linux-stable.git/kernel/bpf/inode.c, branch v4.18.2

bpf: implement dummy fops for bpf objects

2018-06-08T17:58:48+00:00

syzkaller was able to trigger the following warning in
do_dentry_open():

  WARNING: CPU: 1 PID: 4508 at fs/open.c:778 do_dentry_open+0x4ad/0xe40 fs/open.c:778
  Kernel panic - not syncing: panic_on_warn set ...

  CPU: 1 PID: 4508 Comm: syz-executor867 Not tainted 4.17.0+ #90
  Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
  Call Trace:
  [...]
   vfs_open+0x139/0x230 fs/open.c:908
   do_last fs/namei.c:3370 [inline]
   path_openat+0x1717/0x4dc0 fs/namei.c:3511
   do_filp_open+0x249/0x350 fs/namei.c:3545
   do_sys_open+0x56f/0x740 fs/open.c:1101
   __do_sys_openat fs/open.c:1128 [inline]
   __se_sys_openat fs/open.c:1122 [inline]
   __x64_sys_openat+0x9d/0x100 fs/open.c:1122
   do_syscall_64+0x1b1/0x800 arch/x86/entry/common.c:287
   entry_SYSCALL_64_after_hwframe+0x49/0xbe

Problem was that prog and map inodes in bpf fs did not
implement a dummy file open operation that would return an
error. The patch in do_dentry_open() checks whether f_ops
are present and if not bails out with an error. While this
may be fine, we really shouldn't be throwing a warning
though. Thus follow the model similar to bad_file_ops and
reject the request unconditionally with -EIO.

Fixes: b2197755b263 ("bpf: add support for persistent maps/progs")
Reported-by: syzbot+2e7fcab0f56fdbb330b8@syzkaller.appspotmail.com
Signed-off-by: Daniel Borkmann 
Signed-off-by: Alexei Starovoitov

bpf: remove tracepoints from bpf core

2018-04-30T08:55:56+00:00

tracepoints to bpf core were added as a way to provide introspection
to bpf programs and maps, but after some time it became clear that
this approach is inadequate, so prog_id, map_id and corresponding
get_next_id, get_fd_by_id, get_info_by_fd, prog_query APIs were
introduced and fully adopted by bpftool and other applications.
The tracepoints in bpf core started to rot and causing syzbot warnings:
WARNING: CPU: 0 PID: 3008 at kernel/trace/trace_event_perf.c:274
Kernel panic - not syncing: panic_on_warn set ...
perf_trace_bpf_map_keyval+0x260/0xbd0 include/trace/events/bpf.h:228
trace_bpf_map_update_elem include/trace/events/bpf.h:274 [inline]
map_update_elem kernel/bpf/syscall.c:597 [inline]
SYSC_bpf kernel/bpf/syscall.c:1478 [inline]
Hence this patch deletes tracepoints in bpf core.

Reported-by: Eric Biggers 
Reported-by: syzbot 
Signed-off-by: Alexei Starovoitov 
Acked-by: David S. Miller 
Signed-off-by: Daniel Borkmann

bpf: btf: Add pretty print support to the basic arraymap

2018-04-19T19:46:25+00:00

This patch adds pretty print support to the basic arraymap.
Support for other bpf maps can be added later.

This patch adds new attrs to the BPF_MAP_CREATE command to allow
specifying the btf_fd, btf_key_id and btf_value_id.  The
BPF_MAP_CREATE can then associate the btf to the map if
the creating map supports BTF.

A BTF supported map needs to implement two new map ops,
map_seq_show_elem() and map_check_btf().  This patch has
implemented these new map ops for the basic arraymap.

It also adds file_operations, bpffs_map_fops, to the pinned
map such that the pinned map can be opened and read.
After that, the user has an intuitive way to do
"cat bpffs/pathto/a-pinned-map" instead of getting
an error.

bpffs_map_fops should not be extended further to support
other operations.  Other operations (e.g. write/key-lookup...)
should be realized by the userspace tools (e.g. bpftool) through
the BPF_OBJ_GET_INFO_BY_FD, map's lookup/update interface...etc.
Follow up patches will allow the userspace to obtain
the BTF from a map-fd.

Here is a sample output when reading a pinned arraymap
with the following map's value:

struct map_value {
	int count_a;
	int count_b;
};

cat /sys/fs/bpf/pinned_array_map:

0: {1,2}
1: {3,4}
2: {5,6}
...

Signed-off-by: Martin KaFai Lau 
Acked-by: Alexei Starovoitov 
Signed-off-by: Daniel Borkmann

bpf: comment why dots in filenames under BPF virtual FS are not allowed

2018-03-09T09:30:30+00:00

When pinning a file under the BPF virtual file system (traditionally
/sys/fs/bpf), using a dot in the name of the location to pin at is not
allowed. For example, trying to pin at "/sys/fs/bpf/foo.bar" will be
rejected with -EPERM.

This check was introduced at the same time as the BPF file system
itself, with commit b2197755b263 ("bpf: add support for persistent
maps/progs"). At this time, it was checked in a function called
"bpf_dname_reserved()", which made clear that using a dot was reserved
for future extensions.

This function disappeared and the check was moved elsewhere with commit
0c93b7d85d40 ("bpf: reject invalid names right in ->lookup()"), and the
meaning of the dot ban was lost.

The present commit simply adds a comment in the source to explain to the
reader that the usage of dots is reserved for future usage.

Signed-off-by: Quentin Monnet 
Signed-off-by: Daniel Borkmann

Merge branch 'work.mqueue' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

2018-01-31T02:32:21+00:00

Pull mqueue/bpf vfs cleanups from Al Viro:
 "mqueue and bpf go through rather painful and similar contortions to
  create objects in their dentry trees. Provide a primitive for doing
  that without abusing ->mknod(), switch bpf and mqueue to it.

  Another mqueue-related thing that has ended up in that branch is
  on-demand creation of internal mount (based upon the work of Giuseppe
  Scrivano)"

* 'work.mqueue' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  mqueue: switch to on-demand creation of internal mount
  tidy do_mq_open() up a bit
  mqueue: clean prepare_open() up
  do_mq_open(): move all work prior to dentry_open() into a helper
  mqueue: fold mq_attr_ok() into mqueue_get_inode()
  move dentry_open() calls up into do_mq_open()
  mqueue: switch to vfs_mkobj(), quit abusing ->d_fsdata
  bpf_obj_do_pin(): switch to vfs_mkobj(), quit abusing ->mknod()
  new primitive: vfs_mkobj()

bpf_obj_do_pin(): switch to vfs_mkobj(), quit abusing ->mknod()

2018-01-05T16:54:33+00:00

Signed-off-by: Al Viro

fix "netfilter: xt_bpf: Fix XT_BPF_MODE_FD_PINNED mode of 'xt_bpf_info_v1'"

2018-01-05T16:43:39+00:00

Descriptor table is a shared object; it's not a place where you can
stick temporary references to files, especially when we don't need
an opened file at all.

Cc: stable@vger.kernel.org # v4.14
Fixes: 98589a0998b8 ("netfilter: xt_bpf: Fix XT_BPF_MODE_FD_PINNED mode of 'xt_bpf_info_v1'")
Signed-off-by: Al Viro

bpf: Add file mode configuration into bpf maps

2017-10-20T12:32:59+00:00

Introduce the map read/write flags to the eBPF syscalls that returns the
map fd. The flags is used to set up the file mode when construct a new
file descriptor for bpf maps. To not break the backward capability, the
f_flags is set to O_RDWR if the flag passed by syscall is 0. Otherwise
it should be O_RDONLY or O_WRONLY. When the userspace want to modify or
read the map content, it will check the file mode to see if it is
allowed to make the change.

Signed-off-by: Chenbo Feng 
Acked-by: Alexei Starovoitov 
Acked-by: Daniel Borkmann 
Signed-off-by: David S. Miller

netfilter: xt_bpf: Fix XT_BPF_MODE_FD_PINNED mode of 'xt_bpf_info_v1'

2017-10-09T13:18:04+00:00

Commit 2c16d6033264 ("netfilter: xt_bpf: support ebpf") introduced
support for attaching an eBPF object by an fd, with the
'bpf_mt_check_v1' ABI expecting the '.fd' to be specified upon each
IPT_SO_SET_REPLACE call.

However this breaks subsequent iptables calls:

 # iptables -A INPUT -m bpf --object-pinned /sys/fs/bpf/xxx -j ACCEPT
 # iptables -A INPUT -s 5.6.7.8 -j ACCEPT
 iptables: Invalid argument. Run `dmesg' for more information.

That's because iptables works by loading existing rules using
IPT_SO_GET_ENTRIES to userspace, then issuing IPT_SO_SET_REPLACE with
the replacement set.

However, the loaded 'xt_bpf_info_v1' has an arbitrary '.fd' number
(from the initial "iptables -m bpf" invocation) - so when 2nd invocation
occurs, userspace passes a bogus fd number, which leads to
'bpf_mt_check_v1' to fail.

One suggested solution [1] was to hack iptables userspace, to perform a
"entries fixup" immediatley after IPT_SO_GET_ENTRIES, by opening a new,
process-local fd per every 'xt_bpf_info_v1' entry seen.

However, in [2] both Pablo Neira Ayuso and Willem de Bruijn suggested to
depricate the xt_bpf_info_v1 ABI dealing with pinned ebpf objects.

This fix changes the XT_BPF_MODE_FD_PINNED behavior to ignore the given
'.fd' and instead perform an in-kernel lookup for the bpf object given
the provided '.path'.

It also defines an alias for the XT_BPF_MODE_FD_PINNED mode, named
XT_BPF_MODE_PATH_PINNED, to better reflect the fact that the user is
expected to provide the path of the pinned object.

Existing XT_BPF_MODE_FD_ELF behavior (non-pinned fd mode) is preserved.

References: [1] https://marc.info/?l=netfilter-devel&m=150564724607440&w=2
            [2] https://marc.info/?l=netfilter-devel&m=150575727129880&w=2

Reported-by: Rafael Buchbinder 
Signed-off-by: Shmulik Ladkani 
Acked-by: Willem de Bruijn 
Acked-by: Daniel Borkmann 
Signed-off-by: Pablo Neira Ayuso

bpf: Implement show_options

2017-07-06T07:31:46+00:00

Implement the show_options superblock op for bpf as part of a bid to get
rid of s_options and generic_show_options() to make it easier to implement
a context-based mount where the mount options can be passed individually
over a file descriptor.

Signed-off-by: David Howells 
cc: Alexei Starovoitov 
cc: Daniel Borkmann 
cc: netdev@vger.kernel.org
Signed-off-by: Al Viro