Viktor Malik says:
====================
I noticed that the verifier behaves incorrectly when attaching to fentry
of multiple functions of the same name located in different modules (or
in vmlinux). The reason for this is that if the target program is not
specified, the verifier will search kallsyms for the trampoline address
to attach to. The entire kallsyms is always searched, not respecting the
module in which the function to attach to is located.
As Yonghong correctly pointed out, there is yet another issue - the
trampoline acquires the module reference in register_fentry which means
that if the module is unloaded between the place where the address is
found in the verifier and register_fentry, it is possible that another
module is loaded to the same address in the meantime, which may lead to
errors.
This patch fixes the above issues by extracting the module name from the
BTF of the attachment target (which must be specified) and by doing the
search in kallsyms of the correct module. At the same time, the module
reference is acquired right after the address is found and only released
right before the program itself is unloaded.
---
Changes in v10:
- added the new test to DENYLIST.aarch64 (suggested by Andrii)
- renamed the test source file to match the test name
Changes in v9:
- two small changes suggested by Jiri Olsa and Jiri's ack
Changes in v8:
- added module_put to error paths in bpf_check_attach_target after the
module reference is acquired
Changes in v7:
- refactored the module reference manipulation (comments by Jiri Olsa)
- cleaned up the test (comments by Andrii Nakryiko)
Changes in v6:
- storing the module reference inside bpf_prog_aux instead of
bpf_trampoline and releasing it when the program is unloaded
(suggested by Jiri Olsa)
Changes in v5:
- fixed acquiring and releasing of module references by trampolines to
prevent modules being unloaded between address lookup and trampoline
allocation
Changes in v4:
- reworked module kallsyms lookup approach using existing functions,
verifier now calls btf_try_get_module to retrieve the module and
find_kallsyms_symbol_value to get the symbol address (suggested by
Alexei)
- included Jiri Olsa's comments
- improved description of the new test and added it as a comment into
the test source
Changes in v3:
- added trivial implementation for kallsyms_lookup_name_in_module() for
!CONFIG_MODULES (noticed by test robot, fix suggested by Hao Luo)
Changes in v2:
- introduced and used more space-efficient kallsyms lookup function,
suggested by Jiri Olsa
- included Hao Luo's comments
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Adds a new test that tries to attach a program to fentry of two
functions of the same name, one located in vmlinux and the other in
bpf_testmod.
To avoid conflicts with existing tests, a new function
"bpf_fentry_shadow_test" was created both in vmlinux and in bpf_testmod.
The previous commit fixed a bug which caused this test to fail. The
verifier would always use the vmlinux function's address as the target
trampoline address, hence trying to create two trampolines for a single
address, which is forbidden.
The test (similarly to other fentry/fexit tests) is not working on arm64
at the moment.
Signed-off-by: Viktor Malik <vmalik@redhat.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/5fe2f364190b6f79b085066ed7c5989c5bc475fa.1678432753.git.vmalik@redhat.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This resolves two problems with attachment of fentry/fexit/fmod_ret/lsm
to functions located in modules:
1. The verifier tries to find the address to attach to in kallsyms. This
is always done by searching the entire kallsyms, not respecting the
module in which the function is located. Such approach causes an
incorrect attachment address to be computed if the function to attach
to is shadowed by a function of the same name located earlier in
kallsyms.
2. If the address to attach to is located in a module, the module
reference is only acquired in register_fentry. If the module is
unloaded between the place where the address is found
(bpf_check_attach_target in the verifier) and register_fentry, it is
possible that another module is loaded to the same address which may
lead to potential errors.
Since the attachment must contain the BTF of the program to attach to,
we extract the module from it and search for the function address in the
correct module (resolving problem no. 1). Then, the module reference is
taken directly in bpf_check_attach_target and stored in the bpf program
(in bpf_prog_aux). The reference is only released when the program is
unloaded (resolving problem no. 2).
Signed-off-by: Viktor Malik <vmalik@redhat.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Reviewed-by: Luis Chamberlain <mcgrof@kernel.org>
Link: https://lore.kernel.org/r/3f6a9d8ae850532b5ef864ef16327b0f7a669063.1678432753.git.vmalik@redhat.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The commit 332ea1f697 ("bpf: Add bpf_cgroup_from_id() kfunc") added
bpf_cgroup_from_id() which calls current_cgns_cgroup_dfl() through
cgroup_get_from_id(). However, BPF programs may be attached to a point where
current->nsproxy has already been cleared to NULL by exit_task_namespace()
and calling bpf_cgroup_from_id() would cause an oops.
Just return the system-wide root if nsproxy has been cleared. This allows
all cgroups to be looked up after the task passed through
exit_task_namespace(), which semantically makes sense. Given that the only
way to get this behavior is through BPF programs, it seems safe but let's
see what others think.
Fixes: 332ea1f697 ("bpf: Add bpf_cgroup_from_id() kfunc")
Signed-off-by: Tejun Heo <tj@kernel.org>
Link: https://lore.kernel.org/r/ZBDuVWiFj2jiz3i8@slm.duckdns.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
LLVM commit https://reviews.llvm.org/D143726 introduced hoistMinMax optimization
that transformed
(i < VIRTIO_MAX_SGS) && (i < out_sgs)
into
i < MIN(VIRTIO_MAX_SGS, out_sgs)
and caused the verifier to stop recognizing such loop as bounded.
Which resulted in the following test failure:
libbpf: prog 'trace_virtqueue_add_sgs': BPF program load failed: Bad address
libbpf: prog 'trace_virtqueue_add_sgs': -- BEGIN PROG LOAD LOG --
The sequence of 8193 jumps is too complex.
verification time 789206 usec
stack depth 56
processed 156446 insns (limit 1000000) max_states_per_insn 7 total_states 1746 peak_states 1701 mark_read 12
-- END PROG LOAD LOG --
libbpf: prog 'trace_virtqueue_add_sgs': failed to load: -14
libbpf: failed to load object 'loop6.bpf.o'
Workaround the verifier limitation for now with inline asm that
prevents this particular optimization.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Alexander Lobakin says:
====================
Yeah, I still remember that "Who needs cpumap nowadays" (c), but anyway.
__xdp_build_skb_from_frame() missed the moment when the networking stack
became able to recycle skb pages backed by a page_pool. This was making
e.g. cpumap redirect even less effective than simple %XDP_PASS. veth was
also affected in some scenarios.
A lot of drivers use skb_mark_for_recycle() already, it's been almost
two years and seems like there are no issues in using it in the generic
code too. {__,}xdp_release_frame() can be then removed as it losts its
last user.
Page Pool becomes then zero-alloc (or almost) in the abovementioned
cases, too. Other memory type models (who needs them at this point)
have no changes.
Some numbers on 1 Xeon Platinum core bombed with 27 Mpps of 64-byte
IPv6 UDP, iavf w/XDP[0] (CONFIG_PAGE_POOL_STATS is enabled):
Plain %XDP_PASS on baseline, Page Pool driver:
src cpu Rx drops dst cpu Rx
2.1 Mpps N/A 2.1 Mpps
cpumap redirect (cross-core, w/o leaving its NUMA node) on baseline:
6.8 Mpps 5.0 Mpps 1.8 Mpps
cpumap redirect with skb PP recycling:
7.9 Mpps 5.7 Mpps 2.2 Mpps
+22% (from cpumap redir on baseline)
[0] https://github.com/alobakin/linux/commits/iavf-xdp
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
__xdp_build_skb_from_frame() was the last user of
{__,}xdp_release_frame(), which detaches pages from the page_pool.
All the consumers now recycle Page Pool skbs and page, except mlx5,
stmmac and tsnep drivers, which use page_pool_release_page() directly
(might change one day). It's safe to assume this functionality is not
needed anymore and can be removed (in favor of recycling).
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Link: https://lore.kernel.org/r/20230313215553.1045175-5-aleksander.lobakin@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
__xdp_build_skb_from_frame() state(d):
/* Until page_pool get SKB return path, release DMA here */
Page Pool got skb pages recycling in April 2021, but missed this
function.
xdp_release_frame() is relevant only for Page Pool backed frames and it
detaches the page from the corresponding page_pool in order to make it
freeable via page_frag_free(). It can instead just mark the output skb
as eligible for recycling if the frame is backed by a pp. No change for
other memory model types (the same condition check as before).
cpumap redirect and veth on Page Pool drivers now become zero-alloc (or
almost).
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Link: https://lore.kernel.org/r/20230313215553.1045175-4-aleksander.lobakin@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
skb_mark_for_recycle() is guarded with CONFIG_PAGE_POOL, this creates
unneeded complication when using it in the generic code. For now, it's
only used in the drivers always selecting Page Pool, so this works.
Move the guards so that preprocessor will cut out only the operation
itself and the function will still be a noop on !PAGE_POOL systems,
but available there as well.
No functional changes.
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/oe-kbuild-all/202303020342.Wi2PRFFH-lkp@intel.com
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Link: https://lore.kernel.org/r/20230313215553.1045175-3-aleksander.lobakin@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Currently, the test relies on that only dropped ("xmitted") frames will
be recycled and if a frame became an skb, it will be freed later by the
stack and never come back to its page_pool.
So, it easily gets broken by trying to recycle skbs[0]:
test_xdp_do_redirect:PASS:pkt_count_xdp 0 nsec
test_xdp_do_redirect:FAIL:pkt_count_zero unexpected pkt_count_zero:
actual 9936 != expected 2
test_xdp_do_redirect:PASS:pkt_count_tc 0 nsec
That huge mismatch happened because after the TC ingress hook zeroes the
magic, the page gets recycled when skb is freed, not returned to the MM
layer. "Live frames" mode initializes only new pages and keeps the
recycled ones as is by design, so they appear with zeroed magic on the
Rx path again.
Expand the possible magic values from two: 0 (was "xmitted"/dropped or
did hit the TC hook) and 0x42 (hit the input XDP prog) to three: the new
one will mark frames hit the TC hook, so that they will elide both
@pkt_count_zero and @pkt_count_xdp. They can then be recycled to their
page_pool or returned to the page allocator, this won't affect the
counters anyhow. Just make sure to mark them as "input" (0x42) when they
appear on the Rx path again.
Also make an enum from those magics, so that they will be always visible
and can be changed in just one place anytime. This also eases adding any
new marks later on.
Link: https://github.com/kernel-patches/bpf/actions/runs/4386538411/jobs/7681081789
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Link: https://lore.kernel.org/r/20230313215553.1045175-2-aleksander.lobakin@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The verifier rejects the code:
bpf_strncmp(task->comm, 16, "my_task");
with the message:
16: (85) call bpf_strncmp#182
R1 type=trusted_ptr_ expected=fp, pkt, pkt_meta, map_key, map_value, mem, ringbuf_mem, buf
Teach the verifier that such access pattern is safe.
Do not allow untrusted and legacy ptr_to_btf_id to be passed into helpers.
Reported-by: David Vernet <void@manifault.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/r/20230313235845.61029-3-alexei.starovoitov@gmail.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
bpf_strncmp() doesn't write into its first argument.
Make sure that the verifier knows about it.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/r/20230313235845.61029-2-alexei.starovoitov@gmail.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Improve clarity by adding an example of a signed comparison instruction
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Acked-by: David Vernet <void@manifault.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20230310233814.4641-1-dthaler1968@googlemail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The canonical location for the tracefs filesystem is at
/sys/kernel/tracing.
But, from Documentation/trace/ftrace.rst:
Before 4.1, all ftrace tracing control files were within the debugfs
file system, which is typically located at /sys/kernel/debug/tracing.
For backward compatibility, when mounting the debugfs file system,
the tracefs file system will be automatically mounted at:
/sys/kernel/debug/tracing
Many tests in the bpf selftest code still refer to this older debugfs
path, so let's update them to avoid confusion.
Signed-off-by: Ross Zwisler <zwisler@google.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Link: https://lore.kernel.org/r/20230313205628.1058720-3-zwisler@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The canonical location for the tracefs filesystem is at /sys/kernel/tracing.
But, from Documentation/trace/ftrace.rst:
Before 4.1, all ftrace tracing control files were within the debugfs
file system, which is typically located at /sys/kernel/debug/tracing.
For backward compatibility, when mounting the debugfs file system,
the tracefs file system will be automatically mounted at:
/sys/kernel/debug/tracing
Many comments and samples in the bpf code still refer to this older
debugfs path, so let's update them to avoid confusion. There are a few
spots where the bpf code explicitly checks both tracefs and debugfs
(tools/bpf/bpftool/tracelog.c and tools/lib/api/fs/fs.c) and I've left
those alone so that the tools can continue to work with both paths.
Signed-off-by: Ross Zwisler <zwisler@google.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Link: https://lore.kernel.org/r/20230313205628.1058720-2-zwisler@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
When a local kptr is stashed in a map and freed when the map goes away,
currently an error like the below appears:
[ 39.195695] BUG: using smp_processor_id() in preemptible [00000000] code: kworker/u32:15/2875
[ 39.196549] caller is bpf_mem_free+0x56/0xc0
[ 39.196958] CPU: 15 PID: 2875 Comm: kworker/u32:15 Tainted: G O 6.2.0-13016-g22df776a9a86 #4477
[ 39.197897] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
[ 39.198949] Workqueue: events_unbound bpf_map_free_deferred
[ 39.199470] Call Trace:
[ 39.199703] <TASK>
[ 39.199911] dump_stack_lvl+0x60/0x70
[ 39.200267] check_preemption_disabled+0xbf/0xe0
[ 39.200704] bpf_mem_free+0x56/0xc0
[ 39.201032] ? bpf_obj_new_impl+0xa0/0xa0
[ 39.201430] bpf_obj_free_fields+0x1cd/0x200
[ 39.201838] array_map_free+0xad/0x220
[ 39.202193] ? finish_task_switch+0xe5/0x3c0
[ 39.202614] bpf_map_free_deferred+0xea/0x210
[ 39.203006] ? lockdep_hardirqs_on_prepare+0xe/0x220
[ 39.203460] process_one_work+0x64f/0xbe0
[ 39.203822] ? pwq_dec_nr_in_flight+0x110/0x110
[ 39.204264] ? do_raw_spin_lock+0x107/0x1c0
[ 39.204662] ? lockdep_hardirqs_on_prepare+0xe/0x220
[ 39.205107] worker_thread+0x74/0x7a0
[ 39.205451] ? process_one_work+0xbe0/0xbe0
[ 39.205818] kthread+0x171/0x1a0
[ 39.206111] ? kthread_complete_and_exit+0x20/0x20
[ 39.206552] ret_from_fork+0x1f/0x30
[ 39.206886] </TASK>
This happens because the call to __bpf_obj_drop_impl I added in the patch
adding support for stashing local kptrs doesn't disable migration. Prior
to that patch, __bpf_obj_drop_impl logic only ran when called by a BPF
progarm, whereas now it can be called from map free path, so it's
necessary to explicitly disable migration.
Also, refactor a bit to just call __bpf_obj_drop_impl directly instead
of bothering w/ dtor union and setting pointer-to-obj_drop.
Fixes: c8e1875409 ("bpf: Support __kptr to local kptrs")
Reported-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230313214641.3731908-1-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
In commit 3fbd7ee285 ("tasks: Add a count of task RCU users"), a
count on the number of RCU users was added to struct task_struct. This
was done so as to enable the removal of task_rcu_dereference(), and
allow tasks to be protected by RCU even after exiting and being removed
from the runqueue. In this commit, the 'refcount_t rcu_users' field that
keeps track of this refcount was put into a union co-located with
'struct rcu_head rcu', so as to avoid taking up any extra space in
task_struct. This was possible to do safely, because the field was only
ever decremented by a static set of specific callers, and then never
incremented again.
While this restriction of there only being a small, static set of users
of this field has worked fine, it prevents us from leveraging the field
to use RCU to protect tasks in other contexts.
During tracing, for example, it would be useful to be able to collect
some tasks that performed a certain operation, put them in a map, and
then periodically summarize who they are, which cgroup they're in, how
much CPU time they've utilized, etc. While this can currently be done
with 'usage', it becomes tricky when a task is already in a map, or if a
reference should only be taken if a task is valid and will not soon be
reaped. Ideally, we could do something like pass a reference to a map
value, and then try to acquire a reference to the task in an RCU read
region by using refcount_inc_not_zero().
Similarly, in sched_ext, schedulers are using integer pids to remember
tasks, and then looking them up with find_task_by_pid_ns(). This is
slow, error prone, and adds complexity. It would be more convenient and
performant if BPF schedulers could instead store tasks directly in maps,
and then leverage RCU to ensure they can be safely accessed with low
overhead.
Finally, overloading fields like this is error prone. Someone that wants
to use 'rcu_users' could easily overlook the fact that once the rcu
callback is scheduled, the refcount will go back to being nonzero, thus
precluding the use of refcount_inc_not_zero(). Furthermore, as described
below, it's possible to extract the fields of the union without changing
the size of task_struct.
There are several possible ways to enable this:
1. The lightest touch approach is likely the one proposed in this patch,
which is to simply extract 'rcu_users' and 'rcu' from the union, so
that scheduling the 'rcu' callback doesn't overwrite the 'rcu_users'
refcount. If we have a trusted task pointer, this would allow us to
use refcnt_inc_not_zero() inside of an RCU region to determine if we
can safely acquire a reference to the task and store it in a map. As
mentioned below, this can be done without changing the size of
task_struct, by moving the location of the union to another location
that has padding gaps we can fill in.
2. Removing 'refcount_t rcu_users', and instead having the entire task
be freed in an rcu callback. This is likely the most sound overall
design, though it changes the behavioral semantics exposed to
callers, who currently expect that a task that's successfully looked
up in e.g. the pid_list with find_task_by_pid_ns(), can always have a
'usage' reference acquired on them, as it's guaranteed to be >
0 until after the next gp. In order for this approach to work, we'd
have to audit all callers. This approach also slightly changes
behavior observed by user space by not invoking
trace_sched_process_free() until the whole task_struct is actually being
freed, rather than just after it's exited. It also may change
timings, as memory will be freed in an RCU callback rather than
immediately when the final 'usage' refcount drops to 0. This also is
arguably a benefit, as it provides more predictable performance to
callers who are refcounting tasks.
3. There may be other solutions as well that don't require changing the
layout of task_struct. For example, we could possibly do something
complex from the BPF side, such as listen for task exit and remove a
task from a map when the task is exiting. This would likely require
significant custom handling for task_struct in the verifier, so a
more generalizable solution is likely warranted.
As mentioned above, this patch proposes the lightest-touch approach
which allows callers elsewhere in the kernel to use 'rcu_users' to
ensure the lifetime of a task, by extracting 'rcu_users' and 'rcu' from
the union. There is no size change in task_struct with this patch.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: David Vernet <void@manifault.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Link: https://lore.kernel.org/r/20230215233033.889644-1-void@manifault.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Fix wrong order of frame index vs register/slot index in precision
propagation verbose (level 2) output. It's wrong and very confusing as is.
Fixes: 529409ea92 ("bpf: propagate precision across all frames, not just the last one")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20230313184017.4083374-1-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Dave Marchevsky says:
====================
Local kptrs are kptrs allocated via bpf_obj_new with a type specified in program
BTF. A BPF program which creates a local kptr has exclusive control of the
lifetime of the kptr, and, prior to terminating, must:
* free the kptr via bpf_obj_drop
* If the kptr is a {list,rbtree} node, add the node to a {list, rbtree},
thereby passing control of the lifetime to the collection
This series adds a third option:
* stash the kptr in a map value using bpf_kptr_xchg
As indicated by the use of "stash" to describe this behavior, the intended use
of this feature is temporary storage of local kptrs. For example, a sched_ext
([0]) scheduler may want to create an rbtree node for each new cgroup on cgroup
init, but to add that node to the rbtree as part of a separate program which
runs on enqueue. Stashing the node in a map_value allows its lifetime to outlive
the execution of the cgroup_init program.
Behavior:
There is no semantic difference between adding a kptr to a graph collection and
"stashing" it in a map. In both cases exclusive ownership of the kptr's lifetime
is passed to some containing data structure, which is responsible for
bpf_obj_drop'ing it when the container goes away.
Since graph collections also expect exclusive ownership of the nodes they
contain, graph nodes cannot be both stashed in a map_value and contained by
their corresponding collection.
Implementation:
Two observations simplify the verifier changes for this feature. First, kptrs
("referenced kptrs" until a recent renaming) require registration of a
dtor function as part of their acquire/release semantics, so that a referenced
kptr which is placed in a map_value is properly released when the map goes away.
We want this exact behavior for local kptrs, but with bpf_obj_drop as the dtor
instead of a per-btf_id dtor.
The second observation is that, in terms of identification, "referenced kptr"
and "local kptr" already don't interfere with one another. Consider the
following example:
struct node_data {
long key;
long data;
struct bpf_rb_node node;
};
struct map_value {
struct node_data __kptr *node;
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__type(key, int);
__type(value, struct map_value);
__uint(max_entries, 1);
} some_nodes SEC(".maps");
struct map_value *mapval;
struct node_data *res;
int key = 0;
res = bpf_obj_new(typeof(*res));
if (!res) { /* err handling */ }
mapval = bpf_map_lookup_elem(&some_nodes, &key);
if (!mapval) { /* err handling */ }
res = bpf_kptr_xchg(&mapval->node, res);
if (res)
bpf_obj_drop(res);
The __kptr tag identifies map_value's node as a referenced kptr, while the
PTR_TO_BTF_ID which bpf_obj_new returns - a type in some non-vmlinux,
non-module BTF - identifies res as a local kptr. Type tag on the pointer
indicates referenced kptr, while the type of the pointee indicates local kptr.
So using existing facilities we can tell the verifier about a "referenced kptr"
pointer to a "local kptr" pointee.
When kptr_xchg'ing a kptr into a map_value, the verifier can recognize local
kptr types and treat them like referenced kptrs with a properly-typed
bpf_obj_drop as a dtor.
Other implementation notes:
* We don't need to do anything special to enforce "graph nodes cannot be
both stashed in a map_value and contained by their corresponding collection"
* bpf_kptr_xchg both returns and takes as input a (possibly-null) owning
reference. It does not accept non-owning references as input by virtue
of requiring a ref_obj_id. By definition, if a program has an owning
ref to a node, the node isn't in a collection, so it's safe to pass
ownership via bpf_kptr_xchg.
Summary of patches:
* Patch 1 modifies BTF plumbing to support using bpf_obj_drop as a dtor
* Patch 2 adds verifier plumbing to support MEM_ALLOC-flagged param for
bpf_kptr_xchg
* Patch 3 adds selftests exercising the new behavior
Changelog:
v1 -> v2: https://lore.kernel.org/bpf/20230309180111.1618459-1-davemarchevsky@fb.com/
Patch #s used below refer to the patch's position in v1 unless otherwise
specified.
Patches 1-3 were applied and are not included in v2.
Rebase onto latest bpf-next: "libbpf: Revert poisoning of strlcpy"
Patch 4: "bpf: Support __kptr to local kptrs"
* Remove !btf_is_kernel(btf) check, WARN_ON_ONCE instead (Alexei)
Patch 6: "selftests/bpf: Add local kptr stashing test"
* Add test which stashes 2 nodes and later unstashes one of them using a
separate BPF program (Alexei)
* Fix incorrect runner subtest name for original test (was
"rbtree_add_nodes")
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add a new selftest, local_kptr_stash, which uses bpf_kptr_xchg to stash
a bpf_obj_new-allocated object in a map. Test the following scenarios:
* Stash two rb_nodes in an arraymap, don't unstash them, rely on map
free to destruct them
* Stash two rb_nodes in an arraymap, unstash the second one in a
separate program, rely on map free to destruct first
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-4-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The previous patch added necessary plumbing for verifier and runtime to
know what to do with non-kernel PTR_TO_BTF_IDs in map values, but didn't
provide any way to get such local kptrs into a map value. This patch
modifies verifier handling of bpf_kptr_xchg to allow MEM_ALLOC kptr
types.
check_reg_type is modified accept MEM_ALLOC-flagged input to
bpf_kptr_xchg despite such types not being in btf_ptr_types. This could
have been done with a MAYBE_MEM_ALLOC equivalent to MAYBE_NULL, but
bpf_kptr_xchg is the only helper that I can forsee using
MAYBE_MEM_ALLOC, so keep it special-cased for now.
The verifier tags bpf_kptr_xchg retval MEM_ALLOC if and only if the BTF
associated with the retval is not kernel BTF.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-3-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
If a PTR_TO_BTF_ID type comes from program BTF - not vmlinux or module
BTF - it must have been allocated by bpf_obj_new and therefore must be
free'd with bpf_obj_drop. Such a PTR_TO_BTF_ID is considered a "local
kptr" and is tagged with MEM_ALLOC type tag by bpf_obj_new.
This patch adds support for treating __kptr-tagged pointers to "local
kptrs" as having an implicit bpf_obj_drop destructor for referenced kptr
acquire / release semantics. Consider the following example:
struct node_data {
long key;
long data;
struct bpf_rb_node node;
};
struct map_value {
struct node_data __kptr *node;
};
struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
__type(key, int);
__type(value, struct map_value);
__uint(max_entries, 1);
} some_nodes SEC(".maps");
If struct node_data had a matching definition in kernel BTF, the verifier would
expect a destructor for the type to be registered. Since struct node_data does
not match any type in kernel BTF, the verifier knows that there is no kfunc
that provides a PTR_TO_BTF_ID to this type, and that such a PTR_TO_BTF_ID can
only come from bpf_obj_new. So instead of searching for a registered dtor,
a bpf_obj_drop dtor can be assumed.
This allows the runtime to properly destruct such kptrs in
bpf_obj_free_fields, which enables maps to clean up map_vals w/ such
kptrs when going away.
Implementation notes:
* "kernel_btf" variable is renamed to "kptr_btf" in btf_parse_kptr.
Before this patch, the variable would only ever point to vmlinux or
module BTFs, but now it can point to some program BTF for local kptr
type. It's later used to populate the (btf, btf_id) pair in kptr btf
field.
* It's necessary to btf_get the program BTF when populating btf_field
for local kptr. btf_record_free later does a btf_put.
* Behavior for non-local referenced kptrs is not modified, as
bpf_find_btf_id helper only searches vmlinux and module BTFs for
matching BTF type. If such a type is found, btf_field_kptr's btf will
pass btf_is_kernel check, and the associated release function is
some one-argument dtor. If btf_is_kernel check fails, associated
release function is two-arg bpf_obj_drop_impl. Before this patch
only btf_field_kptr's w/ kernel or module BTFs were created.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230310230743.2320707-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add brief text about existence of helper functions, with details to go in
separate psABI text.
Note that text about runtime functions (kfuncs) is part of a separate patch,
not this one.
Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Link: https://lore.kernel.org/r/20230308205303.1308-1-dthaler1968@googlemail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
btf_record_find's 3rd parameter can be multiple enum btf_field_type's
masked together. The function is called with BPF_KPTR in two places in
verifier.c, so it works with masked values already.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230309180111.1618459-4-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This enum was added and used in commit aa3496accc ("bpf: Refactor kptr_off_tab
into btf_record"). Later refactoring in commit db55911782 ("bpf: Consolidate
spin_lock, timer management into btf_record") resulted in the enum
values no longer being used anywhere.
Let's remove them.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230309180111.1618459-3-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
kernel_type_name was introduced in commit 9e15db6613 ("bpf: Implement accurate raw_tp context access via BTF")
with type signature:
const char *kernel_type_name(u32 id)
At that time the function used global btf_vmlinux BTF for all id lookups. Later,
in commit 22dc4a0f5e ("bpf: Remove hard-coded btf_vmlinux assumption from BPF verifier"),
the type signature was changed to:
static const char *kernel_type_name(const struct btf* btf, u32 id)
With the btf parameter used for lookups instead of global btf_vmlinux.
The helper will function as expected for type name lookup using non-kernel BTFs,
and will be used for such in further patches in the series. Let's rename it to
avoid incorrect assumptions that might arise when seeing the current name.
Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
Link: https://lore.kernel.org/r/20230309180111.1618459-2-davemarchevsky@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch tests how many kmallocs is needed to create and free
a batch of UDP sockets and each socket has a 64bytes bpf storage.
It also measures how fast the UDP sockets can be created.
The result is from my qemu setup.
Before bpf_mem_cache_alloc/free:
./bench -p 1 local-storage-create
Setting up benchmark 'local-storage-create'...
Benchmark 'local-storage-create' started.
Iter 0 ( 73.193us): creates 213.552k/s (213.552k/prod), 3.09 kmallocs/create
Iter 1 (-20.724us): creates 211.908k/s (211.908k/prod), 3.09 kmallocs/create
Iter 2 ( 9.280us): creates 212.574k/s (212.574k/prod), 3.12 kmallocs/create
Iter 3 ( 11.039us): creates 213.209k/s (213.209k/prod), 3.12 kmallocs/create
Iter 4 (-11.411us): creates 213.351k/s (213.351k/prod), 3.12 kmallocs/create
Iter 5 ( -7.915us): creates 214.754k/s (214.754k/prod), 3.12 kmallocs/create
Iter 6 ( 11.317us): creates 210.942k/s (210.942k/prod), 3.12 kmallocs/create
Summary: creates 212.789 ± 1.310k/s (212.789k/prod), 3.12 kmallocs/create
After bpf_mem_cache_alloc/free:
./bench -p 1 local-storage-create
Setting up benchmark 'local-storage-create'...
Benchmark 'local-storage-create' started.
Iter 0 ( 68.265us): creates 243.984k/s (243.984k/prod), 1.04 kmallocs/create
Iter 1 ( 30.357us): creates 238.424k/s (238.424k/prod), 1.04 kmallocs/create
Iter 2 (-18.712us): creates 232.963k/s (232.963k/prod), 1.04 kmallocs/create
Iter 3 (-15.885us): creates 238.879k/s (238.879k/prod), 1.04 kmallocs/create
Iter 4 ( 5.590us): creates 237.490k/s (237.490k/prod), 1.04 kmallocs/create
Iter 5 ( 8.577us): creates 237.521k/s (237.521k/prod), 1.04 kmallocs/create
Iter 6 ( -6.263us): creates 238.508k/s (238.508k/prod), 1.04 kmallocs/create
Summary: creates 237.298 ± 2.198k/s (237.298k/prod), 1.04 kmallocs/create
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-18-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch tweats the socket_bind bpf prog to test the
local_storage->smap == NULL case in the bpf_local_storage_free()
code path. The idea is to create the local_storage with
the sk_storage_map's selem first. Then add the sk_storage_map2's selem
and then delete the earlier sk_storeage_map's selem.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-17-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch refactors local_storage freeing logic into
bpf_local_storage_free(). It is a preparation work for a later
patch that uses bpf_mem_cache_alloc/free. The other kfree(local_storage)
cases are also changed to bpf_local_storage_free(..., reuse_now = true).
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-12-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The existing bpf_local_storage_free_rcu is renamed to
bpf_local_storage_free_trace_rcu. A new bpf_local_storage_rcu
callback is added to do the kfree instead of using kfree_rcu.
It is a preparation work for a later patch using
bpf_mem_cache_alloc/free.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-11-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch refactors the selem freeing logic into bpf_selem_free().
It is a preparation work for a later patch using
bpf_mem_cache_alloc/free. The other kfree(selem) cases
are also changed to bpf_selem_free(..., reuse_now = true).
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-10-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add bpf_selem_free_rcu() callback to do the kfree() instead
of using kfree_rcu. It is a preparation work for using
bpf_mem_cache_alloc/free in a later patch.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-9-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch removes the bpf_selem_free_fields*_rcu. The
bpf_obj_free_fields() can be done before the call_rcu_trasks_trace()
and kfree_rcu(). It is needed when a later patch uses
bpf_mem_cache_alloc/free. In bpf hashtab, bpf_obj_free_fields()
is also called before calling bpf_mem_cache_free. The discussion
can be found in
https://lore.kernel.org/bpf/f67021ee-21d9-bfae-6134-4ca542fab843@linux.dev/
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-8-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch re-purpose the use_trace_rcu to mean
if the freed memory can be reused immediately or not.
The use_trace_rcu is renamed to reuse_now. Other than
the boolean test is reversed, it should be a no-op.
The following explains the reason for the rename and how it will
be used in a later patch.
In a later patch, bpf_mem_cache_alloc/free will be used
in the bpf_local_storage. The bpf mem allocator will reuse
the freed memory immediately. Some of the free paths in
bpf_local_storage does not support memory to be reused immediately.
These paths are the "delete" elem cases from the bpf_*_storage_delete()
helper and the map_delete_elem() syscall. Note that "delete" elem
before the owner's (sk/task/cgrp/inode) lifetime ended is not
the common usage for the local storage.
The common free path, bpf_local_storage_destroy(), can reuse the
memory immediately. This common path means the storage stays with
its owner until the owner is destroyed.
The above mentioned "delete" elem paths that cannot
reuse immediately always has the 'use_trace_rcu == true'.
The cases that is safe for immediate reuse always have
'use_trace_rcu == false'. Instead of adding another arg
in a later patch, this patch re-purpose this arg
to reuse_now and have the test logic reversed.
In a later patch, 'reuse_now == true' will free to the
bpf_mem_cache_free() where the memory can be reused
immediately. 'reuse_now == false' will go through the
call_rcu_tasks_trace().
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-7-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch remembers which smap triggers the allocation
of a 'struct bpf_local_storage' object. The local_storage is
allocated during the very first selem added to the owner.
The smap pointer is needed when using the bpf_mem_cache_free
in a later patch because it needs to free to the correct
smap's bpf_mem_alloc object.
When a selem is being removed, it needs to check if it is
the selem that triggers the creation of the local_storage.
If it is, the local_storage->smap pointer will be reset to NULL.
This NULL reset is done under the local_storage->lock in
bpf_selem_unlink_storage_nolock() when a selem is being removed.
Also note that the local_storage may not go away even
local_storage->smap is NULL because there may be other
selem still stored in the local_storage.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-6-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
__bpf_selem_unlink_storage is taking the spin lock and there is
no name collision also. Having the preceding '__' is confusing
when reviewing the later patch.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-5-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
bpf_local_storage_map_alloc() is the only caller of
__bpf_local_storage_map_alloc(). The remaining logic in
bpf_local_storage_map_alloc() is only a one liner setting
the smap->cache_idx.
Remove __bpf_local_storage_map_alloc() to simplify code.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-4-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch first renames bpf_local_storage_unlink_nolock to
bpf_local_storage_destroy(). It better reflects that it is only
used when the storage's owner (sk/task/cgrp/inode) is being kfree().
All bpf_local_storage_destroy's caller is taking the spin lock and
then free the storage. This patch also moves these two steps into
the bpf_local_storage_destroy.
This is a preparation work for a later patch that uses
bpf_mem_cache_alloc/free in the bpf_local_storage.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-3-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch moves the bpf_local_storage_free_rcu() and
bpf_selem_unlink_map() to static because they are
not used outside of bpf_local_storage.c.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230308065936.1550103-2-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The send_signal tracepoint tests are non-deterministically failing in
CI. The test works as follows:
1. Two pairs of file descriptors are created using the pipe() function.
One pair is used to communicate between a parent process -> child
process, and the other for the reverse direction.
2. A child is fork()'ed. The child process registers a signal handler,
notifies its parent that the signal handler is registered, and then
and waits for its parent to have enabled a BPF program that sends a
signal.
3. The parent opens and loads a BPF skeleton with programs that send
signals to the child process. The different programs are triggered by
different perf events (either NMI or normal perf), or by regular
tracepoints. The signal is delivered to the child whenever the child
triggers the program.
4. The child's signal handler is invoked, which sets a flag saying that
the signal handler was reached. The child then signals to the parent
that it received the signal, and the test ends.
The perf testcases (send_signal_perf{_thread} and
send_signal_nmi{_thread}) work 100% of the time, but the tracepoint
testcases fail non-deterministically because the tracepoint is not
always being fired for the child.
There are two tracepoint programs registered in the test:
'tracepoint/sched/sched_switch', and
'tracepoint/syscalls/sys_enter_nanosleep'. The child never intentionally
blocks, nor sleeps, so neither tracepoint is guaranteed to be triggered.
To fix this, we can have the child trigger the nanosleep program with a
usleep().
Before this patch, the test would fail locally every 2-3 runs. Now, it
doesn't fail after more than 1000 runs.
Signed-off-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/r/20230310061909.1420887-1-void@manifault.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
When doing state comparison, if old state has register that is not
marked as REG_LIVE_READ, then we just skip comparison, regardless what's
the state of corresponing register in current state. This is because not
REG_LIVE_READ register is irrelevant for further program execution and
correctness. All good here.
But when we get to precision propagation, after two states were declared
equivalent, we don't take into account old register's liveness, and thus
attempt to propagate precision for register in current state even if
that register in old state was not REG_LIVE_READ anymore. This is bad,
because register in current state could be anything at all and this
could cause -EFAULT due to internal logic bugs.
Fix by taking into account REG_LIVE_READ liveness mark to keep the logic
in state comparison in sync with precision propagation.
Fixes: a3ce685dd0 ("bpf: fix precision tracking")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20230309224131.57449-1-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
State equivalence check and checkpointing performed in is_state_visited()
employs certain heuristics to try to save memory by avoiding state checkpoints
if not enough jumps and instructions happened since last checkpoint. This leads
to unpredictability of whether a particular instruction will be checkpointed
and how regularly. While normally this is not causing much problems (except
inconveniences for predictable verifier tests, which we overcome with
BPF_F_TEST_STATE_FREQ flag), turns out it's not the case for open-coded
iterators.
Checking and saving state checkpoints at iter_next() call is crucial for fast
convergence of open-coded iterator loop logic, so we need to force it. If we
don't do that, is_state_visited() might skip saving a checkpoint, causing
unnecessarily long sequence of not checkpointed instructions and jumps, leading
to exhaustion of jump history buffer, and potentially other undesired outcomes.
It is expected that with correct open-coded iterators convergence will happen
quickly, so we don't run a risk of exhausting memory.
This patch adds, in addition to prune and jump instruction marks, also a
"forced checkpoint" mark, and makes sure that any iter_next() call instruction
is marked as such.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20230310060149.625887-1-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Andrii Nakryiko says:
====================
Make BPF-side compiler flags stricter by adding -Wall. Fix tons of small
issues pointed out by compiler immediately after that. That includes newly
added bpf_for(), bpf_for_each(), and bpf_repeat() macros.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
We recently added -Wuninitialized, but it's not enough to catch various
silly mistakes or omissions. Let's go all the way to -Wall, just like we
do for user-space code.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20230309054015.4068562-5-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Once we enable -Wall for BPF sources, compiler will complain about lots
of unused variables, variables that are set but never read, etc.
Fix all these issues first before enabling -Wall in Makefile.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20230309054015.4068562-4-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add __sink(expr) macro that forces compiler to believe that passed in
expression is both read and written. It used a simple embedded asm for
this. This is useful in a lot of tests where we assign value to some variable
to trigger some action, but later don't read variable, causing compiler
to complain (if corresponding compiler warnings are turned on, which
we'll do in the next patch).
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20230309054015.4068562-3-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
