Brings it closer to more serious actions (adding branching
and allowing for late binding)
Unfortunately this breaks old syntax of the simple action.
But because simple is a pedagogical example unlikely to be used
in production environments (i.e its role is to serve as an example
on how to write actions), then this is ok.
New syntax for simple has new keyword "sdata". Example usage is:
sudo tc actions add action simple sdata "foobar" index 1
or
tc filter add dev $DEV parent ffff: protocol ip prio 1 u32\
match ip dst 17.0.0.1/32 flowid 1:10 action simple sdata "foobar"
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
This action allows for a sending side to encapsulate arbitrary metadata
which is decapsulated by the receiving end.
The sender runs in encoding mode and the receiver in decode mode.
Both sender and receiver must specify the same ethertype.
At some point we hope to have a registered ethertype and we'll
then provide a default so the user doesnt have to specify it.
For now we enforce the user specify it.
Described in netdev01 paper:
"Distributing Linux Traffic Control Classifier-Action Subsystem"
Authors: Jamal Hadi Salim and Damascene M. Joachimpillai
Also refer to IETF draft-ietf-forces-interfelfb-04.txt
Lets show example usage where we encode icmp from a sender towards
a receiver with an skbmark of 17; both sender and receiver use
ethertype of 0xdead to interop.
YYYY: Lets start with Receiver-side policy config:
xxx: add an ingress qdisc
sudo tc qdisc add dev $ETH ingress
xxx: any packets with ethertype 0xdead will be subjected to ife decoding
xxx: we then restart the classification so we can match on icmp at prio 3
sudo $TC filter add dev $ETH parent ffff: prio 2 protocol 0xdead \
u32 match u32 0 0 flowid 1:1 \
action ife decode reclassify
xxx: on restarting the classification from above if it was an icmp
xxx: packet, then match it here and continue to the next rule at prio 4
xxx: which will match based on skb mark of 17
sudo tc filter add dev $ETH parent ffff: prio 3 protocol ip \
u32 match ip protocol 1 0xff flowid 1:1 \
action continue
xxx: match on skbmark of 0x11 (decimal 17) and accept
sudo tc filter add dev $ETH parent ffff: prio 4 protocol ip \
handle 0x11 fw flowid 1:1 \
action ok
xxx: Lets show the decoding policy
sudo tc -s filter ls dev $ETH parent ffff: protocol 0xdead
xxx:
filter pref 2 u32
filter pref 2 u32 fh 800: ht divisor 1
filter pref 2 u32 fh 800::800 order 2048 key ht 800 bkt 0 flowid 1:1 (rule hit 0 success 0)
match 00000000/00000000 at 0 (success 0 )
action order 1: ife decode action reclassify type 0x0
allow mark allow prio
index 11 ref 1 bind 1 installed 45 sec used 45 sec
Action statistics:
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
xxx:
Observe that above lists all metadatum it can decode. Typically these
submodules will already be compiled into a monolithic kernel or
loaded as modules
YYYY: Lets show the sender side now ..
xxx: Add an egress qdisc on the sender netdev
sudo tc qdisc add dev $ETH root handle 1: prio
xxx:
xxx: Match all icmp packets to 192.168.122.237/24, then
xxx: tag the packet with skb mark of decimal 17, then
xxx: Encode it with:
xxx: ethertype 0xdead
xxx: add skb->mark to whitelist of metadatum to send
xxx: rewrite target dst MAC address to 02:15:15:15:15:15
xxx:
sudo $TC filter add dev $ETH parent 1: protocol ip prio 10 u32 \
match ip dst 192.168.122.237/24 \
match ip protocol 1 0xff \
flowid 1:2 \
action skbedit mark 17 \
action ife encode \
type 0xDEAD \
allow mark \
dst 02:15:15:15:15:15
xxx: Lets show the encoding policy
filter pref 10 u32
filter pref 10 u32 fh 800: ht divisor 1
filter pref 10 u32 fh 800::800 order 2048 key ht 800 bkt 0 flowid 1:2 (rule hit 118 success 0)
match c0a87a00/ffffff00 at 16 (success 0 )
match 00010000/00ff0000 at 8 (success 0 )
action order 1: skbedit mark 17
index 11 ref 1 bind 1 installed 3 sec used 3 sec
Action statistics:
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
action order 2: ife encode action pipe type 0xDEAD
allow mark dst 02:15:15:15:15:15
index 12 ref 1 bind 1 installed 3 sec used 3 sec
Action statistics:
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
xxx:
Now test by sending ping from sender to destination
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
We need limits.h for PATH_MAX, fixes:
tc_bpf.c: In function ‘bpf_map_selfcheck_pinned’:
tc_bpf.c:222:12: error: ‘PATH_MAX’ undeclared (first use in this
function)
char file[PATH_MAX], buff[4096];
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Follow-up to kernel commit 6c9059817432 ("bpf: pre-allocate hash map
elements"). Add flags support, so that we can pass in BPF_F_NO_PREALLOC
flag for disallowing preallocation. Update examples accordingly and also
remove the BPF_* map helper macros from them as they were not very useful.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Make it easier to spot issues when loading the object file fails. This
includes reporting in what pinned object specs differ, better indication
when we've reached instruction limits. Don't retry to load a non relo
program once we failed with bpf(2), and report out of bounds tail call key.
Also, add truncation of huge log outputs by default. Sometimes errors are
quite easy to spot by only looking at the tail of the verifier log, but
logs can get huge in size e.g. up to few MB (due to verifier checking all
possible program paths). Thus, by default limit output to the last 4096
bytes and indicate that it's truncated. For the full log, the verbose option
can be used.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
There is only a single user who needs it to be reentrant (not really,
but it's safer like this), add rt_addr_n2a_r() for it to use.
Signed-off-by: Phil Sutter <phil@nwl.cc>
There are only three users which require it to be reentrant, the rest is
fine without. Instead, provide a reentrant format_host_r() for users
which need it.
Signed-off-by: Phil Sutter <phil@nwl.cc>
As Jamal suggested, BRANCH is the wrong name, as these keywords go
beyond simple branch control - e.g. loops are possible, too. Therefore
rename the non-terminal to CONTROL instead which should be more
appropriate.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
The retain value was wrong for u16 and u8 types.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
This was tricky to get right:
- The 'stride' value used for 8 and 16 bit values must behave inverse to
the value's intra word offset to work correctly with big-endian data
act_pedit is editing.
- The 'm' array's values are in host byte order, so they have to be
converted as well (and the ordering was just inverse, for some
reason).
- The only sane way of getting this right is to manipulate value/mask in
host byte order and convert the output.
- TIPV4 (i.e. 'munge ip src/dst') had it's own pitfall: the address
parser converts to network byte order automatically. This patch fixes
this by converting it back before calling pack_key32, which is a hack
but at least does not require to implement a completely separate code
flow.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Break overlong function definitions and remove one extraneous
whitespace.
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Since the IP Header Length field is just half a byte, adjust retain to
only match these bits so the Version field is not overwritten by
accident.
The whole concept is actually broken due to dependency on endianness
which pedit ignores.
Signed-off-by: Phil Sutter <phil@nwl.cc>
This was horribly broken:
* pack_key8() and pack_key16() ...
* missed to invert retain value when applying it to the mask,
* did not sanitize val by ANDing it with retain,
* and ignored the mask which is necessary for 'invert' command.
* pack_key16() did not convert mask to network byte order.
* Changing the retain value for 'invert' or 'retain' operation seems
just plain wrong.
* While here, also got rid of unnecessary offset sanitization in
pack_key32().
* Simplify code a bit by always assigning the local mask variable to
tkey->mask before calling any of the pack_key*() variants.
Signed-off-by: Phil Sutter <phil@nwl.cc>
After lookup of the layered op submodule, pedit would pass argv and argc
including the layered op identifier at first position which confused the
submodule parser. Fix this by calling NEXT_ARG() before calling the
parse_peopt() callback.
Signed-off-by: Phil Sutter <phil@nwl.cc>
This seems to have been a hidden feature, though it's very useful and
necessary at least when combining multiple pedit actions.
Signed-off-by: Phil Sutter <phil@nwl.cc>
b3 buffer has been deleted previously so b2 is followed by b4
which is not consistent.
Signed-off-by: Dmitrii Shcherbakov <fw.dmitrii@yandex.com>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Phil Sutter <phil@nwl.cc>
Remove printing according to the previously used encoding of mpu and
overhead values within the tc_ratespec's mpu field. This encoding is
no longer being used as a separate 'overhead' field in the ratespec
structure has been introduced.
Signed-off-by: Dmitrii Shcherbakov <fw.dmitrii@yandex.com>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Phil Sutter <phil@nwl.cc>
Don't reimplement them and rather use the macros from the gelf header,
that is, GELF_ST_BIND()/GELF_ST_TYPE().
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Provide some more hints to the user/developer when relos have been found
that don't point to ld64 imm instruction. Ran couple of times into relos
generated by clang [1], where the compiler tried to uninline inlined
functions with eBPF and emitted BPF_JMP | BPF_CALL opcodes. If this seems
the case, give a hint that the user should do a work-around to use
always_inline annotation.
[1] https://llvm.org/bugs/show_bug.cgi?id=26243#c3
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
With a bit larger, branchy eBPF programs f.e. already ~BPF_MAXINSNS/7 in
size, it happens rather quickly that bpf(2) rejects also valid programs
when only the verifier log buffer size we have in tc is too small.
Change that, so by default we don't do any logging, and only in error
case we retry with logging enabled. If we should fail providing a
reasonable dump of the verifier analysis, retry few times with a larger
log buffer so that we can at least give the user a chance to debug the
program.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Commit 8f80d450c3 ("tc: fix compilation with old gcc (< 4.6)") was reverted
to ease the merge of the net-next branch.
Here is the new version.
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Add a test that symbol from relocation entry is actually related
to map section and bail out with an error message if it's not the
case; in relation to [1].
[1] https://llvm.org/bugs/show_bug.cgi?id=26243
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
eBPF llvm backend can support different BPF formats, make sure the object
we're trying to load matches with regards to endiannes and while at it, also
check for other attributes related to BPF ELFs.
# llc --version
LLVM (http://llvm.org/):
LLVM version 3.8.0svn
Optimized build.
Built Jan 9 2016 (02:08:10).
Default target: x86_64-unknown-linux-gnu
Host CPU: ivybridge
Registered Targets:
bpf - BPF (host endian)
bpfeb - BPF (big endian)
bpfel - BPF (little endian)
[...]
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
When extracting sections, we better check for name and type. Noticed
that some llvm versions emit .strtab and .shstrtab (e.g. saw it on pre
3.7), while more recent ones only seem to emit .strtab. Thus, make sure
we get the right sections.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Clean it up a bit, we can also get rid of some ugly ifdefs as in our case
TC_H_INGRESS is always defined.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
since all tc classifiers are required to specify ethertype as part of grammar
By not allowing eth_type to be specified we remove contradiction for
example when a user specifies:
tc filter add ... priority xxx protocol ip flower eth_type ipv6
This patch removes that contradiction
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
gcc < 4.6 does not handle C11 syntax for the static initialization of
anonymous struct/union, hence the following error:
tc_bpf.c:260: error: unknown field map_type specified in initializer
Signed-off-by: Julien Floret <julien.floret@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Fix a whitespace in bpf_dump_error() usage, and also a missing closing
bracket in ntohl() macro for eBPF programs.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Since we have all infrastructure in place now, allow atomic live updates
on program arrays. This can be very useful e.g. in case programs that are
being tail-called need to be replaced, f.e. when classifier functionality
needs to be changed, new protocols added/removed during runtime, etc.
Thus, provide a way for in-place code updates, minimal example: Given is
an object file cls.o that contains the entry point in section 'classifier',
has a globally pinned program array 'jmp' with 2 slots and id of 0, and
two tail called programs under section '0/0' (prog array key 0) and '0/1'
(prog array key 1), the section encoding for the loader is <id/key>.
Adding the filter loads everything into cls_bpf:
tc filter add dev foo parent ffff: bpf da obj cls.o
Now, the program under section '0/1' needs to be replaced with an updated
version that resides in the same section (also full path to tc's subfolder
of the mount point can be passed, e.g. /sys/fs/bpf/tc/globals/jmp):
tc exec bpf graft m:globals/jmp obj cls.o sec 0/1
In case the program resides under a different section 'foo', it can also
be injected into the program array like:
tc exec bpf graft m:globals/jmp key 1 obj cls.o sec foo
If the new tail called classifier program is already available as a pinned
object somewhere (here: /sys/fs/bpf/tc/progs/parser), it can be injected
into the prog array like:
tc exec bpf graft m:globals/jmp key 1 fd m:progs/parser
In the kernel, the program on key 1 is being atomically replaced and the
old one's refcount dropped.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
The recently introduced object pinning can be further extended in order
to allow sharing maps beyond tc namespace. F.e. maps that are being pinned
from tracing side, can be accessed through this facility as well.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Make use of the new show_fdinfo() facility and verify that when a
pinned map is being fetched that its basic attributes are the same
as the map we declared from the ELF file. I.e. when placed into the
globalns, collisions could occur. In such a case warn the user and
bail out.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Now that we have the possibility of sharing maps, it's time we get the
ELF loader fully working with regards to tail calls. Since program array
maps are pinned, we can keep them finally alive. I've noticed two bugs
that are being fixed in bpf_fill_prog_arrays() with this patch. Example
code comes as follow-up.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
This larger work addresses one of the bigger remaining issues on
tc's eBPF frontend, that is, to allow for persistent file descriptors.
Whenever tc parses the ELF object, extracts and loads maps into the
kernel, these file descriptors will be out of reach after the tc
instance exits.
Meaning, for simple (unnested) programs which contain one or
multiple maps, the kernel holds a reference, and they will live
on inside the kernel until the program holding them is unloaded,
but they will be out of reach for user space, even worse with
(also multiple nested) tail calls.
For this issue, we introduced the concept of an agent that can
receive the set of file descriptors from the tc instance creating
them, in order to be able to further inspect/update map data for
a specific use case. However, while that is more tied towards
specific applications, it still doesn't easily allow for sharing
maps accross multiple tc instances and would require a daemon to
be running in the background. F.e. when a map should be shared by
two eBPF programs, one attached to ingress, one to egress, this
currently doesn't work with the tc frontend.
This work solves exactly that, i.e. if requested, maps can now be
_arbitrarily_ shared between object files (PIN_GLOBAL_NS) or within
a single object (but various program sections, PIN_OBJECT_NS) without
"loosing" the file descriptor set. To make that happen, we use eBPF
object pinning introduced in kernel commit b2197755b263 ("bpf: add
support for persistent maps/progs") for exactly this purpose.
The shipped examples/bpf/bpf_shared.c code from this patch can be
easily applied, for instance, as:
- classifier-classifier shared:
tc filter add dev foo parent 1: bpf obj shared.o sec egress
tc filter add dev foo parent ffff: bpf obj shared.o sec ingress
- classifier-action shared (here: late binding to a dummy classifier):
tc actions add action bpf obj shared.o sec egress pass index 42
tc filter add dev foo parent ffff: bpf obj shared.o sec ingress
tc filter add dev foo parent 1: bpf bytecode '1,6 0 0 4294967295,' \
action bpf index 42
The toy example increments a shared counter on egress and dumps its
value on ingress (if no sharing (PIN_NONE) would have been chosen,
map value is 0, of course, due to the two map instances being created):
[...]
<idle>-0 [002] ..s. 38264.788234: : map val: 4
<idle>-0 [002] ..s. 38264.788919: : map val: 4
<idle>-0 [002] ..s. 38264.789599: : map val: 5
[...]
... thus if both sections reference the pinned map(s) in question,
tc will take care of fetching the appropriate file descriptor.
The patch has been tested extensively on both, classifier and
action sides.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Jamal Hadi Salim
Gustavo Zacarias
Daniel Borkmann
Jiri Pirko
Stephen Hemminger
Phil Sutter
Luca Lemmo
Dmitrii Shcherbakov
Nicolas Dichtel
Stephen Hemminger
Julien Floret