase the sizes of the volume ID to 64 bits and the vnode ID (inode
number equivalent) to 96 bits to allow the support of YFS.
This requires the iget comparator to check the vnode->fid rather than i_ino
and i_generation as i_ino is not sufficiently capacious. It also requires
this data to be placed into the vnode cache key for fscache.
For the moment, just discard the top 32 bits of the vnode ID when returning
it though stat.
Signed-off-by: David Howells <dhowells@redhat.com>
Improve the error handling in FS server rotation by:
(1) Cache the latest useful error value for the fs operation as a whole in
struct afs_fs_cursor separately from the error cached in the
afs_addr_cursor struct. The one in the address cursor gets clobbered
occasionally. Copy over the error to the fs operation only when it's
something we'd be interested in passing to userspace.
(2) Make it so that EDESTADDRREQ is the default that is seen only if no
addresses are available to be accessed.
(3) When calling utility functions, such as checking a volume status or
probing a fileserver, don't let a successful result clobber the cached
error in the cursor; instead, stash the result in a temporary variable
until it has been assessed.
(4) Don't return ETIMEDOUT or ETIME if a better error, such as
ENETUNREACH, is already cached.
(5) On leaving the rotation loop, turn any remote abort code into a more
useful error than ECONNABORTED.
Fixes: d2ddc776a4 ("afs: Overhaul volume and server record caching and fileserver rotation")
Signed-off-by: David Howells <dhowells@redhat.com>
Sometimes an in-progress call will stop responding on the fileserver when
the fileserver quietly cancels the call with an internally marked abort
(RX_CALL_DEAD), without sending an ABORT to the client.
This causes the client's call to eventually expire from lack of incoming
packets directed its way, which currently leads to it being cancelled
locally with ETIME. Note that it's not currently clear as to why this
happens as it's really hard to reproduce.
The rotation policy implement by kAFS, however, doesn't differentiate
between ETIME meaning we didn't get any response from the server and ETIME
meaning the call got cancelled mid-flow. The latter leads to an oops when
fetching data as the rotation partially resets the afs_read descriptor,
which can result in a cleared page pointer being dereferenced because that
page has already been filled.
Handle this by the following means:
(1) Set a flag on a call when we receive a packet for it.
(2) Store the highest packet serial number so far received for a call
(bearing in mind this may wrap).
(3) If, when the "not received anything recently" timeout expires on a
call, we've received at least one packet for a call and the connection
as a whole has received packets more recently than that call, then
cancel the call locally with ECONNRESET rather than ETIME.
This indicates that the call was definitely in progress on the server.
(4) In kAFS, if the rotation algorithm sees ECONNRESET rather than ETIME,
don't try the next server, but rather abort the call.
This avoids the oops as we don't try to reuse the afs_read struct.
Rather, as-yet ungotten pages will be reread at a later data.
Also:
(5) Add an rxrpc tracepoint to log detection of the call being reset.
Without this, I occasionally see an oops like the following:
general protection fault: 0000 [#1] SMP PTI
...
RIP: 0010:_copy_to_iter+0x204/0x310
RSP: 0018:ffff8800cae0f828 EFLAGS: 00010206
RAX: 0000000000000560 RBX: 0000000000000560 RCX: 0000000000000560
RDX: ffff8800cae0f968 RSI: ffff8800d58b3312 RDI: 0005080000000000
RBP: ffff8800cae0f968 R08: 0000000000000560 R09: ffff8800ca00f400
R10: ffff8800c36f28d4 R11: 00000000000008c4 R12: ffff8800cae0f958
R13: 0000000000000560 R14: ffff8800d58b3312 R15: 0000000000000560
FS: 00007fdaef108080(0000) GS:ffff8800ca680000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007fb28a8fa000 CR3: 00000000d2a76002 CR4: 00000000001606e0
Call Trace:
skb_copy_datagram_iter+0x14e/0x289
rxrpc_recvmsg_data.isra.0+0x6f3/0xf68
? trace_buffer_unlock_commit_regs+0x4f/0x89
rxrpc_kernel_recv_data+0x149/0x421
afs_extract_data+0x1e0/0x798
? afs_wait_for_call_to_complete+0xc9/0x52e
afs_deliver_fs_fetch_data+0x33a/0x5ab
afs_deliver_to_call+0x1ee/0x5e0
? afs_wait_for_call_to_complete+0xc9/0x52e
afs_wait_for_call_to_complete+0x12b/0x52e
? wake_up_q+0x54/0x54
afs_make_call+0x287/0x462
? afs_fs_fetch_data+0x3e6/0x3ed
? rcu_read_lock_sched_held+0x5d/0x63
afs_fs_fetch_data+0x3e6/0x3ed
afs_fetch_data+0xbb/0x14a
afs_readpages+0x317/0x40d
__do_page_cache_readahead+0x203/0x2ba
? ondemand_readahead+0x3a7/0x3c1
ondemand_readahead+0x3a7/0x3c1
generic_file_buffered_read+0x18b/0x62f
__vfs_read+0xdb/0xfe
vfs_read+0xb2/0x137
ksys_read+0x50/0x8c
do_syscall_64+0x7d/0x1a0
entry_SYSCALL_64_after_hwframe+0x49/0xbe
Note the weird value in RDI which is a result of trying to kmap() a NULL
page pointer.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If a volume location record lists multiple file servers for a volume, then
it's possible that due to a misconfiguration or a changing configuration
that one of the file servers doesn't know about it yet and will abort
VNOVOL. Currently, the rotation algorithm will stop with EREMOTEIO.
Fix this by moving on to try the next server if VNOVOL is returned. Once
all the servers have been tried and the record rechecked, the algorithm
will stop with EREMOTEIO or ENOMEDIUM.
Fixes: d2ddc776a4 ("afs: Overhaul volume and server record caching and fileserver rotation")
Reported-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
The server rotation algorithm just gives up if it fails to probe a
fileserver. Fix this by rotating to the next fileserver instead.
Fixes: d2ddc776a4 ("afs: Overhaul volume and server record caching and fileserver rotation")
Signed-off-by: David Howells <dhowells@redhat.com>
The refcounting on afs_cb_interest struct objects in
afs_register_server_cb_interest() is wrong as it uses the server list
entry's call back interest pointer without regard for the fact that it
might be replaced at any time and the object thrown away.
Fix this by:
(1) Put a lock on the afs_server_list struct that can be used to
mediate access to the callback interest pointers in the servers array.
(2) Keep a ref on the callback interest that we get from the entry.
(3) Dropping the old reference held by vnode->cb_interest if we replace
the pointer.
Fixes: c435ee3455 ("afs: Overhaul the callback handling")
Signed-off-by: David Howells <dhowells@redhat.com>
Fix warnings raised by checker, including:
(*) Warnings raised by unequal comparison for the purposes of sorting,
where the endianness doesn't matter:
fs/afs/addr_list.c:246:21: warning: restricted __be16 degrades to integer
fs/afs/addr_list.c:246:30: warning: restricted __be16 degrades to integer
fs/afs/addr_list.c:248:21: warning: restricted __be32 degrades to integer
fs/afs/addr_list.c:248:49: warning: restricted __be32 degrades to integer
fs/afs/addr_list.c:283:21: warning: restricted __be16 degrades to integer
fs/afs/addr_list.c:283:30: warning: restricted __be16 degrades to integer
(*) afs_set_cb_interest() is not actually used and can be removed.
(*) afs_cell_gc_delay() should be provided with a sysctl.
(*) afs_cell_destroy() needs to use rcu_access_pointer() to read
cell->vl_addrs.
(*) afs_init_fs_cursor() should be static.
(*) struct afs_vnode::permit_cache needs to be marked __rcu.
(*) afs_server_rcu() needs to use rcu_access_pointer().
(*) afs_destroy_server() should use rcu_access_pointer() on
server->addresses as the server object is no longer accessible.
(*) afs_find_server() casts __be16/__be32 values to int in order to
directly compare them for the purpose of finding a match in a list,
but is should also annotate the cast with __force to avoid checker
warnings.
(*) afs_check_permit() accesses vnode->permit_cache outside of the RCU
readlock, though it doesn't then access the value; the extraneous
access is deleted.
False positives:
(*) Conditional locking around the code in xdr_decode_AFSFetchStatus. This
can be dealt with in a separate patch.
fs/afs/fsclient.c:148:9: warning: context imbalance in 'xdr_decode_AFSFetchStatus' - different lock contexts for basic block
(*) Incorrect handling of seq-retry lock context balance:
fs/afs/inode.c:455:38: warning: context imbalance in 'afs_getattr' - different
lock contexts for basic block
fs/afs/server.c:52:17: warning: context imbalance in 'afs_find_server' - different lock contexts for basic block
fs/afs/server.c:128:17: warning: context imbalance in 'afs_find_server_by_uuid' - different lock contexts for basic block
Errors:
(*) afs_lookup_cell_rcu() needs to break out of the seq-retry loop, not go
round again if it successfully found the workstation cell.
(*) Fix UUID decode in afs_deliver_cb_probe_uuid().
(*) afs_cache_permit() has a missing rcu_read_unlock() before one of the
jumps to the someone_else_changed_it label. Move the unlock to after
the label.
(*) afs_vl_get_addrs_u() is using ntohl() rather than htonl() when
encoding to XDR.
(*) afs_deliver_yfsvl_get_endpoints() is using htonl() rather than ntohl()
when decoding from XDR.
Signed-off-by: David Howells <dhowells@redhat.com>
Rearrange afs_select_fileserver() a little to put the use_server chunk
before the next_server chunk so that with the removal of a couple of gotos
the main path through the function is all one sequence.
Signed-off-by: David Howells <dhowells@redhat.com>
In afs_select_fileserver(), we need to clear the ->responded flag in the
address list when reusing it. We should also clear it in
afs_select_current_fileserver().
To this end, just memset() the object before initialising it.
Fixes: d2ddc776a4 ("afs: Overhaul volume and server record caching and fileserver rotation")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: stable@vger.kernel.org
afs_select_fileserver() ends the address cursor it is using in the case in
which we get some sort of network error and run out of addresses to iterate
through, before it jumps to try the next server. This also needs to be
done when the server aborts with some sort of error that means we should
try the next server.
Fix this by:
(1) Move the iterate_address afs_end_cursor() call to the next_server
case.
(2) End the cursor in the failed case.
(3) Make afs_end_cursor() clear the ->begun flag and ->addr pointer in the
address cursor.
(4) Make afs_end_cursor() able to be called on an already cleared cursor.
Without this, something like the following oops may occur:
AFS: Assertion failed
18446612134397189888 == 0 is false
0xffff88007c279f00 == 0x0 is false
------------[ cut here ]------------
kernel BUG at fs/afs/rotate.c:360!
RIP: 0010:afs_select_fileserver+0x79b/0xa30 [kafs]
Call Trace:
afs_statfs+0xcc/0x180 [kafs]
? p9_client_statfs+0x9e/0x110 [9pnet]
? _cond_resched+0x19/0x40
statfs_by_dentry+0x6d/0x90
vfs_statfs+0x1b/0xc0
user_statfs+0x4b/0x80
SYSC_statfs+0x15/0x30
SyS_statfs+0xe/0x10
entry_SYSCALL_64_fastpath+0x20/0x83
Fixes: d2ddc776a4 ("afs: Overhaul volume and server record caching and fileserver rotation")
Reported-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: stable@vger.kernel.org
Fix the AFS file locking whereby the use of the big kernel lock (which
could be slept with) was replaced by a spinlock (which couldn't). The
problem is that the AFS code was doing stuff inside the critical section
that might call schedule(), so this is a broken transformation.
Fix this by the following means:
(1) Use a state machine with a proper state that can only be changed under
the spinlock rather than using a collection of bit flags.
(2) Cache the key used for the lock and the lock type in the afs_vnode
struct so that the manager work function doesn't have to refer to a
file_lock struct that's been dequeued. This makes signal handling
safer.
(4) Move the unlock from afs_do_unlk() to afs_fl_release_private() which
means that unlock is achieved in other circumstances too.
(5) Unlock the file on the server before taking the next conflicting lock.
Also change:
(1) Check the permits on a file before actually trying the lock.
(2) fsync the file before effecting an explicit unlock operation. We
don't fsync if the lock is erased otherwise as we might not be in a
context where we can actually do that.
Further fixes:
(1) Fixed-fileserver address rotation is made to work. It's only used by
the locking functions, so couldn't be tested before.
Fixes: 72f98e7255 ("locks: turn lock_flocks into a spinlock")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: jlayton@redhat.com
The current code assumes that volumes and servers are per-cell and are
never shared, but this is not enforced, and, indeed, public cells do exist
that are aliases of each other. Further, an organisation can, say, set up
a public cell and a private cell with overlapping, but not identical, sets
of servers. The difference is purely in the database attached to the VL
servers.
The current code will malfunction if it sees a server in two cells as it
assumes global address -> server record mappings and that each server is in
just one cell.
Further, each server may have multiple addresses - and may have addresses
of different families (IPv4 and IPv6, say).
To this end, the following structural changes are made:
(1) Server record management is overhauled:
(a) Server records are made independent of cell. The namespace keeps
track of them, volume records have lists of them and each vnode
has a server on which its callback interest currently resides.
(b) The cell record no longer keeps a list of servers known to be in
that cell.
(c) The server records are now kept in a flat list because there's no
single address to sort on.
(d) Server records are now keyed by their UUID within the namespace.
(e) The addresses for a server are obtained with the VL.GetAddrsU
rather than with VL.GetEntryByName, using the server's UUID as a
parameter.
(f) Cached server records are garbage collected after a period of
non-use and are counted out of existence before purging is allowed
to complete. This protects the work functions against rmmod.
(g) The servers list is now in /proc/fs/afs/servers.
(2) Volume record management is overhauled:
(a) An RCU-replaceable server list is introduced. This tracks both
servers and their coresponding callback interests.
(b) The superblock is now keyed on cell record and numeric volume ID.
(c) The volume record is now tied to the superblock which mounts it,
and is activated when mounted and deactivated when unmounted.
This makes it easier to handle the cache cookie without causing a
double-use in fscache.
(d) The volume record is loaded from the VLDB using VL.GetEntryByNameU
to get the server UUID list.
(e) The volume name is updated if it is seen to have changed when the
volume is updated (the update is keyed on the volume ID).
(3) The vlocation record is got rid of and VLDB records are no longer
cached. Sufficient information is stored in the volume record, though
an update to a volume record is now no longer shared between related
volumes (volumes come in bundles of three: R/W, R/O and backup).
and the following procedural changes are made:
(1) The fileserver cursor introduced previously is now fleshed out and
used to iterate over fileservers and their addresses.
(2) Volume status is checked during iteration, and the server list is
replaced if a change is detected.
(3) Server status is checked during iteration, and the address list is
replaced if a change is detected.
(4) The abort code is saved into the address list cursor and -ECONNABORTED
returned in afs_make_call() if a remote abort happened rather than
translating the abort into an error message. This allows actions to
be taken depending on the abort code more easily.
(a) If a VMOVED abort is seen then this is handled by rechecking the
volume and restarting the iteration.
(b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is
handled by sleeping for a short period and retrying and/or trying
other servers that might serve that volume. A message is also
displayed once until the condition has cleared.
(c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the
moment.
(d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to
see if it has been deleted; if not, the fileserver is probably
indicating that the volume couldn't be attached and needs
salvaging.
(e) If statfs() sees one of these aborts, it does not sleep, but
rather returns an error, so as not to block the umount program.
(5) The fileserver iteration functions in vnode.c are now merged into
their callers and more heavily macroised around the cursor. vnode.c
is removed.
(6) Operations on a particular vnode are serialised on that vnode because
the server will lock that vnode whilst it operates on it, so a second
op sent will just have to wait.
(7) Fileservers are probed with FS.GetCapabilities before being used.
This is where service upgrade will be done.
(8) A callback interest on a fileserver is set up before an FS operation
is performed and passed through to afs_make_call() so that it can be
set on the vnode if the operation returns a callback. The callback
interest is passed through to afs_iget() also so that it can be set
there too.
In general, record updating is done on an as-needed basis when we try to
access servers, volumes or vnodes rather than offloading it to work items
and special threads.
Notes:
(1) Pre AFS-3.4 servers are no longer supported, though this can be added
back if necessary (AFS-3.4 was released in 1998).
(2) VBUSY is retried forever for the moment at intervals of 1s.
(3) /proc/fs/afs/<cell>/servers no longer exists.
Signed-off-by: David Howells <dhowells@redhat.com>
David Howells