This adds the BTF specification and UAPI bits for supporting BTF Var
and DataSec kinds. This is following LLVM upstream commit ac4082b77e07
("[BPF] Add BTF Var and DataSec Support") which has been merged recently.
Var itself is for describing a global variable and DataSec to describe
ELF sections e.g. data/bss/rodata sections that hold one or multiple
global variables.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Trivial addition to allow '.' aside from '_' as "special" characters
in the object name. Used to allow for substrings in maps from loader
side such as ".bss", ".data", ".rodata", but could also be useful for
other purposes.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

This patch adds a new BPF_MAP_FREEZE command which allows to
"freeze" the map globally as read-only / immutable from syscall
side.

Map permission handling has been refactored into map_get_sys_perms()
and drops FMODE_CAN_WRITE in case of locked map. Main use case is
to allow for setting up .rodata sections from the BPF ELF which
are loaded into the kernel, meaning BPF loader first allocates
map, sets up map value by copying .rodata section into it and once
complete, it calls BPF_MAP_FREEZE on the map fd to prevent further
modifications.

Right now BPF_MAP_FREEZE only takes map fd as argument while remaining
bpf_attr members are required to be zero. I didn't add write-only
locking here as counterpart since I don't have a concrete use-case
for it on my side, and I think it makes probably more sense to wait
once there is actually one. In that case bpf_attr can be extended
as usual with a flag field and/or others where flag 0 means that
we lock the map read-only hence this doesn't prevent to add further
extensions to BPF_MAP_FREEZE upon need.

A map creation flag like BPF_F_WRONCE was not considered for couple
of reasons: i) in case of a generic implementation, a map can consist
of more than just one element, thus there could be multiple map
updates needed to set the map into a state where it can then be
made immutable, ii) WRONCE indicates exact one-time write before
it is then set immutable. A generic implementation would set a bit
atomically on map update entry (if unset), indicating that every
subsequent update from then onwards will need to bail out there.
However, map updates can fail, so upon failure that flag would need
to be unset again and the update attempt would need to be repeated
for it to be eventually made immutable. While this can be made
race-free, this approach feels less clean and in combination with
reason i), it's not generic enough. A dedicated BPF_MAP_FREEZE
command directly sets the flag and caller has the guarantee that
map is immutable from syscall side upon successful return for any
future syscall invocations that would alter the map state, which
is also more intuitive from an API point of view. A command name
such as BPF_MAP_LOCK has been avoided as it's too close with BPF
map spin locks (which already has BPF_F_LOCK flag). BPF_MAP_FREEZE
is so far only enabled for privileged users.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

This work adds two new map creation flags BPF_F_RDONLY_PROG
and BPF_F_WRONLY_PROG in order to allow for read-only or
write-only BPF maps from a BPF program side.

Today we have BPF_F_RDONLY and BPF_F_WRONLY, but this only
applies to system call side, meaning the BPF program has full
read/write access to the map as usual while bpf(2) calls with
map fd can either only read or write into the map depending
on the flags. BPF_F_RDONLY_PROG and BPF_F_WRONLY_PROG allows
for the exact opposite such that verifier is going to reject
program loads if write into a read-only map or a read into a
write-only map is detected. For read-only map case also some
helpers are forbidden for programs that would alter the map
state such as map deletion, update, etc. As opposed to the two
BPF_F_RDONLY / BPF_F_WRONLY flags, BPF_F_RDONLY_PROG as well
as BPF_F_WRONLY_PROG really do correspond to the map lifetime.

We've enabled this generic map extension to various non-special
maps holding normal user data: array, hash, lru, lpm, local
storage, queue and stack. Further generic map types could be
followed up in future depending on use-case. Main use case
here is to forbid writes into .rodata map values from verifier
side.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Both BPF_F_WRONLY / BPF_F_RDONLY flags are tied to the map file
descriptor, but not to the map object itself! Meaning, at map
creation time BPF_F_RDONLY can be set to make the map read-only
from syscall side, but this holds only for the returned fd, so
any other fd either retrieved via bpf file system or via map id
for the very same underlying map object can have read-write access
instead.

Given that, keeping the two flags around in the map_flags attribute
and exposing them to user space upon map dump is misleading and
may lead to false conclusions. Since these two flags are not
tied to the map object lets also not store them as map property.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

This generic extension to BPF maps allows for directly loading
an address residing inside a BPF map value as a single BPF
ldimm64 instruction!

The idea is similar to what BPF_PSEUDO_MAP_FD does today, which
is a special src_reg flag for ldimm64 instruction that indicates
that inside the first part of the double insns's imm field is a
file descriptor which the verifier then replaces as a full 64bit
address of the map into both imm parts. For the newly added
BPF_PSEUDO_MAP_VALUE src_reg flag, the idea is the following:
the first part of the double insns's imm field is again a file
descriptor corresponding to the map, and the second part of the
imm field is an offset into the value. The verifier will then
replace both imm parts with an address that points into the BPF
map value at the given value offset for maps that support this
operation. Currently supported is array map with single entry.
It is possible to support more than just single map element by
reusing both 16bit off fields of the insns as a map index, so
full array map lookup could be expressed that way. It hasn't
been implemented here due to lack of concrete use case, but
could easily be done so in future in a compatible way, since
both off fields right now have to be 0 and would correctly
denote a map index 0.

The BPF_PSEUDO_MAP_VALUE is a distinct flag as otherwise with
BPF_PSEUDO_MAP_FD we could not differ offset 0 between load of
map pointer versus load of map's value at offset 0, and changing
BPF_PSEUDO_MAP_FD's encoding into off by one to differ between
regular map pointer and map value pointer would add unnecessary
complexity and increases barrier for debugability thus less
suitable. Using the second part of the imm field as an offset
into the value does /not/ come with limitations since maximum
possible value size is in u32 universe anyway.

This optimization allows for efficiently retrieving an address
to a map value memory area without having to issue a helper call
which needs to prepare registers according to calling convention,
etc, without needing the extra NULL test, and without having to
add the offset in an additional instruction to the value base
pointer. The verifier then treats the destination register as
PTR_TO_MAP_VALUE with constant reg->off from the user passed
offset from the second imm field, and guarantees that this is
within bounds of the map value. Any subsequent operations are
normally treated as typical map value handling without anything
extra needed from verification side.

The two map operations for direct value access have been added to
array map for now. In future other types could be supported as
well depending on the use case. The main use case for this commit
is to allow for BPF loader support for global variables that
reside in .data/.rodata/.bss sections such that we can directly
load the address of them with minimal additional infrastructure
required. Loader support has been added in subsequent commits for
libbpf library.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

vsprintf() in __base_pr() uses nonliteral format string and it breaks
compilation for those who provide corresponding extra CFLAGS, e.g.:
https://github.com/libbpf/libbpf/issues/27

If libbpf is built with the flags from PR:

  libbpf.c:68:26: error: format string is not a string literal
  [-Werror,-Wformat-nonliteral]
          return vfprintf(stderr, format, args);
                                  ^~~~~~
  1 error generated.

Ignore this warning since the use case in libbpf.c is legit.
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Andrey Ignatov says:

====================
v2->v3:
- sanity check max value for variable offset.

v1->v2:
- rely on meta = NULL to reject var_off stack access to uninit buffer.

This patch set is a follow-up for discussion [1].

It fixes variable offset stack access handling for raw and unprivileged
mode, rejecting both of them, and sanity checks max variable offset value.

Patch 1 handles raw (uninitialized) mode.
Patch 2 adds test for raw mode.
Patch 3 handles unprivileged mode.
Patch 4 adds test for unprivileged mode.
Patch 5 adds sanity check for max value of variable offset.
Patch 6 adds test for variable offset max value checking.
Patch 7 is a minor fix in verbose log.

Unprivileged mode is an interesting case since one (and only?) way to come
up with variable offset is to use pointer arithmetics. Though pointer
arithmetics is already prohibited for unprivileged mode. I'm not sure if
it's enough though and it seems like a good idea to still reject variable
offset for unpriv in check_stack_boundary(). Please see patches 3 and 4
for more details on this.

[1] https://marc.info/?l=linux-netdev&m=155419526427742&w=2
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

check_stack_access() that prints verbose log is used in
adjust_ptr_min_max_vals() that prints its own verbose log and now they
stick together, e.g.:

  variable stack access var_off=(0xfffffffffffffff0; 0x4) off=-16
  size=1R2 stack pointer arithmetic goes out of range, prohibited for
  !root

Add missing newline so that log is more readable:
  variable stack access var_off=(0xfffffffffffffff0; 0x4) off=-16 size=1
  R2 stack pointer arithmetic goes out of range, prohibited for !root

Fixes: f1174f77 ("bpf/verifier: rework value tracking")
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Test the case when reg->smax_value is too small/big and can overflow,
and separately min and max values outside of stack bounds.

Example of output:
  # ./test_verifier
  #856/p indirect variable-offset stack access, unbounded OK
  #857/p indirect variable-offset stack access, max out of bound OK
  #858/p indirect variable-offset stack access, min out of bound OK
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

As discussed in [1] max value of variable offset has to be checked for
overflow on stack access otherwise verifier would accept code like this:

  0: (b7) r2 = 6
  1: (b7) r3 = 28
  2: (7a) *(u64 *)(r10 -16) = 0
  3: (7a) *(u64 *)(r10 -8) = 0
  4: (79) r4 = *(u64 *)(r1 +168)
  5: (c5) if r4 s< 0x0 goto pc+4
   R1=ctx(id=0,off=0,imm=0) R2=inv6 R3=inv28
   R4=inv(id=0,umax_value=9223372036854775807,var_off=(0x0;
   0x7fffffffffffffff)) R10=fp0,call_-1 fp-8=mmmmmmmm fp-16=mmmmmmmm
  6: (17) r4 -= 16
  7: (0f) r4 += r10
  8: (b7) r5 = 8
  9: (85) call bpf_getsockopt#57
  10: (b7) r0 = 0
  11: (95) exit

, where R4 obviosly has unbounded max value.

Fix it by checking that reg->smax_value is inside (-BPF_MAX_VAR_OFF;
BPF_MAX_VAR_OFF) range.

reg->smax_value is used instead of reg->umax_value because stack
pointers are calculated using negative offset from fp. This is opposite
to e.g. map access where offset must be non-negative and where
umax_value is used.

Also dedicated verbose logs are added for both min and max bound check
failures to have diagnostics consistent with variable offset handling in
check_map_access().

[1] https://marc.info/?l=linux-netdev&m=155433357510597&w=2

Fixes: 2011fccf ("bpf: Support variable offset stack access from helpers")
Reported-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Test that verifier rejects indirect stack access with variable offset in
unprivileged mode and accepts same code in privileged mode.

Since pointer arithmetics is prohibited in unprivileged mode verifier
should reject the program even before it gets to helper call that uses
variable offset, at the time when that variable offset is trying to be
constructed.

Example of output:
  # ./test_verifier
  ...
  #859/u indirect variable-offset stack access, priv vs unpriv OK
  #859/p indirect variable-offset stack access, priv vs unpriv OK
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Proper support of indirect stack access with variable offset in
unprivileged mode (!root) requires corresponding support in Spectre
masking for stack ALU in retrieve_ptr_limit().

There are no use-case for variable offset in unprivileged mode though so
make verifier reject such accesses for simplicity.

Pointer arithmetics is one (and only?) way to cause variable offset and
it's already rejected in unpriv mode so that verifier won't even get to
helper function whose argument contains variable offset, e.g.:

  0: (7a) *(u64 *)(r10 -16) = 0
  1: (7a) *(u64 *)(r10 -8) = 0
  2: (61) r2 = *(u32 *)(r1 +0)
  3: (57) r2 &= 4
  4: (17) r2 -= 16
  5: (0f) r2 += r10
  variable stack access var_off=(0xfffffffffffffff0; 0x4) off=-16 size=1R2
  stack pointer arithmetic goes out of range, prohibited for !root

Still it looks like a good idea to reject variable offset indirect stack
access for unprivileged mode in check_stack_boundary() explicitly.

Fixes: 2011fccf ("bpf: Support variable offset stack access from helpers")
Reported-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Test that verifier rejects indirect access to uninitialized stack with
variable offset.

Example of output:
  # ./test_verifier
  ...
  #859/p indirect variable-offset stack access, uninitialized OK
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

It's hard to guarantee that whole memory is marked as initialized on
helper return if uninitialized stack is accessed with variable offset
since specific bounds are unknown to verifier. This may cause
uninitialized stack leaking.

Reject such an access in check_stack_boundary to prevent possible
leaking.

There are no known use-cases for indirect uninitialized stack access
with variable offset so it shouldn't break anything.

Fixes: 2011fccf ("bpf: Support variable offset stack access from helpers")
Reported-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

clang started to error on invalid asm clobber usage in x86 headers
and many bpf program samples failed to build with the message:

  CLANG-bpf  /data/users/ast/bpf-next/samples/bpf/xdp_redirect_kern.o
In file included from /data/users/ast/bpf-next/samples/bpf/xdp_redirect_kern.c:14:
In file included from ../include/linux/in.h:23:
In file included from ../include/uapi/linux/in.h:24:
In file included from ../include/linux/socket.h:8:
In file included from ../include/linux/uio.h:14:
In file included from ../include/crypto/hash.h:16:
In file included from ../include/linux/crypto.h:26:
In file included from ../include/linux/uaccess.h:5:
In file included from ../include/linux/sched.h:15:
In file included from ../include/linux/sem.h:5:
In file included from ../include/uapi/linux/sem.h:5:
In file included from ../include/linux/ipc.h:9:
In file included from ../include/linux/refcount.h:72:
../arch/x86/include/asm/refcount.h:72:36: error: asm-specifier for input or output variable conflicts with asm clobber list
                                         r->refs.counter, e, "er", i, "cx");
                                                                      ^
../arch/x86/include/asm/refcount.h:86:27: error: asm-specifier for input or output variable conflicts with asm clobber list
                                         r->refs.counter, e, "cx");
                                                             ^
2 errors generated.

Override volatile() to workaround the problem.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Since, ksym_search added with verification logic for symbols existence,
it could return NULL when the kernel symbols are not loaded.

This commit will add NULL check logic after ksym_search.
Signed-off-by: Daniel T. Lee <danieltimlee@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Currently, ksym_search located at trace_helpers won't check symbols are
existing or not.

In ksym_search, when symbol is not found, it will return &syms[0](_stext).
But when the kernel symbols are not loaded, it will return NULL, which is
not a desired action.

This commit will add verification logic whether symbols are loaded prior
to the symbol search.
Signed-off-by: Daniel T. Lee <danieltimlee@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Alexei Starovoitov says:

====================
v1->v2:
- fixed typo in patch 1
- added a patch to convert kcalloc to kvcalloc
- added a patch to verbose 16-bit jump offset check
- added a test with 1m insns

This patch set is the first step to be able to accept large programs.
The verifier still suffers from its brute force algorithm and
large programs can easily hit 1M insn_processed limit.
A lot more work is necessary to be able to verify large programs.

v1:
Realize two key ideas to speed up verification speed by ~20 times
1. every 'branching' instructions records all verifier states.
   not all of them are useful for search pruning.
   add a simple heuristic to keep states that were successful in search pruning
   and remove those that were not
2. mark_reg_read walks parentage chain of registers to mark parents as LIVE_READ.
   Once the register is marked there is no need to remark it again in the future.
   Hence stop walking the chain once first LIVE_READ is seen.

1st optimization gives 10x speed up on large programs
and 2nd optimization reduces the cost of mark_reg_read from ~40% of cpu to <1%.
Combined the deliver ~20x speedup on large programs.

Faster and bounded verification time allows to increase insn_processed
limit to 1 million from 130k.
Worst case it takes 1/10 of a second to process that many instructions
and peak memory consumption is peak_states * sizeof(struct bpf_verifier_state)
which is around ~5Mbyte.

Increase insn_per_program limit for root to insn_processed limit.

Add verification stats and stress tests for verifier scalability.
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Add a test to generate 1m ld_imm64 insns to stress the verifier.

Bump the size of fill_ld_abs_vlan_push_pop test from 4k to 29k
and jump_around_ld_abs from 4k to 5.5k.
Larger sizes are not possible due to 16-bit offset encoding
in jump instructions.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Add 3 basic tests that stress verifier scalability.

test_verif_scale1.c calls non-inlined jhash() function 90 times on
different position in the packet.
This test simulates network packet parsing.
jhash function is ~140 instructions and main program is ~1200 insns.

test_verif_scale2.c force inlines jhash() function 90 times.
This program is ~15k instructions long.

test_verif_scale3.c calls non-inlined jhash() function 90 times on
But this time jhash has to process 32-bytes from the packet
instead of 14-bytes in tests 1 and 2.
jhash function is ~230 insns and main program is ~1200 insns.

$ test_progs -s
can be used to see verifier stats.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Allow bpf_prog_load_xattr() to specify log_level for program loading.

Teach libbpf to accept log_level with bit 2 set.

Increase default BPF_LOG_BUF_SIZE from 256k to 16M.
There is no downside to increase it to a maximum allowed by old kernels.
Existing 256k limit caused ENOSPC errors and users were not able to see
verifier error which is printed at the end of the verifier log.

If ENOSPC is hit, double the verifier log and try again to capture
the verifier error.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

The existing 16Mbyte verifier log limit is not enough for log_level=2
even for small programs. Increase it to 1Gbyte.
Note it's not a kernel memory limit.
It's an amount of memory user space provides to store
the verifier log. The kernel populates it 1k at a time.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Large verifier speed improvements allow to increase
verifier complexity limit.
Now regardless of the program composition and its size it takes
little time for the verifier to hit insn_processed limit.
On typical x86 machine non-debug kernel processes 1M instructions
in 1/10 of a second.
(before these speed improvements specially crafted programs
could be hitting multi-second verification times)
Full kasan kernel with debug takes ~1 second for the same 1M insns.
Hence bump the BPF_COMPLEXITY_LIMIT_INSNS limit to 1M.
Also increase the number of instructions per program
from 4k to internal BPF_COMPLEXITY_LIMIT_INSNS limit.
4k limit was confusing to users, since small programs with hundreds
of insns could be hitting BPF_COMPLEXITY_LIMIT_INSNS limit.
Sometimes adding more insns and bpf_trace_printk debug statements
would make the verifier accept the program while removing
code would make the verifier reject it.
Some user space application started to add #define MAX_FOO to
their programs and do:
  MAX_FOO=100;
again:
  compile with MAX_FOO;
  try to load;
  if (fails_to_load) { reduce MAX_FOO; goto again; }
to be able to fit maximum amount of processing into single program.
Other users artificially split their single program into a set of programs
and use all 32 iterations of tail_calls to increase compute limits.
And the most advanced folks used unlimited tc-bpf filter list
to execute many bpf programs.
Essentially the users managed to workaround 4k insn limit.
This patch removes the limit for root programs from uapi.
BPF_COMPLEXITY_LIMIT_INSNS is the kernel internal limit
and success to load the program no longer depends on program size,
but on 'smartness' of the verifier only.
The verifier will continue to get smarter with every kernel release.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Larger programs may trigger 16-bit jump offset overflow check
during instruction patching. Make this error verbose otherwise
users cannot decipher error code without printks in the verifier.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Temporary arrays used during program verification need to be vmalloc-ed
to support large bpf programs.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

With large verifier speed improvement brought by the previous patch
mark_reg_read() becomes the hottest function during verification.
On a typical program it consumes 40% of cpu.
mark_reg_read() walks parentage chain of registers to mark parents as LIVE_READ.
Once the register is marked there is no need to remark it again in the future.
Hence stop walking the chain once first LIVE_READ is seen.
This optimization drops mark_reg_read() time from 40% of cpu to <1%
and overall 2x improvement of verification speed.
For some programs the longest_mark_read_walk counter improves from ~500 to ~5
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Branch instructions, branch targets and calls in a bpf program are
the places where the verifier remembers states that led to successful
verification of the program.
These states are used to prune brute force program analysis.
For unprivileged programs there is a limit of 64 states per such
'branching' instructions (maximum length is tracked by max_states_per_insn
counter introduced in the previous patch).
Simply reducing this threshold to 32 or lower increases insn_processed
metric to the point that small valid programs get rejected.
For root programs there is no limit and cilium programs can have
max_states_per_insn to be 100 or higher.
Walking 100+ states multiplied by number of 'branching' insns during
verification consumes significant amount of cpu time.
Turned out simple LRU-like mechanism can be used to remove states
that unlikely will be helpful in future search pruning.
This patch introduces hit_cnt and miss_cnt counters:
hit_cnt - this many times this state successfully pruned the search
miss_cnt - this many times this state was not equivalent to other states
(and that other states were added to state list)

The heuristic introduced in this patch is:
if (sl->miss_cnt > sl->hit_cnt * 3 + 3)
  /* drop this state from future considerations */

Higher numbers increase max_states_per_insn (allow more states to be
considered for pruning) and slow verification speed, but do not meaningfully
reduce insn_processed metric.
Lower numbers drop too many states and insn_processed increases too much.
Many different formulas were considered.
This one is simple and works well enough in practice.
(the analysis was done on selftests/progs/* and on cilium programs)

The end result is this heuristic improves verification speed by 10 times.
Large synthetic programs that used to take a second more now take
1/10 of a second.
In cases where max_states_per_insn used to be 100 or more, now it's ~10.

There is a slight increase in insn_processed for cilium progs:
                       before   after
bpf_lb-DLB_L3.o 	1831	1838
bpf_lb-DLB_L4.o 	3029	3218
bpf_lb-DUNKNOWN.o 	1064	1064
bpf_lxc-DDROP_ALL.o	26309	26935
bpf_lxc-DUNKNOWN.o	33517	34439
bpf_netdev.o		9713	9721
bpf_overlay.o		6184	6184
bpf_lcx_jit.o		37335	39389
And 2-3 times improvement in the verification speed.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

In order to understand the verifier bottlenecks add various stats
and extend log_level:
log_level 1 and 2 are kept as-is:
bit 0 - level=1 - print every insn and verifier state at branch points
bit 1 - level=2 - print every insn and verifier state at every insn
bit 2 - level=4 - print verifier error and stats at the end of verification

When verifier rejects the program the libbpf is trying to load the program twice.
Once with log_level=0 (no messages, only error code is reported to user space)
and second time with log_level=1 to tell the user why the verifier rejected it.

With introduction of bit 2 - level=4 the libbpf can choose to always use that
level and load programs once, since the verification speed is not affected and
in case of error the verbose message will be available.

Note that the verifier stats are not part of uapi just like all other
verbose messages. They're expected to change in the future.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

This patch adds new config option to trigger generation of BTF type
information from DWARF debuginfo for vmlinux and kernel modules through
pahole, which in turn relies on libbpf for btf_dedup() algorithm.

The intent is to record compact type information of all types used
inside kernel, including all the structs/unions/typedefs/etc. This
enables BPF's compile-once-run-everywhere ([0]) approach, in which
tracing programs that are inspecting kernel's internal data (e.g.,
struct task_struct) can be compiled on a system running some kernel
version, but would be possible to run on other kernel versions (and
configurations) without recompilation, even if the layout of structs
changed and/or some of the fields were added, removed, or renamed.

This is only possible if BPF loader can get kernel type info to adjust
all the offsets correctly. This patch is a first time in this direction,
making sure that BTF type info is part of Linux kernel image in
non-loadable ELF section.

BTF deduplication ([1]) algorithm typically provides 100x savings
compared to DWARF data, so resulting .BTF section is not big as is
typically about 2MB in size.

[0] http://vger.kernel.org/lpc-bpf2018.html#session-2
[1] https://facebookmicrosites.github.io/bpf/blog/2018/11/14/btf-enhancement.html

Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Alexei Starovoitov <ast@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Cc: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: David S. Miller <davem@davemloft.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Stanislav Fomichev says:

====================
This series contains small fixes to make bpf selftests compile cleanly
with clangs. All of them are not real problems, but it's nice to have
an option to use clang for the tests themselves.
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

verifier/ctx_skb.c:708:11: warning: initializer overrides prior initialization of this subobject [-Winitializer-overrides]
        .flags = F_NEEDS_EFFICIENT_UNALIGNED_ACCESS,
                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Use standard C99 %zu for sizeof, not GCC's custom %Zu:
bpf_obj_id.c:76:48: warning: invalid conversion specifier 'Z'
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

flow_dissector_load.c:55:19: warning: format string is not a string literal (potentially insecure)
      [-Wformat-security]
                error(1, errno, command);
                                ^~~~~~~
flow_dissector_load.c:55:19: note: treat the string as an argument to avoid this
                error(1, errno, command);
                                ^
                                "%s",
1 warning generated.
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

This makes sure we don't put headers as input files when doing
compilation, because clang complains about the following:

clang-9: error: cannot specify -o when generating multiple output files
../lib.mk:152: recipe for target 'xxx/tools/testing/selftests/bpf/test_verifier' failed
make: *** [xxx/tools/testing/selftests/bpf/test_verifier] Error 1
make: *** Waiting for unfinished jobs....
clang-9: error: cannot specify -o when generating multiple output files
../lib.mk:152: recipe for target 'xxx/tools/testing/selftests/bpf/test_progs' failed
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

For multiple dimensional arrays like below,
  int a[2][3]
both llvm and pahole generated one BTF_KIND_ARRAY type like
  . element_type: int
  . index_type: unsigned int
  . number of elements: 6

Such a collapsed BTF_KIND_ARRAY type will cause the divergence
in BTF vs. the user code. In the compile-once-run-everywhere
project, the header file is generated from BTF and used for bpf
program, and the definition in the header file will be different
from what user expects.

But the kernel actually supports chained multi-dimensional array
types properly. The above "int a[2][3]" can be represented as
  Type #n:
    . element_type: int
    . index_type: unsigned int
    . number of elements: 3
  Type #(n+1):
    . element_type: type #n
    . index_type: unsigned int
    . number of elements: 2

The following llvm commit
  https://reviews.llvm.org/rL357215
also enables llvm to generated proper chained multi-dimensional arrays.

The test_btf already has a raw test ("struct test #1") for chained
multi-dimensional arrays. This patch added amended bpffs test for
chained multi-dimensional arrays.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Andrey Ignatov says:

====================
The patch set adds support for stack access with variable offset from helpers.

Patch 1 is the main patch in the set and provides more details.
Patch 2 adds selftests for new functionality.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Test different scenarios of indirect variable-offset stack access: out of
bound access (>0), min_off below initialized part of the stack,
max_off+size above initialized part of the stack, initialized stack.

Example of output:
  ...
  #856/p indirect variable-offset stack access, out of bound OK
  #857/p indirect variable-offset stack access, max_off+size > max_initialized OK
  #858/p indirect variable-offset stack access, min_off < min_initialized OK
  #859/p indirect variable-offset stack access, ok OK
  ...
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Currently there is a difference in how verifier checks memory access for
helper arguments for PTR_TO_MAP_VALUE and PTR_TO_STACK with regard to
variable part of offset.

check_map_access, that is used for PTR_TO_MAP_VALUE, can handle variable
offsets just fine, so that BPF program can call a helper like this:

  some_helper(map_value_ptr + off, size);

, where offset is unknown at load time, but is checked by program to be
in a safe rage (off >= 0 && off + size < map_value_size).

But it's not the case for check_stack_boundary, that is used for
PTR_TO_STACK, and same code with pointer to stack is rejected by
verifier:

  some_helper(stack_value_ptr + off, size);

For example:
  0: (7a) *(u64 *)(r10 -16) = 0
  1: (7a) *(u64 *)(r10 -8) = 0
  2: (61) r2 = *(u32 *)(r1 +0)
  3: (57) r2 &= 4
  4: (17) r2 -= 16
  5: (0f) r2 += r10
  6: (18) r1 = 0xffff888111343a80
  8: (85) call bpf_map_lookup_elem#1
  invalid variable stack read R2 var_off=(0xfffffffffffffff0; 0x4)

Add support for variable offset access to check_stack_boundary so that
if offset is checked by program to be in a safe range it's accepted by
verifier.
Signed-off-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Generate a libbpf.pc file at build time so that users can rely
on pkg-config to find the library, its CFLAGS and LDFLAGS.
Signed-off-by: Luca Boccassi <bluca@debian.org>
Acked-by: Andrey Ignatov <rdna@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>