[ Upstream commit a2f611a3dc317d8ea1c98ad6c54b911cf7f93193 ]
The AFS3 FID is three 32-bit unsigned numbers and is represented as three
up-to-8-hex-digit numbers separated by colons to the afs.fid xattr.
However, with the advent of support for YFS, the FID is now a 64-bit volume
number, a 96-bit vnode/inode number and a 32-bit uniquifier (as before).
Whilst the sprintf in afs_xattr_get_fid() has been partially updated (it
currently ignores the upper 32 bits of the 96-bit vnode number), the size
of the stack-based buffer has not been increased to match, thereby allowing
stack corruption to occur.
Fix this by increasing the buffer size appropriately and conditionally
including the upper part of the vnode number if it is non-zero. The latter
requires the lower part to be zero-padded if the upper part is non-zero.
Fixes: 3b6492df4153 ("afs: Increase to 64-bit volume ID and 96-bit vnode ID for YFS")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit eeba1e9cf31d064284dd1fa7bd6cfe01395bd03d ]
The in-kernel afs filesystem client counts the number of server-level
callback invalidation events (CB.InitCallBackState* RPC operations) that it
receives from the server. This is stored in cb_s_break in various
structures, including afs_server and afs_vnode.
If an inode is examined by afs_validate(), say, the afs_server copy is
compared, along with other break counters, to those in afs_vnode, and if
one or more of the counters do not match, it is considered that the
server's callback promise is broken. At points where this happens,
AFS_VNODE_CB_PROMISED is cleared to indicate that the status must be
refetched from the server.
afs_validate() issues an FS.FetchStatus operation to get updated metadata -
and based on the updated data_version may invalidate the pagecache too.
However, the break counters are also used to determine whether to note a
new callback in the vnode (which would set the AFS_VNODE_CB_PROMISED flag)
and whether to cache the permit data included in the YFSFetchStatus record
by the server.
The problem comes when the server sends us a CB.InitCallBackState op. The
first such instance doesn't cause cb_s_break to be incremented, but rather
causes AFS_SERVER_FL_NEW to be cleared - but thereafter, say some hours
after last use and all the volumes have been automatically unmounted and
the server has forgotten about the client[*], this *will* likely cause an
increment.
[*] There are other circumstances too, such as the server restarting or
needing to make space in its callback table.
Note that the server won't send us a CB.InitCallBackState op until we talk
to it again.
So what happens is:
(1) A mount for a new volume is attempted, a inode is created for the root
vnode and vnode->cb_s_break and AFS_VNODE_CB_PROMISED aren't set
immediately, as we don't have a nominated server to talk to yet - and
we may iterate through a few to find one.
(2) Before the operation happens, afs_fetch_status(), say, notes in the
cursor (fc.cb_break) the break counter sum from the vnode, volume and
server counters, but the server->cb_s_break is currently 0.
(3) We send FS.FetchStatus to the server. The server sends us back
CB.InitCallBackState. We increment server->cb_s_break.
(4) Our FS.FetchStatus completes. The reply includes a callback record.
(5) xdr_decode_AFSCallBack()/xdr_decode_YFSCallBack() check to see whether
the callback promise was broken by checking the break counter sum from
step (2) against the current sum.
This fails because of step (3), so we don't set the callback record
and, importantly, don't set AFS_VNODE_CB_PROMISED on the vnode.
This does not preclude the syscall from progressing, and we don't loop here
rechecking the status, but rather assume it's good enough for one round
only and will need to be rechecked next time.
(6) afs_validate() it triggered on the vnode, probably called from
d_revalidate() checking the parent directory.
(7) afs_validate() notes that AFS_VNODE_CB_PROMISED isn't set, so doesn't
update vnode->cb_s_break and assumes the vnode to be invalid.
(8) afs_validate() needs to calls afs_fetch_status(). Go back to step (2)
and repeat, every time the vnode is validated.
This primarily affects volume root dir vnodes. Everything subsequent to
those inherit an already incremented cb_s_break upon mounting.
The issue is that we assume that the callback record and the cached permit
information in a reply from the server can't be trusted after getting a
server break - but this is wrong since the server makes sure things are
done in the right order, holding up our ops if necessary[*].
[*] There is an extremely unlikely scenario where a reply from before the
CB.InitCallBackState could get its delivery deferred till after - at
which point we think we have a promise when we don't. This, however,
requires unlucky mass packet loss to one call.
AFS_SERVER_FL_NEW tries to paper over the cracks for the initial mount from
a server we've never contacted before, but this should be unnecessary.
It's also further insulated from the problem on an initial mount by
querying the server first with FS.GetCapabilities, which triggers the
CB.InitCallBackState.
Fix this by
(1) Remove AFS_SERVER_FL_NEW.
(2) In afs_calc_vnode_cb_break(), don't include cb_s_break in the
calculation.
(3) In afs_cb_is_broken(), don't include cb_s_break in the check.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 21bd68f196ca91fc0f3d9bd1b32f6e530e8c1c88 ] __pagevec_release() complains loudly if any page in the vector is still locked. The pages need to be locked for generic_error_remove_page(), but that function doesn't actually unlock them. Unlock the pages afterwards. Signed-off-by: Marc Dionne <marc.dionne@auristor.com> Signed-off-by: David Howells <dhowells@redhat.com> Tested-by: Jonathan Billings <jsbillin@umich.edu> Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8c7ae38d1ce12a0eaeba655df8562552b3596c7f ]
The marshalling of AFS.StoreData, AFS.StoreData64 and YFS.StoreData64 calls
generated by ->setattr() ops for the purpose of expanding a file is
incorrect due to older documentation incorrectly describing the way the RPC
'FileLength' parameter is meant to work.
The older documentation says that this is the length the file is meant to
end up at the end of the operation; however, it was never implemented this
way in any of the servers, but rather the file is truncated down to this
before the write operation is effected, and never expanded to it (and,
indeed, it was renamed to 'TruncPos' in 2014).
Fix this by setting the position parameter to the new file length and doing
a zero-lengh write there.
The bug causes Xwayland to SIGBUS due to unexpected non-expansion of a file
it then mmaps. This can be tested by giving the following test program a
filename in an AFS directory:
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/mman.h>
int main(int argc, char *argv[])
{
char *p;
int fd;
if (argc != 2) {
fprintf(stderr,
"Format: test-trunc-mmap <file>\n");
exit(2);
}
fd = open(argv[1], O_RDWR | O_CREAT | O_TRUNC);
if (fd < 0) {
perror(argv[1]);
exit(1);
}
if (ftruncate(fd, 0x140008) == -1) {
perror("ftruncate");
exit(1);
}
p = mmap(NULL, 4096, PROT_READ | PROT_WRITE,
MAP_SHARED, fd, 0);
if (p == MAP_FAILED) {
perror("mmap");
exit(1);
}
p[0] = 'a';
if (munmap(p, 4096) < 0) {
perror("munmap");
exit(1);
}
if (close(fd) < 0) {
perror("close");
exit(1);
}
exit(0);
}
Fixes: 31143d5d515e ("AFS: implement basic file write support")
Reported-by: Jonathan Billings <jsbillin@umich.edu>
Tested-by: Jonathan Billings <jsbillin@umich.edu>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin (Microsoft) <sashal@kernel.org>
When a cell with a volume location server list is added manually by
echoing the details into /proc/net/afs/cells, a record is added but the
flag saying it has been looked up isn't set.
This causes the VL server rotation code to wait forever, with the top of
/proc/pid/stack looking like:
afs_select_vlserver+0x3a6/0x6f3
afs_vl_lookup_vldb+0x4b/0x92
afs_create_volume+0x25/0x1b9
...
with the thread stuck in afs_start_vl_iteration() waiting for
AFS_CELL_FL_NO_LOOKUP_YET to be cleared.
Fix this by clearing AFS_CELL_FL_NO_LOOKUP_YET when setting up a record
if that record's details were supplied manually.
Fixes: 0a5143f2f89c ("afs: Implement VL server rotation")
Reported-by: Dave Botsch <dwb7@cornell.edu>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There's a race between afs_make_call() and afs_wake_up_async_call() in the
case that an error is returned from rxrpc_kernel_send_data() after it has
queued the final packet.
afs_make_call() will try and clean up the mess, but the call state may have
been moved on thereby causing afs_process_async_call() to also try and to
delete the call.
Fix this by:
(1) Getting an extra ref for an asynchronous call for the call itself to
hold. This makes sure the call doesn't evaporate on us accidentally
and will allow the call to be retained by the caller in a future
patch. The ref is released on leaving afs_make_call() or
afs_wait_for_call_to_complete().
(2) In the event of an error from rxrpc_kernel_send_data():
(a) Don't set the call state to AFS_CALL_COMPLETE until *after* the
call has been aborted and ended. This prevents
afs_deliver_to_call() from doing anything with any notifications
it gets.
(b) Explicitly end the call immediately to prevent further callbacks.
(c) Cancel any queued async_work and wait for the work if it's
executing. This allows us to be sure the race won't recur when we
change the state. We put the work queue's ref on the call if we
managed to cancel it.
(d) Put the call's ref that we got in (1). This belongs to us as long
as the call is in state AFS_CALL_CL_REQUESTING.
Fixes: 341f741f04be ("afs: Refcount the afs_call struct")
Signed-off-by: David Howells <dhowells@redhat.com>
Provide a function to get a reference on an afs_call struct. Signed-off-by: David Howells <dhowells@redhat.com>
Fix the refcounting of the authentication keys in the file locking code.
The vnode->lock_key member points to a key on which it expects to be
holding a ref, but it isn't always given an extra ref, however.
Fixes: 0fafdc9f888b ("afs: Fix file locking")
Signed-off-by: David Howells <dhowells@redhat.com>
A cb_interest record is not necessarily attached to the vnode on entry to
afs_validate(), which can cause an oops when we try to bring the vnode's
cb_s_break up to date in the default case (ie. no current callback promise
and the vnode has not been deleted).
Fix this by simply removing the line, as vnode->cb_s_break will be set when
needed by afs_register_server_cb_interest() when we next get a callback
promise from RPC call.
The oops looks something like:
BUG: unable to handle kernel NULL pointer dereference at 0000000000000018
...
RIP: 0010:afs_validate+0x66/0x250 [kafs]
...
Call Trace:
afs_d_revalidate+0x8d/0x340 [kafs]
? __d_lookup+0x61/0x150
lookup_dcache+0x44/0x70
? lookup_dcache+0x44/0x70
__lookup_hash+0x24/0xa0
do_unlinkat+0x11d/0x2c0
__x64_sys_unlink+0x23/0x30
do_syscall_64+0x4d/0xf0
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Fixes: ae3b7361dc0e ("afs: Fix validation/callback interaction")
Signed-off-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
A lock type of 0 is "LockRead", which makes the fileserver record an
unintentional read lock on the new file. This will cause problems
later on if the file is the subject of locking operations.
The correct default value should be -1 ("LockNone").
Fix the operation marshalling code to set the value and provide an enum to
symbolise the values whilst we're at it.
Fixes: 30062bd13e36 ("afs: Implement YFS support in the fs client")
Signed-off-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>