This parse_opts will set invalid opts.rfd/wfd in case of failure which
we already check, but it is not clear for readers that parse_opts error
are handled in p9_fd_create: clarify this by explicitely checking the
return value.
Link: https://lkml.kernel.org/r/20220921210921.1654735-1-floridsleeves@gmail.com
Signed-off-by: Li Zhong <floridsleeves@gmail.com>
[Dominique: reworded commit message to clarify this is NOOP]
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
syzbot is reporting hung task at p9_fd_close() [1], for p9_mux_poll_stop()
from p9_conn_destroy() from p9_fd_close() is failing to interrupt already
started kernel_read() from p9_fd_read() from p9_read_work() and/or
kernel_write() from p9_fd_write() from p9_write_work() requests.
Since p9_socket_open() sets O_NONBLOCK flag, p9_mux_poll_stop() does not
need to interrupt kernel_read()/kernel_write(). However, since p9_fd_open()
does not set O_NONBLOCK flag, but pipe blocks unless signal is pending,
p9_mux_poll_stop() needs to interrupt kernel_read()/kernel_write() when
the file descriptor refers to a pipe. In other words, pipe file descriptor
needs to be handled as if socket file descriptor.
We somehow need to interrupt kernel_read()/kernel_write() on pipes.
A minimal change, which this patch is doing, is to set O_NONBLOCK flag
from p9_fd_open(), for O_NONBLOCK flag does not affect reading/writing
of regular files. But this approach changes O_NONBLOCK flag on userspace-
supplied file descriptors (which might break userspace programs), and
O_NONBLOCK flag could be changed by userspace. It would be possible to set
O_NONBLOCK flag every time p9_fd_read()/p9_fd_write() is invoked, but still
remains small race window for clearing O_NONBLOCK flag.
If we don't want to manipulate O_NONBLOCK flag, we might be able to
surround kernel_read()/kernel_write() with set_thread_flag(TIF_SIGPENDING)
and recalc_sigpending(). Since p9_read_work()/p9_write_work() works are
processed by kernel threads which process global system_wq workqueue,
signals could not be delivered from remote threads when p9_mux_poll_stop()
from p9_conn_destroy() from p9_fd_close() is called. Therefore, calling
set_thread_flag(TIF_SIGPENDING)/recalc_sigpending() every time would be
needed if we count on signals for making kernel_read()/kernel_write()
non-blocking.
Link: https://lkml.kernel.org/r/345de429-a88b-7097-d177-adecf9fed342@I-love.SAKURA.ne.jp
Link: https://syzkaller.appspot.com/bug?extid=8b41a1365f1106fd0f33 [1]
Reported-by: syzbot <syzbot+8b41a1365f1106fd0f33@syzkaller.appspotmail.com>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Tested-by: syzbot <syzbot+8b41a1365f1106fd0f33@syzkaller.appspotmail.com>
Reviewed-by: Christian Schoenebeck <linux_oss@crudebyte.com>
[Dominique: add comment at Christian's suggestion]
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
So far 'msize' was simply used for all 9p message types, which is far
too much and slowed down performance tremendously with large values
for user configurable 'msize' option.
Let's stop this waste by using the new p9_msg_buf_size() function for
allocating more appropriate, smaller buffers according to what is
actually sent over the wire.
Only exception: RDMA transport is currently excluded from this message
size optimization - for its response buffers that is - as RDMA transport
would not cope with it, due to its response buffers being pulled from a
shared pool. [1]
Link: https://lore.kernel.org/all/Ys3jjg52EIyITPua@codewreck.org/ [1]
Link: https://lkml.kernel.org/r/3f51590535dc96ed0a165b8218c57639cfa5c36c.1657920926.git.linux_oss@crudebyte.com
Signed-off-by: Christian Schoenebeck <linux_oss@crudebyte.com>
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
This new function calculates a buffer size suitable for holding the
intended 9p request or response. For rather small message types (which
applies to almost all 9p message types actually) simply use hard coded
values. For some variable-length and potentially large message types
calculate a more precise value according to what data is actually
transmitted to avoid unnecessarily huge buffers.
So p9_msg_buf_size() divides the individual 9p message types into 3
message size categories:
- dynamically calculated message size (i.e. potentially large)
- 8k hard coded message size
- 4k hard coded message size
As for the latter two hard coded message types: for most 9p message
types it is pretty obvious whether they would always fit into 4k or
8k. But for some of them it depends on the maximum directory entry
name length allowed by OS and filesystem for determining into which
of the two size categories they would fit into. Currently Linux
supports directory entry names up to NAME_MAX (255), however when
comparing the limitation of individual filesystems, ReiserFS
theoretically supports up to slightly below 4k long names. So in
order to make this code more future proof, and as revisiting it
later on is a bit tedious and has the potential to miss out details,
the decision [1] was made to take 4k as basis as for max. name length.
Link: https://lkml.kernel.org/r/bd6be891cf67e867688e8c8796d06408bfafa0d9.1657920926.git.linux_oss@crudebyte.com
Link: https://lore.kernel.org/all/5564296.oo812IJUPE@silver/ [1]
Signed-off-by: Christian Schoenebeck <linux_oss@crudebyte.com>
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
Pull more iov_iter updates from Al Viro:
- more new_sync_{read,write}() speedups - ITER_UBUF introduction
- ITER_PIPE cleanups
- unification of iov_iter_get_pages/iov_iter_get_pages_alloc and
switching them to advancing semantics
- making ITER_PIPE take high-order pages without splitting them
- handling copy_page_from_iter() for high-order pages properly
* tag 'pull-work.iov_iter-rebased' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (32 commits)
fix copy_page_from_iter() for compound destinations
hugetlbfs: copy_page_to_iter() can deal with compound pages
copy_page_to_iter(): don't split high-order page in case of ITER_PIPE
expand those iov_iter_advance()...
pipe_get_pages(): switch to append_pipe()
get rid of non-advancing variants
ceph: switch the last caller of iov_iter_get_pages_alloc()
9p: convert to advancing variant of iov_iter_get_pages_alloc()
af_alg_make_sg(): switch to advancing variant of iov_iter_get_pages()
iter_to_pipe(): switch to advancing variant of iov_iter_get_pages()
block: convert to advancing variants of iov_iter_get_pages{,_alloc}()
iov_iter: advancing variants of iov_iter_get_pages{,_alloc}()
iov_iter: saner helper for page array allocation
fold __pipe_get_pages() into pipe_get_pages()
ITER_XARRAY: don't open-code DIV_ROUND_UP()
unify the rest of iov_iter_get_pages()/iov_iter_get_pages_alloc() guts
unify xarray_get_pages() and xarray_get_pages_alloc()
unify pipe_get_pages() and pipe_get_pages_alloc()
iov_iter_get_pages(): sanity-check arguments
iov_iter_get_pages_alloc(): lift freeing pages array on failure exits into wrapper
...
Pull 9p updates from Dominique Martinet:
- a couple of fixes
- add a tracepoint for fid refcounting
- some cleanup/followup on fid lookup
- some cleanup around req refcounting
* tag '9p-for-5.20' of https://github.com/martinetd/linux:
net/9p: Initialize the iounit field during fid creation
net: 9p: fix refcount leak in p9_read_work() error handling
9p: roll p9_tag_remove into p9_req_put
9p: Add client parameter to p9_req_put()
9p: Drop kref usage
9p: Fix some kernel-doc comments
9p fid refcount: cleanup p9_fid_put calls
9p fid refcount: add a 9p_fid_ref tracepoint
9p fid refcount: add p9_fid_get/put wrappers
9p: Fix minor typo in code comment
9p: Remove unnecessary variable for old fids while walking from d_parent
9p: Make the path walk logic more clear about when cloning is required
9p: Track the root fid with its own variable during lookups
Ensure that the fid's iounit field is set to zero when a new fid is
created. Certain 9P operations, such as OPEN and CREATE, allow the
server to reply with an iounit size which the client code assigns to the
p9_fid struct shortly after the fid is created by p9_fid_create(). On
the other hand, an XATTRWALK operation doesn't allow for the server to
specify an iounit value. The iounit field of the newly allocated p9_fid
struct remained uninitialized in that case. Depending on allocation
patterns, the iounit value could have been something reasonable that was
carried over from previously freed fids or, in the worst case, could
have been arbitrary values from non-fid related usages of the memory
location.
The bug was detected in the Windows Subsystem for Linux 2 (WSL2) kernel
after the uninitialized iounit field resulted in the typical sequence of
two getxattr(2) syscalls, one to get the size of an xattr and another
after allocating a sufficiently sized buffer to fit the xattr value, to
hit an unexpected ERANGE error in the second call to getxattr(2). An
uninitialized iounit field would sometimes force rsize to be smaller
than the xattr value size in p9_client_read_once() and the 9P server in
WSL refused to chunk up the READ on the attr_fid and, instead, returned
ERANGE to the client. The virtfs server in QEMU seems happy to chunk up
the READ and this problem goes undetected there.
Link: https://lkml.kernel.org/r/20220710141402.803295-1-tyhicks@linux.microsoft.com
Fixes: ebf46264a0 ("fs/9p: Add support user. xattr")
Cc: stable@vger.kernel.org
Signed-off-by: Tyler Hicks <tyhicks@linux.microsoft.com>
Reviewed-by: Christian Schoenebeck <linux_oss@crudebyte.com>
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
mempool prep commit removed the awkward kref usage which didn't
allow passing client pointer easily with the ref, so we no longer
need a separate function to remove the tag from idr.
This has the side benefit that it should be more robust in detecting
leaks: umount will now properly catch unfreed requests as they still
will be in the idr until the last ref is dropped
Link: https://lkml.kernel.org/r/20220712060801.2487140-1-asmadeus@codewreck.org
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
Reviewed-by: Christian Schoenebeck <linux_oss@crudebyte.com>
This adds a tracepoint event for 9p fid lifecycle tracing: when a fid
is created, its reference count increased/decreased, and freed.
The new 9p_fid_ref tracepoint should help anyone wishing to debug any
fid problem such as missing clunk (destroy) or use-after-free.
Link: https://lkml.kernel.org/r/20220612085330.1451496-6-asmadeus@codewreck.org
Signed-off-by: Dominique Martinet <asmadeus@codewreck.org>
p9_client_zc_rpc()/p9_check_zc_errors() are playing fast
and loose with copy_from_iter_full().
Reading from file is done by sending Tread request. Response
consists of fixed-sized header (including the amount of data actually
read) followed by the data itself.
For zero-copy case we arrange the things so that the first
11 bytes of reply go into the fixed-sized buffer, with the rest going
straight into the pages we want to read into.
What makes the things inconvenient is that sglist describing
what should go where has to be set *before* the reply arrives. As
the result, if reply is an error, the things get interesting. On success
we get
size[4] Rread tag[2] count[4] data[count]
For error layout varies depending upon the protocol variant -
in original 9P and 9P2000 it's
size[4] Rerror tag[2] len[2] error[len]
in 9P2000.U
size[4] Rerror tag[2] len[2] error[len] errno[4]
in 9P2000.L
size[4] Rlerror tag[2] errno[4]
The last case is nice and simple - we have an 11-byte response
that fits into the fixed-sized buffer we hoped to get an Rread into.
In other two, though, we get a variable-length string spill into the
pages we'd prepared for the data to be read.
Had that been in fixed-sized buffer (which is actually 4K),
we would've dealt with that the same way we handle non-zerocopy case.
However, for zerocopy it doesn't end up there, so we need to copy it
from those pages.
The trouble is, by the time we get around to that, the
references to pages in question are already dropped. As the result,
p9_zc_check_errors() tries to get the data using copy_from_iter_full().
Unfortunately, the iov_iter it's trying to read from might *NOT* be
capable of that. It is, after all, a data destination, not data source.
In particular, if it's an ITER_PIPE one, copy_from_iter_full() will
simply fail.
In ->zc_request() itself we do have those pages and dealing with
the problem in there would be a simple matter of memcpy_from_page()
into the fixed-sized buffer. Moreover, it isn't hard to recognize
the (rare) case when such copying is needed. That way we get rid of
p9_zc_check_errors() entirely - p9_check_errors() can be used instead
both for zero-copy and non-zero-copy cases.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Instead of a virtual kernel address use a pointer of the associated
struct page as second parameter of gnttab_end_foreign_access().
Most users have that pointer available already and are creating the
virtual address from it, risking problems in case the memory is
located in highmem.
gnttab_end_foreign_access() itself won't need to get the struct page
from the address again.
Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Instead of __get_free_pages() and free_pages() use alloc_pages_exact()
and free_pages_exact(). This is in preparation of a change of
gnttab_end_foreign_access() which will prohibit use of high-order
pages.
By using the local variable "order" instead of ring->intf->ring_order
in the error path of xen_9pfs_front_alloc_dataring() another bug is
fixed, as the error path can be entered before ring->intf->ring_order
is being set.
By using alloc_pages_exact() the size in bytes is specified for the
allocation, which fixes another bug for the case of
order < (PAGE_SHIFT - XEN_PAGE_SHIFT).
This is part of CVE-2022-23041 / XSA-396.
Reported-by: Simon Gaiser <simon@invisiblethingslab.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
V4:
- new patch
Pull virtio updates from Michael Tsirkin:
"virtio,vdpa,qemu_fw_cfg: features, cleanups, and fixes.
- partial support for < MAX_ORDER - 1 granularity for virtio-mem
- driver_override for vdpa
- sysfs ABI documentation for vdpa
- multiqueue config support for mlx5 vdpa
- and misc fixes, cleanups"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (42 commits)
vdpa/mlx5: Fix tracking of current number of VQs
vdpa/mlx5: Fix is_index_valid() to refer to features
vdpa: Protect vdpa reset with cf_mutex
vdpa: Avoid taking cf_mutex lock on get status
vdpa/vdpa_sim_net: Report max device capabilities
vdpa: Use BIT_ULL for bit operations
vdpa/vdpa_sim: Configure max supported virtqueues
vdpa/mlx5: Report max device capabilities
vdpa: Support reporting max device capabilities
vdpa/mlx5: Restore cur_num_vqs in case of failure in change_num_qps()
vdpa: Add support for returning device configuration information
vdpa/mlx5: Support configuring max data virtqueue
vdpa/mlx5: Fix config_attr_mask assignment
vdpa: Allow to configure max data virtqueues
vdpa: Read device configuration only if FEATURES_OK
vdpa: Sync calls set/get config/status with cf_mutex
vdpa/mlx5: Distribute RX virtqueues in RQT object
vdpa: Provide interface to read driver features
vdpa: clean up get_config_size ret value handling
virtio_ring: mark ring unused on error
...
This will enable cleanups down the road.
The idea is to disable cbs, then add "flush_queued_cbs" callback
as a parameter, this way drivers can flush any work
queued after callbacks have been disabled.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Link: https://lore.kernel.org/r/20211013105226.20225-1-mst@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Li Zhong
Xiu Jianfeng
Dominique Martinet
Tetsuo Handa
Christian Schoenebeck
Linus Torvalds
Al Viro
Tyler Hicks
Hangyu Hua
Kent Overstreet
Juergen Gross
Michael S. Tsirkin