linux-stable.git/kernel/bpf, branch v5.11.4 Linux kernel stable tree bpf: Clear subreg_def for global function return values 2021-03-04T11:13:56+00:00 Ilya Leoshkevich iii@linux.ibm.com 2021-02-12T04:04:08+00:00 7c5fdd4de87dcd827cf82f808d5fc3726c1acaae [ Upstream commit 45159b27637b0fef6d5ddb86fc7c46b13c77960f ] test_global_func4 fails on s390 as reported by Yauheni in [1]. The immediate problem is that the zext code includes the instruction, whose result needs to be zero-extended, into the zero-extension patchlet, and if this instruction happens to be a branch, then its delta is not adjusted. As a result, the verifier rejects the program later. However, according to [2], as far as the verifier's algorithm is concerned and as specified by the insn_no_def() function, branching insns do not define anything. This includes call insns, even though one might argue that they define %r0. This means that the real problem is that zero extension kicks in at all. This happens because clear_caller_saved_regs() sets BPF_REG_0's subreg_def after global function calls. This can be fixed in many ways; this patch mimics what helper function call handling already does. [1] https://lore.kernel.org/bpf/20200903140542.156624-1-yauheni.kaliuta@redhat.com/ [2] https://lore.kernel.org/bpf/CAADnVQ+2RPKcftZw8d+B1UwB35cpBhpF5u3OocNh90D9pETPwg@mail.gmail.com/ Fixes: 51c39bb1d5d1 ("bpf: Introduce function-by-function verification") Reported-by: Yauheni Kaliuta <yauheni.kaliuta@redhat.com> Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Link: https://lore.kernel.org/bpf/20210212040408.90109-1-iii@linux.ibm.com Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 45159b27637b0fef6d5ddb86fc7c46b13c77960f ]

test_global_func4 fails on s390 as reported by Yauheni in [1].

The immediate problem is that the zext code includes the instruction,
whose result needs to be zero-extended, into the zero-extension
patchlet, and if this instruction happens to be a branch, then its
delta is not adjusted. As a result, the verifier rejects the program
later.

However, according to [2], as far as the verifier's algorithm is
concerned and as specified by the insn_no_def() function, branching
insns do not define anything. This includes call insns, even though
one might argue that they define %r0.

This means that the real problem is that zero extension kicks in at
all. This happens because clear_caller_saved_regs() sets BPF_REG_0's
subreg_def after global function calls. This can be fixed in many
ways; this patch mimics what helper function call handling already
does.

  [1] https://lore.kernel.org/bpf/20200903140542.156624-1-yauheni.kaliuta@redhat.com/
  [2] https://lore.kernel.org/bpf/CAADnVQ+2RPKcftZw8d+B1UwB35cpBhpF5u3OocNh90D9pETPwg@mail.gmail.com/

Fixes: 51c39bb1d5d1 ("bpf: Introduce function-by-function verification")
Reported-by: Yauheni Kaliuta <yauheni.kaliuta@redhat.com>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210212040408.90109-1-iii@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
bpf, devmap: Use GFP_KERNEL for xdp bulk queue allocation 2021-03-04T11:13:55+00:00 Jun'ichi Nomura junichi.nomura@nec.com 2021-02-09T08:24:52+00:00 7aaa11d8dd0c1cd8d4be21bc786c4e39c226ec1f [ Upstream commit 7d4553b69fb335496c597c31590e982485ebe071 ] The devmap bulk queue is allocated with GFP_ATOMIC and the allocation may fail if there is no available space in existing percpu pool. Since commit 75ccae62cb8d42 ("xdp: Move devmap bulk queue into struct net_device") moved the bulk queue allocation to NETDEV_REGISTER callback, whose context is allowed to sleep, use GFP_KERNEL instead of GFP_ATOMIC to let percpu allocator extend the pool when needed and avoid possible failure of netdev registration. As the required alignment is natural, we can simply use alloc_percpu(). Fixes: 75ccae62cb8d42 ("xdp: Move devmap bulk queue into struct net_device") Signed-off-by: Jun'ichi Nomura <junichi.nomura@nec.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Cc: Toke Høiland-Jørgensen <toke@redhat.com> Link: https://lore.kernel.org/bpf/20210209082451.GA44021@jeru.linux.bs1.fc.nec.co.jp Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 7d4553b69fb335496c597c31590e982485ebe071 ]

The devmap bulk queue is allocated with GFP_ATOMIC and the allocation
may fail if there is no available space in existing percpu pool.

Since commit 75ccae62cb8d42 ("xdp: Move devmap bulk queue into struct net_device")
moved the bulk queue allocation to NETDEV_REGISTER callback, whose context
is allowed to sleep, use GFP_KERNEL instead of GFP_ATOMIC to let percpu
allocator extend the pool when needed and avoid possible failure of netdev
registration.

As the required alignment is natural, we can simply use alloc_percpu().

Fixes: 75ccae62cb8d42 ("xdp: Move devmap bulk queue into struct net_device")
Signed-off-by: Jun'ichi Nomura <junichi.nomura@nec.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20210209082451.GA44021@jeru.linux.bs1.fc.nec.co.jp
Signed-off-by: Sasha Levin <sashal@kernel.org>
bpf: Fix an unitialized value in bpf_iter 2021-03-04T11:13:54+00:00 Yonghong Song yhs@fb.com 2021-02-12T00:59:26+00:00 9e684f7fceed9267b37c2bb5cc379eff453fdac6 [ Upstream commit 17d8beda277a36203585943e70c7909b60775fd5 ] Commit 15d83c4d7cef ("bpf: Allow loading of a bpf_iter program") cached btf_id in struct bpf_iter_target_info so later on if it can be checked cheaply compared to checking registered names. syzbot found a bug that uninitialized value may occur to bpf_iter_target_info->btf_id. This is because we allocated bpf_iter_target_info structure with kmalloc and never initialized field btf_id afterwards. This uninitialized btf_id is typically compared to a u32 bpf program func proto btf_id, and the chance of being equal is extremely slim. This patch fixed the issue by using kzalloc which will also prevent future likely instances due to adding new fields. Fixes: 15d83c4d7cef ("bpf: Allow loading of a bpf_iter program") Reported-by: syzbot+580f4f2a272e452d55cb@syzkaller.appspotmail.com Signed-off-by: Yonghong Song <yhs@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Link: https://lore.kernel.org/bpf/20210212005926.2875002-1-yhs@fb.com Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 17d8beda277a36203585943e70c7909b60775fd5 ]

Commit 15d83c4d7cef ("bpf: Allow loading of a bpf_iter program")
cached btf_id in struct bpf_iter_target_info so later on
if it can be checked cheaply compared to checking registered names.

syzbot found a bug that uninitialized value may occur to
bpf_iter_target_info->btf_id. This is because we allocated
bpf_iter_target_info structure with kmalloc and never initialized
field btf_id afterwards. This uninitialized btf_id is typically
compared to a u32 bpf program func proto btf_id, and the chance
of being equal is extremely slim.

This patch fixed the issue by using kzalloc which will also
prevent future likely instances due to adding new fields.

Fixes: 15d83c4d7cef ("bpf: Allow loading of a bpf_iter program")
Reported-by: syzbot+580f4f2a272e452d55cb@syzkaller.appspotmail.com
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210212005926.2875002-1-yhs@fb.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
bpf_lru_list: Read double-checked variable once without lock 2021-03-04T11:13:49+00:00 Marco Elver elver@google.com 2021-02-09T11:27:01+00:00 2526c18de3ad98f1ddad5aa4baf4432fbc04ed50 [ Upstream commit 6df8fb83301d68ea0a0c0e1cbcc790fcc333ed12 ] For double-checked locking in bpf_common_lru_push_free(), node->type is read outside the critical section and then re-checked under the lock. However, concurrent writes to node->type result in data races. For example, the following concurrent access was observed by KCSAN: write to 0xffff88801521bc22 of 1 bytes by task 10038 on cpu 1: __bpf_lru_node_move_in kernel/bpf/bpf_lru_list.c:91 __local_list_flush kernel/bpf/bpf_lru_list.c:298 ... read to 0xffff88801521bc22 of 1 bytes by task 10043 on cpu 0: bpf_common_lru_push_free kernel/bpf/bpf_lru_list.c:507 bpf_lru_push_free kernel/bpf/bpf_lru_list.c:555 ... Fix the data races where node->type is read outside the critical section (for double-checked locking) by marking the access with READ_ONCE() as well as ensuring the variable is only accessed once. Fixes: 3a08c2fd7634 ("bpf: LRU List") Reported-by: syzbot+3536db46dfa58c573458@syzkaller.appspotmail.com Reported-by: syzbot+516acdb03d3e27d91bcd@syzkaller.appspotmail.com Signed-off-by: Marco Elver <elver@google.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Acked-by: Martin KaFai Lau <kafai@fb.com> Link: https://lore.kernel.org/bpf/20210209112701.3341724-1-elver@google.com Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 6df8fb83301d68ea0a0c0e1cbcc790fcc333ed12 ]

For double-checked locking in bpf_common_lru_push_free(), node->type is
read outside the critical section and then re-checked under the lock.
However, concurrent writes to node->type result in data races.

For example, the following concurrent access was observed by KCSAN:

  write to 0xffff88801521bc22 of 1 bytes by task 10038 on cpu 1:
   __bpf_lru_node_move_in        kernel/bpf/bpf_lru_list.c:91
   __local_list_flush            kernel/bpf/bpf_lru_list.c:298
   ...
  read to 0xffff88801521bc22 of 1 bytes by task 10043 on cpu 0:
   bpf_common_lru_push_free      kernel/bpf/bpf_lru_list.c:507
   bpf_lru_push_free             kernel/bpf/bpf_lru_list.c:555
   ...

Fix the data races where node->type is read outside the critical section
(for double-checked locking) by marking the access with READ_ONCE() as
well as ensuring the variable is only accessed once.

Fixes: 3a08c2fd7634 ("bpf: LRU List")
Reported-by: syzbot+3536db46dfa58c573458@syzkaller.appspotmail.com
Reported-by: syzbot+516acdb03d3e27d91bcd@syzkaller.appspotmail.com
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20210209112701.3341724-1-elver@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
bpf: Fix truncation handling for mod32 dst reg wrt zero 2021-02-26T09:08:37+00:00 Daniel Borkmann daniel@iogearbox.net 2021-02-10T13:14:42+00:00 55c262ea5d0f754648cd25aa73de081adaab07d9 commit 9b00f1b78809309163dda2d044d9e94a3c0248a3 upstream. Recently noticed that when mod32 with a known src reg of 0 is performed, then the dst register is 32-bit truncated in verifier: 0: R1=ctx(id=0,off=0,imm=0) R10=fp0 0: (b7) r0 = 0 1: R0_w=inv0 R1=ctx(id=0,off=0,imm=0) R10=fp0 1: (b7) r1 = -1 2: R0_w=inv0 R1_w=inv-1 R10=fp0 2: (b4) w2 = -1 3: R0_w=inv0 R1_w=inv-1 R2_w=inv4294967295 R10=fp0 3: (9c) w1 %= w0 4: R0_w=inv0 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0 4: (b7) r0 = 1 5: R0_w=inv1 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0 5: (1d) if r1 == r2 goto pc+1 R0_w=inv1 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0 6: R0_w=inv1 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0 6: (b7) r0 = 2 7: R0_w=inv2 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0 7: (95) exit 7: R0=inv1 R1=inv(id=0,umin_value=4294967295,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2=inv4294967295 R10=fp0 7: (95) exit However, as a runtime result, we get 2 instead of 1, meaning the dst register does not contain (u32)-1 in this case. The reason is fairly straight forward given the 0 test leaves the dst register as-is: # ./bpftool p d x i 23 0: (b7) r0 = 0 1: (b7) r1 = -1 2: (b4) w2 = -1 3: (16) if w0 == 0x0 goto pc+1 4: (9c) w1 %= w0 5: (b7) r0 = 1 6: (1d) if r1 == r2 goto pc+1 7: (b7) r0 = 2 8: (95) exit This was originally not an issue given the dst register was marked as completely unknown (aka 64 bit unknown). However, after 468f6eafa6c4 ("bpf: fix 32-bit ALU op verification") the verifier casts the register output to 32 bit, and hence it becomes 32 bit unknown. Note that for the case where the src register is unknown, the dst register is marked 64 bit unknown. After the fix, the register is truncated by the runtime and the test passes: # ./bpftool p d x i 23 0: (b7) r0 = 0 1: (b7) r1 = -1 2: (b4) w2 = -1 3: (16) if w0 == 0x0 goto pc+2 4: (9c) w1 %= w0 5: (05) goto pc+1 6: (bc) w1 = w1 7: (b7) r0 = 1 8: (1d) if r1 == r2 goto pc+1 9: (b7) r0 = 2 10: (95) exit Semantics also match with {R,W}x mod{64,32} 0 -> {R,W}x. Invalid div has always been {R,W}x div{64,32} 0 -> 0. Rewrites are as follows: mod32: mod64: (16) if w0 == 0x0 goto pc+2 (15) if r0 == 0x0 goto pc+1 (9c) w1 %= w0 (9f) r1 %= r0 (05) goto pc+1 (bc) w1 = w1 Fixes: 468f6eafa6c4 ("bpf: fix 32-bit ALU op verification") Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 9b00f1b78809309163dda2d044d9e94a3c0248a3 upstream.

Recently noticed that when mod32 with a known src reg of 0 is performed,
then the dst register is 32-bit truncated in verifier:

  0: R1=ctx(id=0,off=0,imm=0) R10=fp0
  0: (b7) r0 = 0
  1: R0_w=inv0 R1=ctx(id=0,off=0,imm=0) R10=fp0
  1: (b7) r1 = -1
  2: R0_w=inv0 R1_w=inv-1 R10=fp0
  2: (b4) w2 = -1
  3: R0_w=inv0 R1_w=inv-1 R2_w=inv4294967295 R10=fp0
  3: (9c) w1 %= w0
  4: R0_w=inv0 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0
  4: (b7) r0 = 1
  5: R0_w=inv1 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0
  5: (1d) if r1 == r2 goto pc+1
   R0_w=inv1 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0
  6: R0_w=inv1 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0
  6: (b7) r0 = 2
  7: R0_w=inv2 R1_w=inv(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2_w=inv4294967295 R10=fp0
  7: (95) exit
  7: R0=inv1 R1=inv(id=0,umin_value=4294967295,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R2=inv4294967295 R10=fp0
  7: (95) exit

However, as a runtime result, we get 2 instead of 1, meaning the dst
register does not contain (u32)-1 in this case. The reason is fairly
straight forward given the 0 test leaves the dst register as-is:

  # ./bpftool p d x i 23
   0: (b7) r0 = 0
   1: (b7) r1 = -1
   2: (b4) w2 = -1
   3: (16) if w0 == 0x0 goto pc+1
   4: (9c) w1 %= w0
   5: (b7) r0 = 1
   6: (1d) if r1 == r2 goto pc+1
   7: (b7) r0 = 2
   8: (95) exit

This was originally not an issue given the dst register was marked as
completely unknown (aka 64 bit unknown). However, after 468f6eafa6c4
("bpf: fix 32-bit ALU op verification") the verifier casts the register
output to 32 bit, and hence it becomes 32 bit unknown. Note that for
the case where the src register is unknown, the dst register is marked
64 bit unknown. After the fix, the register is truncated by the runtime
and the test passes:

  # ./bpftool p d x i 23
   0: (b7) r0 = 0
   1: (b7) r1 = -1
   2: (b4) w2 = -1
   3: (16) if w0 == 0x0 goto pc+2
   4: (9c) w1 %= w0
   5: (05) goto pc+1
   6: (bc) w1 = w1
   7: (b7) r0 = 1
   8: (1d) if r1 == r2 goto pc+1
   9: (b7) r0 = 2
  10: (95) exit

Semantics also match with {R,W}x mod{64,32} 0 -> {R,W}x. Invalid div
has always been {R,W}x div{64,32} 0 -> 0. Rewrites are as follows:

  mod32:                            mod64:

  (16) if w0 == 0x0 goto pc+2       (15) if r0 == 0x0 goto pc+1
  (9c) w1 %= w0                     (9f) r1 %= r0
  (05) goto pc+1
  (bc) w1 = w1

Fixes: 468f6eafa6c4 ("bpf: fix 32-bit ALU op verification")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
bpf: Fix 32 bit src register truncation on div/mod 2021-02-10T00:32:40+00:00 Daniel Borkmann daniel@iogearbox.net 2021-02-09T18:46:10+00:00 e88b2c6e5a4d9ce30d75391e4d950da74bb2bd90 While reviewing a different fix, John and I noticed an oddity in one of the BPF program dumps that stood out, for example: # bpftool p d x i 13 0: (b7) r0 = 808464450 1: (b4) w4 = 808464432 2: (bc) w0 = w0 3: (15) if r0 == 0x0 goto pc+1 4: (9c) w4 %= w0 [...] In line 2 we noticed that the mov32 would 32 bit truncate the original src register for the div/mod operation. While for the two operations the dst register is typically marked unknown e.g. from adjust_scalar_min_max_vals() the src register is not, and thus verifier keeps tracking original bounds, simplified: 0: R1=ctx(id=0,off=0,imm=0) R10=fp0 0: (b7) r0 = -1 1: R0_w=invP-1 R1=ctx(id=0,off=0,imm=0) R10=fp0 1: (b7) r1 = -1 2: R0_w=invP-1 R1_w=invP-1 R10=fp0 2: (3c) w0 /= w1 3: R0_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R1_w=invP-1 R10=fp0 3: (77) r1 >>= 32 4: R0_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R1_w=invP4294967295 R10=fp0 4: (bf) r0 = r1 5: R0_w=invP4294967295 R1_w=invP4294967295 R10=fp0 5: (95) exit processed 6 insns (limit 1000000) max_states_per_insn 0 total_states 0 peak_states 0 mark_read 0 Runtime result of r0 at exit is 0 instead of expected -1. Remove the verifier mov32 src rewrite in div/mod and replace it with a jmp32 test instead. After the fix, we result in the following code generation when having dividend r1 and divisor r6: div, 64 bit: div, 32 bit: 0: (b7) r6 = 8 0: (b7) r6 = 8 1: (b7) r1 = 8 1: (b7) r1 = 8 2: (55) if r6 != 0x0 goto pc+2 2: (56) if w6 != 0x0 goto pc+2 3: (ac) w1 ^= w1 3: (ac) w1 ^= w1 4: (05) goto pc+1 4: (05) goto pc+1 5: (3f) r1 /= r6 5: (3c) w1 /= w6 6: (b7) r0 = 0 6: (b7) r0 = 0 7: (95) exit 7: (95) exit mod, 64 bit: mod, 32 bit: 0: (b7) r6 = 8 0: (b7) r6 = 8 1: (b7) r1 = 8 1: (b7) r1 = 8 2: (15) if r6 == 0x0 goto pc+1 2: (16) if w6 == 0x0 goto pc+1 3: (9f) r1 %= r6 3: (9c) w1 %= w6 4: (b7) r0 = 0 4: (b7) r0 = 0 5: (95) exit 5: (95) exit x86 in particular can throw a 'divide error' exception for div instruction not only for divisor being zero, but also for the case when the quotient is too large for the designated register. For the edx:eax and rdx:rax dividend pair it is not an issue in x86 BPF JIT since we always zero edx (rdx). Hence really the only protection needed is against divisor being zero. Fixes: 68fda450a7df ("bpf: fix 32-bit divide by zero") Co-developed-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> 
While reviewing a different fix, John and I noticed an oddity in one of the
BPF program dumps that stood out, for example:

  # bpftool p d x i 13
   0: (b7) r0 = 808464450
   1: (b4) w4 = 808464432
   2: (bc) w0 = w0
   3: (15) if r0 == 0x0 goto pc+1
   4: (9c) w4 %= w0
  [...]

In line 2 we noticed that the mov32 would 32 bit truncate the original src
register for the div/mod operation. While for the two operations the dst
register is typically marked unknown e.g. from adjust_scalar_min_max_vals()
the src register is not, and thus verifier keeps tracking original bounds,
simplified:

  0: R1=ctx(id=0,off=0,imm=0) R10=fp0
  0: (b7) r0 = -1
  1: R0_w=invP-1 R1=ctx(id=0,off=0,imm=0) R10=fp0
  1: (b7) r1 = -1
  2: R0_w=invP-1 R1_w=invP-1 R10=fp0
  2: (3c) w0 /= w1
  3: R0_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R1_w=invP-1 R10=fp0
  3: (77) r1 >>= 32
  4: R0_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R1_w=invP4294967295 R10=fp0
  4: (bf) r0 = r1
  5: R0_w=invP4294967295 R1_w=invP4294967295 R10=fp0
  5: (95) exit
  processed 6 insns (limit 1000000) max_states_per_insn 0 total_states 0 peak_states 0 mark_read 0

Runtime result of r0 at exit is 0 instead of expected -1. Remove the
verifier mov32 src rewrite in div/mod and replace it with a jmp32 test
instead. After the fix, we result in the following code generation when
having dividend r1 and divisor r6:

  div, 64 bit:                             div, 32 bit:

   0: (b7) r6 = 8                           0: (b7) r6 = 8
   1: (b7) r1 = 8                           1: (b7) r1 = 8
   2: (55) if r6 != 0x0 goto pc+2           2: (56) if w6 != 0x0 goto pc+2
   3: (ac) w1 ^= w1                         3: (ac) w1 ^= w1
   4: (05) goto pc+1                        4: (05) goto pc+1
   5: (3f) r1 /= r6                         5: (3c) w1 /= w6
   6: (b7) r0 = 0                           6: (b7) r0 = 0
   7: (95) exit                             7: (95) exit

  mod, 64 bit:                             mod, 32 bit:

   0: (b7) r6 = 8                           0: (b7) r6 = 8
   1: (b7) r1 = 8                           1: (b7) r1 = 8
   2: (15) if r6 == 0x0 goto pc+1           2: (16) if w6 == 0x0 goto pc+1
   3: (9f) r1 %= r6                         3: (9c) w1 %= w6
   4: (b7) r0 = 0                           4: (b7) r0 = 0
   5: (95) exit                             5: (95) exit

x86 in particular can throw a 'divide error' exception for div
instruction not only for divisor being zero, but also for the case
when the quotient is too large for the designated register. For the
edx:eax and rdx:rax dividend pair it is not an issue in x86 BPF JIT
since we always zero edx (rdx). Hence really the only protection
needed is against divisor being zero.

Fixes: 68fda450a7df ("bpf: fix 32-bit divide by zero")
Co-developed-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
bpf: Fix verifier jmp32 pruning decision logic 2021-02-10T00:31:46+00:00 Daniel Borkmann daniel@iogearbox.net 2021-02-05T19:48:21+00:00 fd675184fc7abfd1e1c52d23e8e900676b5a1c1a Anatoly has been fuzzing with kBdysch harness and reported a hang in one of the outcomes: func#0 @0 0: R1=ctx(id=0,off=0,imm=0) R10=fp0 0: (b7) r0 = 808464450 1: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R10=fp0 1: (b4) w4 = 808464432 2: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP808464432 R10=fp0 2: (9c) w4 %= w0 3: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R10=fp0 3: (66) if w4 s> 0x30303030 goto pc+0 R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff),s32_max_value=808464432) R10=fp0 4: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff),s32_max_value=808464432) R10=fp0 4: (7f) r0 >>= r0 5: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff),s32_max_value=808464432) R10=fp0 5: (9c) w4 %= w0 6: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 6: (66) if w0 s> 0x3030 goto pc+0 R0_w=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 7: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0 7: (d6) if w0 s<= 0x303030 goto pc+1 9: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0 9: (95) exit propagating r0 from 6 to 7: safe 4: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umin_value=808464433,umax_value=2147483647,var_off=(0x0; 0x7fffffff)) R10=fp0 4: (7f) r0 >>= r0 5: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umin_value=808464433,umax_value=2147483647,var_off=(0x0; 0x7fffffff)) R10=fp0 5: (9c) w4 %= w0 6: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 6: (66) if w0 s> 0x3030 goto pc+0 R0_w=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 propagating r0 7: safe propagating r0 from 6 to 7: safe processed 15 insns (limit 1000000) max_states_per_insn 0 total_states 1 peak_states 1 mark_read 1 The underlying program was xlated as follows: # bpftool p d x i 10 0: (b7) r0 = 808464450 1: (b4) w4 = 808464432 2: (bc) w0 = w0 3: (15) if r0 == 0x0 goto pc+1 4: (9c) w4 %= w0 5: (66) if w4 s> 0x30303030 goto pc+0 6: (7f) r0 >>= r0 7: (bc) w0 = w0 8: (15) if r0 == 0x0 goto pc+1 9: (9c) w4 %= w0 10: (66) if w0 s> 0x3030 goto pc+0 11: (d6) if w0 s<= 0x303030 goto pc+1 12: (05) goto pc-1 13: (95) exit The verifier rewrote original instructions it recognized as dead code with 'goto pc-1', but reality differs from verifier simulation in that we are actually able to trigger a hang due to hitting the 'goto pc-1' instructions. Taking a closer look at the verifier analysis, the reason is that it misjudges its pruning decision at the first 'from 6 to 7: safe' occasion. What happens is that while both old/cur registers are marked as precise, they get misjudged for the jmp32 case as range_within() yields true, meaning that the prior verification path with a wider register bound could be verified successfully and therefore the current path with a narrower register bound is deemed safe as well whereas in reality it's not. R0 old/cur path's bounds compare as follows: old: smin_value=0x8000000000000000,smax_value=0x7fffffffffffffff,umin_value=0x0,umax_value=0xffffffffffffffff,var_off=(0x0; 0xffffffffffffffff) cur: smin_value=0x8000000000000000,smax_value=0x7fffffff7fffffff,umin_value=0x0,umax_value=0xffffffff7fffffff,var_off=(0x0; 0xffffffff7fffffff) old: s32_min_value=0x80000000,s32_max_value=0x00003030,u32_min_value=0x00000000,u32_max_value=0xffffffff cur: s32_min_value=0x00003031,s32_max_value=0x7fffffff,u32_min_value=0x00003031,u32_max_value=0x7fffffff The 64 bit bounds generally look okay and while the information that got propagated from 32 to 64 bit looks correct as well, it's not precise enough for judging a conditional jmp32. Given the latter only operates on subregisters we also need to take these into account as well for a range_within() probe in order to be able to prune paths. Extending the range_within() constraint to both bounds will be able to tell us that the old signed 32 bit bounds are not wider than the cur signed 32 bit bounds. With the fix in place, the program will now verify the 'goto' branch case as it should have been: [...] 6: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 6: (66) if w0 s> 0x3030 goto pc+0 R0_w=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 7: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0 7: (d6) if w0 s<= 0x303030 goto pc+1 9: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0 9: (95) exit 7: R0_w=invP(id=0,smax_value=9223372034707292159,umax_value=18446744071562067967,var_off=(0x0; 0xffffffff7fffffff),s32_min_value=12337,u32_min_value=12337,u32_max_value=2147483647) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 7: (d6) if w0 s<= 0x303030 goto pc+1 R0_w=invP(id=0,smax_value=9223372034707292159,umax_value=18446744071562067967,var_off=(0x0; 0xffffffff7fffffff),s32_min_value=3158065,u32_min_value=3158065,u32_max_value=2147483647) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 8: R0_w=invP(id=0,smax_value=9223372034707292159,umax_value=18446744071562067967,var_off=(0x0; 0xffffffff7fffffff),s32_min_value=3158065,u32_min_value=3158065,u32_max_value=2147483647) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0 8: (30) r0 = *(u8 *)skb[808464432] BPF_LD_[ABS|IND] uses reserved fields processed 11 insns (limit 1000000) max_states_per_insn 1 total_states 1 peak_states 1 mark_read 1 The bug is quite subtle in the sense that when verifier would determine that a given branch is dead code, it would (here: wrongly) remove these instructions from the program and hard-wire the taken branch for privileged programs instead of the 'goto pc-1' rewrites which will cause hard to debug problems. Fixes: 3f50f132d840 ("bpf: Verifier, do explicit ALU32 bounds tracking") Reported-by: Anatoly Trosinenko <anatoly.trosinenko@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Acked-by: Alexei Starovoitov <ast@kernel.org> 
Anatoly has been fuzzing with kBdysch harness and reported a hang in
one of the outcomes:

  func#0 @0
  0: R1=ctx(id=0,off=0,imm=0) R10=fp0
  0: (b7) r0 = 808464450
  1: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R10=fp0
  1: (b4) w4 = 808464432
  2: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP808464432 R10=fp0
  2: (9c) w4 %= w0
  3: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff)) R10=fp0
  3: (66) if w4 s> 0x30303030 goto pc+0
   R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff),s32_max_value=808464432) R10=fp0
  4: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff),s32_max_value=808464432) R10=fp0
  4: (7f) r0 >>= r0
  5: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umax_value=4294967295,var_off=(0x0; 0xffffffff),s32_max_value=808464432) R10=fp0
  5: (9c) w4 %= w0
  6: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  6: (66) if w0 s> 0x3030 goto pc+0
   R0_w=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  7: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0
  7: (d6) if w0 s<= 0x303030 goto pc+1
  9: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0
  9: (95) exit
  propagating r0

  from 6 to 7: safe
  4: R0_w=invP808464450 R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umin_value=808464433,umax_value=2147483647,var_off=(0x0; 0x7fffffff)) R10=fp0
  4: (7f) r0 >>= r0
  5: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0,umin_value=808464433,umax_value=2147483647,var_off=(0x0; 0x7fffffff)) R10=fp0
  5: (9c) w4 %= w0
  6: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  6: (66) if w0 s> 0x3030 goto pc+0
   R0_w=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  propagating r0
  7: safe
  propagating r0

  from 6 to 7: safe
  processed 15 insns (limit 1000000) max_states_per_insn 0 total_states 1 peak_states 1 mark_read 1

The underlying program was xlated as follows:

  # bpftool p d x i 10
   0: (b7) r0 = 808464450
   1: (b4) w4 = 808464432
   2: (bc) w0 = w0
   3: (15) if r0 == 0x0 goto pc+1
   4: (9c) w4 %= w0
   5: (66) if w4 s> 0x30303030 goto pc+0
   6: (7f) r0 >>= r0
   7: (bc) w0 = w0
   8: (15) if r0 == 0x0 goto pc+1
   9: (9c) w4 %= w0
  10: (66) if w0 s> 0x3030 goto pc+0
  11: (d6) if w0 s<= 0x303030 goto pc+1
  12: (05) goto pc-1
  13: (95) exit

The verifier rewrote original instructions it recognized as dead code with
'goto pc-1', but reality differs from verifier simulation in that we are
actually able to trigger a hang due to hitting the 'goto pc-1' instructions.

Taking a closer look at the verifier analysis, the reason is that it misjudges
its pruning decision at the first 'from 6 to 7: safe' occasion. What happens
is that while both old/cur registers are marked as precise, they get misjudged
for the jmp32 case as range_within() yields true, meaning that the prior
verification path with a wider register bound could be verified successfully
and therefore the current path with a narrower register bound is deemed safe
as well whereas in reality it's not. R0 old/cur path's bounds compare as
follows:

  old: smin_value=0x8000000000000000,smax_value=0x7fffffffffffffff,umin_value=0x0,umax_value=0xffffffffffffffff,var_off=(0x0; 0xffffffffffffffff)
  cur: smin_value=0x8000000000000000,smax_value=0x7fffffff7fffffff,umin_value=0x0,umax_value=0xffffffff7fffffff,var_off=(0x0; 0xffffffff7fffffff)

  old: s32_min_value=0x80000000,s32_max_value=0x00003030,u32_min_value=0x00000000,u32_max_value=0xffffffff
  cur: s32_min_value=0x00003031,s32_max_value=0x7fffffff,u32_min_value=0x00003031,u32_max_value=0x7fffffff

The 64 bit bounds generally look okay and while the information that got
propagated from 32 to 64 bit looks correct as well, it's not precise enough
for judging a conditional jmp32. Given the latter only operates on subregisters
we also need to take these into account as well for a range_within() probe
in order to be able to prune paths. Extending the range_within() constraint
to both bounds will be able to tell us that the old signed 32 bit bounds are
not wider than the cur signed 32 bit bounds.

With the fix in place, the program will now verify the 'goto' branch case as
it should have been:

  [...]
  6: R0_w=invP(id=0) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  6: (66) if w0 s> 0x3030 goto pc+0
   R0_w=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  7: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0
  7: (d6) if w0 s<= 0x303030 goto pc+1
  9: R0=invP(id=0,s32_max_value=12336) R1=ctx(id=0,off=0,imm=0) R4=invP(id=0) R10=fp0
  9: (95) exit

  7: R0_w=invP(id=0,smax_value=9223372034707292159,umax_value=18446744071562067967,var_off=(0x0; 0xffffffff7fffffff),s32_min_value=12337,u32_min_value=12337,u32_max_value=2147483647) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  7: (d6) if w0 s<= 0x303030 goto pc+1
   R0_w=invP(id=0,smax_value=9223372034707292159,umax_value=18446744071562067967,var_off=(0x0; 0xffffffff7fffffff),s32_min_value=3158065,u32_min_value=3158065,u32_max_value=2147483647) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  8: R0_w=invP(id=0,smax_value=9223372034707292159,umax_value=18446744071562067967,var_off=(0x0; 0xffffffff7fffffff),s32_min_value=3158065,u32_min_value=3158065,u32_max_value=2147483647) R1=ctx(id=0,off=0,imm=0) R4_w=invP(id=0) R10=fp0
  8: (30) r0 = *(u8 *)skb[808464432]
  BPF_LD_[ABS|IND] uses reserved fields
  processed 11 insns (limit 1000000) max_states_per_insn 1 total_states 1 peak_states 1 mark_read 1

The bug is quite subtle in the sense that when verifier would determine that
a given branch is dead code, it would (here: wrongly) remove these instructions
from the program and hard-wire the taken branch for privileged programs instead
of the 'goto pc-1' rewrites which will cause hard to debug problems.

Fixes: 3f50f132d840 ("bpf: Verifier, do explicit ALU32 bounds tracking")
Reported-by: Anatoly Trosinenko <anatoly.trosinenko@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
bpf: Fix verifier jsgt branch analysis on max bound 2021-02-10T00:31:45+00:00 Daniel Borkmann daniel@iogearbox.net 2021-02-05T16:20:14+00:00 ee114dd64c0071500345439fc79dd5e0f9d106ed Fix incorrect is_branch{32,64}_taken() analysis for the jsgt case. The return code for both will tell the caller whether a given conditional jump is taken or not, e.g. 1 means branch will be taken [for the involved registers] and the goto target will be executed, 0 means branch will not be taken and instead we fall-through to the next insn, and last but not least a -1 denotes that it is not known at verification time whether a branch will be taken or not. Now while the jsgt has the branch-taken case correct with reg->s32_min_value > sval, the branch-not-taken case is off-by-one when testing for reg->s32_max_value < sval since the branch will also be taken for reg->s32_max_value == sval. The jgt branch analysis, for example, gets this right. Fixes: 3f50f132d840 ("bpf: Verifier, do explicit ALU32 bounds tracking") Fixes: 4f7b3e82589e ("bpf: improve verifier branch analysis") Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Acked-by: Alexei Starovoitov <ast@kernel.org> 
Fix incorrect is_branch{32,64}_taken() analysis for the jsgt case. The return
code for both will tell the caller whether a given conditional jump is taken
or not, e.g. 1 means branch will be taken [for the involved registers] and the
goto target will be executed, 0 means branch will not be taken and instead we
fall-through to the next insn, and last but not least a -1 denotes that it is
not known at verification time whether a branch will be taken or not. Now while
the jsgt has the branch-taken case correct with reg->s32_min_value > sval, the
branch-not-taken case is off-by-one when testing for reg->s32_max_value < sval
since the branch will also be taken for reg->s32_max_value == sval. The jgt
branch analysis, for example, gets this right.

Fixes: 3f50f132d840 ("bpf: Verifier, do explicit ALU32 bounds tracking")
Fixes: 4f7b3e82589e ("bpf: improve verifier branch analysis")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
bpf: Check for integer overflow when using roundup_pow_of_two() 2021-02-03T20:45:33+00:00 Bui Quang Minh minhquangbui99@gmail.com 2021-01-27T06:36:53+00:00 6183f4d3a0a2ad230511987c6c362ca43ec0055f On 32-bit architecture, roundup_pow_of_two() can return 0 when the argument has upper most bit set due to resulting 1UL << 32. Add a check for this case. Fixes: d5a3b1f69186 ("bpf: introduce BPF_MAP_TYPE_STACK_TRACE") Signed-off-by: Bui Quang Minh <minhquangbui99@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Link: https://lore.kernel.org/bpf/20210127063653.3576-1-minhquangbui99@gmail.com 
On 32-bit architecture, roundup_pow_of_two() can return 0 when the argument
has upper most bit set due to resulting 1UL << 32. Add a check for this case.

Fixes: d5a3b1f69186 ("bpf: introduce BPF_MAP_TYPE_STACK_TRACE")
Signed-off-by: Bui Quang Minh <minhquangbui99@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210127063653.3576-1-minhquangbui99@gmail.com
bpf, preload: Fix build when $(O) points to a relative path 2021-01-26T22:13:25+00:00 Quentin Monnet quentin@isovalent.com 2021-01-26T16:13:20+00:00 150a27328b681425c8cab239894a48f2aeb870e9 Building the kernel with CONFIG_BPF_PRELOAD, and by providing a relative path for the output directory, may fail with the following error: $ make O=build bindeb-pkg ... /.../linux/tools/scripts/Makefile.include:5: *** O=build does not exist. Stop. make[7]: *** [/.../linux/kernel/bpf/preload/Makefile:9: kernel/bpf/preload/libbpf.a] Error 2 make[6]: *** [/.../linux/scripts/Makefile.build:500: kernel/bpf/preload] Error 2 make[5]: *** [/.../linux/scripts/Makefile.build:500: kernel/bpf] Error 2 make[4]: *** [/.../linux/Makefile:1799: kernel] Error 2 make[4]: *** Waiting for unfinished jobs.... In the case above, for the "bindeb-pkg" target, the error is produced by the "dummy" check in Makefile.include, called from libbpf's Makefile. This check changes directory to $(PWD) before checking for the existence of $(O). But at this step we have $(PWD) pointing to "/.../linux/build", and $(O) pointing to "build". So the Makefile.include tries in fact to assert the existence of a directory named "/.../linux/build/build", which does not exist. Note that the error does not occur for all make targets and architectures combinations. This was observed on x86 for "bindeb-pkg", or for a regular build for UML [0]. Here are some details. The root Makefile recursively calls itself once, after changing directory to $(O). The content for the variable $(PWD) is preserved across recursive calls to make, so it is unchanged at this step. For "bindeb-pkg", $(PWD) is eventually updated because the target writes a new Makefile (as debian/rules) and calls it indirectly through dpkg-buildpackage. This script does not preserve $(PWD), which is reset to the current working directory when the target in debian/rules is called. Although not investigated, it seems likely that something similar causes UML to change its value for $(PWD). Non-trivial fixes could be to remove the use of $(PWD) from the "dummy" check, or to make sure that $(PWD) and $(O) are preserved or updated to always play well and form a valid $(PWD)/$(O) path across the different targets and architectures. Instead, we take a simpler approach and just update $(O) when calling libbpf's Makefile, so it points to an absolute path which should always resolve for the "dummy" check run (through includes) by that Makefile. David Gow previously posted a slightly different version of this patch as a RFC [0], two months ago or so. [0] https://lore.kernel.org/bpf/20201119085022.3606135-1-davidgow@google.com/t/#u Fixes: d71fa5c9763c ("bpf: Add kernel module with user mode driver that populates bpffs.") Reported-by: David Gow <davidgow@google.com> Signed-off-by: Quentin Monnet <quentin@isovalent.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Andrii Nakryiko <andrii@kernel.org> Cc: Brendan Higgins <brendanhiggins@google.com> Cc: Masahiro Yamada <masahiroy@kernel.org> Link: https://lore.kernel.org/bpf/20210126161320.24561-1-quentin@isovalent.com 
Building the kernel with CONFIG_BPF_PRELOAD, and by providing a relative
path for the output directory, may fail with the following error:

  $ make O=build bindeb-pkg
  ...
  /.../linux/tools/scripts/Makefile.include:5: *** O=build does not exist.  Stop.
  make[7]: *** [/.../linux/kernel/bpf/preload/Makefile:9: kernel/bpf/preload/libbpf.a] Error 2
  make[6]: *** [/.../linux/scripts/Makefile.build:500: kernel/bpf/preload] Error 2
  make[5]: *** [/.../linux/scripts/Makefile.build:500: kernel/bpf] Error 2
  make[4]: *** [/.../linux/Makefile:1799: kernel] Error 2
  make[4]: *** Waiting for unfinished jobs....

In the case above, for the "bindeb-pkg" target, the error is produced by
the "dummy" check in Makefile.include, called from libbpf's Makefile.
This check changes directory to $(PWD) before checking for the existence
of $(O). But at this step we have $(PWD) pointing to "/.../linux/build",
and $(O) pointing to "build". So the Makefile.include tries in fact to
assert the existence of a directory named "/.../linux/build/build",
which does not exist.

Note that the error does not occur for all make targets and
architectures combinations. This was observed on x86 for "bindeb-pkg",
or for a regular build for UML [0].

Here are some details. The root Makefile recursively calls itself once,
after changing directory to $(O). The content for the variable $(PWD) is
preserved across recursive calls to make, so it is unchanged at this
step. For "bindeb-pkg", $(PWD) is eventually updated because the target
writes a new Makefile (as debian/rules) and calls it indirectly through
dpkg-buildpackage. This script does not preserve $(PWD), which is reset
to the current working directory when the target in debian/rules is
called.

Although not investigated, it seems likely that something similar causes
UML to change its value for $(PWD).

Non-trivial fixes could be to remove the use of $(PWD) from the "dummy"
check, or to make sure that $(PWD) and $(O) are preserved or updated to
always play well and form a valid $(PWD)/$(O) path across the different
targets and architectures. Instead, we take a simpler approach and just
update $(O) when calling libbpf's Makefile, so it points to an absolute
path which should always resolve for the "dummy" check run (through
includes) by that Makefile.

David Gow previously posted a slightly different version of this patch
as a RFC [0], two months ago or so.

  [0] https://lore.kernel.org/bpf/20201119085022.3606135-1-davidgow@google.com/t/#u

Fixes: d71fa5c9763c ("bpf: Add kernel module with user mode driver that populates bpffs.")
Reported-by: David Gow <davidgow@google.com>
Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Cc: Brendan Higgins <brendanhiggins@google.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Link: https://lore.kernel.org/bpf/20210126161320.24561-1-quentin@isovalent.com